All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH 0/5] Add support for PCIe controller to work in endpoint mode on R-Car SoCs.
@ 2019-11-06 19:36 ` Lad Prabhakar
  0 siblings, 0 replies; 60+ messages in thread
From: Lad Prabhakar @ 2019-11-06 19:36 UTC (permalink / raw)
  To: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Kishon Vijay Abraham I, Marek Vasut,
	Yoshihiro Shimoda, linux-pci
  Cc: Catalin Marinas, Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray, devicetree, linux-kernel,
	linux-arm-kernel, linux-renesas-soc, Chris Paterson, Lad,
	Prabhakar

From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>

This patch series adds support for PCIe controller on rcar to work in endpoint mode,
this also extends the epf framework to handle multiple outbound regions.

Note: 
The cadence/rockchip/designware endpoint drivers are build tested only.

Endpoint configuration:
$ cd /sys/kernel/config/pci_ep
$ mkdir functions/pci_epf_test/func1
$ echo 0x1912 > functions/pci_epf_test/func1/vendorid
$ echo 0x002d > functions/pci_epf_test/func1/deviceid
$ echo 32 > functions/pci_epf_test/func1/msi_interrupts
$ ln -s functions/pci_epf_test/func1/ controllers/fe000000.pcie_ep/
$ echo 1 > controllers/fe000000.pcie_ep/start

Host side results:
lspci output:
01:00.0 Unassigned class [ff00]: Renesas Technology Corp. Device 002d
        Flags: bus master, fast devsel, latency 0, IRQ 103
        Memory at fe200200 (64-bit, non-prefetchable) [size=128]
        Memory at fe200000 (64-bit, non-prefetchable) [size=256]
        Memory at fe200100 (64-bit, non-prefetchable) [size=256]
        Capabilities: [40] Power Management version 3
        Capabilities: [50] MSI: Enable- Count=1/1 Maskable+ 64bit+
        Capabilities: [70] Express Endpoint, MSI 00
        Capabilities: [100] Virtual Channel
        Kernel driver in use: pci-endpoint-test

pcitest results:

BAR tests

BAR0:           OKAY
BAR1:           NOT OKAY
BAR2:           OKAY
BAR3:           NOT OKAY
BAR4:           OKAY
BAR5:           NOT OKAY

Interrupt tests

SET IRQ TYPE TO LEGACY:         OKAY
LEGACY IRQ:     OKAY
Read Tests

SET IRQ TYPE TO LEGACY:         OKAY
READ (      1 bytes):           OKAY
READ (   1024 bytes):           OKAY
READ (   1025 bytes):           OKAY
READ (1024000 bytes):           OKAY
READ (1024001 bytes):           OKAY

Write Tests

WRITE (      1 bytes):          OKAY
WRITE (   1024 bytes):          OKAY
WRITE (   1025 bytes):          OKAY
WRITE (1024000 bytes):          OKAY
WRITE (1024001 bytes):          OKAY

Copy Tests

COPY (      1 bytes):           OKAY
COPY (   1024 bytes):           OKAY
COPY (   1025 bytes):           OKAY
COPY (1024000 bytes):           OKAY
COPY (1024001 bytes):           OKAY

BAR tests for 1/3/5 fail because its configured to use 64bit bars

Lad, Prabhakar (5):
  pci: pcie-rcar: preparation for adding endpoint support
  pci: endpoint: add support to handle multiple base for mapping
    outbound memory
  PCI: rcar: Add R-Car PCIe endpoint device tree bindings
  pci: rcar: add support for rcar pcie controller in endpoint mode
  misc: pci_endpoint_test: add device-id for RZ/G2 pcie controller

 .../devicetree/bindings/pci/rcar-pci-ep.txt   |   43 +
 arch/arm64/configs/defconfig                  |    2 +-
 arch/arm64/configs/renesas_defconfig          |    2 +-
 drivers/misc/pci_endpoint_test.c              |    3 +
 drivers/pci/controller/Kconfig                |   11 +-
 drivers/pci/controller/Makefile               |    3 +-
 .../pci/controller/dwc/pcie-designware-ep.c   |   30 +-
 drivers/pci/controller/pcie-cadence-ep.c      |   11 +-
 drivers/pci/controller/pcie-rcar-ep.c         |  483 +++++++
 drivers/pci/controller/pcie-rcar-host.c       | 1058 ++++++++++++++
 drivers/pci/controller/pcie-rcar.c            | 1231 +----------------
 drivers/pci/controller/pcie-rcar.h            |  129 ++
 drivers/pci/controller/pcie-rockchip-ep.c     |   13 +-
 drivers/pci/endpoint/functions/pci-epf-test.c |   29 +-
 drivers/pci/endpoint/pci-epc-core.c           |    7 +-
 drivers/pci/endpoint/pci-epc-mem.c            |  189 ++-
 include/linux/pci-epc.h                       |   43 +-
 17 files changed, 2016 insertions(+), 1271 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
 create mode 100644 drivers/pci/controller/pcie-rcar-ep.c
 create mode 100644 drivers/pci/controller/pcie-rcar-host.c
 create mode 100644 drivers/pci/controller/pcie-rcar.h

-- 
2.20.1


^ permalink raw reply	[flat|nested] 60+ messages in thread

* [PATCH 0/5] Add support for PCIe controller to work in endpoint mode on R-Car SoCs.
@ 2019-11-06 19:36 ` Lad Prabhakar
  0 siblings, 0 replies; 60+ messages in thread
From: Lad Prabhakar @ 2019-11-06 19:36 UTC (permalink / raw)
  To: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Kishon Vijay Abraham I, Marek Vasut,
	Yoshihiro Shimoda, linux-pci
  Cc: devicetree, Chris Paterson, Lorenzo Pieralisi, Arnd Bergmann,
	Catalin Marinas, linux-kernel, Lad, Prabhakar, linux-renesas-soc,
	Greg Kroah-Hartman, Andrew Murray, Will Deacon, linux-arm-kernel

From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>

This patch series adds support for PCIe controller on rcar to work in endpoint mode,
this also extends the epf framework to handle multiple outbound regions.

Note: 
The cadence/rockchip/designware endpoint drivers are build tested only.

Endpoint configuration:
$ cd /sys/kernel/config/pci_ep
$ mkdir functions/pci_epf_test/func1
$ echo 0x1912 > functions/pci_epf_test/func1/vendorid
$ echo 0x002d > functions/pci_epf_test/func1/deviceid
$ echo 32 > functions/pci_epf_test/func1/msi_interrupts
$ ln -s functions/pci_epf_test/func1/ controllers/fe000000.pcie_ep/
$ echo 1 > controllers/fe000000.pcie_ep/start

Host side results:
lspci output:
01:00.0 Unassigned class [ff00]: Renesas Technology Corp. Device 002d
        Flags: bus master, fast devsel, latency 0, IRQ 103
        Memory at fe200200 (64-bit, non-prefetchable) [size=128]
        Memory at fe200000 (64-bit, non-prefetchable) [size=256]
        Memory at fe200100 (64-bit, non-prefetchable) [size=256]
        Capabilities: [40] Power Management version 3
        Capabilities: [50] MSI: Enable- Count=1/1 Maskable+ 64bit+
        Capabilities: [70] Express Endpoint, MSI 00
        Capabilities: [100] Virtual Channel
        Kernel driver in use: pci-endpoint-test

pcitest results:

BAR tests

BAR0:           OKAY
BAR1:           NOT OKAY
BAR2:           OKAY
BAR3:           NOT OKAY
BAR4:           OKAY
BAR5:           NOT OKAY

Interrupt tests

SET IRQ TYPE TO LEGACY:         OKAY
LEGACY IRQ:     OKAY
Read Tests

SET IRQ TYPE TO LEGACY:         OKAY
READ (      1 bytes):           OKAY
READ (   1024 bytes):           OKAY
READ (   1025 bytes):           OKAY
READ (1024000 bytes):           OKAY
READ (1024001 bytes):           OKAY

Write Tests

WRITE (      1 bytes):          OKAY
WRITE (   1024 bytes):          OKAY
WRITE (   1025 bytes):          OKAY
WRITE (1024000 bytes):          OKAY
WRITE (1024001 bytes):          OKAY

Copy Tests

COPY (      1 bytes):           OKAY
COPY (   1024 bytes):           OKAY
COPY (   1025 bytes):           OKAY
COPY (1024000 bytes):           OKAY
COPY (1024001 bytes):           OKAY

BAR tests for 1/3/5 fail because its configured to use 64bit bars

Lad, Prabhakar (5):
  pci: pcie-rcar: preparation for adding endpoint support
  pci: endpoint: add support to handle multiple base for mapping
    outbound memory
  PCI: rcar: Add R-Car PCIe endpoint device tree bindings
  pci: rcar: add support for rcar pcie controller in endpoint mode
  misc: pci_endpoint_test: add device-id for RZ/G2 pcie controller

 .../devicetree/bindings/pci/rcar-pci-ep.txt   |   43 +
 arch/arm64/configs/defconfig                  |    2 +-
 arch/arm64/configs/renesas_defconfig          |    2 +-
 drivers/misc/pci_endpoint_test.c              |    3 +
 drivers/pci/controller/Kconfig                |   11 +-
 drivers/pci/controller/Makefile               |    3 +-
 .../pci/controller/dwc/pcie-designware-ep.c   |   30 +-
 drivers/pci/controller/pcie-cadence-ep.c      |   11 +-
 drivers/pci/controller/pcie-rcar-ep.c         |  483 +++++++
 drivers/pci/controller/pcie-rcar-host.c       | 1058 ++++++++++++++
 drivers/pci/controller/pcie-rcar.c            | 1231 +----------------
 drivers/pci/controller/pcie-rcar.h            |  129 ++
 drivers/pci/controller/pcie-rockchip-ep.c     |   13 +-
 drivers/pci/endpoint/functions/pci-epf-test.c |   29 +-
 drivers/pci/endpoint/pci-epc-core.c           |    7 +-
 drivers/pci/endpoint/pci-epc-mem.c            |  189 ++-
 include/linux/pci-epc.h                       |   43 +-
 17 files changed, 2016 insertions(+), 1271 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
 create mode 100644 drivers/pci/controller/pcie-rcar-ep.c
 create mode 100644 drivers/pci/controller/pcie-rcar-host.c
 create mode 100644 drivers/pci/controller/pcie-rcar.h

-- 
2.20.1


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 60+ messages in thread

* [PATCH 1/5] pci: pcie-rcar: preparation for adding endpoint support
  2019-11-06 19:36 ` Lad Prabhakar
@ 2019-11-06 19:36   ` Lad Prabhakar
  -1 siblings, 0 replies; 60+ messages in thread
From: Lad Prabhakar @ 2019-11-06 19:36 UTC (permalink / raw)
  To: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Kishon Vijay Abraham I, Marek Vasut,
	Yoshihiro Shimoda, linux-pci
  Cc: Catalin Marinas, Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray, devicetree, linux-kernel,
	linux-arm-kernel, linux-renesas-soc, Chris Paterson, Lad,
	Prabhakar

From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>

this patch prepares for adding endpoint support to rcar controller,
there are no functional changes with this patch, a common file is
created so that it can be shared with endpoint driver.

Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
---
 arch/arm64/configs/defconfig            |    2 +-
 arch/arm64/configs/renesas_defconfig    |    2 +-
 drivers/pci/controller/Kconfig          |    4 +-
 drivers/pci/controller/Makefile         |    2 +-
 drivers/pci/controller/pcie-rcar-host.c | 1058 +++++++++++++++++++
 drivers/pci/controller/pcie-rcar.c      | 1231 +----------------------
 drivers/pci/controller/pcie-rcar.h      |  129 +++
 7 files changed, 1245 insertions(+), 1183 deletions(-)
 create mode 100644 drivers/pci/controller/pcie-rcar-host.c
 create mode 100644 drivers/pci/controller/pcie-rcar.h

diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
index 7d16f10a9251..f67ce0ccea1d 100644
--- a/arch/arm64/configs/defconfig
+++ b/arch/arm64/configs/defconfig
@@ -186,7 +186,7 @@ CONFIG_HOTPLUG_PCI=y
 CONFIG_HOTPLUG_PCI_ACPI=y
 CONFIG_PCI_AARDVARK=y
 CONFIG_PCI_TEGRA=y
-CONFIG_PCIE_RCAR=y
+CONFIG_PCIE_RCAR_HOST=y
 CONFIG_PCI_HOST_GENERIC=y
 CONFIG_PCI_XGENE=y
 CONFIG_PCIE_ALTERA=y
diff --git a/arch/arm64/configs/renesas_defconfig b/arch/arm64/configs/renesas_defconfig
index f7411e8807d6..75fc418377c5 100644
--- a/arch/arm64/configs/renesas_defconfig
+++ b/arch/arm64/configs/renesas_defconfig
@@ -91,7 +91,7 @@ CONFIG_PCIEPORTBUS=y
 CONFIG_HOTPLUG_PCI_PCIE=y
 CONFIG_PCI_IOV=y
 CONFIG_HOTPLUG_PCI=y
-CONFIG_PCIE_RCAR=y
+CONFIG_PCIE_RCAR_HOST=y
 CONFIG_PCI_HOST_GENERIC=y
 CONFIG_DEVTMPFS=y
 CONFIG_DEVTMPFS_MOUNT=y
diff --git a/drivers/pci/controller/Kconfig b/drivers/pci/controller/Kconfig
index 70e078238899..01c238e5ea1e 100644
--- a/drivers/pci/controller/Kconfig
+++ b/drivers/pci/controller/Kconfig
@@ -82,12 +82,12 @@ config PCI_RCAR_GEN2
 	  There are 3 internal PCI controllers available with a single
 	  built-in EHCI/OHCI host controller present on each one.
 
-config PCIE_RCAR
+config PCIE_RCAR_HOST
 	bool "Renesas R-Car PCIe controller"
 	depends on ARCH_RENESAS || COMPILE_TEST
 	depends on PCI_MSI_IRQ_DOMAIN
 	help
-	  Say Y here if you want PCIe controller support on R-Car SoCs.
+	  Say Y here if you want PCIe controller support on R-Car SoCs in host mode.
 
 config PCI_HOST_COMMON
 	bool
diff --git a/drivers/pci/controller/Makefile b/drivers/pci/controller/Makefile
index a2a22c9d91af..3577902a0d40 100644
--- a/drivers/pci/controller/Makefile
+++ b/drivers/pci/controller/Makefile
@@ -9,7 +9,7 @@ obj-$(CONFIG_PCI_MVEBU) += pci-mvebu.o
 obj-$(CONFIG_PCI_AARDVARK) += pci-aardvark.o
 obj-$(CONFIG_PCI_TEGRA) += pci-tegra.o
 obj-$(CONFIG_PCI_RCAR_GEN2) += pci-rcar-gen2.o
-obj-$(CONFIG_PCIE_RCAR) += pcie-rcar.o
+obj-$(CONFIG_PCIE_RCAR_HOST) += pcie-rcar.o pcie-rcar-host.o
 obj-$(CONFIG_PCI_HOST_COMMON) += pci-host-common.o
 obj-$(CONFIG_PCI_HOST_GENERIC) += pci-host-generic.o
 obj-$(CONFIG_PCIE_XILINX) += pcie-xilinx.o
diff --git a/drivers/pci/controller/pcie-rcar-host.c b/drivers/pci/controller/pcie-rcar-host.c
new file mode 100644
index 000000000000..645ef3e41e27
--- /dev/null
+++ b/drivers/pci/controller/pcie-rcar-host.c
@@ -0,0 +1,1058 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * PCIe driver for Renesas R-Car SoCs
+ *  Copyright (C) 2014-2019 Renesas Electronics Europe Ltd
+ *
+ * Based on:
+ *  arch/sh/drivers/pci/pcie-sh7786.c
+ *  arch/sh/drivers/pci/ops-sh7786.c
+ *  Copyright (C) 2009 - 2011  Paul Mundt
+ *
+ * Author: Phil Edworthy <phil.edworthy@renesas.com>
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/msi.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/of_pci.h>
+#include <linux/of_platform.h>
+#include <linux/pci.h>
+#include <linux/phy/phy.h>
+
+#include "pcie-rcar.h"
+
+struct rcar_msi {
+	DECLARE_BITMAP(used, INT_PCI_MSI_NR);
+	struct irq_domain *domain;
+	struct msi_controller chip;
+	unsigned long pages;
+	struct mutex lock; /* serializes msi */
+	int irq1;
+	int irq2;
+};
+
+static inline struct rcar_msi *to_rcar_msi(struct msi_controller *chip)
+{
+	return container_of(chip, struct rcar_msi, chip);
+}
+
+/* Structure representing the PCIe interface */
+struct rcar_pcie {
+	struct device		*dev;
+	struct phy		*phy;
+	void __iomem		*base;
+	struct list_head	resources;
+	int			root_bus_nr;
+	struct clk		*bus_clk;
+	struct			rcar_msi msi;
+};
+
+static u32 rcar_read_conf(struct rcar_pcie *pcie, int where)
+{
+	unsigned int shift = BITS_PER_BYTE * (where & 3);
+	u32 val = rcar_pci_read_reg(pcie->base, where & ~3);
+
+	return val >> shift;
+}
+
+/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
+static int
+rcar_pcie_config_access(struct rcar_pcie *pcie,
+			unsigned char access_type, struct pci_bus *bus,
+			unsigned int devfn, int where, u32 *data)
+{
+	unsigned int dev, func, reg, index;
+
+	dev = PCI_SLOT(devfn);
+	func = PCI_FUNC(devfn);
+	reg = where & ~3;
+	index = reg / 4;
+
+	/*
+	 * While each channel has its own memory-mapped extended config
+	 * space, it's generally only accessible when in endpoint mode.
+	 * When in root complex mode, the controller is unable to target
+	 * itself with either type 0 or type 1 accesses, and indeed, any
+	 * controller initiated target transfer to its own config space
+	 * result in a completer abort.
+	 *
+	 * Each channel effectively only supports a single device, but as
+	 * the same channel <-> device access works for any PCI_SLOT()
+	 * value, we cheat a bit here and bind the controller's config
+	 * space to devfn 0 in order to enable self-enumeration. In this
+	 * case the regular ECAR/ECDR path is sidelined and the mangled
+	 * config access itself is initiated as an internal bus transaction.
+	 */
+	if (pci_is_root_bus(bus)) {
+		if (dev != 0)
+			return PCIBIOS_DEVICE_NOT_FOUND;
+
+		if (access_type == RCAR_PCI_ACCESS_READ) {
+			*data = rcar_pci_read_reg(pcie->base, PCICONF(index));
+		} else {
+			/* Keep an eye out for changes to the root bus number */
+			if (pci_is_root_bus(bus) && reg == PCI_PRIMARY_BUS)
+				pcie->root_bus_nr = *data & 0xff;
+
+			rcar_pci_write_reg(pcie->base, *data, PCICONF(index));
+		}
+
+		return PCIBIOS_SUCCESSFUL;
+	}
+
+	if (pcie->root_bus_nr < 0)
+		return PCIBIOS_DEVICE_NOT_FOUND;
+
+	/* Clear errors */
+	rcar_pci_write_reg(pcie->base, rcar_pci_read_reg(pcie->base, PCIEERRFR),
+			   PCIEERRFR);
+
+	/* Set the PIO address */
+	rcar_pci_write_reg(pcie->base, PCIE_CONF_BUS(bus->number) |
+		PCIE_CONF_DEV(dev) | PCIE_CONF_FUNC(func) | reg, PCIECAR);
+
+	/* Enable the configuration access */
+	if (bus->parent->number == pcie->root_bus_nr)
+		rcar_pci_write_reg(pcie->base,
+				   CONFIG_SEND_ENABLE | TYPE0, PCIECCTLR);
+	else
+		rcar_pci_write_reg(pcie->base,
+				   CONFIG_SEND_ENABLE | TYPE1, PCIECCTLR);
+
+	/* Check for errors */
+	if (rcar_pci_read_reg(pcie->base, PCIEERRFR) & UNSUPPORTED_REQUEST)
+		return PCIBIOS_DEVICE_NOT_FOUND;
+
+	/* Check for master and target aborts */
+	if (rcar_read_conf(pcie, RCONF(PCI_STATUS)) &
+		(PCI_STATUS_REC_MASTER_ABORT | PCI_STATUS_REC_TARGET_ABORT))
+		return PCIBIOS_DEVICE_NOT_FOUND;
+
+	if (access_type == RCAR_PCI_ACCESS_READ)
+		*data = rcar_pci_read_reg(pcie->base, PCIECDR);
+	else
+		rcar_pci_write_reg(pcie->base, *data, PCIECDR);
+
+	/* Disable the configuration access */
+	rcar_pci_write_reg(pcie->base, 0, PCIECCTLR);
+
+	return PCIBIOS_SUCCESSFUL;
+}
+
+static int rcar_pcie_read_conf(struct pci_bus *bus, unsigned int devfn,
+			       int where, int size, u32 *val)
+{
+	struct rcar_pcie *pcie = bus->sysdata;
+	int ret;
+
+	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
+				      bus, devfn, where, val);
+	if (ret != PCIBIOS_SUCCESSFUL) {
+		*val = 0xffffffff;
+		return ret;
+	}
+
+	if (size == 1)
+		*val = (*val >> (BITS_PER_BYTE * (where & 3))) & 0xff;
+	else if (size == 2)
+		*val = (*val >> (BITS_PER_BYTE * (where & 2))) & 0xffff;
+
+	dev_dbg(&bus->dev, "pcie-config-read: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
+		bus->number, devfn, where, size, *val);
+
+	return ret;
+}
+
+/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
+static int rcar_pcie_write_conf(struct pci_bus *bus, unsigned int devfn,
+				int where, int size, u32 val)
+{
+	struct rcar_pcie *pcie = bus->sysdata;
+	unsigned int shift;
+	u32 data;
+	int ret;
+
+	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
+				      bus, devfn, where, &data);
+	if (ret != PCIBIOS_SUCCESSFUL)
+		return ret;
+
+	dev_dbg(&bus->dev, "pcie-config-write: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
+		bus->number, devfn, where, size, val);
+
+	if (size == 1) {
+		shift = BITS_PER_BYTE * (where & 3);
+		data &= ~(0xff << shift);
+		data |= ((val & 0xff) << shift);
+	} else if (size == 2) {
+		shift = BITS_PER_BYTE * (where & 2);
+		data &= ~(0xffff << shift);
+		data |= ((val & 0xffff) << shift);
+	} else {
+		data = val;
+	}
+
+	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_WRITE,
+				      bus, devfn, where, &data);
+
+	return ret;
+}
+
+static struct pci_ops rcar_pcie_ops = {
+	.read	= rcar_pcie_read_conf,
+	.write	= rcar_pcie_write_conf,
+};
+
+static int rcar_pcie_setup(struct list_head *resource, struct rcar_pcie *pci)
+{
+	struct resource_entry *win;
+	int i = 0;
+
+	/* Setup PCI resources */
+	resource_list_for_each_entry(win, &pci->resources) {
+		struct resource *res = win->res;
+
+		if (!res->flags)
+			continue;
+
+		switch (resource_type(res)) {
+		case IORESOURCE_IO:
+		case IORESOURCE_MEM:
+			rcar_pcie_set_outbound(i, pci->base, res, true);
+			i++;
+			break;
+		case IORESOURCE_BUS:
+			pci->root_bus_nr = res->start;
+			break;
+		default:
+			continue;
+		}
+
+		pci_add_resource(resource, res);
+	}
+
+	return 1;
+}
+
+static void rcar_pcie_force_speedup(struct rcar_pcie *pcie)
+{
+	struct device *dev = pcie->dev;
+	unsigned int timeout = 1000;
+	u32 macsr;
+
+	if ((rcar_pci_read_reg(pcie->base, MACS2R) & LINK_SPEED) !=
+		LINK_SPEED_5_0GTS)
+		return;
+
+	if (rcar_pci_read_reg(pcie->base, MACCTLR) & SPEED_CHANGE) {
+		dev_err(dev, "Speed change already in progress\n");
+		return;
+	}
+
+	macsr = rcar_pci_read_reg(pcie->base, MACSR);
+	if ((macsr & LINK_SPEED) == LINK_SPEED_5_0GTS)
+		goto done;
+
+	/* Set target link speed to 5.0 GT/s */
+	rcar_rmw32(pcie->base, EXPCAP(12), PCI_EXP_LNKSTA_CLS,
+		   PCI_EXP_LNKSTA_CLS_5_0GB);
+
+	/* Set speed change reason as intentional factor */
+	rcar_rmw32(pcie->base, MACCGSPSETR, SPCNGRSN, 0);
+
+	/* Clear SPCHGFIN, SPCHGSUC, and SPCHGFAIL */
+	if (macsr & (SPCHGFIN | SPCHGSUC | SPCHGFAIL))
+		rcar_pci_write_reg(pcie->base, macsr, MACSR);
+
+	/* Start link speed change */
+	rcar_rmw32(pcie->base, MACCTLR, SPEED_CHANGE, SPEED_CHANGE);
+
+	while (timeout--) {
+		macsr = rcar_pci_read_reg(pcie->base, MACSR);
+		if (macsr & SPCHGFIN) {
+			/* Clear the interrupt bits */
+			rcar_pci_write_reg(pcie->base, macsr, MACSR);
+
+			if (macsr & SPCHGFAIL)
+				dev_err(dev, "Speed change failed\n");
+
+			goto done;
+		}
+
+		usleep_range(1000, 2000);
+	}
+
+	dev_err(dev, "Speed change timed out\n");
+
+done:
+	dev_info(dev, "Current link speed is %s GT/s\n",
+		 (macsr & LINK_SPEED) == LINK_SPEED_5_0GTS ? "5" : "2.5");
+}
+
+static int rcar_pcie_enable(struct rcar_pcie *pcie)
+{
+	struct pci_host_bridge *bridge = pci_host_bridge_from_priv(pcie);
+	struct device *dev = pcie->dev;
+	struct pci_bus *bus, *child;
+	int ret;
+
+	/* Try setting 5 GT/s link speed */
+	rcar_pcie_force_speedup(pcie);
+
+	rcar_pcie_setup(&bridge->windows, pcie);
+
+	pci_add_flags(PCI_REASSIGN_ALL_BUS);
+
+	bridge->dev.parent = dev;
+	bridge->sysdata = pcie;
+	bridge->busnr = pcie->root_bus_nr;
+	bridge->ops = &rcar_pcie_ops;
+	bridge->map_irq = of_irq_parse_and_map_pci;
+	bridge->swizzle_irq = pci_common_swizzle;
+	if (IS_ENABLED(CONFIG_PCI_MSI))
+		bridge->msi = &pcie->msi.chip;
+
+	ret = pci_scan_root_bus_bridge(bridge);
+	if (ret < 0)
+		return ret;
+
+	bus = bridge->bus;
+
+	pci_bus_size_bridges(bus);
+	pci_bus_assign_resources(bus);
+
+	list_for_each_entry(child, &bus->children, node)
+		pcie_bus_configure_settings(child);
+
+	pci_bus_add_devices(bus);
+
+	return 0;
+}
+
+static int phy_wait_for_ack(struct rcar_pcie *pcie)
+{
+	struct device *dev = pcie->dev;
+	unsigned int timeout = 100;
+
+	while (timeout--) {
+		if (rcar_pci_read_reg(pcie->base, H1_PCIEPHYADRR) & PHY_ACK)
+			return 0;
+
+		usleep_range(90, 110);
+	}
+
+	dev_err(dev, "Access to PCIe phy timed out\n");
+
+	return -ETIMEDOUT;
+}
+
+static void phy_write_reg(struct rcar_pcie *pcie,
+			  unsigned int rate, u32 addr,
+			  unsigned int lane, u32 data)
+{
+	u32 phyaddr;
+
+	phyaddr = WRITE_CMD |
+		((rate & 1) << RATE_POS) |
+		((lane & 0xf) << LANE_POS) |
+		((addr & 0xff) << ADR_POS);
+
+	/* Set write data */
+	rcar_pci_write_reg(pcie->base, data, H1_PCIEPHYDOUTR);
+	rcar_pci_write_reg(pcie->base, phyaddr, H1_PCIEPHYADRR);
+
+	/* Ignore errors as they will be dealt with if the data link is down */
+	phy_wait_for_ack(pcie);
+
+	/* Clear command */
+	rcar_pci_write_reg(pcie->base, 0, H1_PCIEPHYDOUTR);
+	rcar_pci_write_reg(pcie->base, 0, H1_PCIEPHYADRR);
+
+	/* Ignore errors as they will be dealt with if the data link is down */
+	phy_wait_for_ack(pcie);
+}
+
+static int rcar_pcie_hw_init(struct rcar_pcie *pcie)
+{
+	int err;
+
+	/* Begin initialization */
+	rcar_pci_write_reg(pcie->base, 0, PCIETCTLR);
+
+	/* Set mode */
+	rcar_pci_write_reg(pcie->base, 1, PCIEMSR);
+
+	err = rcar_pcie_wait_for_phyrdy(pcie->base);
+	if (err)
+		return err;
+
+	/*
+	 * Initial header for port config space is type 1, set the device
+	 * class to match. Hardware takes care of propagating the IDSETR
+	 * settings, so there is no need to bother with a quirk.
+	 */
+	rcar_pci_write_reg(pcie->base, PCI_CLASS_BRIDGE_PCI << 16, IDSETR1);
+
+	/*
+	 * Setup Secondary Bus Number & Subordinate Bus Number, even though
+	 * they aren't used, to avoid bridge being detected as broken.
+	 */
+	rcar_rmw32(pcie->base, RCONF(PCI_SECONDARY_BUS), 0xff, 1);
+	rcar_rmw32(pcie->base, RCONF(PCI_SUBORDINATE_BUS), 0xff, 1);
+
+	/* Initialize default capabilities. */
+	rcar_rmw32(pcie->base, REXPCAP(0), 0xff, PCI_CAP_ID_EXP);
+	rcar_rmw32(pcie->base, REXPCAP(PCI_EXP_FLAGS),
+		   PCI_EXP_FLAGS_TYPE, PCI_EXP_TYPE_ROOT_PORT << 4);
+	rcar_rmw32(pcie->base, RCONF(PCI_HEADER_TYPE), 0x7f,
+		   PCI_HEADER_TYPE_BRIDGE);
+
+	/* Enable data link layer active state reporting */
+	rcar_rmw32(pcie->base, REXPCAP(PCI_EXP_LNKCAP), PCI_EXP_LNKCAP_DLLLARC,
+		   PCI_EXP_LNKCAP_DLLLARC);
+
+	/* Write out the physical slot number = 0 */
+	rcar_rmw32(pcie->base, REXPCAP(PCI_EXP_SLTCAP), PCI_EXP_SLTCAP_PSN, 0);
+
+	/* Set the completion timer timeout to the maximum 50ms. */
+	rcar_rmw32(pcie->base, TLCTLR + 1, 0x3f, 50);
+
+	/* Terminate list of capabilities (Next Capability Offset=0) */
+	rcar_rmw32(pcie->base, RVCCAP(0), 0xfff00000, 0);
+
+	/* Enable MSI */
+	if (IS_ENABLED(CONFIG_PCI_MSI))
+		rcar_pci_write_reg(pcie->base, 0x801f0000, PCIEMSITXR);
+
+	rcar_rmw32(pcie->base, MACCTLR, MACCTLR_RESERVED, 0);
+
+	/* Finish initialization - establish a PCI Express link */
+	rcar_pci_write_reg(pcie->base, CFINIT, PCIETCTLR);
+
+	/* This will timeout if we don't have a link. */
+	err = rcar_pcie_wait_for_dl(pcie->base);
+	if (err)
+		return err;
+
+	/* Enable INTx interrupts */
+	rcar_rmw32(pcie->base, PCIEINTXR, 0, 0xF << 8);
+
+	/* flush modifications */
+	wmb();
+
+	return 0;
+}
+
+static int rcar_pcie_phy_init_h1(struct rcar_pcie *pcie)
+{
+	/* Initialize the phy */
+	phy_write_reg(pcie, 0, 0x42, 0x1, 0x0EC34191);
+	phy_write_reg(pcie, 1, 0x42, 0x1, 0x0EC34180);
+	phy_write_reg(pcie, 0, 0x43, 0x1, 0x00210188);
+	phy_write_reg(pcie, 1, 0x43, 0x1, 0x00210188);
+	phy_write_reg(pcie, 0, 0x44, 0x1, 0x015C0014);
+	phy_write_reg(pcie, 1, 0x44, 0x1, 0x015C0014);
+	phy_write_reg(pcie, 1, 0x4C, 0x1, 0x786174A0);
+	phy_write_reg(pcie, 1, 0x4D, 0x1, 0x048000BB);
+	phy_write_reg(pcie, 0, 0x51, 0x1, 0x079EC062);
+	phy_write_reg(pcie, 0, 0x52, 0x1, 0x20000000);
+	phy_write_reg(pcie, 1, 0x52, 0x1, 0x20000000);
+	phy_write_reg(pcie, 1, 0x56, 0x1, 0x00003806);
+
+	phy_write_reg(pcie, 0, 0x60, 0x1, 0x004B03A5);
+	phy_write_reg(pcie, 0, 0x64, 0x1, 0x3F0F1F0F);
+	phy_write_reg(pcie, 0, 0x66, 0x1, 0x00008000);
+
+	return 0;
+}
+
+static int rcar_pcie_phy_init_gen2(struct rcar_pcie *pcie)
+{
+	/*
+	 * These settings come from the R-Car Series, 2nd Generation User's
+	 * Manual, section 50.3.1 (2) Initialization of the physical layer.
+	 */
+	rcar_pci_write_reg(pcie->base, 0x000f0030, GEN2_PCIEPHYADDR);
+	rcar_pci_write_reg(pcie->base, 0x00381203, GEN2_PCIEPHYDATA);
+	rcar_pci_write_reg(pcie->base, 0x00000001, GEN2_PCIEPHYCTRL);
+	rcar_pci_write_reg(pcie->base, 0x00000006, GEN2_PCIEPHYCTRL);
+
+	rcar_pci_write_reg(pcie->base, 0x000f0054, GEN2_PCIEPHYADDR);
+	/* The following value is for DC connection, no termination resistor */
+	rcar_pci_write_reg(pcie->base, 0x13802007, GEN2_PCIEPHYDATA);
+	rcar_pci_write_reg(pcie->base, 0x00000001, GEN2_PCIEPHYCTRL);
+	rcar_pci_write_reg(pcie->base, 0x00000006, GEN2_PCIEPHYCTRL);
+
+	return 0;
+}
+
+static int rcar_pcie_phy_init_gen3(struct rcar_pcie *pcie)
+{
+	int err;
+
+	err = phy_init(pcie->phy);
+	if (err)
+		return err;
+
+	err = phy_power_on(pcie->phy);
+	if (err)
+		phy_exit(pcie->phy);
+
+	return err;
+}
+
+static int rcar_msi_alloc(struct rcar_msi *chip)
+{
+	int msi;
+
+	mutex_lock(&chip->lock);
+
+	msi = find_first_zero_bit(chip->used, INT_PCI_MSI_NR);
+	if (msi < INT_PCI_MSI_NR)
+		set_bit(msi, chip->used);
+	else
+		msi = -ENOSPC;
+
+	mutex_unlock(&chip->lock);
+
+	return msi;
+}
+
+static int rcar_msi_alloc_region(struct rcar_msi *chip, int no_irqs)
+{
+	int msi;
+
+	mutex_lock(&chip->lock);
+	msi = bitmap_find_free_region(chip->used, INT_PCI_MSI_NR,
+				      order_base_2(no_irqs));
+	mutex_unlock(&chip->lock);
+
+	return msi;
+}
+
+static void rcar_msi_free(struct rcar_msi *chip, unsigned long irq)
+{
+	mutex_lock(&chip->lock);
+	clear_bit(irq, chip->used);
+	mutex_unlock(&chip->lock);
+}
+
+static irqreturn_t rcar_pcie_msi_irq(int irq, void *data)
+{
+	struct rcar_pcie *pcie = data;
+	struct rcar_msi *msi = &pcie->msi;
+	struct device *dev = pcie->dev;
+	unsigned long reg;
+
+	reg = rcar_pci_read_reg(pcie->base, PCIEMSIFR);
+
+	/* MSI & INTx share an interrupt - we only handle MSI here */
+	if (!reg)
+		return IRQ_NONE;
+
+	while (reg) {
+		unsigned int index = find_first_bit(&reg, 32);
+		unsigned int msi_irq;
+
+		/* clear the interrupt */
+		rcar_pci_write_reg(pcie->base, 1 << index, PCIEMSIFR);
+
+		msi_irq = irq_find_mapping(msi->domain, index);
+		if (msi_irq) {
+			if (test_bit(index, msi->used))
+				generic_handle_irq(msi_irq);
+			else
+				dev_info(dev, "unhandled MSI\n");
+		} else {
+			/* Unknown MSI, just clear it */
+			dev_dbg(dev, "unexpected MSI\n");
+		}
+
+		/* see if there's any more pending in this vector */
+		reg = rcar_pci_read_reg(pcie->base, PCIEMSIFR);
+	}
+
+	return IRQ_HANDLED;
+}
+
+static int rcar_msi_setup_irq(struct msi_controller *chip, struct pci_dev *pdev,
+			      struct msi_desc *desc)
+{
+	struct rcar_msi *msi = to_rcar_msi(chip);
+	struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
+	struct msi_msg msg;
+	unsigned int irq;
+	int hwirq;
+
+	hwirq = rcar_msi_alloc(msi);
+	if (hwirq < 0)
+		return hwirq;
+
+	irq = irq_find_mapping(msi->domain, hwirq);
+	if (!irq) {
+		rcar_msi_free(msi, hwirq);
+		return -EINVAL;
+	}
+
+	irq_set_msi_desc(irq, desc);
+
+	msg.address_lo = rcar_pci_read_reg(pcie->base, PCIEMSIALR) & ~MSIFE;
+	msg.address_hi = rcar_pci_read_reg(pcie->base, PCIEMSIAUR);
+	msg.data = hwirq;
+
+	pci_write_msi_msg(irq, &msg);
+
+	return 0;
+}
+
+static int rcar_msi_setup_irqs(struct msi_controller *chip,
+			       struct pci_dev *pdev, int nvec, int type)
+{
+	struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
+	struct rcar_msi *msi = to_rcar_msi(chip);
+	struct msi_desc *desc;
+	struct msi_msg msg;
+	unsigned int irq;
+	int hwirq;
+	int i;
+
+	/* MSI-X interrupts are not supported */
+	if (type == PCI_CAP_ID_MSIX)
+		return -EINVAL;
+
+	WARN_ON(!list_is_singular(&pdev->dev.msi_list));
+	desc = list_entry(pdev->dev.msi_list.next, struct msi_desc, list);
+
+	hwirq = rcar_msi_alloc_region(msi, nvec);
+	if (hwirq < 0)
+		return -ENOSPC;
+
+	irq = irq_find_mapping(msi->domain, hwirq);
+	if (!irq)
+		return -ENOSPC;
+
+	for (i = 0; i < nvec; i++) {
+		/*
+		 * irq_create_mapping() called from rcar_pcie_probe() pre-
+		 * allocates descs,  so there is no need to allocate descs here.
+		 * We can therefore assume that if irq_find_mapping() above
+		 * returns non-zero, then the descs are also successfully
+		 * allocated.
+		 */
+		if (irq_set_msi_desc_off(irq, i, desc)) {
+			/* TODO: clear */
+			return -EINVAL;
+		}
+	}
+
+	desc->nvec_used = nvec;
+	desc->msi_attrib.multiple = order_base_2(nvec);
+
+	msg.address_lo = rcar_pci_read_reg(pcie->base, PCIEMSIALR) & ~MSIFE;
+	msg.address_hi = rcar_pci_read_reg(pcie->base, PCIEMSIAUR);
+	msg.data = hwirq;
+
+	pci_write_msi_msg(irq, &msg);
+
+	return 0;
+}
+
+static void rcar_msi_teardown_irq(struct msi_controller *chip, unsigned int irq)
+{
+	struct irq_data *d = irq_get_irq_data(irq);
+	struct rcar_msi *msi = to_rcar_msi(chip);
+
+	rcar_msi_free(msi, d->hwirq);
+}
+
+static struct irq_chip rcar_msi_irq_chip = {
+	.name = "R-Car PCIe MSI",
+	.irq_enable = pci_msi_unmask_irq,
+	.irq_disable = pci_msi_mask_irq,
+	.irq_mask = pci_msi_mask_irq,
+	.irq_unmask = pci_msi_unmask_irq,
+};
+
+static int rcar_msi_map(struct irq_domain *domain, unsigned int irq,
+			irq_hw_number_t hwirq)
+{
+	irq_set_chip_and_handler(irq, &rcar_msi_irq_chip, handle_simple_irq);
+	irq_set_chip_data(irq, domain->host_data);
+
+	return 0;
+}
+
+static const struct irq_domain_ops msi_domain_ops = {
+	.map = rcar_msi_map,
+};
+
+static void rcar_pcie_unmap_msi(struct rcar_pcie *pcie)
+{
+	struct rcar_msi *msi = &pcie->msi;
+	int i, irq;
+
+	for (i = 0; i < INT_PCI_MSI_NR; i++) {
+		irq = irq_find_mapping(msi->domain, i);
+		if (irq > 0)
+			irq_dispose_mapping(irq);
+	}
+
+	irq_domain_remove(msi->domain);
+}
+
+static int rcar_pcie_enable_msi(struct rcar_pcie *pcie)
+{
+	struct rcar_msi *msi = &pcie->msi;
+	struct device *dev = pcie->dev;
+	phys_addr_t base;
+	int err, i;
+
+	mutex_init(&msi->lock);
+
+	msi->chip.dev = dev;
+	msi->chip.setup_irq = rcar_msi_setup_irq;
+	msi->chip.setup_irqs = rcar_msi_setup_irqs;
+	msi->chip.teardown_irq = rcar_msi_teardown_irq;
+
+	msi->domain = irq_domain_add_linear(dev->of_node, INT_PCI_MSI_NR,
+					    &msi_domain_ops, &msi->chip);
+	if (!msi->domain) {
+		dev_err(dev, "failed to create IRQ domain\n");
+		return -ENOMEM;
+	}
+
+	for (i = 0; i < INT_PCI_MSI_NR; i++)
+		irq_create_mapping(msi->domain, i);
+
+	/* Two irqs are for MSI, but they are also used for non-MSI irqs */
+	err = devm_request_irq(dev, msi->irq1, rcar_pcie_msi_irq,
+			       IRQF_SHARED | IRQF_NO_THREAD,
+			       rcar_msi_irq_chip.name, pcie);
+	if (err < 0) {
+		dev_err(dev, "failed to request IRQ: %d\n", err);
+		goto err;
+	}
+
+	err = devm_request_irq(dev, msi->irq2, rcar_pcie_msi_irq,
+			       IRQF_SHARED | IRQF_NO_THREAD,
+			       rcar_msi_irq_chip.name, pcie);
+	if (err < 0) {
+		dev_err(dev, "failed to request IRQ: %d\n", err);
+		goto err;
+	}
+
+	/* setup MSI data target */
+	msi->pages = __get_free_pages(GFP_KERNEL, 0);
+	if (!msi->pages) {
+		err = -ENOMEM;
+		goto err;
+	}
+	base = virt_to_phys((void *)msi->pages);
+
+	rcar_pci_write_reg(pcie->base, lower_32_bits(base) | MSIFE, PCIEMSIALR);
+	rcar_pci_write_reg(pcie->base, upper_32_bits(base), PCIEMSIAUR);
+
+	/* enable all MSI interrupts */
+	rcar_pci_write_reg(pcie->base, 0xffffffff, PCIEMSIIER);
+
+	return 0;
+
+err:
+	rcar_pcie_unmap_msi(pcie);
+	return err;
+}
+
+static void rcar_pcie_teardown_msi(struct rcar_pcie *pcie)
+{
+	struct rcar_msi *msi = &pcie->msi;
+
+	/* Disable all MSI interrupts */
+	rcar_pci_write_reg(pcie->base, 0, PCIEMSIIER);
+
+	/* Disable address decoding of the MSI interrupt, MSIFE */
+	rcar_pci_write_reg(pcie->base, 0, PCIEMSIALR);
+
+	free_pages(msi->pages, 0);
+
+	rcar_pcie_unmap_msi(pcie);
+}
+
+static int rcar_pcie_get_resources(struct rcar_pcie *pcie)
+{
+	struct device *dev = pcie->dev;
+	struct resource res;
+	int err, i;
+
+	pcie->phy = devm_phy_optional_get(dev, "pcie");
+	if (IS_ERR(pcie->phy))
+		return PTR_ERR(pcie->phy);
+
+	err = of_address_to_resource(dev->of_node, 0, &res);
+	if (err)
+		return err;
+
+	pcie->base = devm_ioremap_resource(dev, &res);
+	if (IS_ERR(pcie->base))
+		return PTR_ERR(pcie->base);
+
+	pcie->bus_clk = devm_clk_get(dev, "pcie_bus");
+	if (IS_ERR(pcie->bus_clk)) {
+		dev_err(dev, "cannot get pcie bus clock\n");
+		return PTR_ERR(pcie->bus_clk);
+	}
+
+	i = irq_of_parse_and_map(dev->of_node, 0);
+	if (!i) {
+		dev_err(dev, "cannot get platform resources for msi interrupt\n");
+		err = -ENOENT;
+		goto err_irq1;
+	}
+	pcie->msi.irq1 = i;
+
+	i = irq_of_parse_and_map(dev->of_node, 1);
+	if (!i) {
+		dev_err(dev, "cannot get platform resources for msi interrupt\n");
+		err = -ENOENT;
+		goto err_irq2;
+	}
+	pcie->msi.irq2 = i;
+
+	return 0;
+
+err_irq2:
+	irq_dispose_mapping(pcie->msi.irq1);
+err_irq1:
+	return err;
+}
+
+static int rcar_pcie_inbound_ranges(struct rcar_pcie *pcie,
+				    struct of_pci_range *range,
+				    int *index)
+{
+	u64 restype = range->flags;
+	u64 cpu_addr = range->cpu_addr;
+	u64 cpu_end = range->cpu_addr + range->size;
+	u64 pci_addr = range->pci_addr;
+	u32 flags = LAM_64BIT | LAR_ENABLE;
+	u64 mask;
+	u64 size;
+	int idx = *index;
+
+	if (restype & IORESOURCE_PREFETCH)
+		flags |= LAM_PREFETCH;
+
+	/*
+	 * If the size of the range is larger than the alignment of the start
+	 * address, we have to use multiple entries to perform the mapping.
+	 */
+	if (cpu_addr > 0) {
+		unsigned long nr_zeros = __ffs64(cpu_addr);
+		u64 alignment = 1ULL << nr_zeros;
+
+		size = min(range->size, alignment);
+	} else {
+		size = range->size;
+	}
+	/* Hardware supports max 4GiB inbound region */
+	size = min(size, 1ULL << 32);
+
+	mask = roundup_pow_of_two(size) - 1;
+	mask &= ~0xf;
+
+	while (cpu_addr < cpu_end) {
+		rcar_pcie_set_inbound(pcie->base, cpu_addr, pci_addr,
+				      mask | flags, idx, true);
+		pci_addr += size;
+		cpu_addr += size;
+		idx += 2;
+
+		if (idx > MAX_NR_INBOUND_MAPS) {
+			dev_err(pcie->dev, "Failed to map inbound regions!\n");
+			return -EINVAL;
+		}
+	}
+	*index = idx;
+
+	return 0;
+}
+
+static int rcar_pcie_parse_map_dma_ranges(struct rcar_pcie *pcie,
+					  struct device_node *np)
+{
+	struct of_pci_range_parser parser;
+	struct of_pci_range range;
+	int index = 0;
+	int err;
+
+	if (of_pci_dma_range_parser_init(&parser, np))
+		return -EINVAL;
+
+	/* Get the dma-ranges from DT */
+	for_each_of_pci_range(&parser, &range) {
+		u64 end = range.cpu_addr + range.size - 1;
+
+		dev_dbg(pcie->dev, "0x%08x 0x%016llx..0x%016llx -> 0x%016llx\n",
+			range.flags, range.cpu_addr, end, range.pci_addr);
+
+		err = rcar_pcie_inbound_ranges(pcie, &range, &index);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static const struct of_device_id rcar_pcie_of_match[] = {
+	{ .compatible = "renesas,pcie-r8a7779",
+	  .data = rcar_pcie_phy_init_h1 },
+	{ .compatible = "renesas,pcie-r8a7790",
+	  .data = rcar_pcie_phy_init_gen2 },
+	{ .compatible = "renesas,pcie-r8a7791",
+	  .data = rcar_pcie_phy_init_gen2 },
+	{ .compatible = "renesas,pcie-rcar-gen2",
+	  .data = rcar_pcie_phy_init_gen2 },
+	{ .compatible = "renesas,pcie-r8a7795",
+	  .data = rcar_pcie_phy_init_gen3 },
+	{ .compatible = "renesas,pcie-rcar-gen3",
+	  .data = rcar_pcie_phy_init_gen3 },
+	{},
+};
+
+static int rcar_pcie_probe(struct platform_device *pdev)
+{
+	int (*phy_init_fn)(struct rcar_pcie *pcie);
+	struct device *dev = &pdev->dev;
+	struct pci_host_bridge *bridge;
+	struct rcar_pcie *pcie;
+	u32 data;
+	int err;
+
+	bridge = pci_alloc_host_bridge(sizeof(*pcie));
+	if (!bridge)
+		return -ENOMEM;
+
+	pcie = pci_host_bridge_priv(bridge);
+
+	pcie->dev = dev;
+	platform_set_drvdata(pdev, pcie);
+
+	err = pci_parse_request_of_pci_ranges(dev, &pcie->resources, NULL);
+	if (err)
+		goto err_free_bridge;
+
+	pm_runtime_enable(pcie->dev);
+	err = pm_runtime_get_sync(pcie->dev);
+	if (err < 0) {
+		dev_err(pcie->dev, "pm_runtime_get_sync failed\n");
+		goto err_pm_disable;
+	}
+
+	err = rcar_pcie_get_resources(pcie);
+	if (err < 0) {
+		dev_err(dev, "failed to request resources: %d\n", err);
+		goto err_pm_put;
+	}
+
+	err = clk_prepare_enable(pcie->bus_clk);
+	if (err) {
+		dev_err(dev, "failed to enable bus clock: %d\n", err);
+		goto err_unmap_msi_irqs;
+	}
+
+	err = rcar_pcie_parse_map_dma_ranges(pcie, dev->of_node);
+	if (err)
+		goto err_clk_disable;
+
+	phy_init_fn = of_device_get_match_data(dev);
+	err = phy_init_fn(pcie);
+	if (err) {
+		dev_err(dev, "failed to init PCIe PHY\n");
+		goto err_clk_disable;
+	}
+
+	/* Failure to get a link might just be that no cards are inserted */
+	if (rcar_pcie_hw_init(pcie)) {
+		dev_info(dev, "PCIe link down\n");
+		err = -ENODEV;
+		goto err_phy_shutdown;
+	}
+
+	data = rcar_pci_read_reg(pcie->base, MACSR);
+	dev_info(dev, "PCIe x%d: link up\n", (data >> 20) & 0x3f);
+
+	if (IS_ENABLED(CONFIG_PCI_MSI)) {
+		err = rcar_pcie_enable_msi(pcie);
+		if (err < 0) {
+			dev_err(dev,
+				"failed to enable MSI support: %d\n",
+				err);
+			goto err_phy_shutdown;
+		}
+	}
+
+	err = rcar_pcie_enable(pcie);
+	if (err)
+		goto err_msi_teardown;
+
+	return 0;
+
+err_msi_teardown:
+	if (IS_ENABLED(CONFIG_PCI_MSI))
+		rcar_pcie_teardown_msi(pcie);
+
+err_phy_shutdown:
+	if (pcie->phy) {
+		phy_power_off(pcie->phy);
+		phy_exit(pcie->phy);
+	}
+
+err_clk_disable:
+	clk_disable_unprepare(pcie->bus_clk);
+
+err_unmap_msi_irqs:
+	irq_dispose_mapping(pcie->msi.irq2);
+	irq_dispose_mapping(pcie->msi.irq1);
+
+err_pm_put:
+	pm_runtime_put(dev);
+
+err_pm_disable:
+	pm_runtime_disable(dev);
+	pci_free_resource_list(&pcie->resources);
+
+err_free_bridge:
+	pci_free_host_bridge(bridge);
+
+	return err;
+}
+
+static int rcar_pcie_resume_noirq(struct device *dev)
+{
+	struct rcar_pcie *pcie = dev_get_drvdata(dev);
+
+	if (rcar_pci_read_reg(pcie->base, PMSR) &&
+	    !(rcar_pci_read_reg(pcie->base, PCIETCTLR) & DL_DOWN))
+		return 0;
+
+	/* Re-establish the PCIe link */
+	rcar_rmw32(pcie->base, MACCTLR, MACCTLR_RESERVED, 0);
+	rcar_pci_write_reg(pcie->base, CFINIT, PCIETCTLR);
+	return rcar_pcie_wait_for_dl(pcie->base);
+}
+
+static const struct dev_pm_ops rcar_pcie_pm_ops = {
+	.resume_noirq = rcar_pcie_resume_noirq,
+};
+
+static struct platform_driver rcar_pcie_driver = {
+	.driver = {
+		.name = "rcar-pcie",
+		.of_match_table = rcar_pcie_of_match,
+		.pm = &rcar_pcie_pm_ops,
+		.suppress_bind_attrs = true,
+	},
+	.probe = rcar_pcie_probe,
+};
+builtin_platform_driver(rcar_pcie_driver);
diff --git a/drivers/pci/controller/pcie-rcar.c b/drivers/pci/controller/pcie-rcar.c
index e45bb2a7bfa5..1008ae6a896c 100644
--- a/drivers/pci/controller/pcie-rcar.c
+++ b/drivers/pci/controller/pcie-rcar.c
@@ -1,7 +1,7 @@
 // SPDX-License-Identifier: GPL-2.0
 /*
  * PCIe driver for Renesas R-Car SoCs
- *  Copyright (C) 2014 Renesas Electronics Europe Ltd
+ *  Copyright (C) 2014-2019 Renesas Electronics Europe Ltd
  *
  * Based on:
  *  arch/sh/drivers/pci/pcie-sh7786.c
@@ -11,534 +11,80 @@
  * Author: Phil Edworthy <phil.edworthy@renesas.com>
  */
 
-#include <linux/bitops.h>
-#include <linux/clk.h>
 #include <linux/delay.h>
-#include <linux/interrupt.h>
-#include <linux/irq.h>
-#include <linux/irqdomain.h>
-#include <linux/kernel.h>
-#include <linux/init.h>
-#include <linux/msi.h>
-#include <linux/of_address.h>
-#include <linux/of_irq.h>
-#include <linux/of_pci.h>
-#include <linux/of_platform.h>
 #include <linux/pci.h>
-#include <linux/phy/phy.h>
-#include <linux/platform_device.h>
-#include <linux/pm_runtime.h>
-#include <linux/slab.h>
 
-#define PCIECAR			0x000010
-#define PCIECCTLR		0x000018
-#define  CONFIG_SEND_ENABLE	BIT(31)
-#define  TYPE0			(0 << 8)
-#define  TYPE1			BIT(8)
-#define PCIECDR			0x000020
-#define PCIEMSR			0x000028
-#define PCIEINTXR		0x000400
-#define PCIEPHYSR		0x0007f0
-#define  PHYRDY			BIT(0)
-#define PCIEMSITXR		0x000840
+#include "pcie-rcar.h"
 
-/* Transfer control */
-#define PCIETCTLR		0x02000
-#define  DL_DOWN		BIT(3)
-#define  CFINIT			BIT(0)
-#define PCIETSTR		0x02004
-#define  DATA_LINK_ACTIVE	BIT(0)
-#define PCIEERRFR		0x02020
-#define  UNSUPPORTED_REQUEST	BIT(4)
-#define PCIEMSIFR		0x02044
-#define PCIEMSIALR		0x02048
-#define  MSIFE			BIT(0)
-#define PCIEMSIAUR		0x0204c
-#define PCIEMSIIER		0x02050
-
-/* root port address */
-#define PCIEPRAR(x)		(0x02080 + ((x) * 0x4))
-
-/* local address reg & mask */
-#define PCIELAR(x)		(0x02200 + ((x) * 0x20))
-#define PCIELAMR(x)		(0x02208 + ((x) * 0x20))
-#define  LAM_PREFETCH		BIT(3)
-#define  LAM_64BIT		BIT(2)
-#define  LAR_ENABLE		BIT(1)
-
-/* PCIe address reg & mask */
-#define PCIEPALR(x)		(0x03400 + ((x) * 0x20))
-#define PCIEPAUR(x)		(0x03404 + ((x) * 0x20))
-#define PCIEPAMR(x)		(0x03408 + ((x) * 0x20))
-#define PCIEPTCTLR(x)		(0x0340c + ((x) * 0x20))
-#define  PAR_ENABLE		BIT(31)
-#define  IO_SPACE		BIT(8)
-
-/* Configuration */
-#define PCICONF(x)		(0x010000 + ((x) * 0x4))
-#define PMCAP(x)		(0x010040 + ((x) * 0x4))
-#define EXPCAP(x)		(0x010070 + ((x) * 0x4))
-#define VCCAP(x)		(0x010100 + ((x) * 0x4))
-
-/* link layer */
-#define IDSETR1			0x011004
-#define TLCTLR			0x011048
-#define MACSR			0x011054
-#define  SPCHGFIN		BIT(4)
-#define  SPCHGFAIL		BIT(6)
-#define  SPCHGSUC		BIT(7)
-#define  LINK_SPEED		(0xf << 16)
-#define  LINK_SPEED_2_5GTS	(1 << 16)
-#define  LINK_SPEED_5_0GTS	(2 << 16)
-#define MACCTLR			0x011058
-#define  MACCTLR_RESERVED	BIT(0)
-#define  SPEED_CHANGE		BIT(24)
-#define  SCRAMBLE_DISABLE	BIT(27)
-#define PMSR			0x01105c
-#define MACS2R			0x011078
-#define MACCGSPSETR		0x011084
-#define  SPCNGRSN		BIT(31)
-
-/* R-Car H1 PHY */
-#define H1_PCIEPHYADRR		0x04000c
-#define  WRITE_CMD		BIT(16)
-#define  PHY_ACK		BIT(24)
-#define  RATE_POS		12
-#define  LANE_POS		8
-#define  ADR_POS		0
-#define H1_PCIEPHYDOUTR		0x040014
-
-/* R-Car Gen2 PHY */
-#define GEN2_PCIEPHYADDR	0x780
-#define GEN2_PCIEPHYDATA	0x784
-#define GEN2_PCIEPHYCTRL	0x78c
-
-#define INT_PCI_MSI_NR		32
-
-#define RCONF(x)		(PCICONF(0) + (x))
-#define RPMCAP(x)		(PMCAP(0) + (x))
-#define REXPCAP(x)		(EXPCAP(0) + (x))
-#define RVCCAP(x)		(VCCAP(0) + (x))
-
-#define PCIE_CONF_BUS(b)	(((b) & 0xff) << 24)
-#define PCIE_CONF_DEV(d)	(((d) & 0x1f) << 19)
-#define PCIE_CONF_FUNC(f)	(((f) & 0x7) << 16)
-
-#define RCAR_PCI_MAX_RESOURCES	4
-#define MAX_NR_INBOUND_MAPS	6
-
-struct rcar_msi {
-	DECLARE_BITMAP(used, INT_PCI_MSI_NR);
-	struct irq_domain *domain;
-	struct msi_controller chip;
-	unsigned long pages;
-	struct mutex lock;
-	int irq1;
-	int irq2;
-};
-
-static inline struct rcar_msi *to_rcar_msi(struct msi_controller *chip)
-{
-	return container_of(chip, struct rcar_msi, chip);
-}
-
-/* Structure representing the PCIe interface */
-struct rcar_pcie {
-	struct device		*dev;
-	struct phy		*phy;
-	void __iomem		*base;
-	struct list_head	resources;
-	int			root_bus_nr;
-	struct clk		*bus_clk;
-	struct			rcar_msi msi;
-};
-
-static void rcar_pci_write_reg(struct rcar_pcie *pcie, u32 val,
-			       unsigned int reg)
+void rcar_pci_write_reg(void __iomem *base, u32 val, unsigned int reg)
 {
-	writel(val, pcie->base + reg);
+	writel(val, base + reg);
 }
 
-static u32 rcar_pci_read_reg(struct rcar_pcie *pcie, unsigned int reg)
+u32 rcar_pci_read_reg(void __iomem *base, unsigned int reg)
 {
-	return readl(pcie->base + reg);
+	return readl(base + reg);
 }
 
-enum {
-	RCAR_PCI_ACCESS_READ,
-	RCAR_PCI_ACCESS_WRITE,
-};
-
-static void rcar_rmw32(struct rcar_pcie *pcie, int where, u32 mask, u32 data)
+void rcar_rmw32(void __iomem *base, int where, u32 mask, u32 data)
 {
 	unsigned int shift = BITS_PER_BYTE * (where & 3);
-	u32 val = rcar_pci_read_reg(pcie, where & ~3);
+	u32 val = rcar_pci_read_reg(base, where & ~3);
 
 	val &= ~(mask << shift);
 	val |= data << shift;
-	rcar_pci_write_reg(pcie, val, where & ~3);
-}
-
-static u32 rcar_read_conf(struct rcar_pcie *pcie, int where)
-{
-	unsigned int shift = BITS_PER_BYTE * (where & 3);
-	u32 val = rcar_pci_read_reg(pcie, where & ~3);
-
-	return val >> shift;
+	rcar_pci_write_reg(base, val, where & ~3);
 }
 
-/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
-static int rcar_pcie_config_access(struct rcar_pcie *pcie,
-		unsigned char access_type, struct pci_bus *bus,
-		unsigned int devfn, int where, u32 *data)
-{
-	unsigned int dev, func, reg, index;
-
-	dev = PCI_SLOT(devfn);
-	func = PCI_FUNC(devfn);
-	reg = where & ~3;
-	index = reg / 4;
-
-	/*
-	 * While each channel has its own memory-mapped extended config
-	 * space, it's generally only accessible when in endpoint mode.
-	 * When in root complex mode, the controller is unable to target
-	 * itself with either type 0 or type 1 accesses, and indeed, any
-	 * controller initiated target transfer to its own config space
-	 * result in a completer abort.
-	 *
-	 * Each channel effectively only supports a single device, but as
-	 * the same channel <-> device access works for any PCI_SLOT()
-	 * value, we cheat a bit here and bind the controller's config
-	 * space to devfn 0 in order to enable self-enumeration. In this
-	 * case the regular ECAR/ECDR path is sidelined and the mangled
-	 * config access itself is initiated as an internal bus transaction.
-	 */
-	if (pci_is_root_bus(bus)) {
-		if (dev != 0)
-			return PCIBIOS_DEVICE_NOT_FOUND;
-
-		if (access_type == RCAR_PCI_ACCESS_READ) {
-			*data = rcar_pci_read_reg(pcie, PCICONF(index));
-		} else {
-			/* Keep an eye out for changes to the root bus number */
-			if (pci_is_root_bus(bus) && (reg == PCI_PRIMARY_BUS))
-				pcie->root_bus_nr = *data & 0xff;
-
-			rcar_pci_write_reg(pcie, *data, PCICONF(index));
-		}
-
-		return PCIBIOS_SUCCESSFUL;
-	}
-
-	if (pcie->root_bus_nr < 0)
-		return PCIBIOS_DEVICE_NOT_FOUND;
-
-	/* Clear errors */
-	rcar_pci_write_reg(pcie, rcar_pci_read_reg(pcie, PCIEERRFR), PCIEERRFR);
-
-	/* Set the PIO address */
-	rcar_pci_write_reg(pcie, PCIE_CONF_BUS(bus->number) |
-		PCIE_CONF_DEV(dev) | PCIE_CONF_FUNC(func) | reg, PCIECAR);
-
-	/* Enable the configuration access */
-	if (bus->parent->number == pcie->root_bus_nr)
-		rcar_pci_write_reg(pcie, CONFIG_SEND_ENABLE | TYPE0, PCIECCTLR);
-	else
-		rcar_pci_write_reg(pcie, CONFIG_SEND_ENABLE | TYPE1, PCIECCTLR);
-
-	/* Check for errors */
-	if (rcar_pci_read_reg(pcie, PCIEERRFR) & UNSUPPORTED_REQUEST)
-		return PCIBIOS_DEVICE_NOT_FOUND;
-
-	/* Check for master and target aborts */
-	if (rcar_read_conf(pcie, RCONF(PCI_STATUS)) &
-		(PCI_STATUS_REC_MASTER_ABORT | PCI_STATUS_REC_TARGET_ABORT))
-		return PCIBIOS_DEVICE_NOT_FOUND;
-
-	if (access_type == RCAR_PCI_ACCESS_READ)
-		*data = rcar_pci_read_reg(pcie, PCIECDR);
-	else
-		rcar_pci_write_reg(pcie, *data, PCIECDR);
-
-	/* Disable the configuration access */
-	rcar_pci_write_reg(pcie, 0, PCIECCTLR);
-
-	return PCIBIOS_SUCCESSFUL;
-}
-
-static int rcar_pcie_read_conf(struct pci_bus *bus, unsigned int devfn,
-			       int where, int size, u32 *val)
-{
-	struct rcar_pcie *pcie = bus->sysdata;
-	int ret;
-
-	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
-				      bus, devfn, where, val);
-	if (ret != PCIBIOS_SUCCESSFUL) {
-		*val = 0xffffffff;
-		return ret;
-	}
-
-	if (size == 1)
-		*val = (*val >> (BITS_PER_BYTE * (where & 3))) & 0xff;
-	else if (size == 2)
-		*val = (*val >> (BITS_PER_BYTE * (where & 2))) & 0xffff;
-
-	dev_dbg(&bus->dev, "pcie-config-read: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
-		bus->number, devfn, where, size, *val);
-
-	return ret;
-}
-
-/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
-static int rcar_pcie_write_conf(struct pci_bus *bus, unsigned int devfn,
-				int where, int size, u32 val)
-{
-	struct rcar_pcie *pcie = bus->sysdata;
-	unsigned int shift;
-	u32 data;
-	int ret;
-
-	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
-				      bus, devfn, where, &data);
-	if (ret != PCIBIOS_SUCCESSFUL)
-		return ret;
-
-	dev_dbg(&bus->dev, "pcie-config-write: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
-		bus->number, devfn, where, size, val);
-
-	if (size == 1) {
-		shift = BITS_PER_BYTE * (where & 3);
-		data &= ~(0xff << shift);
-		data |= ((val & 0xff) << shift);
-	} else if (size == 2) {
-		shift = BITS_PER_BYTE * (where & 2);
-		data &= ~(0xffff << shift);
-		data |= ((val & 0xffff) << shift);
-	} else
-		data = val;
-
-	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_WRITE,
-				      bus, devfn, where, &data);
-
-	return ret;
-}
-
-static struct pci_ops rcar_pcie_ops = {
-	.read	= rcar_pcie_read_conf,
-	.write	= rcar_pcie_write_conf,
-};
-
-static void rcar_pcie_setup_window(int win, struct rcar_pcie *pcie,
-				   struct resource *res)
+void rcar_pcie_set_outbound(int win, void __iomem *base,
+			    struct resource *res, bool host)
 {
 	/* Setup PCIe address space mappings for each resource */
-	resource_size_t size;
-	resource_size_t res_start;
-	u32 mask;
-
-	rcar_pci_write_reg(pcie, 0x00000000, PCIEPTCTLR(win));
+	resource_size_t size = 0;
+	resource_size_t res_start = 0;
+	u32 mask = 0x0;
 
+	rcar_pci_write_reg(base, mask, PCIEPTCTLR(win));
 	/*
 	 * The PAMR mask is calculated in units of 128Bytes, which
 	 * keeps things pretty simple.
 	 */
 	size = resource_size(res);
-	mask = (roundup_pow_of_two(size) / SZ_128) - 1;
-	rcar_pci_write_reg(pcie, mask << 7, PCIEPAMR(win));
+	if (size > 128)
+		mask = (roundup_pow_of_two(size) / SZ_128) - 1;
+	rcar_pci_write_reg(base, mask << 7, PCIEPAMR(win));
 
-	if (res->flags & IORESOURCE_IO)
-		res_start = pci_pio_to_address(res->start);
-	else
+	if (!host) {
 		res_start = res->start;
+	} else {
+		if (res->flags & IORESOURCE_IO)
+			res_start = pci_pio_to_address(res->start);
+		else
+			res_start = res->start;
+	}
 
-	rcar_pci_write_reg(pcie, upper_32_bits(res_start), PCIEPAUR(win));
-	rcar_pci_write_reg(pcie, lower_32_bits(res_start) & ~0x7F,
+	rcar_pci_write_reg(base, upper_32_bits(res_start), PCIEPAUR(win));
+	rcar_pci_write_reg(base, lower_32_bits(res_start) & ~0x7F,
 			   PCIEPALR(win));
 
-	/* First resource is for IO */
-	mask = PAR_ENABLE;
+	mask = 0x0;
+	if (res->start)
+		mask = PAR_ENABLE;
+
 	if (res->flags & IORESOURCE_IO)
 		mask |= IO_SPACE;
 
-	rcar_pci_write_reg(pcie, mask, PCIEPTCTLR(win));
-}
-
-static int rcar_pcie_setup(struct list_head *resource, struct rcar_pcie *pci)
-{
-	struct resource_entry *win;
-	int i = 0;
-
-	/* Setup PCI resources */
-	resource_list_for_each_entry(win, &pci->resources) {
-		struct resource *res = win->res;
-
-		if (!res->flags)
-			continue;
-
-		switch (resource_type(res)) {
-		case IORESOURCE_IO:
-		case IORESOURCE_MEM:
-			rcar_pcie_setup_window(i, pci, res);
-			i++;
-			break;
-		case IORESOURCE_BUS:
-			pci->root_bus_nr = res->start;
-			break;
-		default:
-			continue;
-		}
-
-		pci_add_resource(resource, res);
-	}
-
-	return 1;
-}
-
-static void rcar_pcie_force_speedup(struct rcar_pcie *pcie)
-{
-	struct device *dev = pcie->dev;
-	unsigned int timeout = 1000;
-	u32 macsr;
-
-	if ((rcar_pci_read_reg(pcie, MACS2R) & LINK_SPEED) != LINK_SPEED_5_0GTS)
-		return;
-
-	if (rcar_pci_read_reg(pcie, MACCTLR) & SPEED_CHANGE) {
-		dev_err(dev, "Speed change already in progress\n");
-		return;
-	}
-
-	macsr = rcar_pci_read_reg(pcie, MACSR);
-	if ((macsr & LINK_SPEED) == LINK_SPEED_5_0GTS)
-		goto done;
-
-	/* Set target link speed to 5.0 GT/s */
-	rcar_rmw32(pcie, EXPCAP(12), PCI_EXP_LNKSTA_CLS,
-		   PCI_EXP_LNKSTA_CLS_5_0GB);
-
-	/* Set speed change reason as intentional factor */
-	rcar_rmw32(pcie, MACCGSPSETR, SPCNGRSN, 0);
-
-	/* Clear SPCHGFIN, SPCHGSUC, and SPCHGFAIL */
-	if (macsr & (SPCHGFIN | SPCHGSUC | SPCHGFAIL))
-		rcar_pci_write_reg(pcie, macsr, MACSR);
-
-	/* Start link speed change */
-	rcar_rmw32(pcie, MACCTLR, SPEED_CHANGE, SPEED_CHANGE);
-
-	while (timeout--) {
-		macsr = rcar_pci_read_reg(pcie, MACSR);
-		if (macsr & SPCHGFIN) {
-			/* Clear the interrupt bits */
-			rcar_pci_write_reg(pcie, macsr, MACSR);
-
-			if (macsr & SPCHGFAIL)
-				dev_err(dev, "Speed change failed\n");
-
-			goto done;
-		}
-
-		msleep(1);
-	}
-
-	dev_err(dev, "Speed change timed out\n");
-
-done:
-	dev_info(dev, "Current link speed is %s GT/s\n",
-		 (macsr & LINK_SPEED) == LINK_SPEED_5_0GTS ? "5" : "2.5");
-}
-
-static int rcar_pcie_enable(struct rcar_pcie *pcie)
-{
-	struct device *dev = pcie->dev;
-	struct pci_host_bridge *bridge = pci_host_bridge_from_priv(pcie);
-	struct pci_bus *bus, *child;
-	int ret;
-
-	/* Try setting 5 GT/s link speed */
-	rcar_pcie_force_speedup(pcie);
-
-	rcar_pcie_setup(&bridge->windows, pcie);
-
-	pci_add_flags(PCI_REASSIGN_ALL_BUS);
-
-	bridge->dev.parent = dev;
-	bridge->sysdata = pcie;
-	bridge->busnr = pcie->root_bus_nr;
-	bridge->ops = &rcar_pcie_ops;
-	bridge->map_irq = of_irq_parse_and_map_pci;
-	bridge->swizzle_irq = pci_common_swizzle;
-	if (IS_ENABLED(CONFIG_PCI_MSI))
-		bridge->msi = &pcie->msi.chip;
-
-	ret = pci_scan_root_bus_bridge(bridge);
-	if (ret < 0)
-		return ret;
-
-	bus = bridge->bus;
-
-	pci_bus_size_bridges(bus);
-	pci_bus_assign_resources(bus);
-
-	list_for_each_entry(child, &bus->children, node)
-		pcie_bus_configure_settings(child);
-
-	pci_bus_add_devices(bus);
-
-	return 0;
-}
-
-static int phy_wait_for_ack(struct rcar_pcie *pcie)
-{
-	struct device *dev = pcie->dev;
-	unsigned int timeout = 100;
-
-	while (timeout--) {
-		if (rcar_pci_read_reg(pcie, H1_PCIEPHYADRR) & PHY_ACK)
-			return 0;
-
-		udelay(100);
-	}
-
-	dev_err(dev, "Access to PCIe phy timed out\n");
-
-	return -ETIMEDOUT;
-}
-
-static void phy_write_reg(struct rcar_pcie *pcie,
-			  unsigned int rate, u32 addr,
-			  unsigned int lane, u32 data)
-{
-	u32 phyaddr;
-
-	phyaddr = WRITE_CMD |
-		((rate & 1) << RATE_POS) |
-		((lane & 0xf) << LANE_POS) |
-		((addr & 0xff) << ADR_POS);
-
-	/* Set write data */
-	rcar_pci_write_reg(pcie, data, H1_PCIEPHYDOUTR);
-	rcar_pci_write_reg(pcie, phyaddr, H1_PCIEPHYADRR);
-
-	/* Ignore errors as they will be dealt with if the data link is down */
-	phy_wait_for_ack(pcie);
-
-	/* Clear command */
-	rcar_pci_write_reg(pcie, 0, H1_PCIEPHYDOUTR);
-	rcar_pci_write_reg(pcie, 0, H1_PCIEPHYADRR);
-
-	/* Ignore errors as they will be dealt with if the data link is down */
-	phy_wait_for_ack(pcie);
+	rcar_pci_write_reg(base, mask, PCIEPTCTLR(win));
+	/* flush modifications */
+	wmb();
 }
 
-static int rcar_pcie_wait_for_phyrdy(struct rcar_pcie *pcie)
+int rcar_pcie_wait_for_phyrdy(void __iomem *base)
 {
 	unsigned int timeout = 10;
 
 	while (timeout--) {
-		if (rcar_pci_read_reg(pcie, PCIEPHYSR) & PHYRDY)
+		if (rcar_pci_read_reg(base, PCIEPHYSR) & PHYRDY)
 			return 0;
 
 		msleep(5);
@@ -547,12 +93,12 @@ static int rcar_pcie_wait_for_phyrdy(struct rcar_pcie *pcie)
 	return -ETIMEDOUT;
 }
 
-static int rcar_pcie_wait_for_dl(struct rcar_pcie *pcie)
+int rcar_pcie_wait_for_dl(void __iomem *base)
 {
 	unsigned int timeout = 10000;
 
 	while (timeout--) {
-		if ((rcar_pci_read_reg(pcie, PCIETSTR) & DATA_LINK_ACTIVE))
+		if ((rcar_pci_read_reg(base, PCIETSTR) & DATA_LINK_ACTIVE))
 			return 0;
 
 		udelay(5);
@@ -562,698 +108,27 @@ static int rcar_pcie_wait_for_dl(struct rcar_pcie *pcie)
 	return -ETIMEDOUT;
 }
 
-static int rcar_pcie_hw_init(struct rcar_pcie *pcie)
-{
-	int err;
-
-	/* Begin initialization */
-	rcar_pci_write_reg(pcie, 0, PCIETCTLR);
-
-	/* Set mode */
-	rcar_pci_write_reg(pcie, 1, PCIEMSR);
-
-	err = rcar_pcie_wait_for_phyrdy(pcie);
-	if (err)
-		return err;
-
-	/*
-	 * Initial header for port config space is type 1, set the device
-	 * class to match. Hardware takes care of propagating the IDSETR
-	 * settings, so there is no need to bother with a quirk.
-	 */
-	rcar_pci_write_reg(pcie, PCI_CLASS_BRIDGE_PCI << 16, IDSETR1);
-
-	/*
-	 * Setup Secondary Bus Number & Subordinate Bus Number, even though
-	 * they aren't used, to avoid bridge being detected as broken.
-	 */
-	rcar_rmw32(pcie, RCONF(PCI_SECONDARY_BUS), 0xff, 1);
-	rcar_rmw32(pcie, RCONF(PCI_SUBORDINATE_BUS), 0xff, 1);
-
-	/* Initialize default capabilities. */
-	rcar_rmw32(pcie, REXPCAP(0), 0xff, PCI_CAP_ID_EXP);
-	rcar_rmw32(pcie, REXPCAP(PCI_EXP_FLAGS),
-		PCI_EXP_FLAGS_TYPE, PCI_EXP_TYPE_ROOT_PORT << 4);
-	rcar_rmw32(pcie, RCONF(PCI_HEADER_TYPE), 0x7f,
-		PCI_HEADER_TYPE_BRIDGE);
-
-	/* Enable data link layer active state reporting */
-	rcar_rmw32(pcie, REXPCAP(PCI_EXP_LNKCAP), PCI_EXP_LNKCAP_DLLLARC,
-		PCI_EXP_LNKCAP_DLLLARC);
-
-	/* Write out the physical slot number = 0 */
-	rcar_rmw32(pcie, REXPCAP(PCI_EXP_SLTCAP), PCI_EXP_SLTCAP_PSN, 0);
-
-	/* Set the completion timer timeout to the maximum 50ms. */
-	rcar_rmw32(pcie, TLCTLR + 1, 0x3f, 50);
-
-	/* Terminate list of capabilities (Next Capability Offset=0) */
-	rcar_rmw32(pcie, RVCCAP(0), 0xfff00000, 0);
-
-	/* Enable MSI */
-	if (IS_ENABLED(CONFIG_PCI_MSI))
-		rcar_pci_write_reg(pcie, 0x801f0000, PCIEMSITXR);
-
-	rcar_rmw32(pcie, MACCTLR, MACCTLR_RESERVED, 0);
-
-	/* Finish initialization - establish a PCI Express link */
-	rcar_pci_write_reg(pcie, CFINIT, PCIETCTLR);
-
-	/* This will timeout if we don't have a link. */
-	err = rcar_pcie_wait_for_dl(pcie);
-	if (err)
-		return err;
-
-	/* Enable INTx interrupts */
-	rcar_rmw32(pcie, PCIEINTXR, 0, 0xF << 8);
-
-	wmb();
-
-	return 0;
-}
-
-static int rcar_pcie_phy_init_h1(struct rcar_pcie *pcie)
-{
-	/* Initialize the phy */
-	phy_write_reg(pcie, 0, 0x42, 0x1, 0x0EC34191);
-	phy_write_reg(pcie, 1, 0x42, 0x1, 0x0EC34180);
-	phy_write_reg(pcie, 0, 0x43, 0x1, 0x00210188);
-	phy_write_reg(pcie, 1, 0x43, 0x1, 0x00210188);
-	phy_write_reg(pcie, 0, 0x44, 0x1, 0x015C0014);
-	phy_write_reg(pcie, 1, 0x44, 0x1, 0x015C0014);
-	phy_write_reg(pcie, 1, 0x4C, 0x1, 0x786174A0);
-	phy_write_reg(pcie, 1, 0x4D, 0x1, 0x048000BB);
-	phy_write_reg(pcie, 0, 0x51, 0x1, 0x079EC062);
-	phy_write_reg(pcie, 0, 0x52, 0x1, 0x20000000);
-	phy_write_reg(pcie, 1, 0x52, 0x1, 0x20000000);
-	phy_write_reg(pcie, 1, 0x56, 0x1, 0x00003806);
-
-	phy_write_reg(pcie, 0, 0x60, 0x1, 0x004B03A5);
-	phy_write_reg(pcie, 0, 0x64, 0x1, 0x3F0F1F0F);
-	phy_write_reg(pcie, 0, 0x66, 0x1, 0x00008000);
-
-	return 0;
-}
-
-static int rcar_pcie_phy_init_gen2(struct rcar_pcie *pcie)
-{
-	/*
-	 * These settings come from the R-Car Series, 2nd Generation User's
-	 * Manual, section 50.3.1 (2) Initialization of the physical layer.
-	 */
-	rcar_pci_write_reg(pcie, 0x000f0030, GEN2_PCIEPHYADDR);
-	rcar_pci_write_reg(pcie, 0x00381203, GEN2_PCIEPHYDATA);
-	rcar_pci_write_reg(pcie, 0x00000001, GEN2_PCIEPHYCTRL);
-	rcar_pci_write_reg(pcie, 0x00000006, GEN2_PCIEPHYCTRL);
-
-	rcar_pci_write_reg(pcie, 0x000f0054, GEN2_PCIEPHYADDR);
-	/* The following value is for DC connection, no termination resistor */
-	rcar_pci_write_reg(pcie, 0x13802007, GEN2_PCIEPHYDATA);
-	rcar_pci_write_reg(pcie, 0x00000001, GEN2_PCIEPHYCTRL);
-	rcar_pci_write_reg(pcie, 0x00000006, GEN2_PCIEPHYCTRL);
-
-	return 0;
-}
-
-static int rcar_pcie_phy_init_gen3(struct rcar_pcie *pcie)
-{
-	int err;
-
-	err = phy_init(pcie->phy);
-	if (err)
-		return err;
-
-	err = phy_power_on(pcie->phy);
-	if (err)
-		phy_exit(pcie->phy);
-
-	return err;
-}
-
-static int rcar_msi_alloc(struct rcar_msi *chip)
-{
-	int msi;
-
-	mutex_lock(&chip->lock);
-
-	msi = find_first_zero_bit(chip->used, INT_PCI_MSI_NR);
-	if (msi < INT_PCI_MSI_NR)
-		set_bit(msi, chip->used);
-	else
-		msi = -ENOSPC;
-
-	mutex_unlock(&chip->lock);
-
-	return msi;
-}
-
-static int rcar_msi_alloc_region(struct rcar_msi *chip, int no_irqs)
-{
-	int msi;
-
-	mutex_lock(&chip->lock);
-	msi = bitmap_find_free_region(chip->used, INT_PCI_MSI_NR,
-				      order_base_2(no_irqs));
-	mutex_unlock(&chip->lock);
-
-	return msi;
-}
-
-static void rcar_msi_free(struct rcar_msi *chip, unsigned long irq)
-{
-	mutex_lock(&chip->lock);
-	clear_bit(irq, chip->used);
-	mutex_unlock(&chip->lock);
-}
-
-static irqreturn_t rcar_pcie_msi_irq(int irq, void *data)
-{
-	struct rcar_pcie *pcie = data;
-	struct rcar_msi *msi = &pcie->msi;
-	struct device *dev = pcie->dev;
-	unsigned long reg;
-
-	reg = rcar_pci_read_reg(pcie, PCIEMSIFR);
-
-	/* MSI & INTx share an interrupt - we only handle MSI here */
-	if (!reg)
-		return IRQ_NONE;
-
-	while (reg) {
-		unsigned int index = find_first_bit(&reg, 32);
-		unsigned int msi_irq;
-
-		/* clear the interrupt */
-		rcar_pci_write_reg(pcie, 1 << index, PCIEMSIFR);
-
-		msi_irq = irq_find_mapping(msi->domain, index);
-		if (msi_irq) {
-			if (test_bit(index, msi->used))
-				generic_handle_irq(msi_irq);
-			else
-				dev_info(dev, "unhandled MSI\n");
-		} else {
-			/* Unknown MSI, just clear it */
-			dev_dbg(dev, "unexpected MSI\n");
-		}
-
-		/* see if there's any more pending in this vector */
-		reg = rcar_pci_read_reg(pcie, PCIEMSIFR);
-	}
-
-	return IRQ_HANDLED;
-}
-
-static int rcar_msi_setup_irq(struct msi_controller *chip, struct pci_dev *pdev,
-			      struct msi_desc *desc)
-{
-	struct rcar_msi *msi = to_rcar_msi(chip);
-	struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
-	struct msi_msg msg;
-	unsigned int irq;
-	int hwirq;
-
-	hwirq = rcar_msi_alloc(msi);
-	if (hwirq < 0)
-		return hwirq;
-
-	irq = irq_find_mapping(msi->domain, hwirq);
-	if (!irq) {
-		rcar_msi_free(msi, hwirq);
-		return -EINVAL;
-	}
-
-	irq_set_msi_desc(irq, desc);
-
-	msg.address_lo = rcar_pci_read_reg(pcie, PCIEMSIALR) & ~MSIFE;
-	msg.address_hi = rcar_pci_read_reg(pcie, PCIEMSIAUR);
-	msg.data = hwirq;
-
-	pci_write_msi_msg(irq, &msg);
-
-	return 0;
-}
-
-static int rcar_msi_setup_irqs(struct msi_controller *chip,
-			       struct pci_dev *pdev, int nvec, int type)
-{
-	struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
-	struct rcar_msi *msi = to_rcar_msi(chip);
-	struct msi_desc *desc;
-	struct msi_msg msg;
-	unsigned int irq;
-	int hwirq;
-	int i;
-
-	/* MSI-X interrupts are not supported */
-	if (type == PCI_CAP_ID_MSIX)
-		return -EINVAL;
-
-	WARN_ON(!list_is_singular(&pdev->dev.msi_list));
-	desc = list_entry(pdev->dev.msi_list.next, struct msi_desc, list);
-
-	hwirq = rcar_msi_alloc_region(msi, nvec);
-	if (hwirq < 0)
-		return -ENOSPC;
-
-	irq = irq_find_mapping(msi->domain, hwirq);
-	if (!irq)
-		return -ENOSPC;
-
-	for (i = 0; i < nvec; i++) {
-		/*
-		 * irq_create_mapping() called from rcar_pcie_probe() pre-
-		 * allocates descs,  so there is no need to allocate descs here.
-		 * We can therefore assume that if irq_find_mapping() above
-		 * returns non-zero, then the descs are also successfully
-		 * allocated.
-		 */
-		if (irq_set_msi_desc_off(irq, i, desc)) {
-			/* TODO: clear */
-			return -EINVAL;
-		}
-	}
-
-	desc->nvec_used = nvec;
-	desc->msi_attrib.multiple = order_base_2(nvec);
-
-	msg.address_lo = rcar_pci_read_reg(pcie, PCIEMSIALR) & ~MSIFE;
-	msg.address_hi = rcar_pci_read_reg(pcie, PCIEMSIAUR);
-	msg.data = hwirq;
-
-	pci_write_msi_msg(irq, &msg);
-
-	return 0;
-}
-
-static void rcar_msi_teardown_irq(struct msi_controller *chip, unsigned int irq)
-{
-	struct rcar_msi *msi = to_rcar_msi(chip);
-	struct irq_data *d = irq_get_irq_data(irq);
-
-	rcar_msi_free(msi, d->hwirq);
-}
-
-static struct irq_chip rcar_msi_irq_chip = {
-	.name = "R-Car PCIe MSI",
-	.irq_enable = pci_msi_unmask_irq,
-	.irq_disable = pci_msi_mask_irq,
-	.irq_mask = pci_msi_mask_irq,
-	.irq_unmask = pci_msi_unmask_irq,
-};
-
-static int rcar_msi_map(struct irq_domain *domain, unsigned int irq,
-			irq_hw_number_t hwirq)
-{
-	irq_set_chip_and_handler(irq, &rcar_msi_irq_chip, handle_simple_irq);
-	irq_set_chip_data(irq, domain->host_data);
-
-	return 0;
-}
-
-static const struct irq_domain_ops msi_domain_ops = {
-	.map = rcar_msi_map,
-};
-
-static void rcar_pcie_unmap_msi(struct rcar_pcie *pcie)
-{
-	struct rcar_msi *msi = &pcie->msi;
-	int i, irq;
-
-	for (i = 0; i < INT_PCI_MSI_NR; i++) {
-		irq = irq_find_mapping(msi->domain, i);
-		if (irq > 0)
-			irq_dispose_mapping(irq);
-	}
-
-	irq_domain_remove(msi->domain);
-}
-
-static int rcar_pcie_enable_msi(struct rcar_pcie *pcie)
-{
-	struct device *dev = pcie->dev;
-	struct rcar_msi *msi = &pcie->msi;
-	phys_addr_t base;
-	int err, i;
-
-	mutex_init(&msi->lock);
-
-	msi->chip.dev = dev;
-	msi->chip.setup_irq = rcar_msi_setup_irq;
-	msi->chip.setup_irqs = rcar_msi_setup_irqs;
-	msi->chip.teardown_irq = rcar_msi_teardown_irq;
-
-	msi->domain = irq_domain_add_linear(dev->of_node, INT_PCI_MSI_NR,
-					    &msi_domain_ops, &msi->chip);
-	if (!msi->domain) {
-		dev_err(dev, "failed to create IRQ domain\n");
-		return -ENOMEM;
-	}
-
-	for (i = 0; i < INT_PCI_MSI_NR; i++)
-		irq_create_mapping(msi->domain, i);
-
-	/* Two irqs are for MSI, but they are also used for non-MSI irqs */
-	err = devm_request_irq(dev, msi->irq1, rcar_pcie_msi_irq,
-			       IRQF_SHARED | IRQF_NO_THREAD,
-			       rcar_msi_irq_chip.name, pcie);
-	if (err < 0) {
-		dev_err(dev, "failed to request IRQ: %d\n", err);
-		goto err;
-	}
-
-	err = devm_request_irq(dev, msi->irq2, rcar_pcie_msi_irq,
-			       IRQF_SHARED | IRQF_NO_THREAD,
-			       rcar_msi_irq_chip.name, pcie);
-	if (err < 0) {
-		dev_err(dev, "failed to request IRQ: %d\n", err);
-		goto err;
-	}
-
-	/* setup MSI data target */
-	msi->pages = __get_free_pages(GFP_KERNEL, 0);
-	if (!msi->pages) {
-		err = -ENOMEM;
-		goto err;
-	}
-	base = virt_to_phys((void *)msi->pages);
-
-	rcar_pci_write_reg(pcie, lower_32_bits(base) | MSIFE, PCIEMSIALR);
-	rcar_pci_write_reg(pcie, upper_32_bits(base), PCIEMSIAUR);
-
-	/* enable all MSI interrupts */
-	rcar_pci_write_reg(pcie, 0xffffffff, PCIEMSIIER);
-
-	return 0;
-
-err:
-	rcar_pcie_unmap_msi(pcie);
-	return err;
-}
-
-static void rcar_pcie_teardown_msi(struct rcar_pcie *pcie)
-{
-	struct rcar_msi *msi = &pcie->msi;
-
-	/* Disable all MSI interrupts */
-	rcar_pci_write_reg(pcie, 0, PCIEMSIIER);
-
-	/* Disable address decoding of the MSI interrupt, MSIFE */
-	rcar_pci_write_reg(pcie, 0, PCIEMSIALR);
-
-	free_pages(msi->pages, 0);
-
-	rcar_pcie_unmap_msi(pcie);
-}
-
-static int rcar_pcie_get_resources(struct rcar_pcie *pcie)
-{
-	struct device *dev = pcie->dev;
-	struct resource res;
-	int err, i;
-
-	pcie->phy = devm_phy_optional_get(dev, "pcie");
-	if (IS_ERR(pcie->phy))
-		return PTR_ERR(pcie->phy);
-
-	err = of_address_to_resource(dev->of_node, 0, &res);
-	if (err)
-		return err;
-
-	pcie->base = devm_ioremap_resource(dev, &res);
-	if (IS_ERR(pcie->base))
-		return PTR_ERR(pcie->base);
-
-	pcie->bus_clk = devm_clk_get(dev, "pcie_bus");
-	if (IS_ERR(pcie->bus_clk)) {
-		dev_err(dev, "cannot get pcie bus clock\n");
-		return PTR_ERR(pcie->bus_clk);
-	}
-
-	i = irq_of_parse_and_map(dev->of_node, 0);
-	if (!i) {
-		dev_err(dev, "cannot get platform resources for msi interrupt\n");
-		err = -ENOENT;
-		goto err_irq1;
-	}
-	pcie->msi.irq1 = i;
-
-	i = irq_of_parse_and_map(dev->of_node, 1);
-	if (!i) {
-		dev_err(dev, "cannot get platform resources for msi interrupt\n");
-		err = -ENOENT;
-		goto err_irq2;
-	}
-	pcie->msi.irq2 = i;
-
-	return 0;
-
-err_irq2:
-	irq_dispose_mapping(pcie->msi.irq1);
-err_irq1:
-	return err;
-}
-
-static int rcar_pcie_inbound_ranges(struct rcar_pcie *pcie,
-				    struct of_pci_range *range,
-				    int *index)
+void rcar_pcie_set_inbound(void __iomem *base,
+			   u64 cpu_addr, u64 pci_addr,
+			   u64 mask, int idx, bool host)
 {
-	u64 restype = range->flags;
-	u64 cpu_addr = range->cpu_addr;
-	u64 cpu_end = range->cpu_addr + range->size;
-	u64 pci_addr = range->pci_addr;
-	u32 flags = LAM_64BIT | LAR_ENABLE;
-	u64 mask;
-	u64 size;
-	int idx = *index;
-
-	if (restype & IORESOURCE_PREFETCH)
-		flags |= LAM_PREFETCH;
-
 	/*
-	 * If the size of the range is larger than the alignment of the start
-	 * address, we have to use multiple entries to perform the mapping.
-	 */
-	if (cpu_addr > 0) {
-		unsigned long nr_zeros = __ffs64(cpu_addr);
-		u64 alignment = 1ULL << nr_zeros;
-
-		size = min(range->size, alignment);
-	} else {
-		size = range->size;
-	}
-	/* Hardware supports max 4GiB inbound region */
-	size = min(size, 1ULL << 32);
-
-	mask = roundup_pow_of_two(size) - 1;
-	mask &= ~0xf;
-
-	while (cpu_addr < cpu_end) {
-		/*
-		 * Set up 64-bit inbound regions as the range parser doesn't
-		 * distinguish between 32 and 64-bit types.
-		 */
-		rcar_pci_write_reg(pcie, lower_32_bits(pci_addr),
+	* Set up 64-bit inbound regions as the range parser doesn't
+	* distinguish between 32 and 64-bit types.
+	*/
+	if (host) {
+		rcar_pci_write_reg(base, lower_32_bits(pci_addr),
 				   PCIEPRAR(idx));
-		rcar_pci_write_reg(pcie, lower_32_bits(cpu_addr), PCIELAR(idx));
-		rcar_pci_write_reg(pcie, lower_32_bits(mask) | flags,
-				   PCIELAMR(idx));
-
-		rcar_pci_write_reg(pcie, upper_32_bits(pci_addr),
+		rcar_pci_write_reg(base, upper_32_bits(pci_addr),
 				   PCIEPRAR(idx + 1));
-		rcar_pci_write_reg(pcie, upper_32_bits(cpu_addr),
-				   PCIELAR(idx + 1));
-		rcar_pci_write_reg(pcie, 0, PCIELAMR(idx + 1));
-
-		pci_addr += size;
-		cpu_addr += size;
-		idx += 2;
-
-		if (idx > MAX_NR_INBOUND_MAPS) {
-			dev_err(pcie->dev, "Failed to map inbound regions!\n");
-			return -EINVAL;
-		}
 	}
-	*index = idx;
-
-	return 0;
-}
 
-static int rcar_pcie_parse_map_dma_ranges(struct rcar_pcie *pcie,
-					  struct device_node *np)
-{
-	struct of_pci_range range;
-	struct of_pci_range_parser parser;
-	int index = 0;
-	int err;
-
-	if (of_pci_dma_range_parser_init(&parser, np))
-		return -EINVAL;
-
-	/* Get the dma-ranges from DT */
-	for_each_of_pci_range(&parser, &range) {
-		u64 end = range.cpu_addr + range.size - 1;
+	rcar_pci_write_reg(base, lower_32_bits(cpu_addr), PCIELAR(idx));
+	rcar_pci_write_reg(base, lower_32_bits(mask), PCIELAMR(idx));
 
-		dev_dbg(pcie->dev, "0x%08x 0x%016llx..0x%016llx -> 0x%016llx\n",
-			range.flags, range.cpu_addr, end, range.pci_addr);
-
-		err = rcar_pcie_inbound_ranges(pcie, &range, &index);
-		if (err)
-			return err;
-	}
+	rcar_pci_write_reg(base, upper_32_bits(cpu_addr), PCIELAR(idx + 1));
+	rcar_pci_write_reg(base, 0, PCIELAMR(idx + 1));
 
-	return 0;
-}
-
-static const struct of_device_id rcar_pcie_of_match[] = {
-	{ .compatible = "renesas,pcie-r8a7779",
-	  .data = rcar_pcie_phy_init_h1 },
-	{ .compatible = "renesas,pcie-r8a7790",
-	  .data = rcar_pcie_phy_init_gen2 },
-	{ .compatible = "renesas,pcie-r8a7791",
-	  .data = rcar_pcie_phy_init_gen2 },
-	{ .compatible = "renesas,pcie-rcar-gen2",
-	  .data = rcar_pcie_phy_init_gen2 },
-	{ .compatible = "renesas,pcie-r8a7795",
-	  .data = rcar_pcie_phy_init_gen3 },
-	{ .compatible = "renesas,pcie-rcar-gen3",
-	  .data = rcar_pcie_phy_init_gen3 },
-	{},
-};
-
-static int rcar_pcie_probe(struct platform_device *pdev)
-{
-	struct device *dev = &pdev->dev;
-	struct rcar_pcie *pcie;
-	u32 data;
-	int err;
-	int (*phy_init_fn)(struct rcar_pcie *);
-	struct pci_host_bridge *bridge;
-
-	bridge = pci_alloc_host_bridge(sizeof(*pcie));
-	if (!bridge)
-		return -ENOMEM;
-
-	pcie = pci_host_bridge_priv(bridge);
-
-	pcie->dev = dev;
-	platform_set_drvdata(pdev, pcie);
-
-	err = pci_parse_request_of_pci_ranges(dev, &pcie->resources, NULL);
-	if (err)
-		goto err_free_bridge;
-
-	pm_runtime_enable(pcie->dev);
-	err = pm_runtime_get_sync(pcie->dev);
-	if (err < 0) {
-		dev_err(pcie->dev, "pm_runtime_get_sync failed\n");
-		goto err_pm_disable;
-	}
-
-	err = rcar_pcie_get_resources(pcie);
-	if (err < 0) {
-		dev_err(dev, "failed to request resources: %d\n", err);
-		goto err_pm_put;
-	}
-
-	err = clk_prepare_enable(pcie->bus_clk);
-	if (err) {
-		dev_err(dev, "failed to enable bus clock: %d\n", err);
-		goto err_unmap_msi_irqs;
-	}
-
-	err = rcar_pcie_parse_map_dma_ranges(pcie, dev->of_node);
-	if (err)
-		goto err_clk_disable;
-
-	phy_init_fn = of_device_get_match_data(dev);
-	err = phy_init_fn(pcie);
-	if (err) {
-		dev_err(dev, "failed to init PCIe PHY\n");
-		goto err_clk_disable;
-	}
-
-	/* Failure to get a link might just be that no cards are inserted */
-	if (rcar_pcie_hw_init(pcie)) {
-		dev_info(dev, "PCIe link down\n");
-		err = -ENODEV;
-		goto err_phy_shutdown;
-	}
-
-	data = rcar_pci_read_reg(pcie, MACSR);
-	dev_info(dev, "PCIe x%d: link up\n", (data >> 20) & 0x3f);
-
-	if (IS_ENABLED(CONFIG_PCI_MSI)) {
-		err = rcar_pcie_enable_msi(pcie);
-		if (err < 0) {
-			dev_err(dev,
-				"failed to enable MSI support: %d\n",
-				err);
-			goto err_phy_shutdown;
-		}
-	}
-
-	err = rcar_pcie_enable(pcie);
-	if (err)
-		goto err_msi_teardown;
-
-	return 0;
-
-err_msi_teardown:
-	if (IS_ENABLED(CONFIG_PCI_MSI))
-		rcar_pcie_teardown_msi(pcie);
-
-err_phy_shutdown:
-	if (pcie->phy) {
-		phy_power_off(pcie->phy);
-		phy_exit(pcie->phy);
-	}
-
-err_clk_disable:
-	clk_disable_unprepare(pcie->bus_clk);
-
-err_unmap_msi_irqs:
-	irq_dispose_mapping(pcie->msi.irq2);
-	irq_dispose_mapping(pcie->msi.irq1);
-
-err_pm_put:
-	pm_runtime_put(dev);
-
-err_pm_disable:
-	pm_runtime_disable(dev);
-	pci_free_resource_list(&pcie->resources);
-
-err_free_bridge:
-	pci_free_host_bridge(bridge);
-
-	return err;
-}
-
-static int rcar_pcie_resume_noirq(struct device *dev)
-{
-	struct rcar_pcie *pcie = dev_get_drvdata(dev);
-
-	if (rcar_pci_read_reg(pcie, PMSR) &&
-	    !(rcar_pci_read_reg(pcie, PCIETCTLR) & DL_DOWN))
-		return 0;
-
-	/* Re-establish the PCIe link */
-	rcar_rmw32(pcie, MACCTLR, MACCTLR_RESERVED, 0);
-	rcar_pci_write_reg(pcie, CFINIT, PCIETCTLR);
-	return rcar_pcie_wait_for_dl(pcie);
+	/* flush modifications */
+	wmb();
 }
-
-static const struct dev_pm_ops rcar_pcie_pm_ops = {
-	.resume_noirq = rcar_pcie_resume_noirq,
-};
-
-static struct platform_driver rcar_pcie_driver = {
-	.driver = {
-		.name = "rcar-pcie",
-		.of_match_table = rcar_pcie_of_match,
-		.pm = &rcar_pcie_pm_ops,
-		.suppress_bind_attrs = true,
-	},
-	.probe = rcar_pcie_probe,
-};
-builtin_platform_driver(rcar_pcie_driver);
diff --git a/drivers/pci/controller/pcie-rcar.h b/drivers/pci/controller/pcie-rcar.h
new file mode 100644
index 000000000000..502621d4e6ac
--- /dev/null
+++ b/drivers/pci/controller/pcie-rcar.h
@@ -0,0 +1,129 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * PCIe driver for Renesas R-Car SoCs
+ *  Copyright (C) 2014-2019 Renesas Electronics Europe Ltd
+ *
+ * Author: Phil Edworthy <phil.edworthy@renesas.com>
+ */
+#ifndef _PCIE_RCAR_H
+#define _PCIE_RCAR_H
+
+#define PCIECAR			0x000010
+#define PCIECCTLR		0x000018
+#define  CONFIG_SEND_ENABLE	BIT(31)
+#define  TYPE0			(0 << 8)
+#define  TYPE1			BIT(8)
+#define PCIECDR			0x000020
+#define PCIEMSR			0x000028
+#define PCIEINTXR		0x000400
+#define PCIEPHYSR		0x0007f0
+#define  PHYRDY			BIT(0)
+#define PCIEMSITXR		0x000840
+
+/* Transfer control */
+#define PCIETCTLR		0x02000
+#define  DL_DOWN		BIT(3)
+#define  CFINIT			BIT(0)
+#define PCIETSTR		0x02004
+#define  DATA_LINK_ACTIVE	BIT(0)
+#define PCIEERRFR		0x02020
+#define  UNSUPPORTED_REQUEST	BIT(4)
+#define PCIEMSIFR		0x02044
+#define PCIEMSIALR		0x02048
+#define  MSIFE			BIT(0)
+#define PCIEMSIAUR		0x0204c
+#define PCIEMSIIER		0x02050
+
+/* root port address */
+#define PCIEPRAR(x)		(0x02080 + ((x) * 0x4))
+
+/* local address reg & mask */
+#define PCIELAR(x)		(0x02200 + ((x) * 0x20))
+#define PCIELAMR(x)		(0x02208 + ((x) * 0x20))
+#define  LAM_PREFETCH		BIT(3)
+#define  LAM_64BIT		BIT(2)
+#define  LAR_ENABLE		BIT(1)
+
+/* PCIe address reg & mask */
+#define PCIEPALR(x)		(0x03400 + ((x) * 0x20))
+#define PCIEPAUR(x)		(0x03404 + ((x) * 0x20))
+#define PCIEPAMR(x)		(0x03408 + ((x) * 0x20))
+#define PCIEPTCTLR(x)		(0x0340c + ((x) * 0x20))
+#define  PAR_ENABLE		BIT(31)
+#define  IO_SPACE		BIT(8)
+
+/* Configuration */
+#define PCICONF(x)		(0x010000 + ((x) * 0x4))
+#define PMCAP(x)		(0x010040 + ((x) * 0x4))
+#define EXPCAP(x)		(0x010070 + ((x) * 0x4))
+#define VCCAP(x)		(0x010100 + ((x) * 0x4))
+
+/* link layer */
+#define IDSETR0			0x011000
+#define  DEVICE_ID_SHFIT	16
+#define IDSETR1			0x011004
+#define SUBIDSETR		0x011024
+#define TLCTLR			0x011048
+#define MACSR			0x011054
+#define  SPCHGFIN		BIT(4)
+#define  SPCHGFAIL		BIT(6)
+#define  SPCHGSUC		BIT(7)
+#define  LINK_SPEED		(0xf << 16)
+#define  LINK_SPEED_2_5GTS	BIT(16)
+#define  LINK_SPEED_5_0GTS	BIT(17)
+#define MACCTLR			0x011058
+#define  MACCTLR_RESERVED	BIT(0)
+#define  SPEED_CHANGE		BIT(24)
+#define  SCRAMBLE_DISABLE	BIT(27)
+#define PMSR			0x01105c
+#define MACS2R			0x011078
+#define MACCGSPSETR		0x011084
+#define  SPCNGRSN		BIT(31)
+
+/* R-Car H1 PHY */
+#define H1_PCIEPHYADRR		0x04000c
+#define  WRITE_CMD		BIT(16)
+#define  PHY_ACK		BIT(24)
+#define  RATE_POS		12
+#define  LANE_POS		8
+#define  ADR_POS		0
+#define H1_PCIEPHYDOUTR		0x040014
+
+/* R-Car Gen2 PHY */
+#define GEN2_PCIEPHYADDR	0x780
+#define GEN2_PCIEPHYDATA	0x784
+#define GEN2_PCIEPHYCTRL	0x78c
+
+#define INT_PCI_MSI_NR		32
+
+#define RCONF(x)		(PCICONF(0) + (x))
+#define RPMCAP(x)		(PMCAP(0) + (x))
+#define REXPCAP(x)		(EXPCAP(0) + (x))
+#define RVCCAP(x)		(VCCAP(0) + (x))
+
+#define INTDIS_SHIFT		BIT(10)
+#define ASTINTX_SHIFT		BIT(16)
+
+#define PCIE_CONF_BUS(b)	(((b) & 0xff) << 24)
+#define PCIE_CONF_DEV(d)	(((d) & 0x1f) << 19)
+#define PCIE_CONF_FUNC(f)	(((f) & 0x7) << 16)
+
+#define RCAR_PCI_MAX_RESOURCES	4
+#define MAX_NR_INBOUND_MAPS	6
+
+enum {
+	RCAR_PCI_ACCESS_READ,
+	RCAR_PCI_ACCESS_WRITE,
+};
+
+void rcar_pci_write_reg(void __iomem *base, u32 val, unsigned int reg);
+u32 rcar_pci_read_reg(void __iomem *base, unsigned int reg);
+void rcar_rmw32(void __iomem *base, int where, u32 mask, u32 data);
+int rcar_pcie_wait_for_phyrdy(void __iomem *base);
+int rcar_pcie_wait_for_dl(void __iomem *base);
+void rcar_pcie_set_outbound(int win, void __iomem *base,
+			    struct resource *res, bool host);
+void rcar_pcie_set_inbound(void __iomem *base, u64 cpu_addr, u64 pci_addr,
+			   u64 mask, int idx, bool host);
+
+#endif
-- 
2.20.1


^ permalink raw reply related	[flat|nested] 60+ messages in thread

* [PATCH 1/5] pci: pcie-rcar: preparation for adding endpoint support
@ 2019-11-06 19:36   ` Lad Prabhakar
  0 siblings, 0 replies; 60+ messages in thread
From: Lad Prabhakar @ 2019-11-06 19:36 UTC (permalink / raw)
  To: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Kishon Vijay Abraham I, Marek Vasut,
	Yoshihiro Shimoda, linux-pci
  Cc: devicetree, Chris Paterson, Lorenzo Pieralisi, Arnd Bergmann,
	Catalin Marinas, linux-kernel, Lad, Prabhakar, linux-renesas-soc,
	Greg Kroah-Hartman, Andrew Murray, Will Deacon, linux-arm-kernel

From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>

this patch prepares for adding endpoint support to rcar controller,
there are no functional changes with this patch, a common file is
created so that it can be shared with endpoint driver.

Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
---
 arch/arm64/configs/defconfig            |    2 +-
 arch/arm64/configs/renesas_defconfig    |    2 +-
 drivers/pci/controller/Kconfig          |    4 +-
 drivers/pci/controller/Makefile         |    2 +-
 drivers/pci/controller/pcie-rcar-host.c | 1058 +++++++++++++++++++
 drivers/pci/controller/pcie-rcar.c      | 1231 +----------------------
 drivers/pci/controller/pcie-rcar.h      |  129 +++
 7 files changed, 1245 insertions(+), 1183 deletions(-)
 create mode 100644 drivers/pci/controller/pcie-rcar-host.c
 create mode 100644 drivers/pci/controller/pcie-rcar.h

diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
index 7d16f10a9251..f67ce0ccea1d 100644
--- a/arch/arm64/configs/defconfig
+++ b/arch/arm64/configs/defconfig
@@ -186,7 +186,7 @@ CONFIG_HOTPLUG_PCI=y
 CONFIG_HOTPLUG_PCI_ACPI=y
 CONFIG_PCI_AARDVARK=y
 CONFIG_PCI_TEGRA=y
-CONFIG_PCIE_RCAR=y
+CONFIG_PCIE_RCAR_HOST=y
 CONFIG_PCI_HOST_GENERIC=y
 CONFIG_PCI_XGENE=y
 CONFIG_PCIE_ALTERA=y
diff --git a/arch/arm64/configs/renesas_defconfig b/arch/arm64/configs/renesas_defconfig
index f7411e8807d6..75fc418377c5 100644
--- a/arch/arm64/configs/renesas_defconfig
+++ b/arch/arm64/configs/renesas_defconfig
@@ -91,7 +91,7 @@ CONFIG_PCIEPORTBUS=y
 CONFIG_HOTPLUG_PCI_PCIE=y
 CONFIG_PCI_IOV=y
 CONFIG_HOTPLUG_PCI=y
-CONFIG_PCIE_RCAR=y
+CONFIG_PCIE_RCAR_HOST=y
 CONFIG_PCI_HOST_GENERIC=y
 CONFIG_DEVTMPFS=y
 CONFIG_DEVTMPFS_MOUNT=y
diff --git a/drivers/pci/controller/Kconfig b/drivers/pci/controller/Kconfig
index 70e078238899..01c238e5ea1e 100644
--- a/drivers/pci/controller/Kconfig
+++ b/drivers/pci/controller/Kconfig
@@ -82,12 +82,12 @@ config PCI_RCAR_GEN2
 	  There are 3 internal PCI controllers available with a single
 	  built-in EHCI/OHCI host controller present on each one.
 
-config PCIE_RCAR
+config PCIE_RCAR_HOST
 	bool "Renesas R-Car PCIe controller"
 	depends on ARCH_RENESAS || COMPILE_TEST
 	depends on PCI_MSI_IRQ_DOMAIN
 	help
-	  Say Y here if you want PCIe controller support on R-Car SoCs.
+	  Say Y here if you want PCIe controller support on R-Car SoCs in host mode.
 
 config PCI_HOST_COMMON
 	bool
diff --git a/drivers/pci/controller/Makefile b/drivers/pci/controller/Makefile
index a2a22c9d91af..3577902a0d40 100644
--- a/drivers/pci/controller/Makefile
+++ b/drivers/pci/controller/Makefile
@@ -9,7 +9,7 @@ obj-$(CONFIG_PCI_MVEBU) += pci-mvebu.o
 obj-$(CONFIG_PCI_AARDVARK) += pci-aardvark.o
 obj-$(CONFIG_PCI_TEGRA) += pci-tegra.o
 obj-$(CONFIG_PCI_RCAR_GEN2) += pci-rcar-gen2.o
-obj-$(CONFIG_PCIE_RCAR) += pcie-rcar.o
+obj-$(CONFIG_PCIE_RCAR_HOST) += pcie-rcar.o pcie-rcar-host.o
 obj-$(CONFIG_PCI_HOST_COMMON) += pci-host-common.o
 obj-$(CONFIG_PCI_HOST_GENERIC) += pci-host-generic.o
 obj-$(CONFIG_PCIE_XILINX) += pcie-xilinx.o
diff --git a/drivers/pci/controller/pcie-rcar-host.c b/drivers/pci/controller/pcie-rcar-host.c
new file mode 100644
index 000000000000..645ef3e41e27
--- /dev/null
+++ b/drivers/pci/controller/pcie-rcar-host.c
@@ -0,0 +1,1058 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * PCIe driver for Renesas R-Car SoCs
+ *  Copyright (C) 2014-2019 Renesas Electronics Europe Ltd
+ *
+ * Based on:
+ *  arch/sh/drivers/pci/pcie-sh7786.c
+ *  arch/sh/drivers/pci/ops-sh7786.c
+ *  Copyright (C) 2009 - 2011  Paul Mundt
+ *
+ * Author: Phil Edworthy <phil.edworthy@renesas.com>
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/msi.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/of_pci.h>
+#include <linux/of_platform.h>
+#include <linux/pci.h>
+#include <linux/phy/phy.h>
+
+#include "pcie-rcar.h"
+
+struct rcar_msi {
+	DECLARE_BITMAP(used, INT_PCI_MSI_NR);
+	struct irq_domain *domain;
+	struct msi_controller chip;
+	unsigned long pages;
+	struct mutex lock; /* serializes msi */
+	int irq1;
+	int irq2;
+};
+
+static inline struct rcar_msi *to_rcar_msi(struct msi_controller *chip)
+{
+	return container_of(chip, struct rcar_msi, chip);
+}
+
+/* Structure representing the PCIe interface */
+struct rcar_pcie {
+	struct device		*dev;
+	struct phy		*phy;
+	void __iomem		*base;
+	struct list_head	resources;
+	int			root_bus_nr;
+	struct clk		*bus_clk;
+	struct			rcar_msi msi;
+};
+
+static u32 rcar_read_conf(struct rcar_pcie *pcie, int where)
+{
+	unsigned int shift = BITS_PER_BYTE * (where & 3);
+	u32 val = rcar_pci_read_reg(pcie->base, where & ~3);
+
+	return val >> shift;
+}
+
+/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
+static int
+rcar_pcie_config_access(struct rcar_pcie *pcie,
+			unsigned char access_type, struct pci_bus *bus,
+			unsigned int devfn, int where, u32 *data)
+{
+	unsigned int dev, func, reg, index;
+
+	dev = PCI_SLOT(devfn);
+	func = PCI_FUNC(devfn);
+	reg = where & ~3;
+	index = reg / 4;
+
+	/*
+	 * While each channel has its own memory-mapped extended config
+	 * space, it's generally only accessible when in endpoint mode.
+	 * When in root complex mode, the controller is unable to target
+	 * itself with either type 0 or type 1 accesses, and indeed, any
+	 * controller initiated target transfer to its own config space
+	 * result in a completer abort.
+	 *
+	 * Each channel effectively only supports a single device, but as
+	 * the same channel <-> device access works for any PCI_SLOT()
+	 * value, we cheat a bit here and bind the controller's config
+	 * space to devfn 0 in order to enable self-enumeration. In this
+	 * case the regular ECAR/ECDR path is sidelined and the mangled
+	 * config access itself is initiated as an internal bus transaction.
+	 */
+	if (pci_is_root_bus(bus)) {
+		if (dev != 0)
+			return PCIBIOS_DEVICE_NOT_FOUND;
+
+		if (access_type == RCAR_PCI_ACCESS_READ) {
+			*data = rcar_pci_read_reg(pcie->base, PCICONF(index));
+		} else {
+			/* Keep an eye out for changes to the root bus number */
+			if (pci_is_root_bus(bus) && reg == PCI_PRIMARY_BUS)
+				pcie->root_bus_nr = *data & 0xff;
+
+			rcar_pci_write_reg(pcie->base, *data, PCICONF(index));
+		}
+
+		return PCIBIOS_SUCCESSFUL;
+	}
+
+	if (pcie->root_bus_nr < 0)
+		return PCIBIOS_DEVICE_NOT_FOUND;
+
+	/* Clear errors */
+	rcar_pci_write_reg(pcie->base, rcar_pci_read_reg(pcie->base, PCIEERRFR),
+			   PCIEERRFR);
+
+	/* Set the PIO address */
+	rcar_pci_write_reg(pcie->base, PCIE_CONF_BUS(bus->number) |
+		PCIE_CONF_DEV(dev) | PCIE_CONF_FUNC(func) | reg, PCIECAR);
+
+	/* Enable the configuration access */
+	if (bus->parent->number == pcie->root_bus_nr)
+		rcar_pci_write_reg(pcie->base,
+				   CONFIG_SEND_ENABLE | TYPE0, PCIECCTLR);
+	else
+		rcar_pci_write_reg(pcie->base,
+				   CONFIG_SEND_ENABLE | TYPE1, PCIECCTLR);
+
+	/* Check for errors */
+	if (rcar_pci_read_reg(pcie->base, PCIEERRFR) & UNSUPPORTED_REQUEST)
+		return PCIBIOS_DEVICE_NOT_FOUND;
+
+	/* Check for master and target aborts */
+	if (rcar_read_conf(pcie, RCONF(PCI_STATUS)) &
+		(PCI_STATUS_REC_MASTER_ABORT | PCI_STATUS_REC_TARGET_ABORT))
+		return PCIBIOS_DEVICE_NOT_FOUND;
+
+	if (access_type == RCAR_PCI_ACCESS_READ)
+		*data = rcar_pci_read_reg(pcie->base, PCIECDR);
+	else
+		rcar_pci_write_reg(pcie->base, *data, PCIECDR);
+
+	/* Disable the configuration access */
+	rcar_pci_write_reg(pcie->base, 0, PCIECCTLR);
+
+	return PCIBIOS_SUCCESSFUL;
+}
+
+static int rcar_pcie_read_conf(struct pci_bus *bus, unsigned int devfn,
+			       int where, int size, u32 *val)
+{
+	struct rcar_pcie *pcie = bus->sysdata;
+	int ret;
+
+	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
+				      bus, devfn, where, val);
+	if (ret != PCIBIOS_SUCCESSFUL) {
+		*val = 0xffffffff;
+		return ret;
+	}
+
+	if (size == 1)
+		*val = (*val >> (BITS_PER_BYTE * (where & 3))) & 0xff;
+	else if (size == 2)
+		*val = (*val >> (BITS_PER_BYTE * (where & 2))) & 0xffff;
+
+	dev_dbg(&bus->dev, "pcie-config-read: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
+		bus->number, devfn, where, size, *val);
+
+	return ret;
+}
+
+/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
+static int rcar_pcie_write_conf(struct pci_bus *bus, unsigned int devfn,
+				int where, int size, u32 val)
+{
+	struct rcar_pcie *pcie = bus->sysdata;
+	unsigned int shift;
+	u32 data;
+	int ret;
+
+	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
+				      bus, devfn, where, &data);
+	if (ret != PCIBIOS_SUCCESSFUL)
+		return ret;
+
+	dev_dbg(&bus->dev, "pcie-config-write: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
+		bus->number, devfn, where, size, val);
+
+	if (size == 1) {
+		shift = BITS_PER_BYTE * (where & 3);
+		data &= ~(0xff << shift);
+		data |= ((val & 0xff) << shift);
+	} else if (size == 2) {
+		shift = BITS_PER_BYTE * (where & 2);
+		data &= ~(0xffff << shift);
+		data |= ((val & 0xffff) << shift);
+	} else {
+		data = val;
+	}
+
+	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_WRITE,
+				      bus, devfn, where, &data);
+
+	return ret;
+}
+
+static struct pci_ops rcar_pcie_ops = {
+	.read	= rcar_pcie_read_conf,
+	.write	= rcar_pcie_write_conf,
+};
+
+static int rcar_pcie_setup(struct list_head *resource, struct rcar_pcie *pci)
+{
+	struct resource_entry *win;
+	int i = 0;
+
+	/* Setup PCI resources */
+	resource_list_for_each_entry(win, &pci->resources) {
+		struct resource *res = win->res;
+
+		if (!res->flags)
+			continue;
+
+		switch (resource_type(res)) {
+		case IORESOURCE_IO:
+		case IORESOURCE_MEM:
+			rcar_pcie_set_outbound(i, pci->base, res, true);
+			i++;
+			break;
+		case IORESOURCE_BUS:
+			pci->root_bus_nr = res->start;
+			break;
+		default:
+			continue;
+		}
+
+		pci_add_resource(resource, res);
+	}
+
+	return 1;
+}
+
+static void rcar_pcie_force_speedup(struct rcar_pcie *pcie)
+{
+	struct device *dev = pcie->dev;
+	unsigned int timeout = 1000;
+	u32 macsr;
+
+	if ((rcar_pci_read_reg(pcie->base, MACS2R) & LINK_SPEED) !=
+		LINK_SPEED_5_0GTS)
+		return;
+
+	if (rcar_pci_read_reg(pcie->base, MACCTLR) & SPEED_CHANGE) {
+		dev_err(dev, "Speed change already in progress\n");
+		return;
+	}
+
+	macsr = rcar_pci_read_reg(pcie->base, MACSR);
+	if ((macsr & LINK_SPEED) == LINK_SPEED_5_0GTS)
+		goto done;
+
+	/* Set target link speed to 5.0 GT/s */
+	rcar_rmw32(pcie->base, EXPCAP(12), PCI_EXP_LNKSTA_CLS,
+		   PCI_EXP_LNKSTA_CLS_5_0GB);
+
+	/* Set speed change reason as intentional factor */
+	rcar_rmw32(pcie->base, MACCGSPSETR, SPCNGRSN, 0);
+
+	/* Clear SPCHGFIN, SPCHGSUC, and SPCHGFAIL */
+	if (macsr & (SPCHGFIN | SPCHGSUC | SPCHGFAIL))
+		rcar_pci_write_reg(pcie->base, macsr, MACSR);
+
+	/* Start link speed change */
+	rcar_rmw32(pcie->base, MACCTLR, SPEED_CHANGE, SPEED_CHANGE);
+
+	while (timeout--) {
+		macsr = rcar_pci_read_reg(pcie->base, MACSR);
+		if (macsr & SPCHGFIN) {
+			/* Clear the interrupt bits */
+			rcar_pci_write_reg(pcie->base, macsr, MACSR);
+
+			if (macsr & SPCHGFAIL)
+				dev_err(dev, "Speed change failed\n");
+
+			goto done;
+		}
+
+		usleep_range(1000, 2000);
+	}
+
+	dev_err(dev, "Speed change timed out\n");
+
+done:
+	dev_info(dev, "Current link speed is %s GT/s\n",
+		 (macsr & LINK_SPEED) == LINK_SPEED_5_0GTS ? "5" : "2.5");
+}
+
+static int rcar_pcie_enable(struct rcar_pcie *pcie)
+{
+	struct pci_host_bridge *bridge = pci_host_bridge_from_priv(pcie);
+	struct device *dev = pcie->dev;
+	struct pci_bus *bus, *child;
+	int ret;
+
+	/* Try setting 5 GT/s link speed */
+	rcar_pcie_force_speedup(pcie);
+
+	rcar_pcie_setup(&bridge->windows, pcie);
+
+	pci_add_flags(PCI_REASSIGN_ALL_BUS);
+
+	bridge->dev.parent = dev;
+	bridge->sysdata = pcie;
+	bridge->busnr = pcie->root_bus_nr;
+	bridge->ops = &rcar_pcie_ops;
+	bridge->map_irq = of_irq_parse_and_map_pci;
+	bridge->swizzle_irq = pci_common_swizzle;
+	if (IS_ENABLED(CONFIG_PCI_MSI))
+		bridge->msi = &pcie->msi.chip;
+
+	ret = pci_scan_root_bus_bridge(bridge);
+	if (ret < 0)
+		return ret;
+
+	bus = bridge->bus;
+
+	pci_bus_size_bridges(bus);
+	pci_bus_assign_resources(bus);
+
+	list_for_each_entry(child, &bus->children, node)
+		pcie_bus_configure_settings(child);
+
+	pci_bus_add_devices(bus);
+
+	return 0;
+}
+
+static int phy_wait_for_ack(struct rcar_pcie *pcie)
+{
+	struct device *dev = pcie->dev;
+	unsigned int timeout = 100;
+
+	while (timeout--) {
+		if (rcar_pci_read_reg(pcie->base, H1_PCIEPHYADRR) & PHY_ACK)
+			return 0;
+
+		usleep_range(90, 110);
+	}
+
+	dev_err(dev, "Access to PCIe phy timed out\n");
+
+	return -ETIMEDOUT;
+}
+
+static void phy_write_reg(struct rcar_pcie *pcie,
+			  unsigned int rate, u32 addr,
+			  unsigned int lane, u32 data)
+{
+	u32 phyaddr;
+
+	phyaddr = WRITE_CMD |
+		((rate & 1) << RATE_POS) |
+		((lane & 0xf) << LANE_POS) |
+		((addr & 0xff) << ADR_POS);
+
+	/* Set write data */
+	rcar_pci_write_reg(pcie->base, data, H1_PCIEPHYDOUTR);
+	rcar_pci_write_reg(pcie->base, phyaddr, H1_PCIEPHYADRR);
+
+	/* Ignore errors as they will be dealt with if the data link is down */
+	phy_wait_for_ack(pcie);
+
+	/* Clear command */
+	rcar_pci_write_reg(pcie->base, 0, H1_PCIEPHYDOUTR);
+	rcar_pci_write_reg(pcie->base, 0, H1_PCIEPHYADRR);
+
+	/* Ignore errors as they will be dealt with if the data link is down */
+	phy_wait_for_ack(pcie);
+}
+
+static int rcar_pcie_hw_init(struct rcar_pcie *pcie)
+{
+	int err;
+
+	/* Begin initialization */
+	rcar_pci_write_reg(pcie->base, 0, PCIETCTLR);
+
+	/* Set mode */
+	rcar_pci_write_reg(pcie->base, 1, PCIEMSR);
+
+	err = rcar_pcie_wait_for_phyrdy(pcie->base);
+	if (err)
+		return err;
+
+	/*
+	 * Initial header for port config space is type 1, set the device
+	 * class to match. Hardware takes care of propagating the IDSETR
+	 * settings, so there is no need to bother with a quirk.
+	 */
+	rcar_pci_write_reg(pcie->base, PCI_CLASS_BRIDGE_PCI << 16, IDSETR1);
+
+	/*
+	 * Setup Secondary Bus Number & Subordinate Bus Number, even though
+	 * they aren't used, to avoid bridge being detected as broken.
+	 */
+	rcar_rmw32(pcie->base, RCONF(PCI_SECONDARY_BUS), 0xff, 1);
+	rcar_rmw32(pcie->base, RCONF(PCI_SUBORDINATE_BUS), 0xff, 1);
+
+	/* Initialize default capabilities. */
+	rcar_rmw32(pcie->base, REXPCAP(0), 0xff, PCI_CAP_ID_EXP);
+	rcar_rmw32(pcie->base, REXPCAP(PCI_EXP_FLAGS),
+		   PCI_EXP_FLAGS_TYPE, PCI_EXP_TYPE_ROOT_PORT << 4);
+	rcar_rmw32(pcie->base, RCONF(PCI_HEADER_TYPE), 0x7f,
+		   PCI_HEADER_TYPE_BRIDGE);
+
+	/* Enable data link layer active state reporting */
+	rcar_rmw32(pcie->base, REXPCAP(PCI_EXP_LNKCAP), PCI_EXP_LNKCAP_DLLLARC,
+		   PCI_EXP_LNKCAP_DLLLARC);
+
+	/* Write out the physical slot number = 0 */
+	rcar_rmw32(pcie->base, REXPCAP(PCI_EXP_SLTCAP), PCI_EXP_SLTCAP_PSN, 0);
+
+	/* Set the completion timer timeout to the maximum 50ms. */
+	rcar_rmw32(pcie->base, TLCTLR + 1, 0x3f, 50);
+
+	/* Terminate list of capabilities (Next Capability Offset=0) */
+	rcar_rmw32(pcie->base, RVCCAP(0), 0xfff00000, 0);
+
+	/* Enable MSI */
+	if (IS_ENABLED(CONFIG_PCI_MSI))
+		rcar_pci_write_reg(pcie->base, 0x801f0000, PCIEMSITXR);
+
+	rcar_rmw32(pcie->base, MACCTLR, MACCTLR_RESERVED, 0);
+
+	/* Finish initialization - establish a PCI Express link */
+	rcar_pci_write_reg(pcie->base, CFINIT, PCIETCTLR);
+
+	/* This will timeout if we don't have a link. */
+	err = rcar_pcie_wait_for_dl(pcie->base);
+	if (err)
+		return err;
+
+	/* Enable INTx interrupts */
+	rcar_rmw32(pcie->base, PCIEINTXR, 0, 0xF << 8);
+
+	/* flush modifications */
+	wmb();
+
+	return 0;
+}
+
+static int rcar_pcie_phy_init_h1(struct rcar_pcie *pcie)
+{
+	/* Initialize the phy */
+	phy_write_reg(pcie, 0, 0x42, 0x1, 0x0EC34191);
+	phy_write_reg(pcie, 1, 0x42, 0x1, 0x0EC34180);
+	phy_write_reg(pcie, 0, 0x43, 0x1, 0x00210188);
+	phy_write_reg(pcie, 1, 0x43, 0x1, 0x00210188);
+	phy_write_reg(pcie, 0, 0x44, 0x1, 0x015C0014);
+	phy_write_reg(pcie, 1, 0x44, 0x1, 0x015C0014);
+	phy_write_reg(pcie, 1, 0x4C, 0x1, 0x786174A0);
+	phy_write_reg(pcie, 1, 0x4D, 0x1, 0x048000BB);
+	phy_write_reg(pcie, 0, 0x51, 0x1, 0x079EC062);
+	phy_write_reg(pcie, 0, 0x52, 0x1, 0x20000000);
+	phy_write_reg(pcie, 1, 0x52, 0x1, 0x20000000);
+	phy_write_reg(pcie, 1, 0x56, 0x1, 0x00003806);
+
+	phy_write_reg(pcie, 0, 0x60, 0x1, 0x004B03A5);
+	phy_write_reg(pcie, 0, 0x64, 0x1, 0x3F0F1F0F);
+	phy_write_reg(pcie, 0, 0x66, 0x1, 0x00008000);
+
+	return 0;
+}
+
+static int rcar_pcie_phy_init_gen2(struct rcar_pcie *pcie)
+{
+	/*
+	 * These settings come from the R-Car Series, 2nd Generation User's
+	 * Manual, section 50.3.1 (2) Initialization of the physical layer.
+	 */
+	rcar_pci_write_reg(pcie->base, 0x000f0030, GEN2_PCIEPHYADDR);
+	rcar_pci_write_reg(pcie->base, 0x00381203, GEN2_PCIEPHYDATA);
+	rcar_pci_write_reg(pcie->base, 0x00000001, GEN2_PCIEPHYCTRL);
+	rcar_pci_write_reg(pcie->base, 0x00000006, GEN2_PCIEPHYCTRL);
+
+	rcar_pci_write_reg(pcie->base, 0x000f0054, GEN2_PCIEPHYADDR);
+	/* The following value is for DC connection, no termination resistor */
+	rcar_pci_write_reg(pcie->base, 0x13802007, GEN2_PCIEPHYDATA);
+	rcar_pci_write_reg(pcie->base, 0x00000001, GEN2_PCIEPHYCTRL);
+	rcar_pci_write_reg(pcie->base, 0x00000006, GEN2_PCIEPHYCTRL);
+
+	return 0;
+}
+
+static int rcar_pcie_phy_init_gen3(struct rcar_pcie *pcie)
+{
+	int err;
+
+	err = phy_init(pcie->phy);
+	if (err)
+		return err;
+
+	err = phy_power_on(pcie->phy);
+	if (err)
+		phy_exit(pcie->phy);
+
+	return err;
+}
+
+static int rcar_msi_alloc(struct rcar_msi *chip)
+{
+	int msi;
+
+	mutex_lock(&chip->lock);
+
+	msi = find_first_zero_bit(chip->used, INT_PCI_MSI_NR);
+	if (msi < INT_PCI_MSI_NR)
+		set_bit(msi, chip->used);
+	else
+		msi = -ENOSPC;
+
+	mutex_unlock(&chip->lock);
+
+	return msi;
+}
+
+static int rcar_msi_alloc_region(struct rcar_msi *chip, int no_irqs)
+{
+	int msi;
+
+	mutex_lock(&chip->lock);
+	msi = bitmap_find_free_region(chip->used, INT_PCI_MSI_NR,
+				      order_base_2(no_irqs));
+	mutex_unlock(&chip->lock);
+
+	return msi;
+}
+
+static void rcar_msi_free(struct rcar_msi *chip, unsigned long irq)
+{
+	mutex_lock(&chip->lock);
+	clear_bit(irq, chip->used);
+	mutex_unlock(&chip->lock);
+}
+
+static irqreturn_t rcar_pcie_msi_irq(int irq, void *data)
+{
+	struct rcar_pcie *pcie = data;
+	struct rcar_msi *msi = &pcie->msi;
+	struct device *dev = pcie->dev;
+	unsigned long reg;
+
+	reg = rcar_pci_read_reg(pcie->base, PCIEMSIFR);
+
+	/* MSI & INTx share an interrupt - we only handle MSI here */
+	if (!reg)
+		return IRQ_NONE;
+
+	while (reg) {
+		unsigned int index = find_first_bit(&reg, 32);
+		unsigned int msi_irq;
+
+		/* clear the interrupt */
+		rcar_pci_write_reg(pcie->base, 1 << index, PCIEMSIFR);
+
+		msi_irq = irq_find_mapping(msi->domain, index);
+		if (msi_irq) {
+			if (test_bit(index, msi->used))
+				generic_handle_irq(msi_irq);
+			else
+				dev_info(dev, "unhandled MSI\n");
+		} else {
+			/* Unknown MSI, just clear it */
+			dev_dbg(dev, "unexpected MSI\n");
+		}
+
+		/* see if there's any more pending in this vector */
+		reg = rcar_pci_read_reg(pcie->base, PCIEMSIFR);
+	}
+
+	return IRQ_HANDLED;
+}
+
+static int rcar_msi_setup_irq(struct msi_controller *chip, struct pci_dev *pdev,
+			      struct msi_desc *desc)
+{
+	struct rcar_msi *msi = to_rcar_msi(chip);
+	struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
+	struct msi_msg msg;
+	unsigned int irq;
+	int hwirq;
+
+	hwirq = rcar_msi_alloc(msi);
+	if (hwirq < 0)
+		return hwirq;
+
+	irq = irq_find_mapping(msi->domain, hwirq);
+	if (!irq) {
+		rcar_msi_free(msi, hwirq);
+		return -EINVAL;
+	}
+
+	irq_set_msi_desc(irq, desc);
+
+	msg.address_lo = rcar_pci_read_reg(pcie->base, PCIEMSIALR) & ~MSIFE;
+	msg.address_hi = rcar_pci_read_reg(pcie->base, PCIEMSIAUR);
+	msg.data = hwirq;
+
+	pci_write_msi_msg(irq, &msg);
+
+	return 0;
+}
+
+static int rcar_msi_setup_irqs(struct msi_controller *chip,
+			       struct pci_dev *pdev, int nvec, int type)
+{
+	struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
+	struct rcar_msi *msi = to_rcar_msi(chip);
+	struct msi_desc *desc;
+	struct msi_msg msg;
+	unsigned int irq;
+	int hwirq;
+	int i;
+
+	/* MSI-X interrupts are not supported */
+	if (type == PCI_CAP_ID_MSIX)
+		return -EINVAL;
+
+	WARN_ON(!list_is_singular(&pdev->dev.msi_list));
+	desc = list_entry(pdev->dev.msi_list.next, struct msi_desc, list);
+
+	hwirq = rcar_msi_alloc_region(msi, nvec);
+	if (hwirq < 0)
+		return -ENOSPC;
+
+	irq = irq_find_mapping(msi->domain, hwirq);
+	if (!irq)
+		return -ENOSPC;
+
+	for (i = 0; i < nvec; i++) {
+		/*
+		 * irq_create_mapping() called from rcar_pcie_probe() pre-
+		 * allocates descs,  so there is no need to allocate descs here.
+		 * We can therefore assume that if irq_find_mapping() above
+		 * returns non-zero, then the descs are also successfully
+		 * allocated.
+		 */
+		if (irq_set_msi_desc_off(irq, i, desc)) {
+			/* TODO: clear */
+			return -EINVAL;
+		}
+	}
+
+	desc->nvec_used = nvec;
+	desc->msi_attrib.multiple = order_base_2(nvec);
+
+	msg.address_lo = rcar_pci_read_reg(pcie->base, PCIEMSIALR) & ~MSIFE;
+	msg.address_hi = rcar_pci_read_reg(pcie->base, PCIEMSIAUR);
+	msg.data = hwirq;
+
+	pci_write_msi_msg(irq, &msg);
+
+	return 0;
+}
+
+static void rcar_msi_teardown_irq(struct msi_controller *chip, unsigned int irq)
+{
+	struct irq_data *d = irq_get_irq_data(irq);
+	struct rcar_msi *msi = to_rcar_msi(chip);
+
+	rcar_msi_free(msi, d->hwirq);
+}
+
+static struct irq_chip rcar_msi_irq_chip = {
+	.name = "R-Car PCIe MSI",
+	.irq_enable = pci_msi_unmask_irq,
+	.irq_disable = pci_msi_mask_irq,
+	.irq_mask = pci_msi_mask_irq,
+	.irq_unmask = pci_msi_unmask_irq,
+};
+
+static int rcar_msi_map(struct irq_domain *domain, unsigned int irq,
+			irq_hw_number_t hwirq)
+{
+	irq_set_chip_and_handler(irq, &rcar_msi_irq_chip, handle_simple_irq);
+	irq_set_chip_data(irq, domain->host_data);
+
+	return 0;
+}
+
+static const struct irq_domain_ops msi_domain_ops = {
+	.map = rcar_msi_map,
+};
+
+static void rcar_pcie_unmap_msi(struct rcar_pcie *pcie)
+{
+	struct rcar_msi *msi = &pcie->msi;
+	int i, irq;
+
+	for (i = 0; i < INT_PCI_MSI_NR; i++) {
+		irq = irq_find_mapping(msi->domain, i);
+		if (irq > 0)
+			irq_dispose_mapping(irq);
+	}
+
+	irq_domain_remove(msi->domain);
+}
+
+static int rcar_pcie_enable_msi(struct rcar_pcie *pcie)
+{
+	struct rcar_msi *msi = &pcie->msi;
+	struct device *dev = pcie->dev;
+	phys_addr_t base;
+	int err, i;
+
+	mutex_init(&msi->lock);
+
+	msi->chip.dev = dev;
+	msi->chip.setup_irq = rcar_msi_setup_irq;
+	msi->chip.setup_irqs = rcar_msi_setup_irqs;
+	msi->chip.teardown_irq = rcar_msi_teardown_irq;
+
+	msi->domain = irq_domain_add_linear(dev->of_node, INT_PCI_MSI_NR,
+					    &msi_domain_ops, &msi->chip);
+	if (!msi->domain) {
+		dev_err(dev, "failed to create IRQ domain\n");
+		return -ENOMEM;
+	}
+
+	for (i = 0; i < INT_PCI_MSI_NR; i++)
+		irq_create_mapping(msi->domain, i);
+
+	/* Two irqs are for MSI, but they are also used for non-MSI irqs */
+	err = devm_request_irq(dev, msi->irq1, rcar_pcie_msi_irq,
+			       IRQF_SHARED | IRQF_NO_THREAD,
+			       rcar_msi_irq_chip.name, pcie);
+	if (err < 0) {
+		dev_err(dev, "failed to request IRQ: %d\n", err);
+		goto err;
+	}
+
+	err = devm_request_irq(dev, msi->irq2, rcar_pcie_msi_irq,
+			       IRQF_SHARED | IRQF_NO_THREAD,
+			       rcar_msi_irq_chip.name, pcie);
+	if (err < 0) {
+		dev_err(dev, "failed to request IRQ: %d\n", err);
+		goto err;
+	}
+
+	/* setup MSI data target */
+	msi->pages = __get_free_pages(GFP_KERNEL, 0);
+	if (!msi->pages) {
+		err = -ENOMEM;
+		goto err;
+	}
+	base = virt_to_phys((void *)msi->pages);
+
+	rcar_pci_write_reg(pcie->base, lower_32_bits(base) | MSIFE, PCIEMSIALR);
+	rcar_pci_write_reg(pcie->base, upper_32_bits(base), PCIEMSIAUR);
+
+	/* enable all MSI interrupts */
+	rcar_pci_write_reg(pcie->base, 0xffffffff, PCIEMSIIER);
+
+	return 0;
+
+err:
+	rcar_pcie_unmap_msi(pcie);
+	return err;
+}
+
+static void rcar_pcie_teardown_msi(struct rcar_pcie *pcie)
+{
+	struct rcar_msi *msi = &pcie->msi;
+
+	/* Disable all MSI interrupts */
+	rcar_pci_write_reg(pcie->base, 0, PCIEMSIIER);
+
+	/* Disable address decoding of the MSI interrupt, MSIFE */
+	rcar_pci_write_reg(pcie->base, 0, PCIEMSIALR);
+
+	free_pages(msi->pages, 0);
+
+	rcar_pcie_unmap_msi(pcie);
+}
+
+static int rcar_pcie_get_resources(struct rcar_pcie *pcie)
+{
+	struct device *dev = pcie->dev;
+	struct resource res;
+	int err, i;
+
+	pcie->phy = devm_phy_optional_get(dev, "pcie");
+	if (IS_ERR(pcie->phy))
+		return PTR_ERR(pcie->phy);
+
+	err = of_address_to_resource(dev->of_node, 0, &res);
+	if (err)
+		return err;
+
+	pcie->base = devm_ioremap_resource(dev, &res);
+	if (IS_ERR(pcie->base))
+		return PTR_ERR(pcie->base);
+
+	pcie->bus_clk = devm_clk_get(dev, "pcie_bus");
+	if (IS_ERR(pcie->bus_clk)) {
+		dev_err(dev, "cannot get pcie bus clock\n");
+		return PTR_ERR(pcie->bus_clk);
+	}
+
+	i = irq_of_parse_and_map(dev->of_node, 0);
+	if (!i) {
+		dev_err(dev, "cannot get platform resources for msi interrupt\n");
+		err = -ENOENT;
+		goto err_irq1;
+	}
+	pcie->msi.irq1 = i;
+
+	i = irq_of_parse_and_map(dev->of_node, 1);
+	if (!i) {
+		dev_err(dev, "cannot get platform resources for msi interrupt\n");
+		err = -ENOENT;
+		goto err_irq2;
+	}
+	pcie->msi.irq2 = i;
+
+	return 0;
+
+err_irq2:
+	irq_dispose_mapping(pcie->msi.irq1);
+err_irq1:
+	return err;
+}
+
+static int rcar_pcie_inbound_ranges(struct rcar_pcie *pcie,
+				    struct of_pci_range *range,
+				    int *index)
+{
+	u64 restype = range->flags;
+	u64 cpu_addr = range->cpu_addr;
+	u64 cpu_end = range->cpu_addr + range->size;
+	u64 pci_addr = range->pci_addr;
+	u32 flags = LAM_64BIT | LAR_ENABLE;
+	u64 mask;
+	u64 size;
+	int idx = *index;
+
+	if (restype & IORESOURCE_PREFETCH)
+		flags |= LAM_PREFETCH;
+
+	/*
+	 * If the size of the range is larger than the alignment of the start
+	 * address, we have to use multiple entries to perform the mapping.
+	 */
+	if (cpu_addr > 0) {
+		unsigned long nr_zeros = __ffs64(cpu_addr);
+		u64 alignment = 1ULL << nr_zeros;
+
+		size = min(range->size, alignment);
+	} else {
+		size = range->size;
+	}
+	/* Hardware supports max 4GiB inbound region */
+	size = min(size, 1ULL << 32);
+
+	mask = roundup_pow_of_two(size) - 1;
+	mask &= ~0xf;
+
+	while (cpu_addr < cpu_end) {
+		rcar_pcie_set_inbound(pcie->base, cpu_addr, pci_addr,
+				      mask | flags, idx, true);
+		pci_addr += size;
+		cpu_addr += size;
+		idx += 2;
+
+		if (idx > MAX_NR_INBOUND_MAPS) {
+			dev_err(pcie->dev, "Failed to map inbound regions!\n");
+			return -EINVAL;
+		}
+	}
+	*index = idx;
+
+	return 0;
+}
+
+static int rcar_pcie_parse_map_dma_ranges(struct rcar_pcie *pcie,
+					  struct device_node *np)
+{
+	struct of_pci_range_parser parser;
+	struct of_pci_range range;
+	int index = 0;
+	int err;
+
+	if (of_pci_dma_range_parser_init(&parser, np))
+		return -EINVAL;
+
+	/* Get the dma-ranges from DT */
+	for_each_of_pci_range(&parser, &range) {
+		u64 end = range.cpu_addr + range.size - 1;
+
+		dev_dbg(pcie->dev, "0x%08x 0x%016llx..0x%016llx -> 0x%016llx\n",
+			range.flags, range.cpu_addr, end, range.pci_addr);
+
+		err = rcar_pcie_inbound_ranges(pcie, &range, &index);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static const struct of_device_id rcar_pcie_of_match[] = {
+	{ .compatible = "renesas,pcie-r8a7779",
+	  .data = rcar_pcie_phy_init_h1 },
+	{ .compatible = "renesas,pcie-r8a7790",
+	  .data = rcar_pcie_phy_init_gen2 },
+	{ .compatible = "renesas,pcie-r8a7791",
+	  .data = rcar_pcie_phy_init_gen2 },
+	{ .compatible = "renesas,pcie-rcar-gen2",
+	  .data = rcar_pcie_phy_init_gen2 },
+	{ .compatible = "renesas,pcie-r8a7795",
+	  .data = rcar_pcie_phy_init_gen3 },
+	{ .compatible = "renesas,pcie-rcar-gen3",
+	  .data = rcar_pcie_phy_init_gen3 },
+	{},
+};
+
+static int rcar_pcie_probe(struct platform_device *pdev)
+{
+	int (*phy_init_fn)(struct rcar_pcie *pcie);
+	struct device *dev = &pdev->dev;
+	struct pci_host_bridge *bridge;
+	struct rcar_pcie *pcie;
+	u32 data;
+	int err;
+
+	bridge = pci_alloc_host_bridge(sizeof(*pcie));
+	if (!bridge)
+		return -ENOMEM;
+
+	pcie = pci_host_bridge_priv(bridge);
+
+	pcie->dev = dev;
+	platform_set_drvdata(pdev, pcie);
+
+	err = pci_parse_request_of_pci_ranges(dev, &pcie->resources, NULL);
+	if (err)
+		goto err_free_bridge;
+
+	pm_runtime_enable(pcie->dev);
+	err = pm_runtime_get_sync(pcie->dev);
+	if (err < 0) {
+		dev_err(pcie->dev, "pm_runtime_get_sync failed\n");
+		goto err_pm_disable;
+	}
+
+	err = rcar_pcie_get_resources(pcie);
+	if (err < 0) {
+		dev_err(dev, "failed to request resources: %d\n", err);
+		goto err_pm_put;
+	}
+
+	err = clk_prepare_enable(pcie->bus_clk);
+	if (err) {
+		dev_err(dev, "failed to enable bus clock: %d\n", err);
+		goto err_unmap_msi_irqs;
+	}
+
+	err = rcar_pcie_parse_map_dma_ranges(pcie, dev->of_node);
+	if (err)
+		goto err_clk_disable;
+
+	phy_init_fn = of_device_get_match_data(dev);
+	err = phy_init_fn(pcie);
+	if (err) {
+		dev_err(dev, "failed to init PCIe PHY\n");
+		goto err_clk_disable;
+	}
+
+	/* Failure to get a link might just be that no cards are inserted */
+	if (rcar_pcie_hw_init(pcie)) {
+		dev_info(dev, "PCIe link down\n");
+		err = -ENODEV;
+		goto err_phy_shutdown;
+	}
+
+	data = rcar_pci_read_reg(pcie->base, MACSR);
+	dev_info(dev, "PCIe x%d: link up\n", (data >> 20) & 0x3f);
+
+	if (IS_ENABLED(CONFIG_PCI_MSI)) {
+		err = rcar_pcie_enable_msi(pcie);
+		if (err < 0) {
+			dev_err(dev,
+				"failed to enable MSI support: %d\n",
+				err);
+			goto err_phy_shutdown;
+		}
+	}
+
+	err = rcar_pcie_enable(pcie);
+	if (err)
+		goto err_msi_teardown;
+
+	return 0;
+
+err_msi_teardown:
+	if (IS_ENABLED(CONFIG_PCI_MSI))
+		rcar_pcie_teardown_msi(pcie);
+
+err_phy_shutdown:
+	if (pcie->phy) {
+		phy_power_off(pcie->phy);
+		phy_exit(pcie->phy);
+	}
+
+err_clk_disable:
+	clk_disable_unprepare(pcie->bus_clk);
+
+err_unmap_msi_irqs:
+	irq_dispose_mapping(pcie->msi.irq2);
+	irq_dispose_mapping(pcie->msi.irq1);
+
+err_pm_put:
+	pm_runtime_put(dev);
+
+err_pm_disable:
+	pm_runtime_disable(dev);
+	pci_free_resource_list(&pcie->resources);
+
+err_free_bridge:
+	pci_free_host_bridge(bridge);
+
+	return err;
+}
+
+static int rcar_pcie_resume_noirq(struct device *dev)
+{
+	struct rcar_pcie *pcie = dev_get_drvdata(dev);
+
+	if (rcar_pci_read_reg(pcie->base, PMSR) &&
+	    !(rcar_pci_read_reg(pcie->base, PCIETCTLR) & DL_DOWN))
+		return 0;
+
+	/* Re-establish the PCIe link */
+	rcar_rmw32(pcie->base, MACCTLR, MACCTLR_RESERVED, 0);
+	rcar_pci_write_reg(pcie->base, CFINIT, PCIETCTLR);
+	return rcar_pcie_wait_for_dl(pcie->base);
+}
+
+static const struct dev_pm_ops rcar_pcie_pm_ops = {
+	.resume_noirq = rcar_pcie_resume_noirq,
+};
+
+static struct platform_driver rcar_pcie_driver = {
+	.driver = {
+		.name = "rcar-pcie",
+		.of_match_table = rcar_pcie_of_match,
+		.pm = &rcar_pcie_pm_ops,
+		.suppress_bind_attrs = true,
+	},
+	.probe = rcar_pcie_probe,
+};
+builtin_platform_driver(rcar_pcie_driver);
diff --git a/drivers/pci/controller/pcie-rcar.c b/drivers/pci/controller/pcie-rcar.c
index e45bb2a7bfa5..1008ae6a896c 100644
--- a/drivers/pci/controller/pcie-rcar.c
+++ b/drivers/pci/controller/pcie-rcar.c
@@ -1,7 +1,7 @@
 // SPDX-License-Identifier: GPL-2.0
 /*
  * PCIe driver for Renesas R-Car SoCs
- *  Copyright (C) 2014 Renesas Electronics Europe Ltd
+ *  Copyright (C) 2014-2019 Renesas Electronics Europe Ltd
  *
  * Based on:
  *  arch/sh/drivers/pci/pcie-sh7786.c
@@ -11,534 +11,80 @@
  * Author: Phil Edworthy <phil.edworthy@renesas.com>
  */
 
-#include <linux/bitops.h>
-#include <linux/clk.h>
 #include <linux/delay.h>
-#include <linux/interrupt.h>
-#include <linux/irq.h>
-#include <linux/irqdomain.h>
-#include <linux/kernel.h>
-#include <linux/init.h>
-#include <linux/msi.h>
-#include <linux/of_address.h>
-#include <linux/of_irq.h>
-#include <linux/of_pci.h>
-#include <linux/of_platform.h>
 #include <linux/pci.h>
-#include <linux/phy/phy.h>
-#include <linux/platform_device.h>
-#include <linux/pm_runtime.h>
-#include <linux/slab.h>
 
-#define PCIECAR			0x000010
-#define PCIECCTLR		0x000018
-#define  CONFIG_SEND_ENABLE	BIT(31)
-#define  TYPE0			(0 << 8)
-#define  TYPE1			BIT(8)
-#define PCIECDR			0x000020
-#define PCIEMSR			0x000028
-#define PCIEINTXR		0x000400
-#define PCIEPHYSR		0x0007f0
-#define  PHYRDY			BIT(0)
-#define PCIEMSITXR		0x000840
+#include "pcie-rcar.h"
 
-/* Transfer control */
-#define PCIETCTLR		0x02000
-#define  DL_DOWN		BIT(3)
-#define  CFINIT			BIT(0)
-#define PCIETSTR		0x02004
-#define  DATA_LINK_ACTIVE	BIT(0)
-#define PCIEERRFR		0x02020
-#define  UNSUPPORTED_REQUEST	BIT(4)
-#define PCIEMSIFR		0x02044
-#define PCIEMSIALR		0x02048
-#define  MSIFE			BIT(0)
-#define PCIEMSIAUR		0x0204c
-#define PCIEMSIIER		0x02050
-
-/* root port address */
-#define PCIEPRAR(x)		(0x02080 + ((x) * 0x4))
-
-/* local address reg & mask */
-#define PCIELAR(x)		(0x02200 + ((x) * 0x20))
-#define PCIELAMR(x)		(0x02208 + ((x) * 0x20))
-#define  LAM_PREFETCH		BIT(3)
-#define  LAM_64BIT		BIT(2)
-#define  LAR_ENABLE		BIT(1)
-
-/* PCIe address reg & mask */
-#define PCIEPALR(x)		(0x03400 + ((x) * 0x20))
-#define PCIEPAUR(x)		(0x03404 + ((x) * 0x20))
-#define PCIEPAMR(x)		(0x03408 + ((x) * 0x20))
-#define PCIEPTCTLR(x)		(0x0340c + ((x) * 0x20))
-#define  PAR_ENABLE		BIT(31)
-#define  IO_SPACE		BIT(8)
-
-/* Configuration */
-#define PCICONF(x)		(0x010000 + ((x) * 0x4))
-#define PMCAP(x)		(0x010040 + ((x) * 0x4))
-#define EXPCAP(x)		(0x010070 + ((x) * 0x4))
-#define VCCAP(x)		(0x010100 + ((x) * 0x4))
-
-/* link layer */
-#define IDSETR1			0x011004
-#define TLCTLR			0x011048
-#define MACSR			0x011054
-#define  SPCHGFIN		BIT(4)
-#define  SPCHGFAIL		BIT(6)
-#define  SPCHGSUC		BIT(7)
-#define  LINK_SPEED		(0xf << 16)
-#define  LINK_SPEED_2_5GTS	(1 << 16)
-#define  LINK_SPEED_5_0GTS	(2 << 16)
-#define MACCTLR			0x011058
-#define  MACCTLR_RESERVED	BIT(0)
-#define  SPEED_CHANGE		BIT(24)
-#define  SCRAMBLE_DISABLE	BIT(27)
-#define PMSR			0x01105c
-#define MACS2R			0x011078
-#define MACCGSPSETR		0x011084
-#define  SPCNGRSN		BIT(31)
-
-/* R-Car H1 PHY */
-#define H1_PCIEPHYADRR		0x04000c
-#define  WRITE_CMD		BIT(16)
-#define  PHY_ACK		BIT(24)
-#define  RATE_POS		12
-#define  LANE_POS		8
-#define  ADR_POS		0
-#define H1_PCIEPHYDOUTR		0x040014
-
-/* R-Car Gen2 PHY */
-#define GEN2_PCIEPHYADDR	0x780
-#define GEN2_PCIEPHYDATA	0x784
-#define GEN2_PCIEPHYCTRL	0x78c
-
-#define INT_PCI_MSI_NR		32
-
-#define RCONF(x)		(PCICONF(0) + (x))
-#define RPMCAP(x)		(PMCAP(0) + (x))
-#define REXPCAP(x)		(EXPCAP(0) + (x))
-#define RVCCAP(x)		(VCCAP(0) + (x))
-
-#define PCIE_CONF_BUS(b)	(((b) & 0xff) << 24)
-#define PCIE_CONF_DEV(d)	(((d) & 0x1f) << 19)
-#define PCIE_CONF_FUNC(f)	(((f) & 0x7) << 16)
-
-#define RCAR_PCI_MAX_RESOURCES	4
-#define MAX_NR_INBOUND_MAPS	6
-
-struct rcar_msi {
-	DECLARE_BITMAP(used, INT_PCI_MSI_NR);
-	struct irq_domain *domain;
-	struct msi_controller chip;
-	unsigned long pages;
-	struct mutex lock;
-	int irq1;
-	int irq2;
-};
-
-static inline struct rcar_msi *to_rcar_msi(struct msi_controller *chip)
-{
-	return container_of(chip, struct rcar_msi, chip);
-}
-
-/* Structure representing the PCIe interface */
-struct rcar_pcie {
-	struct device		*dev;
-	struct phy		*phy;
-	void __iomem		*base;
-	struct list_head	resources;
-	int			root_bus_nr;
-	struct clk		*bus_clk;
-	struct			rcar_msi msi;
-};
-
-static void rcar_pci_write_reg(struct rcar_pcie *pcie, u32 val,
-			       unsigned int reg)
+void rcar_pci_write_reg(void __iomem *base, u32 val, unsigned int reg)
 {
-	writel(val, pcie->base + reg);
+	writel(val, base + reg);
 }
 
-static u32 rcar_pci_read_reg(struct rcar_pcie *pcie, unsigned int reg)
+u32 rcar_pci_read_reg(void __iomem *base, unsigned int reg)
 {
-	return readl(pcie->base + reg);
+	return readl(base + reg);
 }
 
-enum {
-	RCAR_PCI_ACCESS_READ,
-	RCAR_PCI_ACCESS_WRITE,
-};
-
-static void rcar_rmw32(struct rcar_pcie *pcie, int where, u32 mask, u32 data)
+void rcar_rmw32(void __iomem *base, int where, u32 mask, u32 data)
 {
 	unsigned int shift = BITS_PER_BYTE * (where & 3);
-	u32 val = rcar_pci_read_reg(pcie, where & ~3);
+	u32 val = rcar_pci_read_reg(base, where & ~3);
 
 	val &= ~(mask << shift);
 	val |= data << shift;
-	rcar_pci_write_reg(pcie, val, where & ~3);
-}
-
-static u32 rcar_read_conf(struct rcar_pcie *pcie, int where)
-{
-	unsigned int shift = BITS_PER_BYTE * (where & 3);
-	u32 val = rcar_pci_read_reg(pcie, where & ~3);
-
-	return val >> shift;
+	rcar_pci_write_reg(base, val, where & ~3);
 }
 
-/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
-static int rcar_pcie_config_access(struct rcar_pcie *pcie,
-		unsigned char access_type, struct pci_bus *bus,
-		unsigned int devfn, int where, u32 *data)
-{
-	unsigned int dev, func, reg, index;
-
-	dev = PCI_SLOT(devfn);
-	func = PCI_FUNC(devfn);
-	reg = where & ~3;
-	index = reg / 4;
-
-	/*
-	 * While each channel has its own memory-mapped extended config
-	 * space, it's generally only accessible when in endpoint mode.
-	 * When in root complex mode, the controller is unable to target
-	 * itself with either type 0 or type 1 accesses, and indeed, any
-	 * controller initiated target transfer to its own config space
-	 * result in a completer abort.
-	 *
-	 * Each channel effectively only supports a single device, but as
-	 * the same channel <-> device access works for any PCI_SLOT()
-	 * value, we cheat a bit here and bind the controller's config
-	 * space to devfn 0 in order to enable self-enumeration. In this
-	 * case the regular ECAR/ECDR path is sidelined and the mangled
-	 * config access itself is initiated as an internal bus transaction.
-	 */
-	if (pci_is_root_bus(bus)) {
-		if (dev != 0)
-			return PCIBIOS_DEVICE_NOT_FOUND;
-
-		if (access_type == RCAR_PCI_ACCESS_READ) {
-			*data = rcar_pci_read_reg(pcie, PCICONF(index));
-		} else {
-			/* Keep an eye out for changes to the root bus number */
-			if (pci_is_root_bus(bus) && (reg == PCI_PRIMARY_BUS))
-				pcie->root_bus_nr = *data & 0xff;
-
-			rcar_pci_write_reg(pcie, *data, PCICONF(index));
-		}
-
-		return PCIBIOS_SUCCESSFUL;
-	}
-
-	if (pcie->root_bus_nr < 0)
-		return PCIBIOS_DEVICE_NOT_FOUND;
-
-	/* Clear errors */
-	rcar_pci_write_reg(pcie, rcar_pci_read_reg(pcie, PCIEERRFR), PCIEERRFR);
-
-	/* Set the PIO address */
-	rcar_pci_write_reg(pcie, PCIE_CONF_BUS(bus->number) |
-		PCIE_CONF_DEV(dev) | PCIE_CONF_FUNC(func) | reg, PCIECAR);
-
-	/* Enable the configuration access */
-	if (bus->parent->number == pcie->root_bus_nr)
-		rcar_pci_write_reg(pcie, CONFIG_SEND_ENABLE | TYPE0, PCIECCTLR);
-	else
-		rcar_pci_write_reg(pcie, CONFIG_SEND_ENABLE | TYPE1, PCIECCTLR);
-
-	/* Check for errors */
-	if (rcar_pci_read_reg(pcie, PCIEERRFR) & UNSUPPORTED_REQUEST)
-		return PCIBIOS_DEVICE_NOT_FOUND;
-
-	/* Check for master and target aborts */
-	if (rcar_read_conf(pcie, RCONF(PCI_STATUS)) &
-		(PCI_STATUS_REC_MASTER_ABORT | PCI_STATUS_REC_TARGET_ABORT))
-		return PCIBIOS_DEVICE_NOT_FOUND;
-
-	if (access_type == RCAR_PCI_ACCESS_READ)
-		*data = rcar_pci_read_reg(pcie, PCIECDR);
-	else
-		rcar_pci_write_reg(pcie, *data, PCIECDR);
-
-	/* Disable the configuration access */
-	rcar_pci_write_reg(pcie, 0, PCIECCTLR);
-
-	return PCIBIOS_SUCCESSFUL;
-}
-
-static int rcar_pcie_read_conf(struct pci_bus *bus, unsigned int devfn,
-			       int where, int size, u32 *val)
-{
-	struct rcar_pcie *pcie = bus->sysdata;
-	int ret;
-
-	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
-				      bus, devfn, where, val);
-	if (ret != PCIBIOS_SUCCESSFUL) {
-		*val = 0xffffffff;
-		return ret;
-	}
-
-	if (size == 1)
-		*val = (*val >> (BITS_PER_BYTE * (where & 3))) & 0xff;
-	else if (size == 2)
-		*val = (*val >> (BITS_PER_BYTE * (where & 2))) & 0xffff;
-
-	dev_dbg(&bus->dev, "pcie-config-read: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
-		bus->number, devfn, where, size, *val);
-
-	return ret;
-}
-
-/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
-static int rcar_pcie_write_conf(struct pci_bus *bus, unsigned int devfn,
-				int where, int size, u32 val)
-{
-	struct rcar_pcie *pcie = bus->sysdata;
-	unsigned int shift;
-	u32 data;
-	int ret;
-
-	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
-				      bus, devfn, where, &data);
-	if (ret != PCIBIOS_SUCCESSFUL)
-		return ret;
-
-	dev_dbg(&bus->dev, "pcie-config-write: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
-		bus->number, devfn, where, size, val);
-
-	if (size == 1) {
-		shift = BITS_PER_BYTE * (where & 3);
-		data &= ~(0xff << shift);
-		data |= ((val & 0xff) << shift);
-	} else if (size == 2) {
-		shift = BITS_PER_BYTE * (where & 2);
-		data &= ~(0xffff << shift);
-		data |= ((val & 0xffff) << shift);
-	} else
-		data = val;
-
-	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_WRITE,
-				      bus, devfn, where, &data);
-
-	return ret;
-}
-
-static struct pci_ops rcar_pcie_ops = {
-	.read	= rcar_pcie_read_conf,
-	.write	= rcar_pcie_write_conf,
-};
-
-static void rcar_pcie_setup_window(int win, struct rcar_pcie *pcie,
-				   struct resource *res)
+void rcar_pcie_set_outbound(int win, void __iomem *base,
+			    struct resource *res, bool host)
 {
 	/* Setup PCIe address space mappings for each resource */
-	resource_size_t size;
-	resource_size_t res_start;
-	u32 mask;
-
-	rcar_pci_write_reg(pcie, 0x00000000, PCIEPTCTLR(win));
+	resource_size_t size = 0;
+	resource_size_t res_start = 0;
+	u32 mask = 0x0;
 
+	rcar_pci_write_reg(base, mask, PCIEPTCTLR(win));
 	/*
 	 * The PAMR mask is calculated in units of 128Bytes, which
 	 * keeps things pretty simple.
 	 */
 	size = resource_size(res);
-	mask = (roundup_pow_of_two(size) / SZ_128) - 1;
-	rcar_pci_write_reg(pcie, mask << 7, PCIEPAMR(win));
+	if (size > 128)
+		mask = (roundup_pow_of_two(size) / SZ_128) - 1;
+	rcar_pci_write_reg(base, mask << 7, PCIEPAMR(win));
 
-	if (res->flags & IORESOURCE_IO)
-		res_start = pci_pio_to_address(res->start);
-	else
+	if (!host) {
 		res_start = res->start;
+	} else {
+		if (res->flags & IORESOURCE_IO)
+			res_start = pci_pio_to_address(res->start);
+		else
+			res_start = res->start;
+	}
 
-	rcar_pci_write_reg(pcie, upper_32_bits(res_start), PCIEPAUR(win));
-	rcar_pci_write_reg(pcie, lower_32_bits(res_start) & ~0x7F,
+	rcar_pci_write_reg(base, upper_32_bits(res_start), PCIEPAUR(win));
+	rcar_pci_write_reg(base, lower_32_bits(res_start) & ~0x7F,
 			   PCIEPALR(win));
 
-	/* First resource is for IO */
-	mask = PAR_ENABLE;
+	mask = 0x0;
+	if (res->start)
+		mask = PAR_ENABLE;
+
 	if (res->flags & IORESOURCE_IO)
 		mask |= IO_SPACE;
 
-	rcar_pci_write_reg(pcie, mask, PCIEPTCTLR(win));
-}
-
-static int rcar_pcie_setup(struct list_head *resource, struct rcar_pcie *pci)
-{
-	struct resource_entry *win;
-	int i = 0;
-
-	/* Setup PCI resources */
-	resource_list_for_each_entry(win, &pci->resources) {
-		struct resource *res = win->res;
-
-		if (!res->flags)
-			continue;
-
-		switch (resource_type(res)) {
-		case IORESOURCE_IO:
-		case IORESOURCE_MEM:
-			rcar_pcie_setup_window(i, pci, res);
-			i++;
-			break;
-		case IORESOURCE_BUS:
-			pci->root_bus_nr = res->start;
-			break;
-		default:
-			continue;
-		}
-
-		pci_add_resource(resource, res);
-	}
-
-	return 1;
-}
-
-static void rcar_pcie_force_speedup(struct rcar_pcie *pcie)
-{
-	struct device *dev = pcie->dev;
-	unsigned int timeout = 1000;
-	u32 macsr;
-
-	if ((rcar_pci_read_reg(pcie, MACS2R) & LINK_SPEED) != LINK_SPEED_5_0GTS)
-		return;
-
-	if (rcar_pci_read_reg(pcie, MACCTLR) & SPEED_CHANGE) {
-		dev_err(dev, "Speed change already in progress\n");
-		return;
-	}
-
-	macsr = rcar_pci_read_reg(pcie, MACSR);
-	if ((macsr & LINK_SPEED) == LINK_SPEED_5_0GTS)
-		goto done;
-
-	/* Set target link speed to 5.0 GT/s */
-	rcar_rmw32(pcie, EXPCAP(12), PCI_EXP_LNKSTA_CLS,
-		   PCI_EXP_LNKSTA_CLS_5_0GB);
-
-	/* Set speed change reason as intentional factor */
-	rcar_rmw32(pcie, MACCGSPSETR, SPCNGRSN, 0);
-
-	/* Clear SPCHGFIN, SPCHGSUC, and SPCHGFAIL */
-	if (macsr & (SPCHGFIN | SPCHGSUC | SPCHGFAIL))
-		rcar_pci_write_reg(pcie, macsr, MACSR);
-
-	/* Start link speed change */
-	rcar_rmw32(pcie, MACCTLR, SPEED_CHANGE, SPEED_CHANGE);
-
-	while (timeout--) {
-		macsr = rcar_pci_read_reg(pcie, MACSR);
-		if (macsr & SPCHGFIN) {
-			/* Clear the interrupt bits */
-			rcar_pci_write_reg(pcie, macsr, MACSR);
-
-			if (macsr & SPCHGFAIL)
-				dev_err(dev, "Speed change failed\n");
-
-			goto done;
-		}
-
-		msleep(1);
-	}
-
-	dev_err(dev, "Speed change timed out\n");
-
-done:
-	dev_info(dev, "Current link speed is %s GT/s\n",
-		 (macsr & LINK_SPEED) == LINK_SPEED_5_0GTS ? "5" : "2.5");
-}
-
-static int rcar_pcie_enable(struct rcar_pcie *pcie)
-{
-	struct device *dev = pcie->dev;
-	struct pci_host_bridge *bridge = pci_host_bridge_from_priv(pcie);
-	struct pci_bus *bus, *child;
-	int ret;
-
-	/* Try setting 5 GT/s link speed */
-	rcar_pcie_force_speedup(pcie);
-
-	rcar_pcie_setup(&bridge->windows, pcie);
-
-	pci_add_flags(PCI_REASSIGN_ALL_BUS);
-
-	bridge->dev.parent = dev;
-	bridge->sysdata = pcie;
-	bridge->busnr = pcie->root_bus_nr;
-	bridge->ops = &rcar_pcie_ops;
-	bridge->map_irq = of_irq_parse_and_map_pci;
-	bridge->swizzle_irq = pci_common_swizzle;
-	if (IS_ENABLED(CONFIG_PCI_MSI))
-		bridge->msi = &pcie->msi.chip;
-
-	ret = pci_scan_root_bus_bridge(bridge);
-	if (ret < 0)
-		return ret;
-
-	bus = bridge->bus;
-
-	pci_bus_size_bridges(bus);
-	pci_bus_assign_resources(bus);
-
-	list_for_each_entry(child, &bus->children, node)
-		pcie_bus_configure_settings(child);
-
-	pci_bus_add_devices(bus);
-
-	return 0;
-}
-
-static int phy_wait_for_ack(struct rcar_pcie *pcie)
-{
-	struct device *dev = pcie->dev;
-	unsigned int timeout = 100;
-
-	while (timeout--) {
-		if (rcar_pci_read_reg(pcie, H1_PCIEPHYADRR) & PHY_ACK)
-			return 0;
-
-		udelay(100);
-	}
-
-	dev_err(dev, "Access to PCIe phy timed out\n");
-
-	return -ETIMEDOUT;
-}
-
-static void phy_write_reg(struct rcar_pcie *pcie,
-			  unsigned int rate, u32 addr,
-			  unsigned int lane, u32 data)
-{
-	u32 phyaddr;
-
-	phyaddr = WRITE_CMD |
-		((rate & 1) << RATE_POS) |
-		((lane & 0xf) << LANE_POS) |
-		((addr & 0xff) << ADR_POS);
-
-	/* Set write data */
-	rcar_pci_write_reg(pcie, data, H1_PCIEPHYDOUTR);
-	rcar_pci_write_reg(pcie, phyaddr, H1_PCIEPHYADRR);
-
-	/* Ignore errors as they will be dealt with if the data link is down */
-	phy_wait_for_ack(pcie);
-
-	/* Clear command */
-	rcar_pci_write_reg(pcie, 0, H1_PCIEPHYDOUTR);
-	rcar_pci_write_reg(pcie, 0, H1_PCIEPHYADRR);
-
-	/* Ignore errors as they will be dealt with if the data link is down */
-	phy_wait_for_ack(pcie);
+	rcar_pci_write_reg(base, mask, PCIEPTCTLR(win));
+	/* flush modifications */
+	wmb();
 }
 
-static int rcar_pcie_wait_for_phyrdy(struct rcar_pcie *pcie)
+int rcar_pcie_wait_for_phyrdy(void __iomem *base)
 {
 	unsigned int timeout = 10;
 
 	while (timeout--) {
-		if (rcar_pci_read_reg(pcie, PCIEPHYSR) & PHYRDY)
+		if (rcar_pci_read_reg(base, PCIEPHYSR) & PHYRDY)
 			return 0;
 
 		msleep(5);
@@ -547,12 +93,12 @@ static int rcar_pcie_wait_for_phyrdy(struct rcar_pcie *pcie)
 	return -ETIMEDOUT;
 }
 
-static int rcar_pcie_wait_for_dl(struct rcar_pcie *pcie)
+int rcar_pcie_wait_for_dl(void __iomem *base)
 {
 	unsigned int timeout = 10000;
 
 	while (timeout--) {
-		if ((rcar_pci_read_reg(pcie, PCIETSTR) & DATA_LINK_ACTIVE))
+		if ((rcar_pci_read_reg(base, PCIETSTR) & DATA_LINK_ACTIVE))
 			return 0;
 
 		udelay(5);
@@ -562,698 +108,27 @@ static int rcar_pcie_wait_for_dl(struct rcar_pcie *pcie)
 	return -ETIMEDOUT;
 }
 
-static int rcar_pcie_hw_init(struct rcar_pcie *pcie)
-{
-	int err;
-
-	/* Begin initialization */
-	rcar_pci_write_reg(pcie, 0, PCIETCTLR);
-
-	/* Set mode */
-	rcar_pci_write_reg(pcie, 1, PCIEMSR);
-
-	err = rcar_pcie_wait_for_phyrdy(pcie);
-	if (err)
-		return err;
-
-	/*
-	 * Initial header for port config space is type 1, set the device
-	 * class to match. Hardware takes care of propagating the IDSETR
-	 * settings, so there is no need to bother with a quirk.
-	 */
-	rcar_pci_write_reg(pcie, PCI_CLASS_BRIDGE_PCI << 16, IDSETR1);
-
-	/*
-	 * Setup Secondary Bus Number & Subordinate Bus Number, even though
-	 * they aren't used, to avoid bridge being detected as broken.
-	 */
-	rcar_rmw32(pcie, RCONF(PCI_SECONDARY_BUS), 0xff, 1);
-	rcar_rmw32(pcie, RCONF(PCI_SUBORDINATE_BUS), 0xff, 1);
-
-	/* Initialize default capabilities. */
-	rcar_rmw32(pcie, REXPCAP(0), 0xff, PCI_CAP_ID_EXP);
-	rcar_rmw32(pcie, REXPCAP(PCI_EXP_FLAGS),
-		PCI_EXP_FLAGS_TYPE, PCI_EXP_TYPE_ROOT_PORT << 4);
-	rcar_rmw32(pcie, RCONF(PCI_HEADER_TYPE), 0x7f,
-		PCI_HEADER_TYPE_BRIDGE);
-
-	/* Enable data link layer active state reporting */
-	rcar_rmw32(pcie, REXPCAP(PCI_EXP_LNKCAP), PCI_EXP_LNKCAP_DLLLARC,
-		PCI_EXP_LNKCAP_DLLLARC);
-
-	/* Write out the physical slot number = 0 */
-	rcar_rmw32(pcie, REXPCAP(PCI_EXP_SLTCAP), PCI_EXP_SLTCAP_PSN, 0);
-
-	/* Set the completion timer timeout to the maximum 50ms. */
-	rcar_rmw32(pcie, TLCTLR + 1, 0x3f, 50);
-
-	/* Terminate list of capabilities (Next Capability Offset=0) */
-	rcar_rmw32(pcie, RVCCAP(0), 0xfff00000, 0);
-
-	/* Enable MSI */
-	if (IS_ENABLED(CONFIG_PCI_MSI))
-		rcar_pci_write_reg(pcie, 0x801f0000, PCIEMSITXR);
-
-	rcar_rmw32(pcie, MACCTLR, MACCTLR_RESERVED, 0);
-
-	/* Finish initialization - establish a PCI Express link */
-	rcar_pci_write_reg(pcie, CFINIT, PCIETCTLR);
-
-	/* This will timeout if we don't have a link. */
-	err = rcar_pcie_wait_for_dl(pcie);
-	if (err)
-		return err;
-
-	/* Enable INTx interrupts */
-	rcar_rmw32(pcie, PCIEINTXR, 0, 0xF << 8);
-
-	wmb();
-
-	return 0;
-}
-
-static int rcar_pcie_phy_init_h1(struct rcar_pcie *pcie)
-{
-	/* Initialize the phy */
-	phy_write_reg(pcie, 0, 0x42, 0x1, 0x0EC34191);
-	phy_write_reg(pcie, 1, 0x42, 0x1, 0x0EC34180);
-	phy_write_reg(pcie, 0, 0x43, 0x1, 0x00210188);
-	phy_write_reg(pcie, 1, 0x43, 0x1, 0x00210188);
-	phy_write_reg(pcie, 0, 0x44, 0x1, 0x015C0014);
-	phy_write_reg(pcie, 1, 0x44, 0x1, 0x015C0014);
-	phy_write_reg(pcie, 1, 0x4C, 0x1, 0x786174A0);
-	phy_write_reg(pcie, 1, 0x4D, 0x1, 0x048000BB);
-	phy_write_reg(pcie, 0, 0x51, 0x1, 0x079EC062);
-	phy_write_reg(pcie, 0, 0x52, 0x1, 0x20000000);
-	phy_write_reg(pcie, 1, 0x52, 0x1, 0x20000000);
-	phy_write_reg(pcie, 1, 0x56, 0x1, 0x00003806);
-
-	phy_write_reg(pcie, 0, 0x60, 0x1, 0x004B03A5);
-	phy_write_reg(pcie, 0, 0x64, 0x1, 0x3F0F1F0F);
-	phy_write_reg(pcie, 0, 0x66, 0x1, 0x00008000);
-
-	return 0;
-}
-
-static int rcar_pcie_phy_init_gen2(struct rcar_pcie *pcie)
-{
-	/*
-	 * These settings come from the R-Car Series, 2nd Generation User's
-	 * Manual, section 50.3.1 (2) Initialization of the physical layer.
-	 */
-	rcar_pci_write_reg(pcie, 0x000f0030, GEN2_PCIEPHYADDR);
-	rcar_pci_write_reg(pcie, 0x00381203, GEN2_PCIEPHYDATA);
-	rcar_pci_write_reg(pcie, 0x00000001, GEN2_PCIEPHYCTRL);
-	rcar_pci_write_reg(pcie, 0x00000006, GEN2_PCIEPHYCTRL);
-
-	rcar_pci_write_reg(pcie, 0x000f0054, GEN2_PCIEPHYADDR);
-	/* The following value is for DC connection, no termination resistor */
-	rcar_pci_write_reg(pcie, 0x13802007, GEN2_PCIEPHYDATA);
-	rcar_pci_write_reg(pcie, 0x00000001, GEN2_PCIEPHYCTRL);
-	rcar_pci_write_reg(pcie, 0x00000006, GEN2_PCIEPHYCTRL);
-
-	return 0;
-}
-
-static int rcar_pcie_phy_init_gen3(struct rcar_pcie *pcie)
-{
-	int err;
-
-	err = phy_init(pcie->phy);
-	if (err)
-		return err;
-
-	err = phy_power_on(pcie->phy);
-	if (err)
-		phy_exit(pcie->phy);
-
-	return err;
-}
-
-static int rcar_msi_alloc(struct rcar_msi *chip)
-{
-	int msi;
-
-	mutex_lock(&chip->lock);
-
-	msi = find_first_zero_bit(chip->used, INT_PCI_MSI_NR);
-	if (msi < INT_PCI_MSI_NR)
-		set_bit(msi, chip->used);
-	else
-		msi = -ENOSPC;
-
-	mutex_unlock(&chip->lock);
-
-	return msi;
-}
-
-static int rcar_msi_alloc_region(struct rcar_msi *chip, int no_irqs)
-{
-	int msi;
-
-	mutex_lock(&chip->lock);
-	msi = bitmap_find_free_region(chip->used, INT_PCI_MSI_NR,
-				      order_base_2(no_irqs));
-	mutex_unlock(&chip->lock);
-
-	return msi;
-}
-
-static void rcar_msi_free(struct rcar_msi *chip, unsigned long irq)
-{
-	mutex_lock(&chip->lock);
-	clear_bit(irq, chip->used);
-	mutex_unlock(&chip->lock);
-}
-
-static irqreturn_t rcar_pcie_msi_irq(int irq, void *data)
-{
-	struct rcar_pcie *pcie = data;
-	struct rcar_msi *msi = &pcie->msi;
-	struct device *dev = pcie->dev;
-	unsigned long reg;
-
-	reg = rcar_pci_read_reg(pcie, PCIEMSIFR);
-
-	/* MSI & INTx share an interrupt - we only handle MSI here */
-	if (!reg)
-		return IRQ_NONE;
-
-	while (reg) {
-		unsigned int index = find_first_bit(&reg, 32);
-		unsigned int msi_irq;
-
-		/* clear the interrupt */
-		rcar_pci_write_reg(pcie, 1 << index, PCIEMSIFR);
-
-		msi_irq = irq_find_mapping(msi->domain, index);
-		if (msi_irq) {
-			if (test_bit(index, msi->used))
-				generic_handle_irq(msi_irq);
-			else
-				dev_info(dev, "unhandled MSI\n");
-		} else {
-			/* Unknown MSI, just clear it */
-			dev_dbg(dev, "unexpected MSI\n");
-		}
-
-		/* see if there's any more pending in this vector */
-		reg = rcar_pci_read_reg(pcie, PCIEMSIFR);
-	}
-
-	return IRQ_HANDLED;
-}
-
-static int rcar_msi_setup_irq(struct msi_controller *chip, struct pci_dev *pdev,
-			      struct msi_desc *desc)
-{
-	struct rcar_msi *msi = to_rcar_msi(chip);
-	struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
-	struct msi_msg msg;
-	unsigned int irq;
-	int hwirq;
-
-	hwirq = rcar_msi_alloc(msi);
-	if (hwirq < 0)
-		return hwirq;
-
-	irq = irq_find_mapping(msi->domain, hwirq);
-	if (!irq) {
-		rcar_msi_free(msi, hwirq);
-		return -EINVAL;
-	}
-
-	irq_set_msi_desc(irq, desc);
-
-	msg.address_lo = rcar_pci_read_reg(pcie, PCIEMSIALR) & ~MSIFE;
-	msg.address_hi = rcar_pci_read_reg(pcie, PCIEMSIAUR);
-	msg.data = hwirq;
-
-	pci_write_msi_msg(irq, &msg);
-
-	return 0;
-}
-
-static int rcar_msi_setup_irqs(struct msi_controller *chip,
-			       struct pci_dev *pdev, int nvec, int type)
-{
-	struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
-	struct rcar_msi *msi = to_rcar_msi(chip);
-	struct msi_desc *desc;
-	struct msi_msg msg;
-	unsigned int irq;
-	int hwirq;
-	int i;
-
-	/* MSI-X interrupts are not supported */
-	if (type == PCI_CAP_ID_MSIX)
-		return -EINVAL;
-
-	WARN_ON(!list_is_singular(&pdev->dev.msi_list));
-	desc = list_entry(pdev->dev.msi_list.next, struct msi_desc, list);
-
-	hwirq = rcar_msi_alloc_region(msi, nvec);
-	if (hwirq < 0)
-		return -ENOSPC;
-
-	irq = irq_find_mapping(msi->domain, hwirq);
-	if (!irq)
-		return -ENOSPC;
-
-	for (i = 0; i < nvec; i++) {
-		/*
-		 * irq_create_mapping() called from rcar_pcie_probe() pre-
-		 * allocates descs,  so there is no need to allocate descs here.
-		 * We can therefore assume that if irq_find_mapping() above
-		 * returns non-zero, then the descs are also successfully
-		 * allocated.
-		 */
-		if (irq_set_msi_desc_off(irq, i, desc)) {
-			/* TODO: clear */
-			return -EINVAL;
-		}
-	}
-
-	desc->nvec_used = nvec;
-	desc->msi_attrib.multiple = order_base_2(nvec);
-
-	msg.address_lo = rcar_pci_read_reg(pcie, PCIEMSIALR) & ~MSIFE;
-	msg.address_hi = rcar_pci_read_reg(pcie, PCIEMSIAUR);
-	msg.data = hwirq;
-
-	pci_write_msi_msg(irq, &msg);
-
-	return 0;
-}
-
-static void rcar_msi_teardown_irq(struct msi_controller *chip, unsigned int irq)
-{
-	struct rcar_msi *msi = to_rcar_msi(chip);
-	struct irq_data *d = irq_get_irq_data(irq);
-
-	rcar_msi_free(msi, d->hwirq);
-}
-
-static struct irq_chip rcar_msi_irq_chip = {
-	.name = "R-Car PCIe MSI",
-	.irq_enable = pci_msi_unmask_irq,
-	.irq_disable = pci_msi_mask_irq,
-	.irq_mask = pci_msi_mask_irq,
-	.irq_unmask = pci_msi_unmask_irq,
-};
-
-static int rcar_msi_map(struct irq_domain *domain, unsigned int irq,
-			irq_hw_number_t hwirq)
-{
-	irq_set_chip_and_handler(irq, &rcar_msi_irq_chip, handle_simple_irq);
-	irq_set_chip_data(irq, domain->host_data);
-
-	return 0;
-}
-
-static const struct irq_domain_ops msi_domain_ops = {
-	.map = rcar_msi_map,
-};
-
-static void rcar_pcie_unmap_msi(struct rcar_pcie *pcie)
-{
-	struct rcar_msi *msi = &pcie->msi;
-	int i, irq;
-
-	for (i = 0; i < INT_PCI_MSI_NR; i++) {
-		irq = irq_find_mapping(msi->domain, i);
-		if (irq > 0)
-			irq_dispose_mapping(irq);
-	}
-
-	irq_domain_remove(msi->domain);
-}
-
-static int rcar_pcie_enable_msi(struct rcar_pcie *pcie)
-{
-	struct device *dev = pcie->dev;
-	struct rcar_msi *msi = &pcie->msi;
-	phys_addr_t base;
-	int err, i;
-
-	mutex_init(&msi->lock);
-
-	msi->chip.dev = dev;
-	msi->chip.setup_irq = rcar_msi_setup_irq;
-	msi->chip.setup_irqs = rcar_msi_setup_irqs;
-	msi->chip.teardown_irq = rcar_msi_teardown_irq;
-
-	msi->domain = irq_domain_add_linear(dev->of_node, INT_PCI_MSI_NR,
-					    &msi_domain_ops, &msi->chip);
-	if (!msi->domain) {
-		dev_err(dev, "failed to create IRQ domain\n");
-		return -ENOMEM;
-	}
-
-	for (i = 0; i < INT_PCI_MSI_NR; i++)
-		irq_create_mapping(msi->domain, i);
-
-	/* Two irqs are for MSI, but they are also used for non-MSI irqs */
-	err = devm_request_irq(dev, msi->irq1, rcar_pcie_msi_irq,
-			       IRQF_SHARED | IRQF_NO_THREAD,
-			       rcar_msi_irq_chip.name, pcie);
-	if (err < 0) {
-		dev_err(dev, "failed to request IRQ: %d\n", err);
-		goto err;
-	}
-
-	err = devm_request_irq(dev, msi->irq2, rcar_pcie_msi_irq,
-			       IRQF_SHARED | IRQF_NO_THREAD,
-			       rcar_msi_irq_chip.name, pcie);
-	if (err < 0) {
-		dev_err(dev, "failed to request IRQ: %d\n", err);
-		goto err;
-	}
-
-	/* setup MSI data target */
-	msi->pages = __get_free_pages(GFP_KERNEL, 0);
-	if (!msi->pages) {
-		err = -ENOMEM;
-		goto err;
-	}
-	base = virt_to_phys((void *)msi->pages);
-
-	rcar_pci_write_reg(pcie, lower_32_bits(base) | MSIFE, PCIEMSIALR);
-	rcar_pci_write_reg(pcie, upper_32_bits(base), PCIEMSIAUR);
-
-	/* enable all MSI interrupts */
-	rcar_pci_write_reg(pcie, 0xffffffff, PCIEMSIIER);
-
-	return 0;
-
-err:
-	rcar_pcie_unmap_msi(pcie);
-	return err;
-}
-
-static void rcar_pcie_teardown_msi(struct rcar_pcie *pcie)
-{
-	struct rcar_msi *msi = &pcie->msi;
-
-	/* Disable all MSI interrupts */
-	rcar_pci_write_reg(pcie, 0, PCIEMSIIER);
-
-	/* Disable address decoding of the MSI interrupt, MSIFE */
-	rcar_pci_write_reg(pcie, 0, PCIEMSIALR);
-
-	free_pages(msi->pages, 0);
-
-	rcar_pcie_unmap_msi(pcie);
-}
-
-static int rcar_pcie_get_resources(struct rcar_pcie *pcie)
-{
-	struct device *dev = pcie->dev;
-	struct resource res;
-	int err, i;
-
-	pcie->phy = devm_phy_optional_get(dev, "pcie");
-	if (IS_ERR(pcie->phy))
-		return PTR_ERR(pcie->phy);
-
-	err = of_address_to_resource(dev->of_node, 0, &res);
-	if (err)
-		return err;
-
-	pcie->base = devm_ioremap_resource(dev, &res);
-	if (IS_ERR(pcie->base))
-		return PTR_ERR(pcie->base);
-
-	pcie->bus_clk = devm_clk_get(dev, "pcie_bus");
-	if (IS_ERR(pcie->bus_clk)) {
-		dev_err(dev, "cannot get pcie bus clock\n");
-		return PTR_ERR(pcie->bus_clk);
-	}
-
-	i = irq_of_parse_and_map(dev->of_node, 0);
-	if (!i) {
-		dev_err(dev, "cannot get platform resources for msi interrupt\n");
-		err = -ENOENT;
-		goto err_irq1;
-	}
-	pcie->msi.irq1 = i;
-
-	i = irq_of_parse_and_map(dev->of_node, 1);
-	if (!i) {
-		dev_err(dev, "cannot get platform resources for msi interrupt\n");
-		err = -ENOENT;
-		goto err_irq2;
-	}
-	pcie->msi.irq2 = i;
-
-	return 0;
-
-err_irq2:
-	irq_dispose_mapping(pcie->msi.irq1);
-err_irq1:
-	return err;
-}
-
-static int rcar_pcie_inbound_ranges(struct rcar_pcie *pcie,
-				    struct of_pci_range *range,
-				    int *index)
+void rcar_pcie_set_inbound(void __iomem *base,
+			   u64 cpu_addr, u64 pci_addr,
+			   u64 mask, int idx, bool host)
 {
-	u64 restype = range->flags;
-	u64 cpu_addr = range->cpu_addr;
-	u64 cpu_end = range->cpu_addr + range->size;
-	u64 pci_addr = range->pci_addr;
-	u32 flags = LAM_64BIT | LAR_ENABLE;
-	u64 mask;
-	u64 size;
-	int idx = *index;
-
-	if (restype & IORESOURCE_PREFETCH)
-		flags |= LAM_PREFETCH;
-
 	/*
-	 * If the size of the range is larger than the alignment of the start
-	 * address, we have to use multiple entries to perform the mapping.
-	 */
-	if (cpu_addr > 0) {
-		unsigned long nr_zeros = __ffs64(cpu_addr);
-		u64 alignment = 1ULL << nr_zeros;
-
-		size = min(range->size, alignment);
-	} else {
-		size = range->size;
-	}
-	/* Hardware supports max 4GiB inbound region */
-	size = min(size, 1ULL << 32);
-
-	mask = roundup_pow_of_two(size) - 1;
-	mask &= ~0xf;
-
-	while (cpu_addr < cpu_end) {
-		/*
-		 * Set up 64-bit inbound regions as the range parser doesn't
-		 * distinguish between 32 and 64-bit types.
-		 */
-		rcar_pci_write_reg(pcie, lower_32_bits(pci_addr),
+	* Set up 64-bit inbound regions as the range parser doesn't
+	* distinguish between 32 and 64-bit types.
+	*/
+	if (host) {
+		rcar_pci_write_reg(base, lower_32_bits(pci_addr),
 				   PCIEPRAR(idx));
-		rcar_pci_write_reg(pcie, lower_32_bits(cpu_addr), PCIELAR(idx));
-		rcar_pci_write_reg(pcie, lower_32_bits(mask) | flags,
-				   PCIELAMR(idx));
-
-		rcar_pci_write_reg(pcie, upper_32_bits(pci_addr),
+		rcar_pci_write_reg(base, upper_32_bits(pci_addr),
 				   PCIEPRAR(idx + 1));
-		rcar_pci_write_reg(pcie, upper_32_bits(cpu_addr),
-				   PCIELAR(idx + 1));
-		rcar_pci_write_reg(pcie, 0, PCIELAMR(idx + 1));
-
-		pci_addr += size;
-		cpu_addr += size;
-		idx += 2;
-
-		if (idx > MAX_NR_INBOUND_MAPS) {
-			dev_err(pcie->dev, "Failed to map inbound regions!\n");
-			return -EINVAL;
-		}
 	}
-	*index = idx;
-
-	return 0;
-}
 
-static int rcar_pcie_parse_map_dma_ranges(struct rcar_pcie *pcie,
-					  struct device_node *np)
-{
-	struct of_pci_range range;
-	struct of_pci_range_parser parser;
-	int index = 0;
-	int err;
-
-	if (of_pci_dma_range_parser_init(&parser, np))
-		return -EINVAL;
-
-	/* Get the dma-ranges from DT */
-	for_each_of_pci_range(&parser, &range) {
-		u64 end = range.cpu_addr + range.size - 1;
+	rcar_pci_write_reg(base, lower_32_bits(cpu_addr), PCIELAR(idx));
+	rcar_pci_write_reg(base, lower_32_bits(mask), PCIELAMR(idx));
 
-		dev_dbg(pcie->dev, "0x%08x 0x%016llx..0x%016llx -> 0x%016llx\n",
-			range.flags, range.cpu_addr, end, range.pci_addr);
-
-		err = rcar_pcie_inbound_ranges(pcie, &range, &index);
-		if (err)
-			return err;
-	}
+	rcar_pci_write_reg(base, upper_32_bits(cpu_addr), PCIELAR(idx + 1));
+	rcar_pci_write_reg(base, 0, PCIELAMR(idx + 1));
 
-	return 0;
-}
-
-static const struct of_device_id rcar_pcie_of_match[] = {
-	{ .compatible = "renesas,pcie-r8a7779",
-	  .data = rcar_pcie_phy_init_h1 },
-	{ .compatible = "renesas,pcie-r8a7790",
-	  .data = rcar_pcie_phy_init_gen2 },
-	{ .compatible = "renesas,pcie-r8a7791",
-	  .data = rcar_pcie_phy_init_gen2 },
-	{ .compatible = "renesas,pcie-rcar-gen2",
-	  .data = rcar_pcie_phy_init_gen2 },
-	{ .compatible = "renesas,pcie-r8a7795",
-	  .data = rcar_pcie_phy_init_gen3 },
-	{ .compatible = "renesas,pcie-rcar-gen3",
-	  .data = rcar_pcie_phy_init_gen3 },
-	{},
-};
-
-static int rcar_pcie_probe(struct platform_device *pdev)
-{
-	struct device *dev = &pdev->dev;
-	struct rcar_pcie *pcie;
-	u32 data;
-	int err;
-	int (*phy_init_fn)(struct rcar_pcie *);
-	struct pci_host_bridge *bridge;
-
-	bridge = pci_alloc_host_bridge(sizeof(*pcie));
-	if (!bridge)
-		return -ENOMEM;
-
-	pcie = pci_host_bridge_priv(bridge);
-
-	pcie->dev = dev;
-	platform_set_drvdata(pdev, pcie);
-
-	err = pci_parse_request_of_pci_ranges(dev, &pcie->resources, NULL);
-	if (err)
-		goto err_free_bridge;
-
-	pm_runtime_enable(pcie->dev);
-	err = pm_runtime_get_sync(pcie->dev);
-	if (err < 0) {
-		dev_err(pcie->dev, "pm_runtime_get_sync failed\n");
-		goto err_pm_disable;
-	}
-
-	err = rcar_pcie_get_resources(pcie);
-	if (err < 0) {
-		dev_err(dev, "failed to request resources: %d\n", err);
-		goto err_pm_put;
-	}
-
-	err = clk_prepare_enable(pcie->bus_clk);
-	if (err) {
-		dev_err(dev, "failed to enable bus clock: %d\n", err);
-		goto err_unmap_msi_irqs;
-	}
-
-	err = rcar_pcie_parse_map_dma_ranges(pcie, dev->of_node);
-	if (err)
-		goto err_clk_disable;
-
-	phy_init_fn = of_device_get_match_data(dev);
-	err = phy_init_fn(pcie);
-	if (err) {
-		dev_err(dev, "failed to init PCIe PHY\n");
-		goto err_clk_disable;
-	}
-
-	/* Failure to get a link might just be that no cards are inserted */
-	if (rcar_pcie_hw_init(pcie)) {
-		dev_info(dev, "PCIe link down\n");
-		err = -ENODEV;
-		goto err_phy_shutdown;
-	}
-
-	data = rcar_pci_read_reg(pcie, MACSR);
-	dev_info(dev, "PCIe x%d: link up\n", (data >> 20) & 0x3f);
-
-	if (IS_ENABLED(CONFIG_PCI_MSI)) {
-		err = rcar_pcie_enable_msi(pcie);
-		if (err < 0) {
-			dev_err(dev,
-				"failed to enable MSI support: %d\n",
-				err);
-			goto err_phy_shutdown;
-		}
-	}
-
-	err = rcar_pcie_enable(pcie);
-	if (err)
-		goto err_msi_teardown;
-
-	return 0;
-
-err_msi_teardown:
-	if (IS_ENABLED(CONFIG_PCI_MSI))
-		rcar_pcie_teardown_msi(pcie);
-
-err_phy_shutdown:
-	if (pcie->phy) {
-		phy_power_off(pcie->phy);
-		phy_exit(pcie->phy);
-	}
-
-err_clk_disable:
-	clk_disable_unprepare(pcie->bus_clk);
-
-err_unmap_msi_irqs:
-	irq_dispose_mapping(pcie->msi.irq2);
-	irq_dispose_mapping(pcie->msi.irq1);
-
-err_pm_put:
-	pm_runtime_put(dev);
-
-err_pm_disable:
-	pm_runtime_disable(dev);
-	pci_free_resource_list(&pcie->resources);
-
-err_free_bridge:
-	pci_free_host_bridge(bridge);
-
-	return err;
-}
-
-static int rcar_pcie_resume_noirq(struct device *dev)
-{
-	struct rcar_pcie *pcie = dev_get_drvdata(dev);
-
-	if (rcar_pci_read_reg(pcie, PMSR) &&
-	    !(rcar_pci_read_reg(pcie, PCIETCTLR) & DL_DOWN))
-		return 0;
-
-	/* Re-establish the PCIe link */
-	rcar_rmw32(pcie, MACCTLR, MACCTLR_RESERVED, 0);
-	rcar_pci_write_reg(pcie, CFINIT, PCIETCTLR);
-	return rcar_pcie_wait_for_dl(pcie);
+	/* flush modifications */
+	wmb();
 }
-
-static const struct dev_pm_ops rcar_pcie_pm_ops = {
-	.resume_noirq = rcar_pcie_resume_noirq,
-};
-
-static struct platform_driver rcar_pcie_driver = {
-	.driver = {
-		.name = "rcar-pcie",
-		.of_match_table = rcar_pcie_of_match,
-		.pm = &rcar_pcie_pm_ops,
-		.suppress_bind_attrs = true,
-	},
-	.probe = rcar_pcie_probe,
-};
-builtin_platform_driver(rcar_pcie_driver);
diff --git a/drivers/pci/controller/pcie-rcar.h b/drivers/pci/controller/pcie-rcar.h
new file mode 100644
index 000000000000..502621d4e6ac
--- /dev/null
+++ b/drivers/pci/controller/pcie-rcar.h
@@ -0,0 +1,129 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * PCIe driver for Renesas R-Car SoCs
+ *  Copyright (C) 2014-2019 Renesas Electronics Europe Ltd
+ *
+ * Author: Phil Edworthy <phil.edworthy@renesas.com>
+ */
+#ifndef _PCIE_RCAR_H
+#define _PCIE_RCAR_H
+
+#define PCIECAR			0x000010
+#define PCIECCTLR		0x000018
+#define  CONFIG_SEND_ENABLE	BIT(31)
+#define  TYPE0			(0 << 8)
+#define  TYPE1			BIT(8)
+#define PCIECDR			0x000020
+#define PCIEMSR			0x000028
+#define PCIEINTXR		0x000400
+#define PCIEPHYSR		0x0007f0
+#define  PHYRDY			BIT(0)
+#define PCIEMSITXR		0x000840
+
+/* Transfer control */
+#define PCIETCTLR		0x02000
+#define  DL_DOWN		BIT(3)
+#define  CFINIT			BIT(0)
+#define PCIETSTR		0x02004
+#define  DATA_LINK_ACTIVE	BIT(0)
+#define PCIEERRFR		0x02020
+#define  UNSUPPORTED_REQUEST	BIT(4)
+#define PCIEMSIFR		0x02044
+#define PCIEMSIALR		0x02048
+#define  MSIFE			BIT(0)
+#define PCIEMSIAUR		0x0204c
+#define PCIEMSIIER		0x02050
+
+/* root port address */
+#define PCIEPRAR(x)		(0x02080 + ((x) * 0x4))
+
+/* local address reg & mask */
+#define PCIELAR(x)		(0x02200 + ((x) * 0x20))
+#define PCIELAMR(x)		(0x02208 + ((x) * 0x20))
+#define  LAM_PREFETCH		BIT(3)
+#define  LAM_64BIT		BIT(2)
+#define  LAR_ENABLE		BIT(1)
+
+/* PCIe address reg & mask */
+#define PCIEPALR(x)		(0x03400 + ((x) * 0x20))
+#define PCIEPAUR(x)		(0x03404 + ((x) * 0x20))
+#define PCIEPAMR(x)		(0x03408 + ((x) * 0x20))
+#define PCIEPTCTLR(x)		(0x0340c + ((x) * 0x20))
+#define  PAR_ENABLE		BIT(31)
+#define  IO_SPACE		BIT(8)
+
+/* Configuration */
+#define PCICONF(x)		(0x010000 + ((x) * 0x4))
+#define PMCAP(x)		(0x010040 + ((x) * 0x4))
+#define EXPCAP(x)		(0x010070 + ((x) * 0x4))
+#define VCCAP(x)		(0x010100 + ((x) * 0x4))
+
+/* link layer */
+#define IDSETR0			0x011000
+#define  DEVICE_ID_SHFIT	16
+#define IDSETR1			0x011004
+#define SUBIDSETR		0x011024
+#define TLCTLR			0x011048
+#define MACSR			0x011054
+#define  SPCHGFIN		BIT(4)
+#define  SPCHGFAIL		BIT(6)
+#define  SPCHGSUC		BIT(7)
+#define  LINK_SPEED		(0xf << 16)
+#define  LINK_SPEED_2_5GTS	BIT(16)
+#define  LINK_SPEED_5_0GTS	BIT(17)
+#define MACCTLR			0x011058
+#define  MACCTLR_RESERVED	BIT(0)
+#define  SPEED_CHANGE		BIT(24)
+#define  SCRAMBLE_DISABLE	BIT(27)
+#define PMSR			0x01105c
+#define MACS2R			0x011078
+#define MACCGSPSETR		0x011084
+#define  SPCNGRSN		BIT(31)
+
+/* R-Car H1 PHY */
+#define H1_PCIEPHYADRR		0x04000c
+#define  WRITE_CMD		BIT(16)
+#define  PHY_ACK		BIT(24)
+#define  RATE_POS		12
+#define  LANE_POS		8
+#define  ADR_POS		0
+#define H1_PCIEPHYDOUTR		0x040014
+
+/* R-Car Gen2 PHY */
+#define GEN2_PCIEPHYADDR	0x780
+#define GEN2_PCIEPHYDATA	0x784
+#define GEN2_PCIEPHYCTRL	0x78c
+
+#define INT_PCI_MSI_NR		32
+
+#define RCONF(x)		(PCICONF(0) + (x))
+#define RPMCAP(x)		(PMCAP(0) + (x))
+#define REXPCAP(x)		(EXPCAP(0) + (x))
+#define RVCCAP(x)		(VCCAP(0) + (x))
+
+#define INTDIS_SHIFT		BIT(10)
+#define ASTINTX_SHIFT		BIT(16)
+
+#define PCIE_CONF_BUS(b)	(((b) & 0xff) << 24)
+#define PCIE_CONF_DEV(d)	(((d) & 0x1f) << 19)
+#define PCIE_CONF_FUNC(f)	(((f) & 0x7) << 16)
+
+#define RCAR_PCI_MAX_RESOURCES	4
+#define MAX_NR_INBOUND_MAPS	6
+
+enum {
+	RCAR_PCI_ACCESS_READ,
+	RCAR_PCI_ACCESS_WRITE,
+};
+
+void rcar_pci_write_reg(void __iomem *base, u32 val, unsigned int reg);
+u32 rcar_pci_read_reg(void __iomem *base, unsigned int reg);
+void rcar_rmw32(void __iomem *base, int where, u32 mask, u32 data);
+int rcar_pcie_wait_for_phyrdy(void __iomem *base);
+int rcar_pcie_wait_for_dl(void __iomem *base);
+void rcar_pcie_set_outbound(int win, void __iomem *base,
+			    struct resource *res, bool host);
+void rcar_pcie_set_inbound(void __iomem *base, u64 cpu_addr, u64 pci_addr,
+			   u64 mask, int idx, bool host);
+
+#endif
-- 
2.20.1


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply related	[flat|nested] 60+ messages in thread

* [PATCH 2/5] pci: endpoint: add support to handle multiple base for mapping outbound memory
  2019-11-06 19:36 ` Lad Prabhakar
@ 2019-11-06 19:36   ` Lad Prabhakar
  -1 siblings, 0 replies; 60+ messages in thread
From: Lad Prabhakar @ 2019-11-06 19:36 UTC (permalink / raw)
  To: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Kishon Vijay Abraham I, Marek Vasut,
	Yoshihiro Shimoda, linux-pci
  Cc: Catalin Marinas, Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray, devicetree, linux-kernel,
	linux-arm-kernel, linux-renesas-soc, Chris Paterson, Lad,
	Prabhakar, linux-rockchip, Shawn Lin, Heiko Stuebner, Tom Joseph,
	Jingoo Han, Gustavo Pimentel

From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>

rcar pcie controller has support to map multiple memory regions
for mapping the outbound memory in local system, this feature
inspires to add support for handling multiple memory bases in
endpoint framework. In case of multiple memory regions only chunk
or complete region can be mapped and this window needs to be
passed to the controller driver.

Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Cc: <linux-rockchip@lists.infradead.org>
Cc: Shawn Lin <shawn.lin@rock-chips.com>
Cc: Heiko Stuebner <heiko@sntech.de>
Cc: Tom Joseph <tjoseph@cadence.com>
Cc: Jingoo Han <jingoohan1@gmail.com>
Cc: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
---
 .../pci/controller/dwc/pcie-designware-ep.c   |  30 ++-
 drivers/pci/controller/pcie-cadence-ep.c      |  11 +-
 drivers/pci/controller/pcie-rockchip-ep.c     |  13 +-
 drivers/pci/endpoint/functions/pci-epf-test.c |  29 +--
 drivers/pci/endpoint/pci-epc-core.c           |   7 +-
 drivers/pci/endpoint/pci-epc-mem.c            | 189 ++++++++++++++----
 include/linux/pci-epc.h                       |  43 ++--
 7 files changed, 234 insertions(+), 88 deletions(-)

diff --git a/drivers/pci/controller/dwc/pcie-designware-ep.c b/drivers/pci/controller/dwc/pcie-designware-ep.c
index 3dd2e2697294..8d23c20b9afd 100644
--- a/drivers/pci/controller/dwc/pcie-designware-ep.c
+++ b/drivers/pci/controller/dwc/pcie-designware-ep.c
@@ -195,7 +195,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, u8 func_no,
 }
 
 static int dw_pcie_ep_map_addr(struct pci_epc *epc, u8 func_no,
-			       phys_addr_t addr,
+			       phys_addr_t addr, int window,
 			       u64 pci_addr, size_t size)
 {
 	int ret;
@@ -367,6 +367,7 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
 	unsigned int aligned_offset;
 	u16 msg_ctrl, msg_data;
 	u32 msg_addr_lower, msg_addr_upper, reg;
+	int window = PCI_EPC_DEFAULT_WINDOW;
 	u64 msg_addr;
 	bool has_upper;
 	int ret;
@@ -390,11 +391,11 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
 		reg = ep->msi_cap + PCI_MSI_DATA_32;
 		msg_data = dw_pcie_readw_dbi(pci, reg);
 	}
-	aligned_offset = msg_addr_lower & (epc->mem->page_size - 1);
+	aligned_offset = msg_addr_lower & (epc->mem[window]->page_size - 1);
 	msg_addr = ((u64)msg_addr_upper) << 32 |
 			(msg_addr_lower & ~aligned_offset);
-	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
-				  epc->mem->page_size);
+	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
+				  msg_addr, epc->mem[window]->page_size);
 	if (ret)
 		return ret;
 
@@ -416,6 +417,7 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
 	u32 reg, msg_data, vec_ctrl;
 	u64 tbl_addr, msg_addr, reg_u64;
 	void __iomem *msix_tbl;
+	int window = PCI_EPC_DEFAULT_WINDOW;
 	int ret;
 
 	reg = ep->msix_cap + PCI_MSIX_TABLE;
@@ -452,8 +454,8 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
 		return -EPERM;
 	}
 
-	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
-				  epc->mem->page_size);
+	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
+				  msg_addr, epc->mem[window]->page_size);
 	if (ret)
 		return ret;
 
@@ -466,10 +468,11 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
 
 void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
 {
+	int window = PCI_EPC_DEFAULT_WINDOW;
 	struct pci_epc *epc = ep->epc;
 
 	pci_epc_mem_free_addr(epc, ep->msi_mem_phys, ep->msi_mem,
-			      epc->mem->page_size);
+			      epc->mem[window]->page_size);
 
 	pci_epc_mem_exit(epc);
 }
@@ -499,9 +502,12 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
 	u32 reg;
 	void *addr;
 	u8 hdr_type;
+	int window;
 	unsigned int nbars;
 	unsigned int offset;
 	struct pci_epc *epc;
+	size_t msi_page_size;
+	struct pci_epc_mem_window mem_window;
 	struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
 	struct device *dev = pci->dev;
 	struct device_node *np = dev->of_node;
@@ -574,15 +580,17 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
 	if (ret < 0)
 		epc->max_functions = 1;
 
-	ret = __pci_epc_mem_init(epc, ep->phys_base, ep->addr_size,
-				 ep->page_size);
+	mem_window.phys_base = ep->phys_base;
+	mem_window.size = ep->addr_size;
+	ret = __pci_epc_mem_init(epc, &mem_window, 1, ep->page_size);
 	if (ret < 0) {
 		dev_err(dev, "Failed to initialize address space\n");
 		return ret;
 	}
 
-	ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys,
-					     epc->mem->page_size);
+	msi_page_size = epc->mem[PCI_EPC_DEFAULT_WINDOW]->page_size;
+	ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys, &window,
+					     msi_page_size);
 	if (!ep->msi_mem) {
 		dev_err(dev, "Failed to reserve memory for MSI/MSI-X\n");
 		return -ENOMEM;
diff --git a/drivers/pci/controller/pcie-cadence-ep.c b/drivers/pci/controller/pcie-cadence-ep.c
index def7820cb824..7991b38a5350 100644
--- a/drivers/pci/controller/pcie-cadence-ep.c
+++ b/drivers/pci/controller/pcie-cadence-ep.c
@@ -172,7 +172,7 @@ static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
 }
 
 static int cdns_pcie_ep_map_addr(struct pci_epc *epc, u8 fn, phys_addr_t addr,
-				 u64 pci_addr, size_t size)
+				 int window, u64 pci_addr, size_t size)
 {
 	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
 	struct cdns_pcie *pcie = &ep->pcie;
@@ -434,12 +434,14 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
 {
 	struct device *dev = &pdev->dev;
 	struct device_node *np = dev->of_node;
+	struct pci_epc_mem_window mem_window;
 	struct cdns_pcie_ep *ep;
 	struct cdns_pcie *pcie;
 	struct pci_epc *epc;
 	struct resource *res;
 	int ret;
 	int phy_count;
+	int window;
 
 	ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
 	if (!ep)
@@ -502,15 +504,16 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
 	if (of_property_read_u8(np, "max-functions", &epc->max_functions) < 0)
 		epc->max_functions = 1;
 
-	ret = pci_epc_mem_init(epc, pcie->mem_res->start,
-			       resource_size(pcie->mem_res));
+	mem_window.phys_base = pcie->mem_res->start;
+	mem_window.size = resource_size(pcie->mem_res);
+	ret = pci_epc_mem_init(epc, &mem_window, 1);
 	if (ret < 0) {
 		dev_err(dev, "failed to initialize the memory space\n");
 		goto err_init;
 	}
 
 	ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
-						  SZ_128K);
+						  &window, SZ_128K);
 	if (!ep->irq_cpu_addr) {
 		dev_err(dev, "failed to reserve memory space for MSI\n");
 		ret = -ENOMEM;
diff --git a/drivers/pci/controller/pcie-rockchip-ep.c b/drivers/pci/controller/pcie-rockchip-ep.c
index d743b0a48988..d59e85c8d319 100644
--- a/drivers/pci/controller/pcie-rockchip-ep.c
+++ b/drivers/pci/controller/pcie-rockchip-ep.c
@@ -256,8 +256,8 @@ static void rockchip_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
 }
 
 static int rockchip_pcie_ep_map_addr(struct pci_epc *epc, u8 fn,
-				     phys_addr_t addr, u64 pci_addr,
-				     size_t size)
+				     phys_addr_t addr, int window,
+				     u64 pci_addr, size_t size)
 {
 	struct rockchip_pcie_ep *ep = epc_get_drvdata(epc);
 	struct rockchip_pcie *pcie = &ep->rockchip;
@@ -562,11 +562,13 @@ static const struct of_device_id rockchip_pcie_ep_of_match[] = {
 
 static int rockchip_pcie_ep_probe(struct platform_device *pdev)
 {
+	struct pci_epc_mem_window mem_window;
 	struct device *dev = &pdev->dev;
 	struct rockchip_pcie_ep *ep;
 	struct rockchip_pcie *rockchip;
 	struct pci_epc *epc;
 	size_t max_regions;
+	int window;
 	int err;
 
 	ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
@@ -614,15 +616,16 @@ static int rockchip_pcie_ep_probe(struct platform_device *pdev)
 	/* Only enable function 0 by default */
 	rockchip_pcie_write(rockchip, BIT(0), PCIE_CORE_PHY_FUNC_CFG);
 
-	err = pci_epc_mem_init(epc, rockchip->mem_res->start,
-			       resource_size(rockchip->mem_res));
+	mem_window.phys_base = rockchip->mem_res->start;
+	mem_window.size = resource_size(rockchip->mem_res);
+	err = pci_epc_mem_init(epc, &mem_window, 1);
 	if (err < 0) {
 		dev_err(dev, "failed to initialize the memory space\n");
 		goto err_uninit_port;
 	}
 
 	ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
-						  SZ_128K);
+						  &window, SZ_128K);
 	if (!ep->irq_cpu_addr) {
 		dev_err(dev, "failed to reserve memory space for MSI\n");
 		err = -ENOMEM;
diff --git a/drivers/pci/endpoint/functions/pci-epf-test.c b/drivers/pci/endpoint/functions/pci-epf-test.c
index 5d74f81ddfe4..475228011703 100644
--- a/drivers/pci/endpoint/functions/pci-epf-test.c
+++ b/drivers/pci/endpoint/functions/pci-epf-test.c
@@ -84,8 +84,10 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
 	struct pci_epc *epc = epf->epc;
 	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
 	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
+	int window;
 
-	src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr, reg->size);
+	src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr,
+					  &window, reg->size);
 	if (!src_addr) {
 		dev_err(dev, "Failed to allocate source address\n");
 		reg->status = STATUS_SRC_ADDR_INVALID;
@@ -93,15 +95,16 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
 		goto err;
 	}
 
-	ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, reg->src_addr,
-			       reg->size);
+	ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, window,
+			       reg->src_addr, reg->size);
 	if (ret) {
 		dev_err(dev, "Failed to map source address\n");
 		reg->status = STATUS_SRC_ADDR_INVALID;
 		goto err_src_addr;
 	}
 
-	dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr, reg->size);
+	dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr,
+					  &window, reg->size);
 	if (!dst_addr) {
 		dev_err(dev, "Failed to allocate destination address\n");
 		reg->status = STATUS_DST_ADDR_INVALID;
@@ -109,8 +112,8 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
 		goto err_src_map_addr;
 	}
 
-	ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, reg->dst_addr,
-			       reg->size);
+	ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, window,
+			       reg->dst_addr, reg->size);
 	if (ret) {
 		dev_err(dev, "Failed to map destination address\n");
 		reg->status = STATUS_DST_ADDR_INVALID;
@@ -146,8 +149,9 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
 	struct pci_epc *epc = epf->epc;
 	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
 	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
+	int window;
 
-	src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
+	src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
 	if (!src_addr) {
 		dev_err(dev, "Failed to allocate address\n");
 		reg->status = STATUS_SRC_ADDR_INVALID;
@@ -155,8 +159,8 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
 		goto err;
 	}
 
-	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->src_addr,
-			       reg->size);
+	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
+			       reg->src_addr, reg->size);
 	if (ret) {
 		dev_err(dev, "Failed to map address\n");
 		reg->status = STATUS_SRC_ADDR_INVALID;
@@ -193,13 +197,14 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
 	void __iomem *dst_addr;
 	void *buf;
 	phys_addr_t phys_addr;
+	int window;
 	struct pci_epf *epf = epf_test->epf;
 	struct device *dev = &epf->dev;
 	struct pci_epc *epc = epf->epc;
 	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
 	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
 
-	dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
+	dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
 	if (!dst_addr) {
 		dev_err(dev, "Failed to allocate address\n");
 		reg->status = STATUS_DST_ADDR_INVALID;
@@ -207,8 +212,8 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
 		goto err;
 	}
 
-	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->dst_addr,
-			       reg->size);
+	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
+			       reg->dst_addr, reg->size);
 	if (ret) {
 		dev_err(dev, "Failed to map address\n");
 		reg->status = STATUS_DST_ADDR_INVALID;
diff --git a/drivers/pci/endpoint/pci-epc-core.c b/drivers/pci/endpoint/pci-epc-core.c
index 2091508c1620..289c266c2d90 100644
--- a/drivers/pci/endpoint/pci-epc-core.c
+++ b/drivers/pci/endpoint/pci-epc-core.c
@@ -358,13 +358,15 @@ EXPORT_SYMBOL_GPL(pci_epc_unmap_addr);
  * @epc: the EPC device on which address is allocated
  * @func_no: the endpoint function number in the EPC device
  * @phys_addr: physical address of the local system
+ * @window: index to the window region where PCI address will be mapped
  * @pci_addr: PCI address to which the physical address should be mapped
  * @size: the size of the allocation
  *
  * Invoke to map CPU address with PCI address.
  */
 int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
-		     phys_addr_t phys_addr, u64 pci_addr, size_t size)
+		     phys_addr_t phys_addr, int window,
+		     u64 pci_addr, size_t size)
 {
 	int ret;
 	unsigned long flags;
@@ -376,7 +378,8 @@ int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
 		return 0;
 
 	spin_lock_irqsave(&epc->lock, flags);
-	ret = epc->ops->map_addr(epc, func_no, phys_addr, pci_addr, size);
+	ret = epc->ops->map_addr(epc, func_no, phys_addr,
+				 window, pci_addr, size);
 	spin_unlock_irqrestore(&epc->lock, flags);
 
 	return ret;
diff --git a/drivers/pci/endpoint/pci-epc-mem.c b/drivers/pci/endpoint/pci-epc-mem.c
index d2b174ce15de..c955f2c97944 100644
--- a/drivers/pci/endpoint/pci-epc-mem.c
+++ b/drivers/pci/endpoint/pci-epc-mem.c
@@ -39,56 +39,77 @@ static int pci_epc_mem_get_order(struct pci_epc_mem *mem, size_t size)
  * __pci_epc_mem_init() - initialize the pci_epc_mem structure
  * @epc: the EPC device that invoked pci_epc_mem_init
  * @phys_base: the physical address of the base
- * @size: the size of the address space
+ * @num_windows: number of windows device supports
  * @page_size: size of each page
  *
  * Invoke to initialize the pci_epc_mem structure used by the
  * endpoint functions to allocate mapped PCI address.
  */
-int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_base, size_t size,
-		       size_t page_size)
+int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *windows,
+		       int num_windows, size_t page_size)
 {
-	int ret;
-	struct pci_epc_mem *mem;
-	unsigned long *bitmap;
+	struct pci_epc_mem *mem = NULL;
+	unsigned long *bitmap = NULL;
 	unsigned int page_shift;
-	int pages;
 	int bitmap_size;
+	int pages;
+	int ret;
+	int i;
+
+	epc->mem_windows = 0;
+
+	if (!windows)
+		return -EINVAL;
+
+	if (num_windows <= 0)
+		return -EINVAL;
 
 	if (page_size < PAGE_SIZE)
 		page_size = PAGE_SIZE;
 
 	page_shift = ilog2(page_size);
-	pages = size >> page_shift;
-	bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
 
-	mem = kzalloc(sizeof(*mem), GFP_KERNEL);
-	if (!mem) {
-		ret = -ENOMEM;
-		goto err;
-	}
+	epc->mem = kcalloc(num_windows, sizeof(*mem), GFP_KERNEL);
+	if (!epc->mem)
+		return -EINVAL;
 
-	bitmap = kzalloc(bitmap_size, GFP_KERNEL);
-	if (!bitmap) {
-		ret = -ENOMEM;
-		goto err_mem;
-	}
+	for (i = 0; i < num_windows; i++) {
+		pages = windows[i].phys_base >> page_shift;
+		bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
 
-	mem->bitmap = bitmap;
-	mem->phys_base = phys_base;
-	mem->page_size = page_size;
-	mem->pages = pages;
-	mem->size = size;
+		mem = kzalloc(sizeof(*mem), GFP_KERNEL);
+		if (!mem) {
+			ret = -ENOMEM;
+			goto err_mem;
+		}
 
-	epc->mem = mem;
+		bitmap = kzalloc(bitmap_size, GFP_KERNEL);
+		if (!bitmap) {
+			ret = -ENOMEM;
+			goto err_mem;
+		}
+
+		mem->bitmap = bitmap;
+		mem->window.phys_base = windows[i].phys_base;
+		mem->page_size = page_size;
+		mem->pages = pages;
+		mem->window.size = windows[i].size;
+		mem->window.map_size = 0;
+
+		epc->mem[i] = mem;
+	}
+	epc->mem_windows = num_windows;
 
 	return 0;
 
 err_mem:
-	kfree(mem);
+	for (; i >= 0; i--) {
+		kfree(mem->bitmap);
+		kfree(epc->mem[i]);
+	}
+	kfree(epc->mem);
 
-err:
-return ret;
+	return ret;
 }
 EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
 
@@ -101,48 +122,126 @@ EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
  */
 void pci_epc_mem_exit(struct pci_epc *epc)
 {
-	struct pci_epc_mem *mem = epc->mem;
+	struct pci_epc_mem *mem;
+	int i;
+
+	if (!epc->mem_windows)
+		return;
+
+	for (i = 0; i <= epc->mem_windows; i--) {
+		mem = epc->mem[i];
+		kfree(mem->bitmap);
+		kfree(epc->mem[i]);
+	}
+	kfree(epc->mem);
 
 	epc->mem = NULL;
-	kfree(mem->bitmap);
-	kfree(mem);
+	epc->mem_windows = 0;
 }
 EXPORT_SYMBOL_GPL(pci_epc_mem_exit);
 
+static int pci_epc_find_best_fit_window(struct pci_epc *epc, size_t size)
+{
+	size_t window_least_size = 0;
+	int best_fit_window = -1;
+	struct pci_epc_mem *mem;
+	size_t actual_size;
+	int i;
+
+	for (i = 0; i < epc->mem_windows; i++) {
+		mem = epc->mem[i];
+
+		/* if chunk from this region is already used skip it */
+		if (mem->window.map_size)
+			continue;
+
+		actual_size = ALIGN(size, mem->page_size);
+
+		if (best_fit_window == -1) {
+			best_fit_window = i;
+			window_least_size = mem->window.size;
+		} else {
+			if (actual_size <= mem->window.size &&
+			    mem->window.size < window_least_size) {
+				best_fit_window = i;
+				window_least_size = mem->window.size;
+			}
+		}
+	}
+
+	return best_fit_window;
+}
+
 /**
  * pci_epc_mem_alloc_addr() - allocate memory address from EPC addr space
  * @epc: the EPC device on which memory has to be allocated
  * @phys_addr: populate the allocated physical address here
+ * @window: populate the window here which will be used to map PCI address
  * @size: the size of the address space that has to be allocated
  *
  * Invoke to allocate memory address from the EPC address space. This
  * is usually done to map the remote RC address into the local system.
  */
 void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
-				     phys_addr_t *phys_addr, size_t size)
+				     phys_addr_t *phys_addr,
+				     int *window, size_t size)
 {
+	int best_fit = PCI_EPC_DEFAULT_WINDOW;
+	void __iomem *virt_addr = NULL;
+	struct pci_epc_mem *mem;
+	unsigned int page_shift;
 	int pageno;
-	void __iomem *virt_addr;
-	struct pci_epc_mem *mem = epc->mem;
-	unsigned int page_shift = ilog2(mem->page_size);
 	int order;
 
+	if (epc->mem_windows <= 0)
+		return NULL;
+
+	if (epc->mem_windows > 1) {
+		best_fit = pci_epc_find_best_fit_window(epc, size);
+		if (best_fit < 0)
+			return NULL;
+	}
+
+	mem = epc->mem[best_fit];
 	size = ALIGN(size, mem->page_size);
+	if (size > (mem->window.size - mem->window.map_size))
+		return NULL;
+	page_shift = ilog2(mem->page_size);
 	order = pci_epc_mem_get_order(mem, size);
 
 	pageno = bitmap_find_free_region(mem->bitmap, mem->pages, order);
 	if (pageno < 0)
 		return NULL;
 
-	*phys_addr = mem->phys_base + ((phys_addr_t)pageno << page_shift);
+	*phys_addr = mem->window.phys_base + (pageno << page_shift);
 	virt_addr = ioremap(*phys_addr, size);
-	if (!virt_addr)
+	if (!virt_addr) {
 		bitmap_release_region(mem->bitmap, pageno, order);
+	} else {
+		mem->window.map_size += size;
+		*window = best_fit;
+	}
 
 	return virt_addr;
 }
 EXPORT_SYMBOL_GPL(pci_epc_mem_alloc_addr);
 
+static int pci_epc_get_matching_window(struct pci_epc *epc,
+				       phys_addr_t phys_addr)
+{
+	struct pci_epc_mem *mem;
+	int i;
+
+	for (i = 0; i < epc->mem_windows; i++) {
+		mem = epc->mem[i];
+
+		if (mem->window.phys_base == phys_addr)
+			return i;
+	}
+
+	return -EINVAL;
+}
+
 /**
  * pci_epc_mem_free_addr() - free the allocated memory address
  * @epc: the EPC device on which memory was allocated
@@ -155,16 +254,26 @@ EXPORT_SYMBOL_GPL(pci_epc_mem_alloc_addr);
 void pci_epc_mem_free_addr(struct pci_epc *epc, phys_addr_t phys_addr,
 			   void __iomem *virt_addr, size_t size)
 {
+	struct pci_epc_mem *mem;
+	unsigned int page_shift;
+	int window = 0;
 	int pageno;
-	struct pci_epc_mem *mem = epc->mem;
-	unsigned int page_shift = ilog2(mem->page_size);
 	int order;
 
+	if (epc->mem_windows > 1) {
+		window = pci_epc_get_matching_window(epc, phys_addr);
+		if (window < 0)
+			return;
+	}
+
+	mem = epc->mem[window];
+	page_shift = ilog2(mem->page_size);
 	iounmap(virt_addr);
-	pageno = (phys_addr - mem->phys_base) >> page_shift;
+	pageno = (phys_addr - mem->window.phys_base) >> page_shift;
 	size = ALIGN(size, mem->page_size);
 	order = pci_epc_mem_get_order(mem, size);
 	bitmap_release_region(mem->bitmap, pageno, order);
+	mem->window.map_size -= size;
 }
 EXPORT_SYMBOL_GPL(pci_epc_mem_free_addr);
 
diff --git a/include/linux/pci-epc.h b/include/linux/pci-epc.h
index 56f1846b9d39..5e86d22dd5c5 100644
--- a/include/linux/pci-epc.h
+++ b/include/linux/pci-epc.h
@@ -48,7 +48,8 @@ struct pci_epc_ops {
 	void	(*clear_bar)(struct pci_epc *epc, u8 func_no,
 			     struct pci_epf_bar *epf_bar);
 	int	(*map_addr)(struct pci_epc *epc, u8 func_no,
-			    phys_addr_t addr, u64 pci_addr, size_t size);
+			    phys_addr_t addr, int window,
+			    u64 pci_addr, size_t size);
 	void	(*unmap_addr)(struct pci_epc *epc, u8 func_no,
 			      phys_addr_t addr);
 	int	(*set_msi)(struct pci_epc *epc, u8 func_no, u8 interrupts);
@@ -64,17 +65,28 @@ struct pci_epc_ops {
 	struct module *owner;
 };
 
+#define PCI_EPC_DEFAULT_WINDOW		0
+
+/**
+ * struct pci_epc_mem_window - address window of the endpoint controller
+ * @phys_base: physical base address of the PCI address window
+ * @size: the size of the PCI address window
+ */
+struct pci_epc_mem_window {
+	phys_addr_t	phys_base;
+	size_t		size;
+	size_t		map_size;
+};
+
 /**
  * struct pci_epc_mem - address space of the endpoint controller
- * @phys_base: physical base address of the PCI address space
- * @size: the size of the PCI address space
+ * @window: address window of the endpoint controller
  * @bitmap: bitmap to manage the PCI address space
- * @pages: number of bits representing the address region
  * @page_size: size of each page
+ * @pages: number of bits representing the address region
  */
 struct pci_epc_mem {
-	phys_addr_t	phys_base;
-	size_t		size;
+	struct pci_epc_mem_window window;
 	unsigned long	*bitmap;
 	size_t		page_size;
 	int		pages;
@@ -85,7 +97,8 @@ struct pci_epc_mem {
  * @dev: PCI EPC device
  * @pci_epf: list of endpoint functions present in this EPC device
  * @ops: function pointers for performing endpoint operations
- * @mem: address space of the endpoint controller
+ * @mem: array of address space of the endpoint controller
+ * @mem_windows: number of windows supported by device
  * @max_functions: max number of functions that can be configured in this EPC
  * @group: configfs group representing the PCI EPC device
  * @lock: spinlock to protect pci_epc ops
@@ -94,7 +107,8 @@ struct pci_epc {
 	struct device			dev;
 	struct list_head		pci_epf;
 	const struct pci_epc_ops	*ops;
-	struct pci_epc_mem		*mem;
+	struct pci_epc_mem		**mem;
+	int				mem_windows;
 	u8				max_functions;
 	struct config_group		*group;
 	/* spinlock to protect against concurrent access of EP controller */
@@ -128,8 +142,8 @@ struct pci_epc_features {
 #define devm_pci_epc_create(dev, ops)    \
 		__devm_pci_epc_create((dev), (ops), THIS_MODULE)
 
-#define pci_epc_mem_init(epc, phys_addr, size)	\
-		__pci_epc_mem_init((epc), (phys_addr), (size), PAGE_SIZE)
+#define pci_epc_mem_init(epc, windows, num_windows)	\
+		__pci_epc_mem_init((epc), windows, num_windows, PAGE_SIZE)
 
 static inline void epc_set_drvdata(struct pci_epc *epc, void *data)
 {
@@ -159,7 +173,7 @@ int pci_epc_set_bar(struct pci_epc *epc, u8 func_no,
 void pci_epc_clear_bar(struct pci_epc *epc, u8 func_no,
 		       struct pci_epf_bar *epf_bar);
 int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
-		     phys_addr_t phys_addr,
+		     phys_addr_t phys_addr, int window,
 		     u64 pci_addr, size_t size);
 void pci_epc_unmap_addr(struct pci_epc *epc, u8 func_no,
 			phys_addr_t phys_addr);
@@ -178,11 +192,12 @@ unsigned int pci_epc_get_first_free_bar(const struct pci_epc_features
 struct pci_epc *pci_epc_get(const char *epc_name);
 void pci_epc_put(struct pci_epc *epc);
 
-int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_addr, size_t size,
-		       size_t page_size);
+int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *window,
+		       int num_windows, size_t page_size);
 void pci_epc_mem_exit(struct pci_epc *epc);
 void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
-				     phys_addr_t *phys_addr, size_t size);
+				     phys_addr_t *phys_addr,
+				     int *window, size_t size);
 void pci_epc_mem_free_addr(struct pci_epc *epc, phys_addr_t phys_addr,
 			   void __iomem *virt_addr, size_t size);
 #endif /* __LINUX_PCI_EPC_H */
-- 
2.20.1


^ permalink raw reply related	[flat|nested] 60+ messages in thread

* [PATCH 2/5] pci: endpoint: add support to handle multiple base for mapping outbound memory
@ 2019-11-06 19:36   ` Lad Prabhakar
  0 siblings, 0 replies; 60+ messages in thread
From: Lad Prabhakar @ 2019-11-06 19:36 UTC (permalink / raw)
  To: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Kishon Vijay Abraham I, Marek Vasut,
	Yoshihiro Shimoda, linux-pci
  Cc: devicetree, Chris Paterson, Lorenzo Pieralisi, Heiko Stuebner,
	Arnd Bergmann, Jingoo Han, Catalin Marinas, Shawn Lin,
	linux-kernel, Lad, Prabhakar, linux-renesas-soc, linux-rockchip,
	Tom Joseph, Greg Kroah-Hartman, Andrew Murray, Gustavo Pimentel,
	Will Deacon, linux-arm-kernel

From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>

rcar pcie controller has support to map multiple memory regions
for mapping the outbound memory in local system, this feature
inspires to add support for handling multiple memory bases in
endpoint framework. In case of multiple memory regions only chunk
or complete region can be mapped and this window needs to be
passed to the controller driver.

Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Cc: <linux-rockchip@lists.infradead.org>
Cc: Shawn Lin <shawn.lin@rock-chips.com>
Cc: Heiko Stuebner <heiko@sntech.de>
Cc: Tom Joseph <tjoseph@cadence.com>
Cc: Jingoo Han <jingoohan1@gmail.com>
Cc: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
---
 .../pci/controller/dwc/pcie-designware-ep.c   |  30 ++-
 drivers/pci/controller/pcie-cadence-ep.c      |  11 +-
 drivers/pci/controller/pcie-rockchip-ep.c     |  13 +-
 drivers/pci/endpoint/functions/pci-epf-test.c |  29 +--
 drivers/pci/endpoint/pci-epc-core.c           |   7 +-
 drivers/pci/endpoint/pci-epc-mem.c            | 189 ++++++++++++++----
 include/linux/pci-epc.h                       |  43 ++--
 7 files changed, 234 insertions(+), 88 deletions(-)

diff --git a/drivers/pci/controller/dwc/pcie-designware-ep.c b/drivers/pci/controller/dwc/pcie-designware-ep.c
index 3dd2e2697294..8d23c20b9afd 100644
--- a/drivers/pci/controller/dwc/pcie-designware-ep.c
+++ b/drivers/pci/controller/dwc/pcie-designware-ep.c
@@ -195,7 +195,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, u8 func_no,
 }
 
 static int dw_pcie_ep_map_addr(struct pci_epc *epc, u8 func_no,
-			       phys_addr_t addr,
+			       phys_addr_t addr, int window,
 			       u64 pci_addr, size_t size)
 {
 	int ret;
@@ -367,6 +367,7 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
 	unsigned int aligned_offset;
 	u16 msg_ctrl, msg_data;
 	u32 msg_addr_lower, msg_addr_upper, reg;
+	int window = PCI_EPC_DEFAULT_WINDOW;
 	u64 msg_addr;
 	bool has_upper;
 	int ret;
@@ -390,11 +391,11 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
 		reg = ep->msi_cap + PCI_MSI_DATA_32;
 		msg_data = dw_pcie_readw_dbi(pci, reg);
 	}
-	aligned_offset = msg_addr_lower & (epc->mem->page_size - 1);
+	aligned_offset = msg_addr_lower & (epc->mem[window]->page_size - 1);
 	msg_addr = ((u64)msg_addr_upper) << 32 |
 			(msg_addr_lower & ~aligned_offset);
-	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
-				  epc->mem->page_size);
+	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
+				  msg_addr, epc->mem[window]->page_size);
 	if (ret)
 		return ret;
 
@@ -416,6 +417,7 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
 	u32 reg, msg_data, vec_ctrl;
 	u64 tbl_addr, msg_addr, reg_u64;
 	void __iomem *msix_tbl;
+	int window = PCI_EPC_DEFAULT_WINDOW;
 	int ret;
 
 	reg = ep->msix_cap + PCI_MSIX_TABLE;
@@ -452,8 +454,8 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
 		return -EPERM;
 	}
 
-	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
-				  epc->mem->page_size);
+	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
+				  msg_addr, epc->mem[window]->page_size);
 	if (ret)
 		return ret;
 
@@ -466,10 +468,11 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
 
 void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
 {
+	int window = PCI_EPC_DEFAULT_WINDOW;
 	struct pci_epc *epc = ep->epc;
 
 	pci_epc_mem_free_addr(epc, ep->msi_mem_phys, ep->msi_mem,
-			      epc->mem->page_size);
+			      epc->mem[window]->page_size);
 
 	pci_epc_mem_exit(epc);
 }
@@ -499,9 +502,12 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
 	u32 reg;
 	void *addr;
 	u8 hdr_type;
+	int window;
 	unsigned int nbars;
 	unsigned int offset;
 	struct pci_epc *epc;
+	size_t msi_page_size;
+	struct pci_epc_mem_window mem_window;
 	struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
 	struct device *dev = pci->dev;
 	struct device_node *np = dev->of_node;
@@ -574,15 +580,17 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
 	if (ret < 0)
 		epc->max_functions = 1;
 
-	ret = __pci_epc_mem_init(epc, ep->phys_base, ep->addr_size,
-				 ep->page_size);
+	mem_window.phys_base = ep->phys_base;
+	mem_window.size = ep->addr_size;
+	ret = __pci_epc_mem_init(epc, &mem_window, 1, ep->page_size);
 	if (ret < 0) {
 		dev_err(dev, "Failed to initialize address space\n");
 		return ret;
 	}
 
-	ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys,
-					     epc->mem->page_size);
+	msi_page_size = epc->mem[PCI_EPC_DEFAULT_WINDOW]->page_size;
+	ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys, &window,
+					     msi_page_size);
 	if (!ep->msi_mem) {
 		dev_err(dev, "Failed to reserve memory for MSI/MSI-X\n");
 		return -ENOMEM;
diff --git a/drivers/pci/controller/pcie-cadence-ep.c b/drivers/pci/controller/pcie-cadence-ep.c
index def7820cb824..7991b38a5350 100644
--- a/drivers/pci/controller/pcie-cadence-ep.c
+++ b/drivers/pci/controller/pcie-cadence-ep.c
@@ -172,7 +172,7 @@ static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
 }
 
 static int cdns_pcie_ep_map_addr(struct pci_epc *epc, u8 fn, phys_addr_t addr,
-				 u64 pci_addr, size_t size)
+				 int window, u64 pci_addr, size_t size)
 {
 	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
 	struct cdns_pcie *pcie = &ep->pcie;
@@ -434,12 +434,14 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
 {
 	struct device *dev = &pdev->dev;
 	struct device_node *np = dev->of_node;
+	struct pci_epc_mem_window mem_window;
 	struct cdns_pcie_ep *ep;
 	struct cdns_pcie *pcie;
 	struct pci_epc *epc;
 	struct resource *res;
 	int ret;
 	int phy_count;
+	int window;
 
 	ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
 	if (!ep)
@@ -502,15 +504,16 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
 	if (of_property_read_u8(np, "max-functions", &epc->max_functions) < 0)
 		epc->max_functions = 1;
 
-	ret = pci_epc_mem_init(epc, pcie->mem_res->start,
-			       resource_size(pcie->mem_res));
+	mem_window.phys_base = pcie->mem_res->start;
+	mem_window.size = resource_size(pcie->mem_res);
+	ret = pci_epc_mem_init(epc, &mem_window, 1);
 	if (ret < 0) {
 		dev_err(dev, "failed to initialize the memory space\n");
 		goto err_init;
 	}
 
 	ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
-						  SZ_128K);
+						  &window, SZ_128K);
 	if (!ep->irq_cpu_addr) {
 		dev_err(dev, "failed to reserve memory space for MSI\n");
 		ret = -ENOMEM;
diff --git a/drivers/pci/controller/pcie-rockchip-ep.c b/drivers/pci/controller/pcie-rockchip-ep.c
index d743b0a48988..d59e85c8d319 100644
--- a/drivers/pci/controller/pcie-rockchip-ep.c
+++ b/drivers/pci/controller/pcie-rockchip-ep.c
@@ -256,8 +256,8 @@ static void rockchip_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
 }
 
 static int rockchip_pcie_ep_map_addr(struct pci_epc *epc, u8 fn,
-				     phys_addr_t addr, u64 pci_addr,
-				     size_t size)
+				     phys_addr_t addr, int window,
+				     u64 pci_addr, size_t size)
 {
 	struct rockchip_pcie_ep *ep = epc_get_drvdata(epc);
 	struct rockchip_pcie *pcie = &ep->rockchip;
@@ -562,11 +562,13 @@ static const struct of_device_id rockchip_pcie_ep_of_match[] = {
 
 static int rockchip_pcie_ep_probe(struct platform_device *pdev)
 {
+	struct pci_epc_mem_window mem_window;
 	struct device *dev = &pdev->dev;
 	struct rockchip_pcie_ep *ep;
 	struct rockchip_pcie *rockchip;
 	struct pci_epc *epc;
 	size_t max_regions;
+	int window;
 	int err;
 
 	ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
@@ -614,15 +616,16 @@ static int rockchip_pcie_ep_probe(struct platform_device *pdev)
 	/* Only enable function 0 by default */
 	rockchip_pcie_write(rockchip, BIT(0), PCIE_CORE_PHY_FUNC_CFG);
 
-	err = pci_epc_mem_init(epc, rockchip->mem_res->start,
-			       resource_size(rockchip->mem_res));
+	mem_window.phys_base = rockchip->mem_res->start;
+	mem_window.size = resource_size(rockchip->mem_res);
+	err = pci_epc_mem_init(epc, &mem_window, 1);
 	if (err < 0) {
 		dev_err(dev, "failed to initialize the memory space\n");
 		goto err_uninit_port;
 	}
 
 	ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
-						  SZ_128K);
+						  &window, SZ_128K);
 	if (!ep->irq_cpu_addr) {
 		dev_err(dev, "failed to reserve memory space for MSI\n");
 		err = -ENOMEM;
diff --git a/drivers/pci/endpoint/functions/pci-epf-test.c b/drivers/pci/endpoint/functions/pci-epf-test.c
index 5d74f81ddfe4..475228011703 100644
--- a/drivers/pci/endpoint/functions/pci-epf-test.c
+++ b/drivers/pci/endpoint/functions/pci-epf-test.c
@@ -84,8 +84,10 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
 	struct pci_epc *epc = epf->epc;
 	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
 	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
+	int window;
 
-	src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr, reg->size);
+	src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr,
+					  &window, reg->size);
 	if (!src_addr) {
 		dev_err(dev, "Failed to allocate source address\n");
 		reg->status = STATUS_SRC_ADDR_INVALID;
@@ -93,15 +95,16 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
 		goto err;
 	}
 
-	ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, reg->src_addr,
-			       reg->size);
+	ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, window,
+			       reg->src_addr, reg->size);
 	if (ret) {
 		dev_err(dev, "Failed to map source address\n");
 		reg->status = STATUS_SRC_ADDR_INVALID;
 		goto err_src_addr;
 	}
 
-	dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr, reg->size);
+	dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr,
+					  &window, reg->size);
 	if (!dst_addr) {
 		dev_err(dev, "Failed to allocate destination address\n");
 		reg->status = STATUS_DST_ADDR_INVALID;
@@ -109,8 +112,8 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
 		goto err_src_map_addr;
 	}
 
-	ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, reg->dst_addr,
-			       reg->size);
+	ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, window,
+			       reg->dst_addr, reg->size);
 	if (ret) {
 		dev_err(dev, "Failed to map destination address\n");
 		reg->status = STATUS_DST_ADDR_INVALID;
@@ -146,8 +149,9 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
 	struct pci_epc *epc = epf->epc;
 	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
 	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
+	int window;
 
-	src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
+	src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
 	if (!src_addr) {
 		dev_err(dev, "Failed to allocate address\n");
 		reg->status = STATUS_SRC_ADDR_INVALID;
@@ -155,8 +159,8 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
 		goto err;
 	}
 
-	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->src_addr,
-			       reg->size);
+	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
+			       reg->src_addr, reg->size);
 	if (ret) {
 		dev_err(dev, "Failed to map address\n");
 		reg->status = STATUS_SRC_ADDR_INVALID;
@@ -193,13 +197,14 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
 	void __iomem *dst_addr;
 	void *buf;
 	phys_addr_t phys_addr;
+	int window;
 	struct pci_epf *epf = epf_test->epf;
 	struct device *dev = &epf->dev;
 	struct pci_epc *epc = epf->epc;
 	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
 	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
 
-	dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
+	dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
 	if (!dst_addr) {
 		dev_err(dev, "Failed to allocate address\n");
 		reg->status = STATUS_DST_ADDR_INVALID;
@@ -207,8 +212,8 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
 		goto err;
 	}
 
-	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->dst_addr,
-			       reg->size);
+	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
+			       reg->dst_addr, reg->size);
 	if (ret) {
 		dev_err(dev, "Failed to map address\n");
 		reg->status = STATUS_DST_ADDR_INVALID;
diff --git a/drivers/pci/endpoint/pci-epc-core.c b/drivers/pci/endpoint/pci-epc-core.c
index 2091508c1620..289c266c2d90 100644
--- a/drivers/pci/endpoint/pci-epc-core.c
+++ b/drivers/pci/endpoint/pci-epc-core.c
@@ -358,13 +358,15 @@ EXPORT_SYMBOL_GPL(pci_epc_unmap_addr);
  * @epc: the EPC device on which address is allocated
  * @func_no: the endpoint function number in the EPC device
  * @phys_addr: physical address of the local system
+ * @window: index to the window region where PCI address will be mapped
  * @pci_addr: PCI address to which the physical address should be mapped
  * @size: the size of the allocation
  *
  * Invoke to map CPU address with PCI address.
  */
 int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
-		     phys_addr_t phys_addr, u64 pci_addr, size_t size)
+		     phys_addr_t phys_addr, int window,
+		     u64 pci_addr, size_t size)
 {
 	int ret;
 	unsigned long flags;
@@ -376,7 +378,8 @@ int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
 		return 0;
 
 	spin_lock_irqsave(&epc->lock, flags);
-	ret = epc->ops->map_addr(epc, func_no, phys_addr, pci_addr, size);
+	ret = epc->ops->map_addr(epc, func_no, phys_addr,
+				 window, pci_addr, size);
 	spin_unlock_irqrestore(&epc->lock, flags);
 
 	return ret;
diff --git a/drivers/pci/endpoint/pci-epc-mem.c b/drivers/pci/endpoint/pci-epc-mem.c
index d2b174ce15de..c955f2c97944 100644
--- a/drivers/pci/endpoint/pci-epc-mem.c
+++ b/drivers/pci/endpoint/pci-epc-mem.c
@@ -39,56 +39,77 @@ static int pci_epc_mem_get_order(struct pci_epc_mem *mem, size_t size)
  * __pci_epc_mem_init() - initialize the pci_epc_mem structure
  * @epc: the EPC device that invoked pci_epc_mem_init
  * @phys_base: the physical address of the base
- * @size: the size of the address space
+ * @num_windows: number of windows device supports
  * @page_size: size of each page
  *
  * Invoke to initialize the pci_epc_mem structure used by the
  * endpoint functions to allocate mapped PCI address.
  */
-int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_base, size_t size,
-		       size_t page_size)
+int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *windows,
+		       int num_windows, size_t page_size)
 {
-	int ret;
-	struct pci_epc_mem *mem;
-	unsigned long *bitmap;
+	struct pci_epc_mem *mem = NULL;
+	unsigned long *bitmap = NULL;
 	unsigned int page_shift;
-	int pages;
 	int bitmap_size;
+	int pages;
+	int ret;
+	int i;
+
+	epc->mem_windows = 0;
+
+	if (!windows)
+		return -EINVAL;
+
+	if (num_windows <= 0)
+		return -EINVAL;
 
 	if (page_size < PAGE_SIZE)
 		page_size = PAGE_SIZE;
 
 	page_shift = ilog2(page_size);
-	pages = size >> page_shift;
-	bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
 
-	mem = kzalloc(sizeof(*mem), GFP_KERNEL);
-	if (!mem) {
-		ret = -ENOMEM;
-		goto err;
-	}
+	epc->mem = kcalloc(num_windows, sizeof(*mem), GFP_KERNEL);
+	if (!epc->mem)
+		return -EINVAL;
 
-	bitmap = kzalloc(bitmap_size, GFP_KERNEL);
-	if (!bitmap) {
-		ret = -ENOMEM;
-		goto err_mem;
-	}
+	for (i = 0; i < num_windows; i++) {
+		pages = windows[i].phys_base >> page_shift;
+		bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
 
-	mem->bitmap = bitmap;
-	mem->phys_base = phys_base;
-	mem->page_size = page_size;
-	mem->pages = pages;
-	mem->size = size;
+		mem = kzalloc(sizeof(*mem), GFP_KERNEL);
+		if (!mem) {
+			ret = -ENOMEM;
+			goto err_mem;
+		}
 
-	epc->mem = mem;
+		bitmap = kzalloc(bitmap_size, GFP_KERNEL);
+		if (!bitmap) {
+			ret = -ENOMEM;
+			goto err_mem;
+		}
+
+		mem->bitmap = bitmap;
+		mem->window.phys_base = windows[i].phys_base;
+		mem->page_size = page_size;
+		mem->pages = pages;
+		mem->window.size = windows[i].size;
+		mem->window.map_size = 0;
+
+		epc->mem[i] = mem;
+	}
+	epc->mem_windows = num_windows;
 
 	return 0;
 
 err_mem:
-	kfree(mem);
+	for (; i >= 0; i--) {
+		kfree(mem->bitmap);
+		kfree(epc->mem[i]);
+	}
+	kfree(epc->mem);
 
-err:
-return ret;
+	return ret;
 }
 EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
 
@@ -101,48 +122,126 @@ EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
  */
 void pci_epc_mem_exit(struct pci_epc *epc)
 {
-	struct pci_epc_mem *mem = epc->mem;
+	struct pci_epc_mem *mem;
+	int i;
+
+	if (!epc->mem_windows)
+		return;
+
+	for (i = 0; i <= epc->mem_windows; i--) {
+		mem = epc->mem[i];
+		kfree(mem->bitmap);
+		kfree(epc->mem[i]);
+	}
+	kfree(epc->mem);
 
 	epc->mem = NULL;
-	kfree(mem->bitmap);
-	kfree(mem);
+	epc->mem_windows = 0;
 }
 EXPORT_SYMBOL_GPL(pci_epc_mem_exit);
 
+static int pci_epc_find_best_fit_window(struct pci_epc *epc, size_t size)
+{
+	size_t window_least_size = 0;
+	int best_fit_window = -1;
+	struct pci_epc_mem *mem;
+	size_t actual_size;
+	int i;
+
+	for (i = 0; i < epc->mem_windows; i++) {
+		mem = epc->mem[i];
+
+		/* if chunk from this region is already used skip it */
+		if (mem->window.map_size)
+			continue;
+
+		actual_size = ALIGN(size, mem->page_size);
+
+		if (best_fit_window == -1) {
+			best_fit_window = i;
+			window_least_size = mem->window.size;
+		} else {
+			if (actual_size <= mem->window.size &&
+			    mem->window.size < window_least_size) {
+				best_fit_window = i;
+				window_least_size = mem->window.size;
+			}
+		}
+	}
+
+	return best_fit_window;
+}
+
 /**
  * pci_epc_mem_alloc_addr() - allocate memory address from EPC addr space
  * @epc: the EPC device on which memory has to be allocated
  * @phys_addr: populate the allocated physical address here
+ * @window: populate the window here which will be used to map PCI address
  * @size: the size of the address space that has to be allocated
  *
  * Invoke to allocate memory address from the EPC address space. This
  * is usually done to map the remote RC address into the local system.
  */
 void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
-				     phys_addr_t *phys_addr, size_t size)
+				     phys_addr_t *phys_addr,
+				     int *window, size_t size)
 {
+	int best_fit = PCI_EPC_DEFAULT_WINDOW;
+	void __iomem *virt_addr = NULL;
+	struct pci_epc_mem *mem;
+	unsigned int page_shift;
 	int pageno;
-	void __iomem *virt_addr;
-	struct pci_epc_mem *mem = epc->mem;
-	unsigned int page_shift = ilog2(mem->page_size);
 	int order;
 
+	if (epc->mem_windows <= 0)
+		return NULL;
+
+	if (epc->mem_windows > 1) {
+		best_fit = pci_epc_find_best_fit_window(epc, size);
+		if (best_fit < 0)
+			return NULL;
+	}
+
+	mem = epc->mem[best_fit];
 	size = ALIGN(size, mem->page_size);
+	if (size > (mem->window.size - mem->window.map_size))
+		return NULL;
+	page_shift = ilog2(mem->page_size);
 	order = pci_epc_mem_get_order(mem, size);
 
 	pageno = bitmap_find_free_region(mem->bitmap, mem->pages, order);
 	if (pageno < 0)
 		return NULL;
 
-	*phys_addr = mem->phys_base + ((phys_addr_t)pageno << page_shift);
+	*phys_addr = mem->window.phys_base + (pageno << page_shift);
 	virt_addr = ioremap(*phys_addr, size);
-	if (!virt_addr)
+	if (!virt_addr) {
 		bitmap_release_region(mem->bitmap, pageno, order);
+	} else {
+		mem->window.map_size += size;
+		*window = best_fit;
+	}
 
 	return virt_addr;
 }
 EXPORT_SYMBOL_GPL(pci_epc_mem_alloc_addr);
 
+static int pci_epc_get_matching_window(struct pci_epc *epc,
+				       phys_addr_t phys_addr)
+{
+	struct pci_epc_mem *mem;
+	int i;
+
+	for (i = 0; i < epc->mem_windows; i++) {
+		mem = epc->mem[i];
+
+		if (mem->window.phys_base == phys_addr)
+			return i;
+	}
+
+	return -EINVAL;
+}
+
 /**
  * pci_epc_mem_free_addr() - free the allocated memory address
  * @epc: the EPC device on which memory was allocated
@@ -155,16 +254,26 @@ EXPORT_SYMBOL_GPL(pci_epc_mem_alloc_addr);
 void pci_epc_mem_free_addr(struct pci_epc *epc, phys_addr_t phys_addr,
 			   void __iomem *virt_addr, size_t size)
 {
+	struct pci_epc_mem *mem;
+	unsigned int page_shift;
+	int window = 0;
 	int pageno;
-	struct pci_epc_mem *mem = epc->mem;
-	unsigned int page_shift = ilog2(mem->page_size);
 	int order;
 
+	if (epc->mem_windows > 1) {
+		window = pci_epc_get_matching_window(epc, phys_addr);
+		if (window < 0)
+			return;
+	}
+
+	mem = epc->mem[window];
+	page_shift = ilog2(mem->page_size);
 	iounmap(virt_addr);
-	pageno = (phys_addr - mem->phys_base) >> page_shift;
+	pageno = (phys_addr - mem->window.phys_base) >> page_shift;
 	size = ALIGN(size, mem->page_size);
 	order = pci_epc_mem_get_order(mem, size);
 	bitmap_release_region(mem->bitmap, pageno, order);
+	mem->window.map_size -= size;
 }
 EXPORT_SYMBOL_GPL(pci_epc_mem_free_addr);
 
diff --git a/include/linux/pci-epc.h b/include/linux/pci-epc.h
index 56f1846b9d39..5e86d22dd5c5 100644
--- a/include/linux/pci-epc.h
+++ b/include/linux/pci-epc.h
@@ -48,7 +48,8 @@ struct pci_epc_ops {
 	void	(*clear_bar)(struct pci_epc *epc, u8 func_no,
 			     struct pci_epf_bar *epf_bar);
 	int	(*map_addr)(struct pci_epc *epc, u8 func_no,
-			    phys_addr_t addr, u64 pci_addr, size_t size);
+			    phys_addr_t addr, int window,
+			    u64 pci_addr, size_t size);
 	void	(*unmap_addr)(struct pci_epc *epc, u8 func_no,
 			      phys_addr_t addr);
 	int	(*set_msi)(struct pci_epc *epc, u8 func_no, u8 interrupts);
@@ -64,17 +65,28 @@ struct pci_epc_ops {
 	struct module *owner;
 };
 
+#define PCI_EPC_DEFAULT_WINDOW		0
+
+/**
+ * struct pci_epc_mem_window - address window of the endpoint controller
+ * @phys_base: physical base address of the PCI address window
+ * @size: the size of the PCI address window
+ */
+struct pci_epc_mem_window {
+	phys_addr_t	phys_base;
+	size_t		size;
+	size_t		map_size;
+};
+
 /**
  * struct pci_epc_mem - address space of the endpoint controller
- * @phys_base: physical base address of the PCI address space
- * @size: the size of the PCI address space
+ * @window: address window of the endpoint controller
  * @bitmap: bitmap to manage the PCI address space
- * @pages: number of bits representing the address region
  * @page_size: size of each page
+ * @pages: number of bits representing the address region
  */
 struct pci_epc_mem {
-	phys_addr_t	phys_base;
-	size_t		size;
+	struct pci_epc_mem_window window;
 	unsigned long	*bitmap;
 	size_t		page_size;
 	int		pages;
@@ -85,7 +97,8 @@ struct pci_epc_mem {
  * @dev: PCI EPC device
  * @pci_epf: list of endpoint functions present in this EPC device
  * @ops: function pointers for performing endpoint operations
- * @mem: address space of the endpoint controller
+ * @mem: array of address space of the endpoint controller
+ * @mem_windows: number of windows supported by device
  * @max_functions: max number of functions that can be configured in this EPC
  * @group: configfs group representing the PCI EPC device
  * @lock: spinlock to protect pci_epc ops
@@ -94,7 +107,8 @@ struct pci_epc {
 	struct device			dev;
 	struct list_head		pci_epf;
 	const struct pci_epc_ops	*ops;
-	struct pci_epc_mem		*mem;
+	struct pci_epc_mem		**mem;
+	int				mem_windows;
 	u8				max_functions;
 	struct config_group		*group;
 	/* spinlock to protect against concurrent access of EP controller */
@@ -128,8 +142,8 @@ struct pci_epc_features {
 #define devm_pci_epc_create(dev, ops)    \
 		__devm_pci_epc_create((dev), (ops), THIS_MODULE)
 
-#define pci_epc_mem_init(epc, phys_addr, size)	\
-		__pci_epc_mem_init((epc), (phys_addr), (size), PAGE_SIZE)
+#define pci_epc_mem_init(epc, windows, num_windows)	\
+		__pci_epc_mem_init((epc), windows, num_windows, PAGE_SIZE)
 
 static inline void epc_set_drvdata(struct pci_epc *epc, void *data)
 {
@@ -159,7 +173,7 @@ int pci_epc_set_bar(struct pci_epc *epc, u8 func_no,
 void pci_epc_clear_bar(struct pci_epc *epc, u8 func_no,
 		       struct pci_epf_bar *epf_bar);
 int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
-		     phys_addr_t phys_addr,
+		     phys_addr_t phys_addr, int window,
 		     u64 pci_addr, size_t size);
 void pci_epc_unmap_addr(struct pci_epc *epc, u8 func_no,
 			phys_addr_t phys_addr);
@@ -178,11 +192,12 @@ unsigned int pci_epc_get_first_free_bar(const struct pci_epc_features
 struct pci_epc *pci_epc_get(const char *epc_name);
 void pci_epc_put(struct pci_epc *epc);
 
-int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_addr, size_t size,
-		       size_t page_size);
+int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *window,
+		       int num_windows, size_t page_size);
 void pci_epc_mem_exit(struct pci_epc *epc);
 void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
-				     phys_addr_t *phys_addr, size_t size);
+				     phys_addr_t *phys_addr,
+				     int *window, size_t size);
 void pci_epc_mem_free_addr(struct pci_epc *epc, phys_addr_t phys_addr,
 			   void __iomem *virt_addr, size_t size);
 #endif /* __LINUX_PCI_EPC_H */
-- 
2.20.1


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply related	[flat|nested] 60+ messages in thread

* [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
  2019-11-06 19:36 ` Lad Prabhakar
@ 2019-11-06 19:36   ` Lad Prabhakar
  -1 siblings, 0 replies; 60+ messages in thread
From: Lad Prabhakar @ 2019-11-06 19:36 UTC (permalink / raw)
  To: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Kishon Vijay Abraham I, Marek Vasut,
	Yoshihiro Shimoda, linux-pci
  Cc: Catalin Marinas, Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray, devicetree, linux-kernel,
	linux-arm-kernel, linux-renesas-soc, Chris Paterson, Lad,
	Prabhakar

From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>

This patch adds the bindings for the R-Car PCIe endpoint driver.

Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
---
 .../devicetree/bindings/pci/rcar-pci-ep.txt   | 43 +++++++++++++++++++
 1 file changed, 43 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/pci/rcar-pci-ep.txt

diff --git a/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
new file mode 100644
index 000000000000..b8c8616ca007
--- /dev/null
+++ b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
@@ -0,0 +1,43 @@
+* Renesas R-Car PCIe Endpoint Controller DT description
+
+Required properties:
+	    "renesas,pcie-ep-r8a774c0" for the R8A774C0 SoC;
+	    "renesas,pcie-ep-rcar-gen3" for a generic R-Car Gen3 or
+				     RZ/G2 compatible device.
+
+	    When compatible with the generic version, nodes must list the
+	    SoC-specific version corresponding to the platform first
+	    followed by the generic version.
+
+- reg: Five register ranges as listed in the reg-names property
+- reg-names: Must include the following names
+	- "apb-base"
+	- "memory0"
+	- "memory1"
+	- "memory2"
+	- "memory3"
+- resets: Must contain phandles to PCIe-related reset lines exposed by IP block
+- clocks: from common clock binding: clock specifiers for the PCIe controller
+	 clock.
+- clock-names: from common clock binding: should be "pcie".
+
+Optional Property:
+- max-functions: Maximum number of functions that can be configured (default 1).
+
+Example:
+
+SoC-specific DT Entry:
+
+	pcie_ep: pcie_ep@fe000000 {
+		compatible = "renesas,pcie-r8a7791", "renesas,pcie-rcar-gen2";
+		reg = <0 0xfe000000 0 0x80000>,
+			<0x0 0xfe100000 0 0x100000>,
+			<0x0 0xfe200000 0 0x200000>,
+			<0x0 0x30000000 0 0x8000000>,
+			<0x0 0x38000000 0 0x8000000>;
+		reg-names = "apb-base", "memory0", "memory1", "memory2", "memory3";
+		clocks = <&cpg CPG_MOD 319>;
+		clock-names = "pcie";
+		power-domains = <&sysc R8A774C0_PD_ALWAYS_ON>;
+		resets = <&cpg 319>;
+	};
-- 
2.20.1


^ permalink raw reply related	[flat|nested] 60+ messages in thread

* [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
@ 2019-11-06 19:36   ` Lad Prabhakar
  0 siblings, 0 replies; 60+ messages in thread
From: Lad Prabhakar @ 2019-11-06 19:36 UTC (permalink / raw)
  To: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Kishon Vijay Abraham I, Marek Vasut,
	Yoshihiro Shimoda, linux-pci
  Cc: devicetree, Chris Paterson, Lorenzo Pieralisi, Arnd Bergmann,
	Catalin Marinas, linux-kernel, Lad, Prabhakar, linux-renesas-soc,
	Greg Kroah-Hartman, Andrew Murray, Will Deacon, linux-arm-kernel

From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>

This patch adds the bindings for the R-Car PCIe endpoint driver.

Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
---
 .../devicetree/bindings/pci/rcar-pci-ep.txt   | 43 +++++++++++++++++++
 1 file changed, 43 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/pci/rcar-pci-ep.txt

diff --git a/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
new file mode 100644
index 000000000000..b8c8616ca007
--- /dev/null
+++ b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
@@ -0,0 +1,43 @@
+* Renesas R-Car PCIe Endpoint Controller DT description
+
+Required properties:
+	    "renesas,pcie-ep-r8a774c0" for the R8A774C0 SoC;
+	    "renesas,pcie-ep-rcar-gen3" for a generic R-Car Gen3 or
+				     RZ/G2 compatible device.
+
+	    When compatible with the generic version, nodes must list the
+	    SoC-specific version corresponding to the platform first
+	    followed by the generic version.
+
+- reg: Five register ranges as listed in the reg-names property
+- reg-names: Must include the following names
+	- "apb-base"
+	- "memory0"
+	- "memory1"
+	- "memory2"
+	- "memory3"
+- resets: Must contain phandles to PCIe-related reset lines exposed by IP block
+- clocks: from common clock binding: clock specifiers for the PCIe controller
+	 clock.
+- clock-names: from common clock binding: should be "pcie".
+
+Optional Property:
+- max-functions: Maximum number of functions that can be configured (default 1).
+
+Example:
+
+SoC-specific DT Entry:
+
+	pcie_ep: pcie_ep@fe000000 {
+		compatible = "renesas,pcie-r8a7791", "renesas,pcie-rcar-gen2";
+		reg = <0 0xfe000000 0 0x80000>,
+			<0x0 0xfe100000 0 0x100000>,
+			<0x0 0xfe200000 0 0x200000>,
+			<0x0 0x30000000 0 0x8000000>,
+			<0x0 0x38000000 0 0x8000000>;
+		reg-names = "apb-base", "memory0", "memory1", "memory2", "memory3";
+		clocks = <&cpg CPG_MOD 319>;
+		clock-names = "pcie";
+		power-domains = <&sysc R8A774C0_PD_ALWAYS_ON>;
+		resets = <&cpg 319>;
+	};
-- 
2.20.1


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply related	[flat|nested] 60+ messages in thread

* [PATCH 4/5] pci: rcar: add support for rcar pcie controller in endpoint mode
  2019-11-06 19:36 ` Lad Prabhakar
@ 2019-11-06 19:36   ` Lad Prabhakar
  -1 siblings, 0 replies; 60+ messages in thread
From: Lad Prabhakar @ 2019-11-06 19:36 UTC (permalink / raw)
  To: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Kishon Vijay Abraham I, Marek Vasut,
	Yoshihiro Shimoda, linux-pci
  Cc: Catalin Marinas, Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray, devicetree, linux-kernel,
	linux-arm-kernel, linux-renesas-soc, Chris Paterson, Lad,
	Prabhakar

From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>

Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
---
 drivers/pci/controller/Kconfig        |   7 +
 drivers/pci/controller/Makefile       |   1 +
 drivers/pci/controller/pcie-rcar-ep.c | 483 ++++++++++++++++++++++++++
 3 files changed, 491 insertions(+)
 create mode 100644 drivers/pci/controller/pcie-rcar-ep.c

diff --git a/drivers/pci/controller/Kconfig b/drivers/pci/controller/Kconfig
index 01c238e5ea1e..94c541267bbc 100644
--- a/drivers/pci/controller/Kconfig
+++ b/drivers/pci/controller/Kconfig
@@ -89,6 +89,13 @@ config PCIE_RCAR_HOST
 	help
 	  Say Y here if you want PCIe controller support on R-Car SoCs in host mode.
 
+config PCIE_RCAR_EP
+	bool "Renesas R-Car PCIe endpoint controller"
+	depends on ARCH_RENESAS || COMPILE_TEST
+	depends on PCI_ENDPOINT
+	help
+	  Say Y here if you want PCIe controller support on R-Car SoCs in endpoint mode.
+
 config PCI_HOST_COMMON
 	bool
 	select PCI_ECAM
diff --git a/drivers/pci/controller/Makefile b/drivers/pci/controller/Makefile
index 3577902a0d40..16aa1f7cdc1c 100644
--- a/drivers/pci/controller/Makefile
+++ b/drivers/pci/controller/Makefile
@@ -10,6 +10,7 @@ obj-$(CONFIG_PCI_AARDVARK) += pci-aardvark.o
 obj-$(CONFIG_PCI_TEGRA) += pci-tegra.o
 obj-$(CONFIG_PCI_RCAR_GEN2) += pci-rcar-gen2.o
 obj-$(CONFIG_PCIE_RCAR_HOST) += pcie-rcar.o pcie-rcar-host.o
+obj-$(CONFIG_PCIE_RCAR_EP) += pcie-rcar.o pcie-rcar-ep.o
 obj-$(CONFIG_PCI_HOST_COMMON) += pci-host-common.o
 obj-$(CONFIG_PCI_HOST_GENERIC) += pci-host-generic.o
 obj-$(CONFIG_PCIE_XILINX) += pcie-xilinx.o
diff --git a/drivers/pci/controller/pcie-rcar-ep.c b/drivers/pci/controller/pcie-rcar-ep.c
new file mode 100644
index 000000000000..ef4115d17052
--- /dev/null
+++ b/drivers/pci/controller/pcie-rcar-ep.c
@@ -0,0 +1,483 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * PCIe endpoint driver for Renesas R-Car SoCs
+ *  Copyright (c) 2019 Renesas Electronics Europe GmbH
+ *
+ * Author: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/of_pci.h>
+#include <linux/of_platform.h>
+#include <linux/pci.h>
+#include <linux/pci-epc.h>
+#include <linux/phy/phy.h>
+#include <linux/platform_device.h>
+
+#include "pcie-rcar.h"
+
+#define MAX_NR_INBOUND_MAPS	6
+#define MAX_NR_OUTBOUND_MAPS	4
+
+/* Structure representing the PCIe interface */
+struct rcar_pcie {
+	phys_addr_t		*ob_addr;
+	struct pci_epc_mem_window *ob_window;
+	struct pci_epc		*epc;
+	struct device		*dev;
+	void __iomem		*base;
+	u8			max_functions;
+	unsigned int		bar_to_atu[MAX_NR_INBOUND_MAPS];
+	unsigned long		*ib_window_map;
+	u32			num_ib_windows;
+	u32			num_ob_windows;
+};
+
+static void rcar_pcie_ep_hw_init(struct rcar_pcie *pcie)
+{
+	u32 val;
+
+	rcar_pci_write_reg(pcie->base, 0, PCIETCTLR);
+
+	/* Set endpoint mode */
+	rcar_pci_write_reg(pcie->base, 0, PCIEMSR);
+
+	/* Initialize default capabilities. */
+	rcar_rmw32(pcie->base, REXPCAP(0), 0xff, PCI_CAP_ID_EXP);
+	rcar_rmw32(pcie->base, REXPCAP(PCI_EXP_FLAGS),
+		   PCI_EXP_FLAGS_TYPE, PCI_EXP_TYPE_ENDPOINT << 4);
+	rcar_rmw32(pcie->base, RCONF(PCI_HEADER_TYPE), 0x7f,
+		   PCI_HEADER_TYPE_NORMAL);
+
+	/* Write out the physical slot number = 0 */
+	rcar_rmw32(pcie->base, REXPCAP(PCI_EXP_SLTCAP), PCI_EXP_SLTCAP_PSN, 0);
+
+	val = rcar_pci_read_reg(pcie->base, EXPCAP(1));
+	/* device supports fixed 128 bytes MPSS */
+	val &= ~GENMASK(2, 0);
+	/* L1 to L0 transition latency no time limit */
+	val |= GENMASK(11, 9);
+	/* L0s to L0 transistion no time limit */
+	val |= GENMASK(8, 6);
+	rcar_pci_write_reg(pcie->base, val, EXPCAP(1));
+
+	val = rcar_pci_read_reg(pcie->base, EXPCAP(2));
+	/* read requests size 128 bytes */
+	val &= ~GENMASK(14, 12);
+	/* payload size 128 bytes */
+	val &= ~GENMASK(7, 5);
+	/* disable relaxed ordering transaction */
+	val &= ~BIT(4);
+	rcar_pci_write_reg(pcie->base, val, EXPCAP(2));
+
+	val = rcar_pci_read_reg(pcie->base, EXPCAP(4));
+	/* disable ASPM control */
+	val &= ~GENMASK(1, 0);
+	rcar_pci_write_reg(pcie->base, val, EXPCAP(4));
+
+	/* Set target link speed to 5.0 GT/s */
+	rcar_rmw32(pcie->base, EXPCAP(12), PCI_EXP_LNKSTA_CLS,
+		   PCI_EXP_LNKSTA_CLS_5_0GB);
+
+	/* Set the completion timer timeout to the maximum 50ms. */
+	rcar_rmw32(pcie->base, TLCTLR + 1, 0x3f, 50);
+
+	/* Terminate list of capabilities (Next Capability Offset=0) */
+	rcar_rmw32(pcie->base, RVCCAP(0), 0xfff00000, 0);
+
+	/* flush modifications */
+	wmb();
+}
+
+static int rcar_pcie_ep_get_pdata(struct rcar_pcie *pcie,
+				  struct platform_device *pdev)
+{
+	struct pci_epc_mem_window *window;
+	struct device *dev = pcie->dev;
+	char outbound_name[10];
+	struct resource *res;
+	unsigned int i;
+	int err;
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "apb-base");
+	pcie->base = devm_ioremap_resource(dev, res);
+	if (IS_ERR(pcie->base)) {
+		dev_err(dev, "pcie apb-base missing\n");
+		return PTR_ERR(pcie->base);
+	}
+
+	pcie->ob_window = devm_kcalloc(dev, MAX_NR_OUTBOUND_MAPS,
+				       sizeof(*window), GFP_KERNEL);
+	if (!pcie->ob_window)
+		return -ENOMEM;
+
+	pcie->num_ob_windows = 0;
+	for (i = 0; i < MAX_NR_OUTBOUND_MAPS; i++) {
+		sprintf(outbound_name, "memory%u", i);
+		res = platform_get_resource_byname(pdev,
+						   IORESOURCE_MEM,
+						   outbound_name);
+		if (!res) {
+			dev_err(dev, "missing outbound window %u\n", i);
+			return -EINVAL;
+		}
+		if (!devm_request_mem_region(&pdev->dev, res->start,
+					     resource_size(res),
+					     outbound_name)) {
+			dev_err(dev,
+				"Cannot request memory region %s.\n",
+				outbound_name);
+			return -EIO;
+		}
+		pcie->ob_window[i].phys_base = res->start;
+		pcie->ob_window[i].size = resource_size(res);
+		pcie->num_ob_windows++;
+	}
+
+	err = of_property_read_u8(dev->of_node, "max-functions",
+				  &pcie->max_functions);
+	if (err < 0)
+		pcie->max_functions = 1;
+
+	return 0;
+}
+
+static int rcar_pcie_ep_write_header(struct pci_epc *epc, u8 fn,
+				     struct pci_epf_header *hdr)
+{
+	struct rcar_pcie *ep = epc_get_drvdata(epc);
+	u32 val;
+
+	if (!fn)
+		val = hdr->vendorid;
+	else
+		val = rcar_pci_read_reg(ep->base, IDSETR0);
+	val |= hdr->deviceid << DEVICE_ID_SHFIT;
+	rcar_pci_write_reg(ep->base, val, IDSETR0);
+
+	val = hdr->revid;
+	val |= hdr->progif_code << 8;
+	val |= hdr->subclass_code << 16;
+	val |= hdr->baseclass_code << 24;
+	rcar_pci_write_reg(ep->base, val, IDSETR1);
+
+	if (!fn)
+		val = hdr->subsys_vendor_id;
+	else
+		val = rcar_pci_read_reg(ep->base, SUBIDSETR);
+	val |= hdr->subsys_id << 16;
+	rcar_pci_write_reg(ep->base, val, SUBIDSETR);
+
+	if (hdr->interrupt_pin > PCI_INTERRUPT_INTA)
+		return -EINVAL;
+	val = rcar_pci_read_reg(ep->base, PCICONF(15));
+	val |= (hdr->interrupt_pin << 8);
+	rcar_pci_write_reg(ep->base, val, PCICONF(15));
+
+	return 0;
+}
+
+static int rcar_pcie_ep_set_bar(struct pci_epc *epc, u8 func_no,
+				struct pci_epf_bar *epf_bar)
+{
+	struct rcar_pcie *ep = epc_get_drvdata(epc);
+	dma_addr_t cpu_addr = epf_bar->phys_addr;
+	int flags = epf_bar->flags | LAR_ENABLE | LAM_64BIT;
+	enum pci_barno bar = epf_bar->barno;
+	u64 size = 1ULL << fls64(epf_bar->size - 1);
+	u32 mask;
+	int idx;
+	int err;
+
+	idx = find_first_zero_bit(ep->ib_window_map, ep->num_ib_windows);
+	if (idx >= ep->num_ib_windows) {
+		dev_err(ep->dev, "no free inbound window\n");
+		return -EINVAL;
+	}
+
+	if ((flags & PCI_BASE_ADDRESS_SPACE) == PCI_BASE_ADDRESS_SPACE_IO)
+		flags |= IO_SPACE;
+
+	ep->bar_to_atu[bar] = idx;
+	/* use 64 bit bars */
+	set_bit(idx, ep->ib_window_map);
+	set_bit(idx + 1, ep->ib_window_map);
+
+	if (cpu_addr > 0) {
+		unsigned long nr_zeros = __ffs64(cpu_addr);
+		u64 alignment = 1ULL << nr_zeros;
+
+		size = min(size, alignment);
+	} else {
+		size = size;
+	}
+
+	size = min(size, 1ULL << 32);
+
+	mask = roundup_pow_of_two(size) - 1;
+	mask &= ~0xf;
+
+	rcar_pcie_set_inbound(ep->base, cpu_addr,
+			      0x0, mask | flags, idx, false);
+
+	err = rcar_pcie_wait_for_phyrdy(ep->base);
+	if (err) {
+		dev_err(ep->dev, "phy not ready\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static void rcar_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
+				   struct pci_epf_bar *epf_bar)
+{
+	struct rcar_pcie *ep = epc_get_drvdata(epc);
+	enum pci_barno bar = epf_bar->barno;
+	u32 atu_index = ep->bar_to_atu[bar];
+
+	rcar_pcie_set_inbound(ep->base, 0x0, 0x0, 0x0, bar, false);
+
+	clear_bit(atu_index, ep->ib_window_map);
+	clear_bit(atu_index + 1, ep->ib_window_map);
+}
+
+static int rcar_pcie_ep_map_addr(struct pci_epc *epc, u8 fn,
+				 phys_addr_t addr, int window,
+				 u64 pci_addr, size_t size)
+{
+	struct rcar_pcie *ep = epc_get_drvdata(epc);
+	struct resource res;
+	int err;
+
+	/* check if we have a link. */
+	err = rcar_pcie_wait_for_dl(ep->base);
+	if (err) {
+		dev_err(ep->dev, "link not up\n");
+		memset(&res, 0x0, sizeof(res));
+		rcar_pcie_set_outbound(window, ep->base, &res, false);
+		return err;
+	}
+
+	if (window >= ep->num_ob_windows) {
+		dev_err(ep->dev, "no free outbound window\n");
+		return -EINVAL;
+	}
+
+	memset(&res, 0x0, sizeof(res));
+	res.start = pci_addr;
+	res.end = pci_addr + size - 1;
+	res.flags = IORESOURCE_MEM;
+
+	rcar_pcie_set_outbound(window, ep->base, &res, false);
+
+	ep->ob_addr[window] = addr;
+
+	return 0;
+}
+
+static void rcar_pcie_ep_unmap_addr(struct pci_epc *epc, u8 fn,
+				    phys_addr_t addr)
+{
+	struct rcar_pcie *ep = epc_get_drvdata(epc);
+	struct resource res;
+	int idx;
+
+	for (idx = 0; idx < ep->num_ob_windows; idx++)
+		if (ep->ob_addr[idx] == addr)
+			break;
+
+	if (idx >= ep->num_ob_windows)
+		return;
+
+	memset(&res, 0x0, sizeof(res));
+	rcar_pcie_set_outbound(idx, ep->base, &res, false);
+
+	ep->ob_addr[idx] = 0;
+}
+
+static int rcar_pcie_ep_assert_intx(struct rcar_pcie *ep, u8 fn, u8 intx)
+{
+	u32 val;
+
+	val = rcar_pci_read_reg(ep->base, PCIEMSITXR);
+	if ((val & PCI_MSI_FLAGS_ENABLE)) {
+		dev_err(ep->dev, "MSI is enabled, cannot assert INTx\n");
+		return -EINVAL;
+	}
+
+	val = rcar_pci_read_reg(ep->base, PCICONF(1));
+	if ((val & INTDIS_SHIFT)) {
+		dev_err(ep->dev, "INTx message transmission is disabled\n");
+		return -EINVAL;
+	}
+
+	val = rcar_pci_read_reg(ep->base, PCIEINTXR);
+	if ((val & ASTINTX_SHIFT)) {
+		dev_err(ep->dev, "INTx is already asserted\n");
+		return -EINVAL;
+	}
+
+	val |= ASTINTX_SHIFT;
+	rcar_pci_write_reg(ep->base, val, PCIEINTXR);
+	mdelay(1);
+	val = rcar_pci_read_reg(ep->base, PCIEINTXR);
+	val &= ~ASTINTX_SHIFT;
+	rcar_pci_write_reg(ep->base, val, PCIEINTXR);
+
+	return 0;
+}
+
+static int rcar_pcie_ep_raise_irq(struct pci_epc *epc, u8 fn,
+				  enum pci_epc_irq_type type,
+				  u16 interrupt_num)
+{
+	struct rcar_pcie *ep = epc_get_drvdata(epc);
+
+	switch (type) {
+	case PCI_EPC_IRQ_LEGACY:
+		return rcar_pcie_ep_assert_intx(ep, fn, 0);
+	default:
+		return -EINVAL;
+	}
+}
+
+static int rcar_pcie_ep_start(struct pci_epc *epc)
+{
+	struct rcar_pcie *ep = epc_get_drvdata(epc);
+
+	rcar_pci_write_reg(ep->base, CFINIT, PCIETCTLR);
+
+	return 0;
+}
+
+static void rcar_pcie_ep_stop(struct pci_epc *epc)
+{
+	struct rcar_pcie *ep = epc_get_drvdata(epc);
+
+	rcar_pci_write_reg(ep->base, 0, PCIETCTLR);
+}
+
+static const struct pci_epc_features rcar_pcie_epc_features = {
+	.linkup_notifier = false,
+	.msi_capable = false,
+	.msix_capable = false,
+	/* use 64-bit bars so mark bar1/3/5 as reserved */
+	.reserved_bar = 1 << BAR_1 | 1 << BAR_3 | 1 << BAR_5,
+	.bar_fixed_64bit =  (1 << BAR_0) | (1 << BAR_2) | (1 << BAR_4),
+	.bar_fixed_size[0] = 128,
+	.bar_fixed_size[2] = 256,
+	.bar_fixed_size[4] = 256,
+};
+
+static const struct pci_epc_features*
+rcar_pcie_ep_get_features(struct pci_epc *epc, u8 func_no)
+{
+	return &rcar_pcie_epc_features;
+}
+
+static const struct pci_epc_ops rcar_pcie_epc_ops = {
+	.write_header	= rcar_pcie_ep_write_header,
+	.set_bar	= rcar_pcie_ep_set_bar,
+	.clear_bar	= rcar_pcie_ep_clear_bar,
+	.map_addr	= rcar_pcie_ep_map_addr,
+	.unmap_addr	= rcar_pcie_ep_unmap_addr,
+	.raise_irq	= rcar_pcie_ep_raise_irq,
+	.start		= rcar_pcie_ep_start,
+	.stop		= rcar_pcie_ep_stop,
+	.get_features	= rcar_pcie_ep_get_features,
+};
+
+static const struct of_device_id rcar_pcie_ep_of_match[] = {
+	{
+		.compatible = "renesas,pcie-ep-rcar-gen3"
+	},
+	{},
+};
+
+static int rcar_pcie_ep_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct rcar_pcie *pcie;
+	struct pci_epc *epc;
+	int err;
+
+	pcie = devm_kzalloc(dev, sizeof(*pcie), GFP_KERNEL);
+	if (!pcie)
+		return -ENOMEM;
+
+	pcie->dev = dev;
+
+	pm_runtime_enable(pcie->dev);
+	err = pm_runtime_get_sync(pcie->dev);
+	if (err < 0) {
+		dev_err(pcie->dev, "pm_runtime_get_sync failed\n");
+		goto err_pm_disable;
+	}
+
+	err = rcar_pcie_ep_get_pdata(pcie, pdev);
+	if (err < 0) {
+		dev_err(dev, "failed to request resources: %d\n", err);
+		goto err_pm_put;
+	}
+
+	pcie->num_ib_windows = MAX_NR_INBOUND_MAPS;
+	pcie->ib_window_map =
+			devm_kcalloc(dev, BITS_TO_LONGS(pcie->num_ib_windows),
+				     sizeof(long), GFP_KERNEL);
+	if (!pcie->ib_window_map) {
+		err = -ENOMEM;
+		dev_err(dev, "failed to allocate memory for inbound map\n");
+		goto err_pm_put;
+	}
+
+	pcie->ob_addr = devm_kcalloc(dev, pcie->num_ob_windows,
+				     sizeof(*pcie->ob_addr), GFP_KERNEL);
+	if (!pcie->ob_addr) {
+		err = -ENOMEM;
+		dev_err(dev, "failed to allocate memory for outbound memory pointers\n");
+		goto err_pm_put;
+	}
+
+	epc = devm_pci_epc_create(dev, &rcar_pcie_epc_ops);
+	if (IS_ERR(epc)) {
+		dev_err(dev, "failed to create epc device\n");
+		err = PTR_ERR(epc);
+		goto err_pm_put;
+	}
+
+	epc->max_functions = pcie->max_functions;
+	pcie->epc = epc;
+	epc_set_drvdata(epc, pcie);
+
+	err = pci_epc_mem_init(epc, pcie->ob_window, pcie->num_ob_windows);
+	if (err < 0) {
+		dev_err(dev, "failed to initialize the epc memory space\n");
+		goto err_pm_put;
+	}
+
+	rcar_pcie_ep_hw_init(pcie);
+
+	return 0;
+
+err_pm_put:
+	pm_runtime_put(dev);
+
+err_pm_disable:
+	pm_runtime_disable(dev);
+
+	return err;
+}
+
+static struct platform_driver rcar_pcie_ep_driver = {
+	.driver = {
+		.name = "rcar-pcie-ep",
+		.of_match_table = rcar_pcie_ep_of_match,
+		.suppress_bind_attrs = true,
+	},
+	.probe = rcar_pcie_ep_probe,
+};
+builtin_platform_driver(rcar_pcie_ep_driver);
-- 
2.20.1


^ permalink raw reply related	[flat|nested] 60+ messages in thread

* [PATCH 4/5] pci: rcar: add support for rcar pcie controller in endpoint mode
@ 2019-11-06 19:36   ` Lad Prabhakar
  0 siblings, 0 replies; 60+ messages in thread
From: Lad Prabhakar @ 2019-11-06 19:36 UTC (permalink / raw)
  To: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Kishon Vijay Abraham I, Marek Vasut,
	Yoshihiro Shimoda, linux-pci
  Cc: devicetree, Chris Paterson, Lorenzo Pieralisi, Arnd Bergmann,
	Catalin Marinas, linux-kernel, Lad, Prabhakar, linux-renesas-soc,
	Greg Kroah-Hartman, Andrew Murray, Will Deacon, linux-arm-kernel

From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>

Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
---
 drivers/pci/controller/Kconfig        |   7 +
 drivers/pci/controller/Makefile       |   1 +
 drivers/pci/controller/pcie-rcar-ep.c | 483 ++++++++++++++++++++++++++
 3 files changed, 491 insertions(+)
 create mode 100644 drivers/pci/controller/pcie-rcar-ep.c

diff --git a/drivers/pci/controller/Kconfig b/drivers/pci/controller/Kconfig
index 01c238e5ea1e..94c541267bbc 100644
--- a/drivers/pci/controller/Kconfig
+++ b/drivers/pci/controller/Kconfig
@@ -89,6 +89,13 @@ config PCIE_RCAR_HOST
 	help
 	  Say Y here if you want PCIe controller support on R-Car SoCs in host mode.
 
+config PCIE_RCAR_EP
+	bool "Renesas R-Car PCIe endpoint controller"
+	depends on ARCH_RENESAS || COMPILE_TEST
+	depends on PCI_ENDPOINT
+	help
+	  Say Y here if you want PCIe controller support on R-Car SoCs in endpoint mode.
+
 config PCI_HOST_COMMON
 	bool
 	select PCI_ECAM
diff --git a/drivers/pci/controller/Makefile b/drivers/pci/controller/Makefile
index 3577902a0d40..16aa1f7cdc1c 100644
--- a/drivers/pci/controller/Makefile
+++ b/drivers/pci/controller/Makefile
@@ -10,6 +10,7 @@ obj-$(CONFIG_PCI_AARDVARK) += pci-aardvark.o
 obj-$(CONFIG_PCI_TEGRA) += pci-tegra.o
 obj-$(CONFIG_PCI_RCAR_GEN2) += pci-rcar-gen2.o
 obj-$(CONFIG_PCIE_RCAR_HOST) += pcie-rcar.o pcie-rcar-host.o
+obj-$(CONFIG_PCIE_RCAR_EP) += pcie-rcar.o pcie-rcar-ep.o
 obj-$(CONFIG_PCI_HOST_COMMON) += pci-host-common.o
 obj-$(CONFIG_PCI_HOST_GENERIC) += pci-host-generic.o
 obj-$(CONFIG_PCIE_XILINX) += pcie-xilinx.o
diff --git a/drivers/pci/controller/pcie-rcar-ep.c b/drivers/pci/controller/pcie-rcar-ep.c
new file mode 100644
index 000000000000..ef4115d17052
--- /dev/null
+++ b/drivers/pci/controller/pcie-rcar-ep.c
@@ -0,0 +1,483 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * PCIe endpoint driver for Renesas R-Car SoCs
+ *  Copyright (c) 2019 Renesas Electronics Europe GmbH
+ *
+ * Author: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/of_pci.h>
+#include <linux/of_platform.h>
+#include <linux/pci.h>
+#include <linux/pci-epc.h>
+#include <linux/phy/phy.h>
+#include <linux/platform_device.h>
+
+#include "pcie-rcar.h"
+
+#define MAX_NR_INBOUND_MAPS	6
+#define MAX_NR_OUTBOUND_MAPS	4
+
+/* Structure representing the PCIe interface */
+struct rcar_pcie {
+	phys_addr_t		*ob_addr;
+	struct pci_epc_mem_window *ob_window;
+	struct pci_epc		*epc;
+	struct device		*dev;
+	void __iomem		*base;
+	u8			max_functions;
+	unsigned int		bar_to_atu[MAX_NR_INBOUND_MAPS];
+	unsigned long		*ib_window_map;
+	u32			num_ib_windows;
+	u32			num_ob_windows;
+};
+
+static void rcar_pcie_ep_hw_init(struct rcar_pcie *pcie)
+{
+	u32 val;
+
+	rcar_pci_write_reg(pcie->base, 0, PCIETCTLR);
+
+	/* Set endpoint mode */
+	rcar_pci_write_reg(pcie->base, 0, PCIEMSR);
+
+	/* Initialize default capabilities. */
+	rcar_rmw32(pcie->base, REXPCAP(0), 0xff, PCI_CAP_ID_EXP);
+	rcar_rmw32(pcie->base, REXPCAP(PCI_EXP_FLAGS),
+		   PCI_EXP_FLAGS_TYPE, PCI_EXP_TYPE_ENDPOINT << 4);
+	rcar_rmw32(pcie->base, RCONF(PCI_HEADER_TYPE), 0x7f,
+		   PCI_HEADER_TYPE_NORMAL);
+
+	/* Write out the physical slot number = 0 */
+	rcar_rmw32(pcie->base, REXPCAP(PCI_EXP_SLTCAP), PCI_EXP_SLTCAP_PSN, 0);
+
+	val = rcar_pci_read_reg(pcie->base, EXPCAP(1));
+	/* device supports fixed 128 bytes MPSS */
+	val &= ~GENMASK(2, 0);
+	/* L1 to L0 transition latency no time limit */
+	val |= GENMASK(11, 9);
+	/* L0s to L0 transistion no time limit */
+	val |= GENMASK(8, 6);
+	rcar_pci_write_reg(pcie->base, val, EXPCAP(1));
+
+	val = rcar_pci_read_reg(pcie->base, EXPCAP(2));
+	/* read requests size 128 bytes */
+	val &= ~GENMASK(14, 12);
+	/* payload size 128 bytes */
+	val &= ~GENMASK(7, 5);
+	/* disable relaxed ordering transaction */
+	val &= ~BIT(4);
+	rcar_pci_write_reg(pcie->base, val, EXPCAP(2));
+
+	val = rcar_pci_read_reg(pcie->base, EXPCAP(4));
+	/* disable ASPM control */
+	val &= ~GENMASK(1, 0);
+	rcar_pci_write_reg(pcie->base, val, EXPCAP(4));
+
+	/* Set target link speed to 5.0 GT/s */
+	rcar_rmw32(pcie->base, EXPCAP(12), PCI_EXP_LNKSTA_CLS,
+		   PCI_EXP_LNKSTA_CLS_5_0GB);
+
+	/* Set the completion timer timeout to the maximum 50ms. */
+	rcar_rmw32(pcie->base, TLCTLR + 1, 0x3f, 50);
+
+	/* Terminate list of capabilities (Next Capability Offset=0) */
+	rcar_rmw32(pcie->base, RVCCAP(0), 0xfff00000, 0);
+
+	/* flush modifications */
+	wmb();
+}
+
+static int rcar_pcie_ep_get_pdata(struct rcar_pcie *pcie,
+				  struct platform_device *pdev)
+{
+	struct pci_epc_mem_window *window;
+	struct device *dev = pcie->dev;
+	char outbound_name[10];
+	struct resource *res;
+	unsigned int i;
+	int err;
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "apb-base");
+	pcie->base = devm_ioremap_resource(dev, res);
+	if (IS_ERR(pcie->base)) {
+		dev_err(dev, "pcie apb-base missing\n");
+		return PTR_ERR(pcie->base);
+	}
+
+	pcie->ob_window = devm_kcalloc(dev, MAX_NR_OUTBOUND_MAPS,
+				       sizeof(*window), GFP_KERNEL);
+	if (!pcie->ob_window)
+		return -ENOMEM;
+
+	pcie->num_ob_windows = 0;
+	for (i = 0; i < MAX_NR_OUTBOUND_MAPS; i++) {
+		sprintf(outbound_name, "memory%u", i);
+		res = platform_get_resource_byname(pdev,
+						   IORESOURCE_MEM,
+						   outbound_name);
+		if (!res) {
+			dev_err(dev, "missing outbound window %u\n", i);
+			return -EINVAL;
+		}
+		if (!devm_request_mem_region(&pdev->dev, res->start,
+					     resource_size(res),
+					     outbound_name)) {
+			dev_err(dev,
+				"Cannot request memory region %s.\n",
+				outbound_name);
+			return -EIO;
+		}
+		pcie->ob_window[i].phys_base = res->start;
+		pcie->ob_window[i].size = resource_size(res);
+		pcie->num_ob_windows++;
+	}
+
+	err = of_property_read_u8(dev->of_node, "max-functions",
+				  &pcie->max_functions);
+	if (err < 0)
+		pcie->max_functions = 1;
+
+	return 0;
+}
+
+static int rcar_pcie_ep_write_header(struct pci_epc *epc, u8 fn,
+				     struct pci_epf_header *hdr)
+{
+	struct rcar_pcie *ep = epc_get_drvdata(epc);
+	u32 val;
+
+	if (!fn)
+		val = hdr->vendorid;
+	else
+		val = rcar_pci_read_reg(ep->base, IDSETR0);
+	val |= hdr->deviceid << DEVICE_ID_SHFIT;
+	rcar_pci_write_reg(ep->base, val, IDSETR0);
+
+	val = hdr->revid;
+	val |= hdr->progif_code << 8;
+	val |= hdr->subclass_code << 16;
+	val |= hdr->baseclass_code << 24;
+	rcar_pci_write_reg(ep->base, val, IDSETR1);
+
+	if (!fn)
+		val = hdr->subsys_vendor_id;
+	else
+		val = rcar_pci_read_reg(ep->base, SUBIDSETR);
+	val |= hdr->subsys_id << 16;
+	rcar_pci_write_reg(ep->base, val, SUBIDSETR);
+
+	if (hdr->interrupt_pin > PCI_INTERRUPT_INTA)
+		return -EINVAL;
+	val = rcar_pci_read_reg(ep->base, PCICONF(15));
+	val |= (hdr->interrupt_pin << 8);
+	rcar_pci_write_reg(ep->base, val, PCICONF(15));
+
+	return 0;
+}
+
+static int rcar_pcie_ep_set_bar(struct pci_epc *epc, u8 func_no,
+				struct pci_epf_bar *epf_bar)
+{
+	struct rcar_pcie *ep = epc_get_drvdata(epc);
+	dma_addr_t cpu_addr = epf_bar->phys_addr;
+	int flags = epf_bar->flags | LAR_ENABLE | LAM_64BIT;
+	enum pci_barno bar = epf_bar->barno;
+	u64 size = 1ULL << fls64(epf_bar->size - 1);
+	u32 mask;
+	int idx;
+	int err;
+
+	idx = find_first_zero_bit(ep->ib_window_map, ep->num_ib_windows);
+	if (idx >= ep->num_ib_windows) {
+		dev_err(ep->dev, "no free inbound window\n");
+		return -EINVAL;
+	}
+
+	if ((flags & PCI_BASE_ADDRESS_SPACE) == PCI_BASE_ADDRESS_SPACE_IO)
+		flags |= IO_SPACE;
+
+	ep->bar_to_atu[bar] = idx;
+	/* use 64 bit bars */
+	set_bit(idx, ep->ib_window_map);
+	set_bit(idx + 1, ep->ib_window_map);
+
+	if (cpu_addr > 0) {
+		unsigned long nr_zeros = __ffs64(cpu_addr);
+		u64 alignment = 1ULL << nr_zeros;
+
+		size = min(size, alignment);
+	} else {
+		size = size;
+	}
+
+	size = min(size, 1ULL << 32);
+
+	mask = roundup_pow_of_two(size) - 1;
+	mask &= ~0xf;
+
+	rcar_pcie_set_inbound(ep->base, cpu_addr,
+			      0x0, mask | flags, idx, false);
+
+	err = rcar_pcie_wait_for_phyrdy(ep->base);
+	if (err) {
+		dev_err(ep->dev, "phy not ready\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static void rcar_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
+				   struct pci_epf_bar *epf_bar)
+{
+	struct rcar_pcie *ep = epc_get_drvdata(epc);
+	enum pci_barno bar = epf_bar->barno;
+	u32 atu_index = ep->bar_to_atu[bar];
+
+	rcar_pcie_set_inbound(ep->base, 0x0, 0x0, 0x0, bar, false);
+
+	clear_bit(atu_index, ep->ib_window_map);
+	clear_bit(atu_index + 1, ep->ib_window_map);
+}
+
+static int rcar_pcie_ep_map_addr(struct pci_epc *epc, u8 fn,
+				 phys_addr_t addr, int window,
+				 u64 pci_addr, size_t size)
+{
+	struct rcar_pcie *ep = epc_get_drvdata(epc);
+	struct resource res;
+	int err;
+
+	/* check if we have a link. */
+	err = rcar_pcie_wait_for_dl(ep->base);
+	if (err) {
+		dev_err(ep->dev, "link not up\n");
+		memset(&res, 0x0, sizeof(res));
+		rcar_pcie_set_outbound(window, ep->base, &res, false);
+		return err;
+	}
+
+	if (window >= ep->num_ob_windows) {
+		dev_err(ep->dev, "no free outbound window\n");
+		return -EINVAL;
+	}
+
+	memset(&res, 0x0, sizeof(res));
+	res.start = pci_addr;
+	res.end = pci_addr + size - 1;
+	res.flags = IORESOURCE_MEM;
+
+	rcar_pcie_set_outbound(window, ep->base, &res, false);
+
+	ep->ob_addr[window] = addr;
+
+	return 0;
+}
+
+static void rcar_pcie_ep_unmap_addr(struct pci_epc *epc, u8 fn,
+				    phys_addr_t addr)
+{
+	struct rcar_pcie *ep = epc_get_drvdata(epc);
+	struct resource res;
+	int idx;
+
+	for (idx = 0; idx < ep->num_ob_windows; idx++)
+		if (ep->ob_addr[idx] == addr)
+			break;
+
+	if (idx >= ep->num_ob_windows)
+		return;
+
+	memset(&res, 0x0, sizeof(res));
+	rcar_pcie_set_outbound(idx, ep->base, &res, false);
+
+	ep->ob_addr[idx] = 0;
+}
+
+static int rcar_pcie_ep_assert_intx(struct rcar_pcie *ep, u8 fn, u8 intx)
+{
+	u32 val;
+
+	val = rcar_pci_read_reg(ep->base, PCIEMSITXR);
+	if ((val & PCI_MSI_FLAGS_ENABLE)) {
+		dev_err(ep->dev, "MSI is enabled, cannot assert INTx\n");
+		return -EINVAL;
+	}
+
+	val = rcar_pci_read_reg(ep->base, PCICONF(1));
+	if ((val & INTDIS_SHIFT)) {
+		dev_err(ep->dev, "INTx message transmission is disabled\n");
+		return -EINVAL;
+	}
+
+	val = rcar_pci_read_reg(ep->base, PCIEINTXR);
+	if ((val & ASTINTX_SHIFT)) {
+		dev_err(ep->dev, "INTx is already asserted\n");
+		return -EINVAL;
+	}
+
+	val |= ASTINTX_SHIFT;
+	rcar_pci_write_reg(ep->base, val, PCIEINTXR);
+	mdelay(1);
+	val = rcar_pci_read_reg(ep->base, PCIEINTXR);
+	val &= ~ASTINTX_SHIFT;
+	rcar_pci_write_reg(ep->base, val, PCIEINTXR);
+
+	return 0;
+}
+
+static int rcar_pcie_ep_raise_irq(struct pci_epc *epc, u8 fn,
+				  enum pci_epc_irq_type type,
+				  u16 interrupt_num)
+{
+	struct rcar_pcie *ep = epc_get_drvdata(epc);
+
+	switch (type) {
+	case PCI_EPC_IRQ_LEGACY:
+		return rcar_pcie_ep_assert_intx(ep, fn, 0);
+	default:
+		return -EINVAL;
+	}
+}
+
+static int rcar_pcie_ep_start(struct pci_epc *epc)
+{
+	struct rcar_pcie *ep = epc_get_drvdata(epc);
+
+	rcar_pci_write_reg(ep->base, CFINIT, PCIETCTLR);
+
+	return 0;
+}
+
+static void rcar_pcie_ep_stop(struct pci_epc *epc)
+{
+	struct rcar_pcie *ep = epc_get_drvdata(epc);
+
+	rcar_pci_write_reg(ep->base, 0, PCIETCTLR);
+}
+
+static const struct pci_epc_features rcar_pcie_epc_features = {
+	.linkup_notifier = false,
+	.msi_capable = false,
+	.msix_capable = false,
+	/* use 64-bit bars so mark bar1/3/5 as reserved */
+	.reserved_bar = 1 << BAR_1 | 1 << BAR_3 | 1 << BAR_5,
+	.bar_fixed_64bit =  (1 << BAR_0) | (1 << BAR_2) | (1 << BAR_4),
+	.bar_fixed_size[0] = 128,
+	.bar_fixed_size[2] = 256,
+	.bar_fixed_size[4] = 256,
+};
+
+static const struct pci_epc_features*
+rcar_pcie_ep_get_features(struct pci_epc *epc, u8 func_no)
+{
+	return &rcar_pcie_epc_features;
+}
+
+static const struct pci_epc_ops rcar_pcie_epc_ops = {
+	.write_header	= rcar_pcie_ep_write_header,
+	.set_bar	= rcar_pcie_ep_set_bar,
+	.clear_bar	= rcar_pcie_ep_clear_bar,
+	.map_addr	= rcar_pcie_ep_map_addr,
+	.unmap_addr	= rcar_pcie_ep_unmap_addr,
+	.raise_irq	= rcar_pcie_ep_raise_irq,
+	.start		= rcar_pcie_ep_start,
+	.stop		= rcar_pcie_ep_stop,
+	.get_features	= rcar_pcie_ep_get_features,
+};
+
+static const struct of_device_id rcar_pcie_ep_of_match[] = {
+	{
+		.compatible = "renesas,pcie-ep-rcar-gen3"
+	},
+	{},
+};
+
+static int rcar_pcie_ep_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct rcar_pcie *pcie;
+	struct pci_epc *epc;
+	int err;
+
+	pcie = devm_kzalloc(dev, sizeof(*pcie), GFP_KERNEL);
+	if (!pcie)
+		return -ENOMEM;
+
+	pcie->dev = dev;
+
+	pm_runtime_enable(pcie->dev);
+	err = pm_runtime_get_sync(pcie->dev);
+	if (err < 0) {
+		dev_err(pcie->dev, "pm_runtime_get_sync failed\n");
+		goto err_pm_disable;
+	}
+
+	err = rcar_pcie_ep_get_pdata(pcie, pdev);
+	if (err < 0) {
+		dev_err(dev, "failed to request resources: %d\n", err);
+		goto err_pm_put;
+	}
+
+	pcie->num_ib_windows = MAX_NR_INBOUND_MAPS;
+	pcie->ib_window_map =
+			devm_kcalloc(dev, BITS_TO_LONGS(pcie->num_ib_windows),
+				     sizeof(long), GFP_KERNEL);
+	if (!pcie->ib_window_map) {
+		err = -ENOMEM;
+		dev_err(dev, "failed to allocate memory for inbound map\n");
+		goto err_pm_put;
+	}
+
+	pcie->ob_addr = devm_kcalloc(dev, pcie->num_ob_windows,
+				     sizeof(*pcie->ob_addr), GFP_KERNEL);
+	if (!pcie->ob_addr) {
+		err = -ENOMEM;
+		dev_err(dev, "failed to allocate memory for outbound memory pointers\n");
+		goto err_pm_put;
+	}
+
+	epc = devm_pci_epc_create(dev, &rcar_pcie_epc_ops);
+	if (IS_ERR(epc)) {
+		dev_err(dev, "failed to create epc device\n");
+		err = PTR_ERR(epc);
+		goto err_pm_put;
+	}
+
+	epc->max_functions = pcie->max_functions;
+	pcie->epc = epc;
+	epc_set_drvdata(epc, pcie);
+
+	err = pci_epc_mem_init(epc, pcie->ob_window, pcie->num_ob_windows);
+	if (err < 0) {
+		dev_err(dev, "failed to initialize the epc memory space\n");
+		goto err_pm_put;
+	}
+
+	rcar_pcie_ep_hw_init(pcie);
+
+	return 0;
+
+err_pm_put:
+	pm_runtime_put(dev);
+
+err_pm_disable:
+	pm_runtime_disable(dev);
+
+	return err;
+}
+
+static struct platform_driver rcar_pcie_ep_driver = {
+	.driver = {
+		.name = "rcar-pcie-ep",
+		.of_match_table = rcar_pcie_ep_of_match,
+		.suppress_bind_attrs = true,
+	},
+	.probe = rcar_pcie_ep_probe,
+};
+builtin_platform_driver(rcar_pcie_ep_driver);
-- 
2.20.1


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply related	[flat|nested] 60+ messages in thread

* [PATCH 5/5] misc: pci_endpoint_test: add device-id for RZ/G2 pcie controller
  2019-11-06 19:36 ` Lad Prabhakar
@ 2019-11-06 19:36   ` Lad Prabhakar
  -1 siblings, 0 replies; 60+ messages in thread
From: Lad Prabhakar @ 2019-11-06 19:36 UTC (permalink / raw)
  To: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Kishon Vijay Abraham I, Marek Vasut,
	Yoshihiro Shimoda, linux-pci
  Cc: Catalin Marinas, Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray, devicetree, linux-kernel,
	linux-arm-kernel, linux-renesas-soc, Chris Paterson, Lad,
	Prabhakar

From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>

Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
---
 drivers/misc/pci_endpoint_test.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/misc/pci_endpoint_test.c b/drivers/misc/pci_endpoint_test.c
index a5e317073d95..346b03b6e4b9 100644
--- a/drivers/misc/pci_endpoint_test.c
+++ b/drivers/misc/pci_endpoint_test.c
@@ -66,6 +66,8 @@
 
 #define PCI_DEVICE_ID_TI_AM654			0xb00c
 
+#define PCI_DEVICE_ID_RENESAS_RZG2		0x002d
+
 #define is_am654_pci_dev(pdev)		\
 		((pdev)->device == PCI_DEVICE_ID_TI_AM654)
 
@@ -797,6 +799,7 @@ static const struct pci_device_id pci_endpoint_test_tbl[] = {
 	{ PCI_DEVICE(PCI_VENDOR_ID_TI, PCI_DEVICE_ID_TI_AM654),
 	  .driver_data = (kernel_ulong_t)&am654_data
 	},
+	{ PCI_DEVICE(PCI_VENDOR_ID_RENESAS, PCI_DEVICE_ID_RENESAS_RZG2) },
 	{ }
 };
 MODULE_DEVICE_TABLE(pci, pci_endpoint_test_tbl);
-- 
2.20.1


^ permalink raw reply related	[flat|nested] 60+ messages in thread

* [PATCH 5/5] misc: pci_endpoint_test: add device-id for RZ/G2 pcie controller
@ 2019-11-06 19:36   ` Lad Prabhakar
  0 siblings, 0 replies; 60+ messages in thread
From: Lad Prabhakar @ 2019-11-06 19:36 UTC (permalink / raw)
  To: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Kishon Vijay Abraham I, Marek Vasut,
	Yoshihiro Shimoda, linux-pci
  Cc: devicetree, Chris Paterson, Lorenzo Pieralisi, Arnd Bergmann,
	Catalin Marinas, linux-kernel, Lad, Prabhakar, linux-renesas-soc,
	Greg Kroah-Hartman, Andrew Murray, Will Deacon, linux-arm-kernel

From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>

Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
---
 drivers/misc/pci_endpoint_test.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/misc/pci_endpoint_test.c b/drivers/misc/pci_endpoint_test.c
index a5e317073d95..346b03b6e4b9 100644
--- a/drivers/misc/pci_endpoint_test.c
+++ b/drivers/misc/pci_endpoint_test.c
@@ -66,6 +66,8 @@
 
 #define PCI_DEVICE_ID_TI_AM654			0xb00c
 
+#define PCI_DEVICE_ID_RENESAS_RZG2		0x002d
+
 #define is_am654_pci_dev(pdev)		\
 		((pdev)->device == PCI_DEVICE_ID_TI_AM654)
 
@@ -797,6 +799,7 @@ static const struct pci_device_id pci_endpoint_test_tbl[] = {
 	{ PCI_DEVICE(PCI_VENDOR_ID_TI, PCI_DEVICE_ID_TI_AM654),
 	  .driver_data = (kernel_ulong_t)&am654_data
 	},
+	{ PCI_DEVICE(PCI_VENDOR_ID_RENESAS, PCI_DEVICE_ID_RENESAS_RZG2) },
 	{ }
 };
 MODULE_DEVICE_TABLE(pci, pci_endpoint_test_tbl);
-- 
2.20.1


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply related	[flat|nested] 60+ messages in thread

* RE: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
  2019-11-06 19:36   ` Lad Prabhakar
@ 2019-11-07  7:39     ` Biju Das
  -1 siblings, 0 replies; 60+ messages in thread
From: Biju Das @ 2019-11-07  7:39 UTC (permalink / raw)
  To: Lad Prabhakar, Bjorn Helgaas, Rob Herring, Mark Rutland,
	Geert Uytterhoeven, Magnus Damm, Kishon Vijay Abraham I,
	Marek Vasut, Yoshihiro Shimoda, linux-pci
  Cc: Catalin Marinas, Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray, devicetree, linux-kernel,
	linux-arm-kernel, linux-renesas-soc, Chris Paterson,
	Prabhakar Mahadev Lad

Hi Prabhakar,

Thanks for the patch

> Subject: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
> 
> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> 
> This patch adds the bindings for the R-Car PCIe endpoint driver.
> 
> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> ---
>  .../devicetree/bindings/pci/rcar-pci-ep.txt   | 43 +++++++++++++++++++
>  1 file changed, 43 insertions(+)
>  create mode 100644 Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> 
> diff --git a/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> new file mode 100644
> index 000000000000..b8c8616ca007
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> @@ -0,0 +1,43 @@
> +* Renesas R-Car PCIe Endpoint Controller DT description
> +
> +Required properties:
> +	    "renesas,pcie-ep-r8a774c0" for the R8A774C0 SoC;
> +	    "renesas,pcie-ep-rcar-gen3" for a generic R-Car Gen3 or
> +				     RZ/G2 compatible device.
> +
> +	    When compatible with the generic version, nodes must list the
> +	    SoC-specific version corresponding to the platform first
> +	    followed by the generic version.
> +
> +- reg: Five register ranges as listed in the reg-names property
> +- reg-names: Must include the following names
> +	- "apb-base"
> +	- "memory0"
> +	- "memory1"
> +	- "memory2"
> +	- "memory3"
> +- resets: Must contain phandles to PCIe-related reset lines exposed by IP
> block
> +- clocks: from common clock binding: clock specifiers for the PCIe controller
> +	 clock.
> +- clock-names: from common clock binding: should be "pcie".
> +
> +Optional Property:
> +- max-functions: Maximum number of functions that can be configured
> (default 1).
> +
> +Example:
> +
> +SoC-specific DT Entry:
> +
> +	pcie_ep: pcie_ep@fe000000 {
> +		compatible = "renesas,pcie-r8a7791", "renesas,pcie-rcar-
> gen2";

I believe it is currently tested on RZ/G2E. so please use the same.

Cheers,
Biju

> +		reg = <0 0xfe000000 0 0x80000>,
> +			<0x0 0xfe100000 0 0x100000>,
> +			<0x0 0xfe200000 0 0x200000>,
> +			<0x0 0x30000000 0 0x8000000>,
> +			<0x0 0x38000000 0 0x8000000>;
> +		reg-names = "apb-base", "memory0", "memory1",
> "memory2", "memory3";
> +		clocks = <&cpg CPG_MOD 319>;
> +		clock-names = "pcie";
> +		power-domains = <&sysc R8A774C0_PD_ALWAYS_ON>;
> +		resets = <&cpg 319>;
> +	};
> --
> 2.20.1


^ permalink raw reply	[flat|nested] 60+ messages in thread

* RE: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
@ 2019-11-07  7:39     ` Biju Das
  0 siblings, 0 replies; 60+ messages in thread
From: Biju Das @ 2019-11-07  7:39 UTC (permalink / raw)
  To: Lad Prabhakar, Bjorn Helgaas, Rob Herring, Mark Rutland,
	Geert Uytterhoeven, Magnus Damm, Kishon Vijay Abraham I,
	Marek Vasut, Yoshihiro Shimoda, linux-pci
  Cc: devicetree, Chris Paterson, Lorenzo Pieralisi, Arnd Bergmann,
	Catalin Marinas, linux-kernel, Prabhakar Mahadev Lad,
	linux-renesas-soc, Greg Kroah-Hartman, Andrew Murray,
	Will Deacon, linux-arm-kernel

Hi Prabhakar,

Thanks for the patch

> Subject: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
> 
> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> 
> This patch adds the bindings for the R-Car PCIe endpoint driver.
> 
> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> ---
>  .../devicetree/bindings/pci/rcar-pci-ep.txt   | 43 +++++++++++++++++++
>  1 file changed, 43 insertions(+)
>  create mode 100644 Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> 
> diff --git a/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> new file mode 100644
> index 000000000000..b8c8616ca007
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> @@ -0,0 +1,43 @@
> +* Renesas R-Car PCIe Endpoint Controller DT description
> +
> +Required properties:
> +	    "renesas,pcie-ep-r8a774c0" for the R8A774C0 SoC;
> +	    "renesas,pcie-ep-rcar-gen3" for a generic R-Car Gen3 or
> +				     RZ/G2 compatible device.
> +
> +	    When compatible with the generic version, nodes must list the
> +	    SoC-specific version corresponding to the platform first
> +	    followed by the generic version.
> +
> +- reg: Five register ranges as listed in the reg-names property
> +- reg-names: Must include the following names
> +	- "apb-base"
> +	- "memory0"
> +	- "memory1"
> +	- "memory2"
> +	- "memory3"
> +- resets: Must contain phandles to PCIe-related reset lines exposed by IP
> block
> +- clocks: from common clock binding: clock specifiers for the PCIe controller
> +	 clock.
> +- clock-names: from common clock binding: should be "pcie".
> +
> +Optional Property:
> +- max-functions: Maximum number of functions that can be configured
> (default 1).
> +
> +Example:
> +
> +SoC-specific DT Entry:
> +
> +	pcie_ep: pcie_ep@fe000000 {
> +		compatible = "renesas,pcie-r8a7791", "renesas,pcie-rcar-
> gen2";

I believe it is currently tested on RZ/G2E. so please use the same.

Cheers,
Biju

> +		reg = <0 0xfe000000 0 0x80000>,
> +			<0x0 0xfe100000 0 0x100000>,
> +			<0x0 0xfe200000 0 0x200000>,
> +			<0x0 0x30000000 0 0x8000000>,
> +			<0x0 0x38000000 0 0x8000000>;
> +		reg-names = "apb-base", "memory0", "memory1",
> "memory2", "memory3";
> +		clocks = <&cpg CPG_MOD 319>;
> +		clock-names = "pcie";
> +		power-domains = <&sysc R8A774C0_PD_ALWAYS_ON>;
> +		resets = <&cpg 319>;
> +	};
> --
> 2.20.1


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
  2019-11-07  7:39     ` Biju Das
@ 2019-11-07  8:10       ` Lad, Prabhakar
  -1 siblings, 0 replies; 60+ messages in thread
From: Lad, Prabhakar @ 2019-11-07  8:10 UTC (permalink / raw)
  To: Biju Das
  Cc: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Kishon Vijay Abraham I, Marek Vasut,
	Yoshihiro Shimoda, linux-pci, Catalin Marinas, Will Deacon,
	Lorenzo Pieralisi, Arnd Bergmann, Greg Kroah-Hartman,
	Andrew Murray, devicetree, linux-kernel, linux-arm-kernel,
	linux-renesas-soc, Chris Paterson, Prabhakar Mahadev Lad

Hi Biju,

Thank you for the review.

On Thu, Nov 7, 2019 at 7:39 AM Biju Das <biju.das@bp.renesas.com> wrote:
>
> Hi Prabhakar,
>
> Thanks for the patch
>
> > Subject: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
> >
> > From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >
> > This patch adds the bindings for the R-Car PCIe endpoint driver.
> >
> > Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > ---
> >  .../devicetree/bindings/pci/rcar-pci-ep.txt   | 43 +++++++++++++++++++
> >  1 file changed, 43 insertions(+)
> >  create mode 100644 Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> >
> > diff --git a/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> > b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> > new file mode 100644
> > index 000000000000..b8c8616ca007
> > --- /dev/null
> > +++ b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> > @@ -0,0 +1,43 @@
> > +* Renesas R-Car PCIe Endpoint Controller DT description
> > +
> > +Required properties:
> > +         "renesas,pcie-ep-r8a774c0" for the R8A774C0 SoC;
> > +         "renesas,pcie-ep-rcar-gen3" for a generic R-Car Gen3 or
> > +                                  RZ/G2 compatible device.
> > +
> > +         When compatible with the generic version, nodes must list the
> > +         SoC-specific version corresponding to the platform first
> > +         followed by the generic version.
> > +
> > +- reg: Five register ranges as listed in the reg-names property
> > +- reg-names: Must include the following names
> > +     - "apb-base"
> > +     - "memory0"
> > +     - "memory1"
> > +     - "memory2"
> > +     - "memory3"
> > +- resets: Must contain phandles to PCIe-related reset lines exposed by IP
> > block
> > +- clocks: from common clock binding: clock specifiers for the PCIe controller
> > +      clock.
> > +- clock-names: from common clock binding: should be "pcie".
> > +
> > +Optional Property:
> > +- max-functions: Maximum number of functions that can be configured
> > (default 1).
> > +
> > +Example:
> > +
> > +SoC-specific DT Entry:
> > +
> > +     pcie_ep: pcie_ep@fe000000 {
> > +             compatible = "renesas,pcie-r8a7791", "renesas,pcie-rcar-
> > gen2";
>
> I believe it is currently tested on RZ/G2E. so please use the same.
>
Yes you are correct its tested on RZ/G2E board, ill fix it in next iteration.

Cheers,
--Prabhakar

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
@ 2019-11-07  8:10       ` Lad, Prabhakar
  0 siblings, 0 replies; 60+ messages in thread
From: Lad, Prabhakar @ 2019-11-07  8:10 UTC (permalink / raw)
  To: Biju Das
  Cc: Mark Rutland, devicetree, Chris Paterson, Lorenzo Pieralisi,
	Prabhakar Mahadev Lad, Arnd Bergmann, Geert Uytterhoeven,
	Greg Kroah-Hartman, linux-pci, Yoshihiro Shimoda, Magnus Damm,
	linux-kernel, Kishon Vijay Abraham I, linux-renesas-soc,
	Rob Herring, Catalin Marinas, Bjorn Helgaas, Andrew Murray,
	Will Deacon, linux-arm-kernel, Marek Vasut

Hi Biju,

Thank you for the review.

On Thu, Nov 7, 2019 at 7:39 AM Biju Das <biju.das@bp.renesas.com> wrote:
>
> Hi Prabhakar,
>
> Thanks for the patch
>
> > Subject: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
> >
> > From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >
> > This patch adds the bindings for the R-Car PCIe endpoint driver.
> >
> > Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > ---
> >  .../devicetree/bindings/pci/rcar-pci-ep.txt   | 43 +++++++++++++++++++
> >  1 file changed, 43 insertions(+)
> >  create mode 100644 Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> >
> > diff --git a/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> > b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> > new file mode 100644
> > index 000000000000..b8c8616ca007
> > --- /dev/null
> > +++ b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> > @@ -0,0 +1,43 @@
> > +* Renesas R-Car PCIe Endpoint Controller DT description
> > +
> > +Required properties:
> > +         "renesas,pcie-ep-r8a774c0" for the R8A774C0 SoC;
> > +         "renesas,pcie-ep-rcar-gen3" for a generic R-Car Gen3 or
> > +                                  RZ/G2 compatible device.
> > +
> > +         When compatible with the generic version, nodes must list the
> > +         SoC-specific version corresponding to the platform first
> > +         followed by the generic version.
> > +
> > +- reg: Five register ranges as listed in the reg-names property
> > +- reg-names: Must include the following names
> > +     - "apb-base"
> > +     - "memory0"
> > +     - "memory1"
> > +     - "memory2"
> > +     - "memory3"
> > +- resets: Must contain phandles to PCIe-related reset lines exposed by IP
> > block
> > +- clocks: from common clock binding: clock specifiers for the PCIe controller
> > +      clock.
> > +- clock-names: from common clock binding: should be "pcie".
> > +
> > +Optional Property:
> > +- max-functions: Maximum number of functions that can be configured
> > (default 1).
> > +
> > +Example:
> > +
> > +SoC-specific DT Entry:
> > +
> > +     pcie_ep: pcie_ep@fe000000 {
> > +             compatible = "renesas,pcie-r8a7791", "renesas,pcie-rcar-
> > gen2";
>
> I believe it is currently tested on RZ/G2E. so please use the same.
>
Yes you are correct its tested on RZ/G2E board, ill fix it in next iteration.

Cheers,
--Prabhakar

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
  2019-11-06 19:36   ` Lad Prabhakar
@ 2019-11-07  8:44     ` Geert Uytterhoeven
  -1 siblings, 0 replies; 60+ messages in thread
From: Geert Uytterhoeven @ 2019-11-07  8:44 UTC (permalink / raw)
  To: Lad Prabhakar
  Cc: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Kishon Vijay Abraham I, Marek Vasut,
	Yoshihiro Shimoda, linux-pci, Catalin Marinas, Will Deacon,
	Lorenzo Pieralisi, Arnd Bergmann, Greg Kroah-Hartman,
	Andrew Murray,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Linux Kernel Mailing List, Linux ARM, Linux-Renesas,
	Chris Paterson, Lad, Prabhakar

Hi Prabhakar,

On Wed, Nov 6, 2019 at 8:36 PM Lad Prabhakar <prabhakar.csengg@gmail.com> wrote:
> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
>
> This patch adds the bindings for the R-Car PCIe endpoint driver.
>
> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>

Thanks for your patch!

> --- /dev/null
> +++ b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> @@ -0,0 +1,43 @@
> +* Renesas R-Car PCIe Endpoint Controller DT description
> +
> +Required properties:
> +           "renesas,pcie-ep-r8a774c0" for the R8A774C0 SoC;
> +           "renesas,pcie-ep-rcar-gen3" for a generic R-Car Gen3 or
> +                                    RZ/G2 compatible device.

Unless I'm missing something, this is for the exact same hardware block as
Documentation/devicetree/bindings/pci/rcar-pci.txt?
So shouldn't you amend those bindings, instead of adding new compatible
values?
Please remember that DT describes hardware, not software policy.
So IMHO choosing between host and endpoint is purely a configuration
issue, and could be indicated by the presence or lack of some DT properties.
E.g. host mode requires both "bus-range" and "device_type" properties,
so their absence could indicate endpoint mode.

> +- reg: Five register ranges as listed in the reg-names property
> +- reg-names: Must include the following names
> +       - "apb-base"
> +       - "memory0"
> +       - "memory1"
> +       - "memory2"
> +       - "memory3"

What is the purpose of the last 4 regions?
Can they be chosen by the driver, at runtime?

> +- resets: Must contain phandles to PCIe-related reset lines exposed by IP block
> +- clocks: from common clock binding: clock specifiers for the PCIe controller
> +        clock.
> +- clock-names: from common clock binding: should be "pcie".
> +
> +Optional Property:
> +- max-functions: Maximum number of functions that can be configured (default 1).
> +
> +Example:
> +
> +SoC-specific DT Entry:
> +
> +       pcie_ep: pcie_ep@fe000000 {
> +               compatible = "renesas,pcie-r8a7791", "renesas,pcie-rcar-gen2";

These compatible values do not match with the ones above
(but they match with what I'd like to see ;-)

> +               reg = <0 0xfe000000 0 0x80000>,
> +                       <0x0 0xfe100000 0 0x100000>,
> +                       <0x0 0xfe200000 0 0x200000>,
> +                       <0x0 0x30000000 0 0x8000000>,
> +                       <0x0 0x38000000 0 0x8000000>;
> +               reg-names = "apb-base", "memory0", "memory1", "memory2", "memory3";
> +               clocks = <&cpg CPG_MOD 319>;
> +               clock-names = "pcie";
> +               power-domains = <&sysc R8A774C0_PD_ALWAYS_ON>;
> +               resets = <&cpg 319>;
> +       };

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
@ 2019-11-07  8:44     ` Geert Uytterhoeven
  0 siblings, 0 replies; 60+ messages in thread
From: Geert Uytterhoeven @ 2019-11-07  8:44 UTC (permalink / raw)
  To: Lad Prabhakar
  Cc: Mark Rutland,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Chris Paterson, Lorenzo Pieralisi, Lad, Prabhakar, Arnd Bergmann,
	Geert Uytterhoeven, Greg Kroah-Hartman, linux-pci,
	Yoshihiro Shimoda, Magnus Damm, Linux Kernel Mailing List,
	Kishon Vijay Abraham I, Linux-Renesas, Rob Herring,
	Catalin Marinas, Bjorn Helgaas, Andrew Murray, Will Deacon,
	Linux ARM, Marek Vasut

Hi Prabhakar,

On Wed, Nov 6, 2019 at 8:36 PM Lad Prabhakar <prabhakar.csengg@gmail.com> wrote:
> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
>
> This patch adds the bindings for the R-Car PCIe endpoint driver.
>
> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>

Thanks for your patch!

> --- /dev/null
> +++ b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> @@ -0,0 +1,43 @@
> +* Renesas R-Car PCIe Endpoint Controller DT description
> +
> +Required properties:
> +           "renesas,pcie-ep-r8a774c0" for the R8A774C0 SoC;
> +           "renesas,pcie-ep-rcar-gen3" for a generic R-Car Gen3 or
> +                                    RZ/G2 compatible device.

Unless I'm missing something, this is for the exact same hardware block as
Documentation/devicetree/bindings/pci/rcar-pci.txt?
So shouldn't you amend those bindings, instead of adding new compatible
values?
Please remember that DT describes hardware, not software policy.
So IMHO choosing between host and endpoint is purely a configuration
issue, and could be indicated by the presence or lack of some DT properties.
E.g. host mode requires both "bus-range" and "device_type" properties,
so their absence could indicate endpoint mode.

> +- reg: Five register ranges as listed in the reg-names property
> +- reg-names: Must include the following names
> +       - "apb-base"
> +       - "memory0"
> +       - "memory1"
> +       - "memory2"
> +       - "memory3"

What is the purpose of the last 4 regions?
Can they be chosen by the driver, at runtime?

> +- resets: Must contain phandles to PCIe-related reset lines exposed by IP block
> +- clocks: from common clock binding: clock specifiers for the PCIe controller
> +        clock.
> +- clock-names: from common clock binding: should be "pcie".
> +
> +Optional Property:
> +- max-functions: Maximum number of functions that can be configured (default 1).
> +
> +Example:
> +
> +SoC-specific DT Entry:
> +
> +       pcie_ep: pcie_ep@fe000000 {
> +               compatible = "renesas,pcie-r8a7791", "renesas,pcie-rcar-gen2";

These compatible values do not match with the ones above
(but they match with what I'd like to see ;-)

> +               reg = <0 0xfe000000 0 0x80000>,
> +                       <0x0 0xfe100000 0 0x100000>,
> +                       <0x0 0xfe200000 0 0x200000>,
> +                       <0x0 0x30000000 0 0x8000000>,
> +                       <0x0 0x38000000 0 0x8000000>;
> +               reg-names = "apb-base", "memory0", "memory1", "memory2", "memory3";
> +               clocks = <&cpg CPG_MOD 319>;
> +               clock-names = "pcie";
> +               power-domains = <&sysc R8A774C0_PD_ALWAYS_ON>;
> +               resets = <&cpg 319>;
> +       };

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
  2019-11-07  8:44     ` Geert Uytterhoeven
@ 2019-11-07  9:25       ` Lad, Prabhakar
  -1 siblings, 0 replies; 60+ messages in thread
From: Lad, Prabhakar @ 2019-11-07  9:25 UTC (permalink / raw)
  To: Geert Uytterhoeven
  Cc: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Kishon Vijay Abraham I, Marek Vasut,
	Yoshihiro Shimoda, linux-pci, Catalin Marinas, Will Deacon,
	Lorenzo Pieralisi, Arnd Bergmann, Greg Kroah-Hartman,
	Andrew Murray,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Linux Kernel Mailing List, Linux ARM, Linux-Renesas,
	Chris Paterson, Lad, Prabhakar

Hi Geert,

Thank you for the review.

On Thu, Nov 7, 2019 at 8:44 AM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
>
> Hi Prabhakar,
>
> On Wed, Nov 6, 2019 at 8:36 PM Lad Prabhakar <prabhakar.csengg@gmail.com> wrote:
> > From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >
> > This patch adds the bindings for the R-Car PCIe endpoint driver.
> >
> > Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
>
> Thanks for your patch!
>
> > --- /dev/null
> > +++ b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> > @@ -0,0 +1,43 @@
> > +* Renesas R-Car PCIe Endpoint Controller DT description
> > +
> > +Required properties:
> > +           "renesas,pcie-ep-r8a774c0" for the R8A774C0 SoC;
> > +           "renesas,pcie-ep-rcar-gen3" for a generic R-Car Gen3 or
> > +                                    RZ/G2 compatible device.
>
> Unless I'm missing something, this is for the exact same hardware block as
> Documentation/devicetree/bindings/pci/rcar-pci.txt?
> So shouldn't you amend those bindings, instead of adding new compatible
> values?
> Please remember that DT describes hardware, not software policy.
> So IMHO choosing between host and endpoint is purely a configuration
> issue, and could be indicated by the presence or lack of some DT properties.
> E.g. host mode requires both "bus-range" and "device_type" properties,
> so their absence could indicate endpoint mode.
>
yes its the same hardware block as described in the rcar-pci.txt, I
did think about amending it
but  it might turn out to be bit messy,

required properties host ======required properties Endpoint
====================||==================
1: reg                                || reg
2:bus-range                      || reg names
3: device_type                  || resets
4: ranges                          || clocks
5: dma-ranges                  || clock-names
6: interrupts                      ||
7: interrupt-cells               ||
8: interrupt-map-mask     ||
9: clocks                          ||
10: clock-names             ||

and if I go ahead with the same compatible string that would mean to
add support for endpoint
mode in the host driver itself. I did follow the examples of
rockchip/cadence/designware where
its the same hardware block but has two different binding files one
for host mode and other for
endpoint mode.

> > +- reg: Five register ranges as listed in the reg-names property
> > +- reg-names: Must include the following names
> > +       - "apb-base"
> > +       - "memory0"
> > +       - "memory1"
> > +       - "memory2"
> > +       - "memory3"
>
> What is the purpose of the last 4 regions?
> Can they be chosen by the driver, at runtime?
>
no the driver cannot choose them at runtime, as these are the only
PCIE memory(0/1/2/3) ranges
in the AXI address space where host memory can be mapped.

> > +- resets: Must contain phandles to PCIe-related reset lines exposed by IP block
> > +- clocks: from common clock binding: clock specifiers for the PCIe controller
> > +        clock.
> > +- clock-names: from common clock binding: should be "pcie".
> > +
> > +Optional Property:
> > +- max-functions: Maximum number of functions that can be configured (default 1).
> > +
> > +Example:
> > +
> > +SoC-specific DT Entry:
> > +
> > +       pcie_ep: pcie_ep@fe000000 {
> > +               compatible = "renesas,pcie-r8a7791", "renesas,pcie-rcar-gen2";
>
> These compatible values do not match with the ones above
> (but they match with what I'd like to see ;-)
>
my bad I'll update them to reflect the above.

Cheers,
--Prabhakar

> > +               reg = <0 0xfe000000 0 0x80000>,
> > +                       <0x0 0xfe100000 0 0x100000>,
> > +                       <0x0 0xfe200000 0 0x200000>,
> > +                       <0x0 0x30000000 0 0x8000000>,
> > +                       <0x0 0x38000000 0 0x8000000>;
> > +               reg-names = "apb-base", "memory0", "memory1", "memory2", "memory3";
> > +               clocks = <&cpg CPG_MOD 319>;
> > +               clock-names = "pcie";
> > +               power-domains = <&sysc R8A774C0_PD_ALWAYS_ON>;
> > +               resets = <&cpg 319>;
> > +       };
>
> Gr{oetje,eeting}s,
>
>                         Geert
>
> --
> Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
>
> In personal conversations with technical people, I call myself a hacker. But
> when I'm talking to journalists I just say "programmer" or something like that.
>                                 -- Linus Torvalds

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
@ 2019-11-07  9:25       ` Lad, Prabhakar
  0 siblings, 0 replies; 60+ messages in thread
From: Lad, Prabhakar @ 2019-11-07  9:25 UTC (permalink / raw)
  To: Geert Uytterhoeven
  Cc: Mark Rutland,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Chris Paterson, Lorenzo Pieralisi, Lad, Prabhakar, Arnd Bergmann,
	Geert Uytterhoeven, Greg Kroah-Hartman, linux-pci,
	Yoshihiro Shimoda, Magnus Damm, Linux Kernel Mailing List,
	Kishon Vijay Abraham I, Linux-Renesas, Rob Herring,
	Catalin Marinas, Bjorn Helgaas, Andrew Murray, Will Deacon,
	Linux ARM, Marek Vasut

Hi Geert,

Thank you for the review.

On Thu, Nov 7, 2019 at 8:44 AM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
>
> Hi Prabhakar,
>
> On Wed, Nov 6, 2019 at 8:36 PM Lad Prabhakar <prabhakar.csengg@gmail.com> wrote:
> > From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >
> > This patch adds the bindings for the R-Car PCIe endpoint driver.
> >
> > Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
>
> Thanks for your patch!
>
> > --- /dev/null
> > +++ b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> > @@ -0,0 +1,43 @@
> > +* Renesas R-Car PCIe Endpoint Controller DT description
> > +
> > +Required properties:
> > +           "renesas,pcie-ep-r8a774c0" for the R8A774C0 SoC;
> > +           "renesas,pcie-ep-rcar-gen3" for a generic R-Car Gen3 or
> > +                                    RZ/G2 compatible device.
>
> Unless I'm missing something, this is for the exact same hardware block as
> Documentation/devicetree/bindings/pci/rcar-pci.txt?
> So shouldn't you amend those bindings, instead of adding new compatible
> values?
> Please remember that DT describes hardware, not software policy.
> So IMHO choosing between host and endpoint is purely a configuration
> issue, and could be indicated by the presence or lack of some DT properties.
> E.g. host mode requires both "bus-range" and "device_type" properties,
> so their absence could indicate endpoint mode.
>
yes its the same hardware block as described in the rcar-pci.txt, I
did think about amending it
but  it might turn out to be bit messy,

required properties host ======required properties Endpoint
====================||==================
1: reg                                || reg
2:bus-range                      || reg names
3: device_type                  || resets
4: ranges                          || clocks
5: dma-ranges                  || clock-names
6: interrupts                      ||
7: interrupt-cells               ||
8: interrupt-map-mask     ||
9: clocks                          ||
10: clock-names             ||

and if I go ahead with the same compatible string that would mean to
add support for endpoint
mode in the host driver itself. I did follow the examples of
rockchip/cadence/designware where
its the same hardware block but has two different binding files one
for host mode and other for
endpoint mode.

> > +- reg: Five register ranges as listed in the reg-names property
> > +- reg-names: Must include the following names
> > +       - "apb-base"
> > +       - "memory0"
> > +       - "memory1"
> > +       - "memory2"
> > +       - "memory3"
>
> What is the purpose of the last 4 regions?
> Can they be chosen by the driver, at runtime?
>
no the driver cannot choose them at runtime, as these are the only
PCIE memory(0/1/2/3) ranges
in the AXI address space where host memory can be mapped.

> > +- resets: Must contain phandles to PCIe-related reset lines exposed by IP block
> > +- clocks: from common clock binding: clock specifiers for the PCIe controller
> > +        clock.
> > +- clock-names: from common clock binding: should be "pcie".
> > +
> > +Optional Property:
> > +- max-functions: Maximum number of functions that can be configured (default 1).
> > +
> > +Example:
> > +
> > +SoC-specific DT Entry:
> > +
> > +       pcie_ep: pcie_ep@fe000000 {
> > +               compatible = "renesas,pcie-r8a7791", "renesas,pcie-rcar-gen2";
>
> These compatible values do not match with the ones above
> (but they match with what I'd like to see ;-)
>
my bad I'll update them to reflect the above.

Cheers,
--Prabhakar

> > +               reg = <0 0xfe000000 0 0x80000>,
> > +                       <0x0 0xfe100000 0 0x100000>,
> > +                       <0x0 0xfe200000 0 0x200000>,
> > +                       <0x0 0x30000000 0 0x8000000>,
> > +                       <0x0 0x38000000 0 0x8000000>;
> > +               reg-names = "apb-base", "memory0", "memory1", "memory2", "memory3";
> > +               clocks = <&cpg CPG_MOD 319>;
> > +               clock-names = "pcie";
> > +               power-domains = <&sysc R8A774C0_PD_ALWAYS_ON>;
> > +               resets = <&cpg 319>;
> > +       };
>
> Gr{oetje,eeting}s,
>
>                         Geert
>
> --
> Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
>
> In personal conversations with technical people, I call myself a hacker. But
> when I'm talking to journalists I just say "programmer" or something like that.
>                                 -- Linus Torvalds

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
  2019-11-07  9:25       ` Lad, Prabhakar
@ 2019-11-07 20:08         ` Geert Uytterhoeven
  -1 siblings, 0 replies; 60+ messages in thread
From: Geert Uytterhoeven @ 2019-11-07 20:08 UTC (permalink / raw)
  To: Lad, Prabhakar
  Cc: Bjorn Helgaas, Rob Herring, Mark Rutland, Magnus Damm,
	Kishon Vijay Abraham I, Marek Vasut, Yoshihiro Shimoda,
	linux-pci, Catalin Marinas, Will Deacon, Lorenzo Pieralisi,
	Arnd Bergmann, Greg Kroah-Hartman, Andrew Murray,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Linux Kernel Mailing List, Linux ARM, Linux-Renesas,
	Chris Paterson, Lad, Prabhakar

Hi Prabhakar,

On Thu, Nov 7, 2019 at 10:26 AM Lad, Prabhakar
<prabhakar.csengg@gmail.com> wrote:
> On Thu, Nov 7, 2019 at 8:44 AM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> > On Wed, Nov 6, 2019 at 8:36 PM Lad Prabhakar <prabhakar.csengg@gmail.com> wrote:
> > > From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > >
> > > This patch adds the bindings for the R-Car PCIe endpoint driver.
> > >
> > > Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >
> > Thanks for your patch!
> >
> > > --- /dev/null
> > > +++ b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> > > @@ -0,0 +1,43 @@
> > > +* Renesas R-Car PCIe Endpoint Controller DT description
> > > +
> > > +Required properties:
> > > +           "renesas,pcie-ep-r8a774c0" for the R8A774C0 SoC;
> > > +           "renesas,pcie-ep-rcar-gen3" for a generic R-Car Gen3 or
> > > +                                    RZ/G2 compatible device.
> >
> > Unless I'm missing something, this is for the exact same hardware block as
> > Documentation/devicetree/bindings/pci/rcar-pci.txt?
> > So shouldn't you amend those bindings, instead of adding new compatible
> > values?
> > Please remember that DT describes hardware, not software policy.
> > So IMHO choosing between host and endpoint is purely a configuration
> > issue, and could be indicated by the presence or lack of some DT properties.
> > E.g. host mode requires both "bus-range" and "device_type" properties,
> > so their absence could indicate endpoint mode.
> >
> yes its the same hardware block as described in the rcar-pci.txt, I
> did think about amending it
> but  it might turn out to be bit messy,
>
> required properties host ======required properties Endpoint
> ====================||==================
> 1: reg                                || reg
> 2:bus-range                      || reg names
> 3: device_type                  || resets
> 4: ranges                          || clocks
> 5: dma-ranges                  || clock-names
> 6: interrupts                      ||
> 7: interrupt-cells               ||
> 8: interrupt-map-mask     ||
> 9: clocks                          ||
> 10: clock-names             ||

We have a similar situation with SPI, where a controller can operate in
master or slave mode, based on the absence or presence of the
"spi-slave" DT property.

> and if I go ahead with the same compatible string that would mean to
> add support for endpoint
> mode in the host driver itself. I did follow the examples of

You can still have two separate drivers, binding against the same
compatible value.  Just let the .probe() function return -ENODEV if it
discovers (by looking at DT properties) if the node is configured for
the other mode.
Which brings us to my next questions: is there any code that could be
shared between the drivers for the two modes?

> rockchip/cadence/designware where
> its the same hardware block but has two different binding files one
> for host mode and other for
> endpoint mode.

Having two separate DT binding documents sounds fine to me, if unifying
them makes things too complex.
However, I think they should use the same compatible value, because the
hardware block is the same, but just used in a different mode.

Rob/Mark: Any input from the DT maintainers?

> > > +- reg: Five register ranges as listed in the reg-names property
> > > +- reg-names: Must include the following names
> > > +       - "apb-base"
> > > +       - "memory0"
> > > +       - "memory1"
> > > +       - "memory2"
> > > +       - "memory3"
> >
> > What is the purpose of the last 4 regions?
> > Can they be chosen by the driver, at runtime?
> >
> no the driver cannot choose them at runtime, as these are the only
> PCIE memory(0/1/2/3) ranges
> in the AXI address space where host memory can be mapped.

Are they fixed by the PCIe hardware, i.e. could they be looked up by the
driver based on the compatible value?

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
@ 2019-11-07 20:08         ` Geert Uytterhoeven
  0 siblings, 0 replies; 60+ messages in thread
From: Geert Uytterhoeven @ 2019-11-07 20:08 UTC (permalink / raw)
  To: Lad, Prabhakar
  Cc: Mark Rutland,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Chris Paterson, Lorenzo Pieralisi, Lad, Prabhakar, Arnd Bergmann,
	Greg Kroah-Hartman, linux-pci, Yoshihiro Shimoda, Magnus Damm,
	Linux Kernel Mailing List, Kishon Vijay Abraham I, Linux-Renesas,
	Rob Herring, Catalin Marinas, Bjorn Helgaas, Andrew Murray,
	Will Deacon, Linux ARM, Marek Vasut

Hi Prabhakar,

On Thu, Nov 7, 2019 at 10:26 AM Lad, Prabhakar
<prabhakar.csengg@gmail.com> wrote:
> On Thu, Nov 7, 2019 at 8:44 AM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> > On Wed, Nov 6, 2019 at 8:36 PM Lad Prabhakar <prabhakar.csengg@gmail.com> wrote:
> > > From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > >
> > > This patch adds the bindings for the R-Car PCIe endpoint driver.
> > >
> > > Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >
> > Thanks for your patch!
> >
> > > --- /dev/null
> > > +++ b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> > > @@ -0,0 +1,43 @@
> > > +* Renesas R-Car PCIe Endpoint Controller DT description
> > > +
> > > +Required properties:
> > > +           "renesas,pcie-ep-r8a774c0" for the R8A774C0 SoC;
> > > +           "renesas,pcie-ep-rcar-gen3" for a generic R-Car Gen3 or
> > > +                                    RZ/G2 compatible device.
> >
> > Unless I'm missing something, this is for the exact same hardware block as
> > Documentation/devicetree/bindings/pci/rcar-pci.txt?
> > So shouldn't you amend those bindings, instead of adding new compatible
> > values?
> > Please remember that DT describes hardware, not software policy.
> > So IMHO choosing between host and endpoint is purely a configuration
> > issue, and could be indicated by the presence or lack of some DT properties.
> > E.g. host mode requires both "bus-range" and "device_type" properties,
> > so their absence could indicate endpoint mode.
> >
> yes its the same hardware block as described in the rcar-pci.txt, I
> did think about amending it
> but  it might turn out to be bit messy,
>
> required properties host ======required properties Endpoint
> ====================||==================
> 1: reg                                || reg
> 2:bus-range                      || reg names
> 3: device_type                  || resets
> 4: ranges                          || clocks
> 5: dma-ranges                  || clock-names
> 6: interrupts                      ||
> 7: interrupt-cells               ||
> 8: interrupt-map-mask     ||
> 9: clocks                          ||
> 10: clock-names             ||

We have a similar situation with SPI, where a controller can operate in
master or slave mode, based on the absence or presence of the
"spi-slave" DT property.

> and if I go ahead with the same compatible string that would mean to
> add support for endpoint
> mode in the host driver itself. I did follow the examples of

You can still have two separate drivers, binding against the same
compatible value.  Just let the .probe() function return -ENODEV if it
discovers (by looking at DT properties) if the node is configured for
the other mode.
Which brings us to my next questions: is there any code that could be
shared between the drivers for the two modes?

> rockchip/cadence/designware where
> its the same hardware block but has two different binding files one
> for host mode and other for
> endpoint mode.

Having two separate DT binding documents sounds fine to me, if unifying
them makes things too complex.
However, I think they should use the same compatible value, because the
hardware block is the same, but just used in a different mode.

Rob/Mark: Any input from the DT maintainers?

> > > +- reg: Five register ranges as listed in the reg-names property
> > > +- reg-names: Must include the following names
> > > +       - "apb-base"
> > > +       - "memory0"
> > > +       - "memory1"
> > > +       - "memory2"
> > > +       - "memory3"
> >
> > What is the purpose of the last 4 regions?
> > Can they be chosen by the driver, at runtime?
> >
> no the driver cannot choose them at runtime, as these are the only
> PCIE memory(0/1/2/3) ranges
> in the AXI address space where host memory can be mapped.

Are they fixed by the PCIe hardware, i.e. could they be looked up by the
driver based on the compatible value?

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
  2019-11-07 20:08         ` Geert Uytterhoeven
@ 2019-11-07 22:46           ` Lad, Prabhakar
  -1 siblings, 0 replies; 60+ messages in thread
From: Lad, Prabhakar @ 2019-11-07 22:46 UTC (permalink / raw)
  To: Geert Uytterhoeven
  Cc: Bjorn Helgaas, Rob Herring, Mark Rutland, Magnus Damm,
	Kishon Vijay Abraham I, Marek Vasut, Yoshihiro Shimoda,
	linux-pci, Catalin Marinas, Will Deacon, Lorenzo Pieralisi,
	Arnd Bergmann, Greg Kroah-Hartman, Andrew Murray,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Linux Kernel Mailing List, Linux ARM, Linux-Renesas,
	Chris Paterson, Lad, Prabhakar

Hi Geert,

On Thu, Nov 7, 2019 at 8:08 PM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
>
> Hi Prabhakar,
>
> On Thu, Nov 7, 2019 at 10:26 AM Lad, Prabhakar
> <prabhakar.csengg@gmail.com> wrote:
> > On Thu, Nov 7, 2019 at 8:44 AM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> > > On Wed, Nov 6, 2019 at 8:36 PM Lad Prabhakar <prabhakar.csengg@gmail.com> wrote:
> > > > From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > > >
> > > > This patch adds the bindings for the R-Car PCIe endpoint driver.
> > > >
> > > > Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > >
> > > Thanks for your patch!
> > >
> > > > --- /dev/null
> > > > +++ b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> > > > @@ -0,0 +1,43 @@
> > > > +* Renesas R-Car PCIe Endpoint Controller DT description
> > > > +
> > > > +Required properties:
> > > > +           "renesas,pcie-ep-r8a774c0" for the R8A774C0 SoC;
> > > > +           "renesas,pcie-ep-rcar-gen3" for a generic R-Car Gen3 or
> > > > +                                    RZ/G2 compatible device.
> > >
> > > Unless I'm missing something, this is for the exact same hardware block as
> > > Documentation/devicetree/bindings/pci/rcar-pci.txt?
> > > So shouldn't you amend those bindings, instead of adding new compatible
> > > values?
> > > Please remember that DT describes hardware, not software policy.
> > > So IMHO choosing between host and endpoint is purely a configuration
> > > issue, and could be indicated by the presence or lack of some DT properties.
> > > E.g. host mode requires both "bus-range" and "device_type" properties,
> > > so their absence could indicate endpoint mode.
> > >
> > yes its the same hardware block as described in the rcar-pci.txt, I
> > did think about amending it
> > but  it might turn out to be bit messy,
> >
> > required properties host ======required properties Endpoint
> > ====================||==================
> > 1: reg                                || reg
> > 2:bus-range                      || reg names
> > 3: device_type                  || resets
> > 4: ranges                          || clocks
> > 5: dma-ranges                  || clock-names
> > 6: interrupts                      ||
> > 7: interrupt-cells               ||
> > 8: interrupt-map-mask     ||
> > 9: clocks                          ||
> > 10: clock-names             ||
>
> We have a similar situation with SPI, where a controller can operate in
> master or slave mode, based on the absence or presence of the
> "spi-slave" DT property.
>
> > and if I go ahead with the same compatible string that would mean to
> > add support for endpoint
> > mode in the host driver itself. I did follow the examples of
>
> You can still have two separate drivers, binding against the same
> compatible value.  Just let the .probe() function return -ENODEV if it
> discovers (by looking at DT properties) if the node is configured for
> the other mode.
> Which brings us to my next questions: is there any code that could be
> shared between the drivers for the two modes?
>
agreed driver probe could handle depending on the DT properties.
yes there is bit of common code and the first patch of the series prepares
for handling host and endpoint mode.

> > rockchip/cadence/designware where
> > its the same hardware block but has two different binding files one
> > for host mode and other for
> > endpoint mode.
>
> Having two separate DT binding documents sounds fine to me, if unifying
> them makes things too complex.
> However, I think they should use the same compatible value, because the
> hardware block is the same, but just used in a different mode.
>
agreed.

> Rob/Mark: Any input from the DT maintainers?
>
> > > > +- reg: Five register ranges as listed in the reg-names property
> > > > +- reg-names: Must include the following names
> > > > +       - "apb-base"
> > > > +       - "memory0"
> > > > +       - "memory1"
> > > > +       - "memory2"
> > > > +       - "memory3"
> > >
> > > What is the purpose of the last 4 regions?
> > > Can they be chosen by the driver, at runtime?
> > >
> > no the driver cannot choose them at runtime, as these are the only
> > PCIE memory(0/1/2/3) ranges
> > in the AXI address space where host memory can be mapped.
>
> Are they fixed by the PCIe hardware, i.e. could they be looked up by the
> driver based on the compatible value?
>
yes they are fixed by the PCIe hardware and could be looked up by the driver
based on the compatible value.

Cheers,
--Prabhakar

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
@ 2019-11-07 22:46           ` Lad, Prabhakar
  0 siblings, 0 replies; 60+ messages in thread
From: Lad, Prabhakar @ 2019-11-07 22:46 UTC (permalink / raw)
  To: Geert Uytterhoeven
  Cc: Mark Rutland,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Chris Paterson, Lorenzo Pieralisi, Lad, Prabhakar, Arnd Bergmann,
	Greg Kroah-Hartman, linux-pci, Yoshihiro Shimoda, Magnus Damm,
	Linux Kernel Mailing List, Kishon Vijay Abraham I, Linux-Renesas,
	Rob Herring, Catalin Marinas, Bjorn Helgaas, Andrew Murray,
	Will Deacon, Linux ARM, Marek Vasut

Hi Geert,

On Thu, Nov 7, 2019 at 8:08 PM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
>
> Hi Prabhakar,
>
> On Thu, Nov 7, 2019 at 10:26 AM Lad, Prabhakar
> <prabhakar.csengg@gmail.com> wrote:
> > On Thu, Nov 7, 2019 at 8:44 AM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> > > On Wed, Nov 6, 2019 at 8:36 PM Lad Prabhakar <prabhakar.csengg@gmail.com> wrote:
> > > > From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > > >
> > > > This patch adds the bindings for the R-Car PCIe endpoint driver.
> > > >
> > > > Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > >
> > > Thanks for your patch!
> > >
> > > > --- /dev/null
> > > > +++ b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> > > > @@ -0,0 +1,43 @@
> > > > +* Renesas R-Car PCIe Endpoint Controller DT description
> > > > +
> > > > +Required properties:
> > > > +           "renesas,pcie-ep-r8a774c0" for the R8A774C0 SoC;
> > > > +           "renesas,pcie-ep-rcar-gen3" for a generic R-Car Gen3 or
> > > > +                                    RZ/G2 compatible device.
> > >
> > > Unless I'm missing something, this is for the exact same hardware block as
> > > Documentation/devicetree/bindings/pci/rcar-pci.txt?
> > > So shouldn't you amend those bindings, instead of adding new compatible
> > > values?
> > > Please remember that DT describes hardware, not software policy.
> > > So IMHO choosing between host and endpoint is purely a configuration
> > > issue, and could be indicated by the presence or lack of some DT properties.
> > > E.g. host mode requires both "bus-range" and "device_type" properties,
> > > so their absence could indicate endpoint mode.
> > >
> > yes its the same hardware block as described in the rcar-pci.txt, I
> > did think about amending it
> > but  it might turn out to be bit messy,
> >
> > required properties host ======required properties Endpoint
> > ====================||==================
> > 1: reg                                || reg
> > 2:bus-range                      || reg names
> > 3: device_type                  || resets
> > 4: ranges                          || clocks
> > 5: dma-ranges                  || clock-names
> > 6: interrupts                      ||
> > 7: interrupt-cells               ||
> > 8: interrupt-map-mask     ||
> > 9: clocks                          ||
> > 10: clock-names             ||
>
> We have a similar situation with SPI, where a controller can operate in
> master or slave mode, based on the absence or presence of the
> "spi-slave" DT property.
>
> > and if I go ahead with the same compatible string that would mean to
> > add support for endpoint
> > mode in the host driver itself. I did follow the examples of
>
> You can still have two separate drivers, binding against the same
> compatible value.  Just let the .probe() function return -ENODEV if it
> discovers (by looking at DT properties) if the node is configured for
> the other mode.
> Which brings us to my next questions: is there any code that could be
> shared between the drivers for the two modes?
>
agreed driver probe could handle depending on the DT properties.
yes there is bit of common code and the first patch of the series prepares
for handling host and endpoint mode.

> > rockchip/cadence/designware where
> > its the same hardware block but has two different binding files one
> > for host mode and other for
> > endpoint mode.
>
> Having two separate DT binding documents sounds fine to me, if unifying
> them makes things too complex.
> However, I think they should use the same compatible value, because the
> hardware block is the same, but just used in a different mode.
>
agreed.

> Rob/Mark: Any input from the DT maintainers?
>
> > > > +- reg: Five register ranges as listed in the reg-names property
> > > > +- reg-names: Must include the following names
> > > > +       - "apb-base"
> > > > +       - "memory0"
> > > > +       - "memory1"
> > > > +       - "memory2"
> > > > +       - "memory3"
> > >
> > > What is the purpose of the last 4 regions?
> > > Can they be chosen by the driver, at runtime?
> > >
> > no the driver cannot choose them at runtime, as these are the only
> > PCIE memory(0/1/2/3) ranges
> > in the AXI address space where host memory can be mapped.
>
> Are they fixed by the PCIe hardware, i.e. could they be looked up by the
> driver based on the compatible value?
>
yes they are fixed by the PCIe hardware and could be looked up by the driver
based on the compatible value.

Cheers,
--Prabhakar

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
  2019-11-07 22:46           ` Lad, Prabhakar
@ 2019-11-08  8:36             ` Geert Uytterhoeven
  -1 siblings, 0 replies; 60+ messages in thread
From: Geert Uytterhoeven @ 2019-11-08  8:36 UTC (permalink / raw)
  To: Lad, Prabhakar
  Cc: Bjorn Helgaas, Rob Herring, Mark Rutland, Magnus Damm,
	Kishon Vijay Abraham I, Marek Vasut, Yoshihiro Shimoda,
	linux-pci, Catalin Marinas, Will Deacon, Lorenzo Pieralisi,
	Arnd Bergmann, Greg Kroah-Hartman, Andrew Murray,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Linux Kernel Mailing List, Linux ARM, Linux-Renesas,
	Chris Paterson, Lad, Prabhakar

Hi Prabhakar,

On Thu, Nov 7, 2019 at 11:46 PM Lad, Prabhakar
<prabhakar.csengg@gmail.com> wrote:
> On Thu, Nov 7, 2019 at 8:08 PM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> > On Thu, Nov 7, 2019 at 10:26 AM Lad, Prabhakar
> > <prabhakar.csengg@gmail.com> wrote:
> > > On Thu, Nov 7, 2019 at 8:44 AM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> > > > On Wed, Nov 6, 2019 at 8:36 PM Lad Prabhakar <prabhakar.csengg@gmail.com> wrote:
> > > > > From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > > > >
> > > > > This patch adds the bindings for the R-Car PCIe endpoint driver.
> > > > >
> > > > > Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > > >
> > > > > --- /dev/null
> > > > > +++ b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt

> > > > > +- reg: Five register ranges as listed in the reg-names property
> > > > > +- reg-names: Must include the following names
> > > > > +       - "apb-base"
> > > > > +       - "memory0"
> > > > > +       - "memory1"
> > > > > +       - "memory2"
> > > > > +       - "memory3"
> > > >
> > > > What is the purpose of the last 4 regions?
> > > > Can they be chosen by the driver, at runtime?
> > > >
> > > no the driver cannot choose them at runtime, as these are the only
> > > PCIE memory(0/1/2/3) ranges
> > > in the AXI address space where host memory can be mapped.
> >
> > Are they fixed by the PCIe hardware, i.e. could they be looked up by the
> > driver based on the compatible value?
> >
> yes they are fixed by the PCIe hardware and could be looked up by the driver
> based on the compatible value.

Thanks, so we don't need to describe them in DT.

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
@ 2019-11-08  8:36             ` Geert Uytterhoeven
  0 siblings, 0 replies; 60+ messages in thread
From: Geert Uytterhoeven @ 2019-11-08  8:36 UTC (permalink / raw)
  To: Lad, Prabhakar
  Cc: Mark Rutland,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Chris Paterson, Lorenzo Pieralisi, Lad, Prabhakar, Arnd Bergmann,
	Greg Kroah-Hartman, linux-pci, Yoshihiro Shimoda, Magnus Damm,
	Linux Kernel Mailing List, Kishon Vijay Abraham I, Linux-Renesas,
	Rob Herring, Catalin Marinas, Bjorn Helgaas, Andrew Murray,
	Will Deacon, Linux ARM, Marek Vasut

Hi Prabhakar,

On Thu, Nov 7, 2019 at 11:46 PM Lad, Prabhakar
<prabhakar.csengg@gmail.com> wrote:
> On Thu, Nov 7, 2019 at 8:08 PM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> > On Thu, Nov 7, 2019 at 10:26 AM Lad, Prabhakar
> > <prabhakar.csengg@gmail.com> wrote:
> > > On Thu, Nov 7, 2019 at 8:44 AM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> > > > On Wed, Nov 6, 2019 at 8:36 PM Lad Prabhakar <prabhakar.csengg@gmail.com> wrote:
> > > > > From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > > > >
> > > > > This patch adds the bindings for the R-Car PCIe endpoint driver.
> > > > >
> > > > > Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > > >
> > > > > --- /dev/null
> > > > > +++ b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt

> > > > > +- reg: Five register ranges as listed in the reg-names property
> > > > > +- reg-names: Must include the following names
> > > > > +       - "apb-base"
> > > > > +       - "memory0"
> > > > > +       - "memory1"
> > > > > +       - "memory2"
> > > > > +       - "memory3"
> > > >
> > > > What is the purpose of the last 4 regions?
> > > > Can they be chosen by the driver, at runtime?
> > > >
> > > no the driver cannot choose them at runtime, as these are the only
> > > PCIE memory(0/1/2/3) ranges
> > > in the AXI address space where host memory can be mapped.
> >
> > Are they fixed by the PCIe hardware, i.e. could they be looked up by the
> > driver based on the compatible value?
> >
> yes they are fixed by the PCIe hardware and could be looked up by the driver
> based on the compatible value.

Thanks, so we don't need to describe them in DT.

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
  2019-11-07 20:08         ` Geert Uytterhoeven
@ 2019-11-13  4:08           ` Rob Herring
  -1 siblings, 0 replies; 60+ messages in thread
From: Rob Herring @ 2019-11-13  4:08 UTC (permalink / raw)
  To: Geert Uytterhoeven
  Cc: Lad, Prabhakar, Bjorn Helgaas, Mark Rutland, Magnus Damm,
	Kishon Vijay Abraham I, Marek Vasut, Yoshihiro Shimoda,
	linux-pci, Catalin Marinas, Will Deacon, Lorenzo Pieralisi,
	Arnd Bergmann, Greg Kroah-Hartman, Andrew Murray,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Linux Kernel Mailing List, Linux ARM, Linux-Renesas,
	Chris Paterson, Lad, Prabhakar

On Thu, Nov 07, 2019 at 09:08:35PM +0100, Geert Uytterhoeven wrote:
> Hi Prabhakar,
> 
> On Thu, Nov 7, 2019 at 10:26 AM Lad, Prabhakar
> <prabhakar.csengg@gmail.com> wrote:
> > On Thu, Nov 7, 2019 at 8:44 AM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> > > On Wed, Nov 6, 2019 at 8:36 PM Lad Prabhakar <prabhakar.csengg@gmail.com> wrote:
> > > > From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > > >
> > > > This patch adds the bindings for the R-Car PCIe endpoint driver.
> > > >
> > > > Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > >
> > > Thanks for your patch!
> > >
> > > > --- /dev/null
> > > > +++ b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> > > > @@ -0,0 +1,43 @@
> > > > +* Renesas R-Car PCIe Endpoint Controller DT description
> > > > +
> > > > +Required properties:
> > > > +           "renesas,pcie-ep-r8a774c0" for the R8A774C0 SoC;
> > > > +           "renesas,pcie-ep-rcar-gen3" for a generic R-Car Gen3 or
> > > > +                                    RZ/G2 compatible device.
> > >
> > > Unless I'm missing something, this is for the exact same hardware block as
> > > Documentation/devicetree/bindings/pci/rcar-pci.txt?
> > > So shouldn't you amend those bindings, instead of adding new compatible
> > > values?
> > > Please remember that DT describes hardware, not software policy.
> > > So IMHO choosing between host and endpoint is purely a configuration
> > > issue, and could be indicated by the presence or lack of some DT properties.
> > > E.g. host mode requires both "bus-range" and "device_type" properties,
> > > so their absence could indicate endpoint mode.
> > >
> > yes its the same hardware block as described in the rcar-pci.txt, I
> > did think about amending it
> > but  it might turn out to be bit messy,
> >
> > required properties host ======required properties Endpoint
> > ====================||==================
> > 1: reg                                || reg
> > 2:bus-range                      || reg names
> > 3: device_type                  || resets
> > 4: ranges                          || clocks
> > 5: dma-ranges                  || clock-names
> > 6: interrupts                      ||
> > 7: interrupt-cells               ||
> > 8: interrupt-map-mask     ||
> > 9: clocks                          ||
> > 10: clock-names             ||
> 
> We have a similar situation with SPI, where a controller can operate in
> master or slave mode, based on the absence or presence of the
> "spi-slave" DT property.
> 
> > and if I go ahead with the same compatible string that would mean to
> > add support for endpoint
> > mode in the host driver itself. I did follow the examples of
> 
> You can still have two separate drivers, binding against the same
> compatible value.  Just let the .probe() function return -ENODEV if it
> discovers (by looking at DT properties) if the node is configured for
> the other mode.
> Which brings us to my next questions: is there any code that could be
> shared between the drivers for the two modes?
> 
> > rockchip/cadence/designware where
> > its the same hardware block but has two different binding files one
> > for host mode and other for
> > endpoint mode.
> 
> Having two separate DT binding documents sounds fine to me, if unifying
> them makes things too complex.
> However, I think they should use the same compatible value, because the
> hardware block is the same, but just used in a different mode.
> 
> Rob/Mark: Any input from the DT maintainers?

Separate files makes sense because different modes will want to 
include different common schemas. We've generally been doing different 
compatibles too which makes validating the node has the right set of 
properties easier.
 
> > > > +- reg: Five register ranges as listed in the reg-names property
> > > > +- reg-names: Must include the following names
> > > > +       - "apb-base"
> > > > +       - "memory0"
> > > > +       - "memory1"
> > > > +       - "memory2"
> > > > +       - "memory3"
> > >
> > > What is the purpose of the last 4 regions?
> > > Can they be chosen by the driver, at runtime?
> > >
> > no the driver cannot choose them at runtime, as these are the only
> > PCIE memory(0/1/2/3) ranges
> > in the AXI address space where host memory can be mapped.
> 
> Are they fixed by the PCIe hardware, i.e. could they be looked up by the
> driver based on the compatible value?

That would be strange for a memory range.

Sounds like like 'ranges' though I'm not sure if 'ranges' for an EP 
makes sense or what that should look like.

Rob

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
@ 2019-11-13  4:08           ` Rob Herring
  0 siblings, 0 replies; 60+ messages in thread
From: Rob Herring @ 2019-11-13  4:08 UTC (permalink / raw)
  To: Geert Uytterhoeven
  Cc: Mark Rutland,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Chris Paterson, Lorenzo Pieralisi, Lad, Prabhakar, Arnd Bergmann,
	Greg Kroah-Hartman, linux-pci, Yoshihiro Shimoda, Magnus Damm,
	Linux Kernel Mailing List, Kishon Vijay Abraham I, Linux-Renesas,
	Lad, Prabhakar, Catalin Marinas, Bjorn Helgaas, Andrew Murray,
	Will Deacon, Linux ARM, Marek Vasut

On Thu, Nov 07, 2019 at 09:08:35PM +0100, Geert Uytterhoeven wrote:
> Hi Prabhakar,
> 
> On Thu, Nov 7, 2019 at 10:26 AM Lad, Prabhakar
> <prabhakar.csengg@gmail.com> wrote:
> > On Thu, Nov 7, 2019 at 8:44 AM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> > > On Wed, Nov 6, 2019 at 8:36 PM Lad Prabhakar <prabhakar.csengg@gmail.com> wrote:
> > > > From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > > >
> > > > This patch adds the bindings for the R-Car PCIe endpoint driver.
> > > >
> > > > Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > >
> > > Thanks for your patch!
> > >
> > > > --- /dev/null
> > > > +++ b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> > > > @@ -0,0 +1,43 @@
> > > > +* Renesas R-Car PCIe Endpoint Controller DT description
> > > > +
> > > > +Required properties:
> > > > +           "renesas,pcie-ep-r8a774c0" for the R8A774C0 SoC;
> > > > +           "renesas,pcie-ep-rcar-gen3" for a generic R-Car Gen3 or
> > > > +                                    RZ/G2 compatible device.
> > >
> > > Unless I'm missing something, this is for the exact same hardware block as
> > > Documentation/devicetree/bindings/pci/rcar-pci.txt?
> > > So shouldn't you amend those bindings, instead of adding new compatible
> > > values?
> > > Please remember that DT describes hardware, not software policy.
> > > So IMHO choosing between host and endpoint is purely a configuration
> > > issue, and could be indicated by the presence or lack of some DT properties.
> > > E.g. host mode requires both "bus-range" and "device_type" properties,
> > > so their absence could indicate endpoint mode.
> > >
> > yes its the same hardware block as described in the rcar-pci.txt, I
> > did think about amending it
> > but  it might turn out to be bit messy,
> >
> > required properties host ======required properties Endpoint
> > ====================||==================
> > 1: reg                                || reg
> > 2:bus-range                      || reg names
> > 3: device_type                  || resets
> > 4: ranges                          || clocks
> > 5: dma-ranges                  || clock-names
> > 6: interrupts                      ||
> > 7: interrupt-cells               ||
> > 8: interrupt-map-mask     ||
> > 9: clocks                          ||
> > 10: clock-names             ||
> 
> We have a similar situation with SPI, where a controller can operate in
> master or slave mode, based on the absence or presence of the
> "spi-slave" DT property.
> 
> > and if I go ahead with the same compatible string that would mean to
> > add support for endpoint
> > mode in the host driver itself. I did follow the examples of
> 
> You can still have two separate drivers, binding against the same
> compatible value.  Just let the .probe() function return -ENODEV if it
> discovers (by looking at DT properties) if the node is configured for
> the other mode.
> Which brings us to my next questions: is there any code that could be
> shared between the drivers for the two modes?
> 
> > rockchip/cadence/designware where
> > its the same hardware block but has two different binding files one
> > for host mode and other for
> > endpoint mode.
> 
> Having two separate DT binding documents sounds fine to me, if unifying
> them makes things too complex.
> However, I think they should use the same compatible value, because the
> hardware block is the same, but just used in a different mode.
> 
> Rob/Mark: Any input from the DT maintainers?

Separate files makes sense because different modes will want to 
include different common schemas. We've generally been doing different 
compatibles too which makes validating the node has the right set of 
properties easier.
 
> > > > +- reg: Five register ranges as listed in the reg-names property
> > > > +- reg-names: Must include the following names
> > > > +       - "apb-base"
> > > > +       - "memory0"
> > > > +       - "memory1"
> > > > +       - "memory2"
> > > > +       - "memory3"
> > >
> > > What is the purpose of the last 4 regions?
> > > Can they be chosen by the driver, at runtime?
> > >
> > no the driver cannot choose them at runtime, as these are the only
> > PCIE memory(0/1/2/3) ranges
> > in the AXI address space where host memory can be mapped.
> 
> Are they fixed by the PCIe hardware, i.e. could they be looked up by the
> driver based on the compatible value?

That would be strange for a memory range.

Sounds like like 'ranges' though I'm not sure if 'ranges' for an EP 
makes sense or what that should look like.

Rob

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 0/5] Add support for PCIe controller to work in endpoint mode on R-Car SoCs.
  2019-11-06 19:36 ` Lad Prabhakar
@ 2019-11-26 14:33   ` Lad, Prabhakar
  -1 siblings, 0 replies; 60+ messages in thread
From: Lad, Prabhakar @ 2019-11-26 14:33 UTC (permalink / raw)
  To: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Kishon Vijay Abraham I, Marek Vasut,
	Yoshihiro Shimoda, linux-pci
  Cc: Catalin Marinas, Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS, LKML,
	LAK, Linux-Renesas, Chris Paterson, Lad, Prabhakar

Hi Bjorn/Kishon,

Gentle ping for this patch set.

On Wed, Nov 6, 2019 at 7:36 PM Lad Prabhakar <prabhakar.csengg@gmail.com> wrote:
>
> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
>
> This patch series adds support for PCIe controller on rcar to work in endpoint mode,
> this also extends the epf framework to handle multiple outbound regions.
>
Cheers,
--Prabhakar Lad


> Note:
> The cadence/rockchip/designware endpoint drivers are build tested only.
>
> Endpoint configuration:
> $ cd /sys/kernel/config/pci_ep
> $ mkdir functions/pci_epf_test/func1
> $ echo 0x1912 > functions/pci_epf_test/func1/vendorid
> $ echo 0x002d > functions/pci_epf_test/func1/deviceid
> $ echo 32 > functions/pci_epf_test/func1/msi_interrupts
> $ ln -s functions/pci_epf_test/func1/ controllers/fe000000.pcie_ep/
> $ echo 1 > controllers/fe000000.pcie_ep/start
>
> Host side results:
> lspci output:
> 01:00.0 Unassigned class [ff00]: Renesas Technology Corp. Device 002d
>         Flags: bus master, fast devsel, latency 0, IRQ 103
>         Memory at fe200200 (64-bit, non-prefetchable) [size=128]
>         Memory at fe200000 (64-bit, non-prefetchable) [size=256]
>         Memory at fe200100 (64-bit, non-prefetchable) [size=256]
>         Capabilities: [40] Power Management version 3
>         Capabilities: [50] MSI: Enable- Count=1/1 Maskable+ 64bit+
>         Capabilities: [70] Express Endpoint, MSI 00
>         Capabilities: [100] Virtual Channel
>         Kernel driver in use: pci-endpoint-test
>
> pcitest results:
>
> BAR tests
>
> BAR0:           OKAY
> BAR1:           NOT OKAY
> BAR2:           OKAY
> BAR3:           NOT OKAY
> BAR4:           OKAY
> BAR5:           NOT OKAY
>
> Interrupt tests
>
> SET IRQ TYPE TO LEGACY:         OKAY
> LEGACY IRQ:     OKAY
> Read Tests
>
> SET IRQ TYPE TO LEGACY:         OKAY
> READ (      1 bytes):           OKAY
> READ (   1024 bytes):           OKAY
> READ (   1025 bytes):           OKAY
> READ (1024000 bytes):           OKAY
> READ (1024001 bytes):           OKAY
>
> Write Tests
>
> WRITE (      1 bytes):          OKAY
> WRITE (   1024 bytes):          OKAY
> WRITE (   1025 bytes):          OKAY
> WRITE (1024000 bytes):          OKAY
> WRITE (1024001 bytes):          OKAY
>
> Copy Tests
>
> COPY (      1 bytes):           OKAY
> COPY (   1024 bytes):           OKAY
> COPY (   1025 bytes):           OKAY
> COPY (1024000 bytes):           OKAY
> COPY (1024001 bytes):           OKAY
>
> BAR tests for 1/3/5 fail because its configured to use 64bit bars
>
> Lad, Prabhakar (5):
>   pci: pcie-rcar: preparation for adding endpoint support
>   pci: endpoint: add support to handle multiple base for mapping
>     outbound memory
>   PCI: rcar: Add R-Car PCIe endpoint device tree bindings
>   pci: rcar: add support for rcar pcie controller in endpoint mode
>   misc: pci_endpoint_test: add device-id for RZ/G2 pcie controller
>
>  .../devicetree/bindings/pci/rcar-pci-ep.txt   |   43 +
>  arch/arm64/configs/defconfig                  |    2 +-
>  arch/arm64/configs/renesas_defconfig          |    2 +-
>  drivers/misc/pci_endpoint_test.c              |    3 +
>  drivers/pci/controller/Kconfig                |   11 +-
>  drivers/pci/controller/Makefile               |    3 +-
>  .../pci/controller/dwc/pcie-designware-ep.c   |   30 +-
>  drivers/pci/controller/pcie-cadence-ep.c      |   11 +-
>  drivers/pci/controller/pcie-rcar-ep.c         |  483 +++++++
>  drivers/pci/controller/pcie-rcar-host.c       | 1058 ++++++++++++++
>  drivers/pci/controller/pcie-rcar.c            | 1231 +----------------
>  drivers/pci/controller/pcie-rcar.h            |  129 ++
>  drivers/pci/controller/pcie-rockchip-ep.c     |   13 +-
>  drivers/pci/endpoint/functions/pci-epf-test.c |   29 +-
>  drivers/pci/endpoint/pci-epc-core.c           |    7 +-
>  drivers/pci/endpoint/pci-epc-mem.c            |  189 ++-
>  include/linux/pci-epc.h                       |   43 +-
>  17 files changed, 2016 insertions(+), 1271 deletions(-)
>  create mode 100644 Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
>  create mode 100644 drivers/pci/controller/pcie-rcar-ep.c
>  create mode 100644 drivers/pci/controller/pcie-rcar-host.c
>  create mode 100644 drivers/pci/controller/pcie-rcar.h
>
> --
> 2.20.1
>

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 0/5] Add support for PCIe controller to work in endpoint mode on R-Car SoCs.
@ 2019-11-26 14:33   ` Lad, Prabhakar
  0 siblings, 0 replies; 60+ messages in thread
From: Lad, Prabhakar @ 2019-11-26 14:33 UTC (permalink / raw)
  To: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Kishon Vijay Abraham I, Marek Vasut,
	Yoshihiro Shimoda, linux-pci
  Cc: open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Chris Paterson, Lorenzo Pieralisi, Arnd Bergmann,
	Catalin Marinas, LKML, Lad, Prabhakar, Linux-Renesas,
	Greg Kroah-Hartman, Andrew Murray, Will Deacon, LAK

Hi Bjorn/Kishon,

Gentle ping for this patch set.

On Wed, Nov 6, 2019 at 7:36 PM Lad Prabhakar <prabhakar.csengg@gmail.com> wrote:
>
> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
>
> This patch series adds support for PCIe controller on rcar to work in endpoint mode,
> this also extends the epf framework to handle multiple outbound regions.
>
Cheers,
--Prabhakar Lad


> Note:
> The cadence/rockchip/designware endpoint drivers are build tested only.
>
> Endpoint configuration:
> $ cd /sys/kernel/config/pci_ep
> $ mkdir functions/pci_epf_test/func1
> $ echo 0x1912 > functions/pci_epf_test/func1/vendorid
> $ echo 0x002d > functions/pci_epf_test/func1/deviceid
> $ echo 32 > functions/pci_epf_test/func1/msi_interrupts
> $ ln -s functions/pci_epf_test/func1/ controllers/fe000000.pcie_ep/
> $ echo 1 > controllers/fe000000.pcie_ep/start
>
> Host side results:
> lspci output:
> 01:00.0 Unassigned class [ff00]: Renesas Technology Corp. Device 002d
>         Flags: bus master, fast devsel, latency 0, IRQ 103
>         Memory at fe200200 (64-bit, non-prefetchable) [size=128]
>         Memory at fe200000 (64-bit, non-prefetchable) [size=256]
>         Memory at fe200100 (64-bit, non-prefetchable) [size=256]
>         Capabilities: [40] Power Management version 3
>         Capabilities: [50] MSI: Enable- Count=1/1 Maskable+ 64bit+
>         Capabilities: [70] Express Endpoint, MSI 00
>         Capabilities: [100] Virtual Channel
>         Kernel driver in use: pci-endpoint-test
>
> pcitest results:
>
> BAR tests
>
> BAR0:           OKAY
> BAR1:           NOT OKAY
> BAR2:           OKAY
> BAR3:           NOT OKAY
> BAR4:           OKAY
> BAR5:           NOT OKAY
>
> Interrupt tests
>
> SET IRQ TYPE TO LEGACY:         OKAY
> LEGACY IRQ:     OKAY
> Read Tests
>
> SET IRQ TYPE TO LEGACY:         OKAY
> READ (      1 bytes):           OKAY
> READ (   1024 bytes):           OKAY
> READ (   1025 bytes):           OKAY
> READ (1024000 bytes):           OKAY
> READ (1024001 bytes):           OKAY
>
> Write Tests
>
> WRITE (      1 bytes):          OKAY
> WRITE (   1024 bytes):          OKAY
> WRITE (   1025 bytes):          OKAY
> WRITE (1024000 bytes):          OKAY
> WRITE (1024001 bytes):          OKAY
>
> Copy Tests
>
> COPY (      1 bytes):           OKAY
> COPY (   1024 bytes):           OKAY
> COPY (   1025 bytes):           OKAY
> COPY (1024000 bytes):           OKAY
> COPY (1024001 bytes):           OKAY
>
> BAR tests for 1/3/5 fail because its configured to use 64bit bars
>
> Lad, Prabhakar (5):
>   pci: pcie-rcar: preparation for adding endpoint support
>   pci: endpoint: add support to handle multiple base for mapping
>     outbound memory
>   PCI: rcar: Add R-Car PCIe endpoint device tree bindings
>   pci: rcar: add support for rcar pcie controller in endpoint mode
>   misc: pci_endpoint_test: add device-id for RZ/G2 pcie controller
>
>  .../devicetree/bindings/pci/rcar-pci-ep.txt   |   43 +
>  arch/arm64/configs/defconfig                  |    2 +-
>  arch/arm64/configs/renesas_defconfig          |    2 +-
>  drivers/misc/pci_endpoint_test.c              |    3 +
>  drivers/pci/controller/Kconfig                |   11 +-
>  drivers/pci/controller/Makefile               |    3 +-
>  .../pci/controller/dwc/pcie-designware-ep.c   |   30 +-
>  drivers/pci/controller/pcie-cadence-ep.c      |   11 +-
>  drivers/pci/controller/pcie-rcar-ep.c         |  483 +++++++
>  drivers/pci/controller/pcie-rcar-host.c       | 1058 ++++++++++++++
>  drivers/pci/controller/pcie-rcar.c            | 1231 +----------------
>  drivers/pci/controller/pcie-rcar.h            |  129 ++
>  drivers/pci/controller/pcie-rockchip-ep.c     |   13 +-
>  drivers/pci/endpoint/functions/pci-epf-test.c |   29 +-
>  drivers/pci/endpoint/pci-epc-core.c           |    7 +-
>  drivers/pci/endpoint/pci-epc-mem.c            |  189 ++-
>  include/linux/pci-epc.h                       |   43 +-
>  17 files changed, 2016 insertions(+), 1271 deletions(-)
>  create mode 100644 Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
>  create mode 100644 drivers/pci/controller/pcie-rcar-ep.c
>  create mode 100644 drivers/pci/controller/pcie-rcar-host.c
>  create mode 100644 drivers/pci/controller/pcie-rcar.h
>
> --
> 2.20.1
>

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 1/5] pci: pcie-rcar: preparation for adding endpoint support
  2019-11-06 19:36   ` Lad Prabhakar
@ 2019-11-27  4:55     ` Kishon Vijay Abraham I
  -1 siblings, 0 replies; 60+ messages in thread
From: Kishon Vijay Abraham I @ 2019-11-27  4:55 UTC (permalink / raw)
  To: Lad Prabhakar, Bjorn Helgaas, Rob Herring, Mark Rutland,
	Geert Uytterhoeven, Magnus Damm, Marek Vasut, Yoshihiro Shimoda,
	linux-pci
  Cc: Catalin Marinas, Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray, devicetree, linux-kernel,
	linux-arm-kernel, linux-renesas-soc, Chris Paterson, Lad,
	Prabhakar

Hi Prabhakar,

On 07/11/19 1:06 AM, Lad Prabhakar wrote:
> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> 
> this patch prepares for adding endpoint support to rcar controller,
> there are no functional changes with this patch, a common file is
> created so that it can be shared with endpoint driver.
> 
> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> ---
>  arch/arm64/configs/defconfig            |    2 +-
>  arch/arm64/configs/renesas_defconfig    |    2 +-
>  drivers/pci/controller/Kconfig          |    4 +-
>  drivers/pci/controller/Makefile         |    2 +-
>  drivers/pci/controller/pcie-rcar-host.c | 1058 +++++++++++++++++++
>  drivers/pci/controller/pcie-rcar.c      | 1231 +----------------------

Can you create this patch using "git format-patch -M", so that only whatever
changed appears here.

Thanks
Kishon

>  drivers/pci/controller/pcie-rcar.h      |  129 +++
>  7 files changed, 1245 insertions(+), 1183 deletions(-)
>  create mode 100644 drivers/pci/controller/pcie-rcar-host.c
>  create mode 100644 drivers/pci/controller/pcie-rcar.h
> 
> diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
> index 7d16f10a9251..f67ce0ccea1d 100644
> --- a/arch/arm64/configs/defconfig
> +++ b/arch/arm64/configs/defconfig
> @@ -186,7 +186,7 @@ CONFIG_HOTPLUG_PCI=y
>  CONFIG_HOTPLUG_PCI_ACPI=y
>  CONFIG_PCI_AARDVARK=y
>  CONFIG_PCI_TEGRA=y
> -CONFIG_PCIE_RCAR=y
> +CONFIG_PCIE_RCAR_HOST=y
>  CONFIG_PCI_HOST_GENERIC=y
>  CONFIG_PCI_XGENE=y
>  CONFIG_PCIE_ALTERA=y
> diff --git a/arch/arm64/configs/renesas_defconfig b/arch/arm64/configs/renesas_defconfig
> index f7411e8807d6..75fc418377c5 100644
> --- a/arch/arm64/configs/renesas_defconfig
> +++ b/arch/arm64/configs/renesas_defconfig
> @@ -91,7 +91,7 @@ CONFIG_PCIEPORTBUS=y
>  CONFIG_HOTPLUG_PCI_PCIE=y
>  CONFIG_PCI_IOV=y
>  CONFIG_HOTPLUG_PCI=y
> -CONFIG_PCIE_RCAR=y
> +CONFIG_PCIE_RCAR_HOST=y
>  CONFIG_PCI_HOST_GENERIC=y
>  CONFIG_DEVTMPFS=y
>  CONFIG_DEVTMPFS_MOUNT=y
> diff --git a/drivers/pci/controller/Kconfig b/drivers/pci/controller/Kconfig
> index 70e078238899..01c238e5ea1e 100644
> --- a/drivers/pci/controller/Kconfig
> +++ b/drivers/pci/controller/Kconfig
> @@ -82,12 +82,12 @@ config PCI_RCAR_GEN2
>  	  There are 3 internal PCI controllers available with a single
>  	  built-in EHCI/OHCI host controller present on each one.
>  
> -config PCIE_RCAR
> +config PCIE_RCAR_HOST
>  	bool "Renesas R-Car PCIe controller"
>  	depends on ARCH_RENESAS || COMPILE_TEST
>  	depends on PCI_MSI_IRQ_DOMAIN
>  	help
> -	  Say Y here if you want PCIe controller support on R-Car SoCs.
> +	  Say Y here if you want PCIe controller support on R-Car SoCs in host mode.
>  
>  config PCI_HOST_COMMON
>  	bool
> diff --git a/drivers/pci/controller/Makefile b/drivers/pci/controller/Makefile
> index a2a22c9d91af..3577902a0d40 100644
> --- a/drivers/pci/controller/Makefile
> +++ b/drivers/pci/controller/Makefile
> @@ -9,7 +9,7 @@ obj-$(CONFIG_PCI_MVEBU) += pci-mvebu.o
>  obj-$(CONFIG_PCI_AARDVARK) += pci-aardvark.o
>  obj-$(CONFIG_PCI_TEGRA) += pci-tegra.o
>  obj-$(CONFIG_PCI_RCAR_GEN2) += pci-rcar-gen2.o
> -obj-$(CONFIG_PCIE_RCAR) += pcie-rcar.o
> +obj-$(CONFIG_PCIE_RCAR_HOST) += pcie-rcar.o pcie-rcar-host.o
>  obj-$(CONFIG_PCI_HOST_COMMON) += pci-host-common.o
>  obj-$(CONFIG_PCI_HOST_GENERIC) += pci-host-generic.o
>  obj-$(CONFIG_PCIE_XILINX) += pcie-xilinx.o
> diff --git a/drivers/pci/controller/pcie-rcar-host.c b/drivers/pci/controller/pcie-rcar-host.c
> new file mode 100644
> index 000000000000..645ef3e41e27
> --- /dev/null
> +++ b/drivers/pci/controller/pcie-rcar-host.c
> @@ -0,0 +1,1058 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/*
> + * PCIe driver for Renesas R-Car SoCs
> + *  Copyright (C) 2014-2019 Renesas Electronics Europe Ltd
> + *
> + * Based on:
> + *  arch/sh/drivers/pci/pcie-sh7786.c
> + *  arch/sh/drivers/pci/ops-sh7786.c
> + *  Copyright (C) 2009 - 2011  Paul Mundt
> + *
> + * Author: Phil Edworthy <phil.edworthy@renesas.com>
> + */
> +
> +#include <linux/clk.h>
> +#include <linux/delay.h>
> +#include <linux/msi.h>
> +#include <linux/of_address.h>
> +#include <linux/of_irq.h>
> +#include <linux/of_pci.h>
> +#include <linux/of_platform.h>
> +#include <linux/pci.h>
> +#include <linux/phy/phy.h>
> +
> +#include "pcie-rcar.h"
> +
> +struct rcar_msi {
> +	DECLARE_BITMAP(used, INT_PCI_MSI_NR);
> +	struct irq_domain *domain;
> +	struct msi_controller chip;
> +	unsigned long pages;
> +	struct mutex lock; /* serializes msi */
> +	int irq1;
> +	int irq2;
> +};
> +
> +static inline struct rcar_msi *to_rcar_msi(struct msi_controller *chip)
> +{
> +	return container_of(chip, struct rcar_msi, chip);
> +}
> +
> +/* Structure representing the PCIe interface */
> +struct rcar_pcie {
> +	struct device		*dev;
> +	struct phy		*phy;
> +	void __iomem		*base;
> +	struct list_head	resources;
> +	int			root_bus_nr;
> +	struct clk		*bus_clk;
> +	struct			rcar_msi msi;
> +};
> +
> +static u32 rcar_read_conf(struct rcar_pcie *pcie, int where)
> +{
> +	unsigned int shift = BITS_PER_BYTE * (where & 3);
> +	u32 val = rcar_pci_read_reg(pcie->base, where & ~3);
> +
> +	return val >> shift;
> +}
> +
> +/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
> +static int
> +rcar_pcie_config_access(struct rcar_pcie *pcie,
> +			unsigned char access_type, struct pci_bus *bus,
> +			unsigned int devfn, int where, u32 *data)
> +{
> +	unsigned int dev, func, reg, index;
> +
> +	dev = PCI_SLOT(devfn);
> +	func = PCI_FUNC(devfn);
> +	reg = where & ~3;
> +	index = reg / 4;
> +
> +	/*
> +	 * While each channel has its own memory-mapped extended config
> +	 * space, it's generally only accessible when in endpoint mode.
> +	 * When in root complex mode, the controller is unable to target
> +	 * itself with either type 0 or type 1 accesses, and indeed, any
> +	 * controller initiated target transfer to its own config space
> +	 * result in a completer abort.
> +	 *
> +	 * Each channel effectively only supports a single device, but as
> +	 * the same channel <-> device access works for any PCI_SLOT()
> +	 * value, we cheat a bit here and bind the controller's config
> +	 * space to devfn 0 in order to enable self-enumeration. In this
> +	 * case the regular ECAR/ECDR path is sidelined and the mangled
> +	 * config access itself is initiated as an internal bus transaction.
> +	 */
> +	if (pci_is_root_bus(bus)) {
> +		if (dev != 0)
> +			return PCIBIOS_DEVICE_NOT_FOUND;
> +
> +		if (access_type == RCAR_PCI_ACCESS_READ) {
> +			*data = rcar_pci_read_reg(pcie->base, PCICONF(index));
> +		} else {
> +			/* Keep an eye out for changes to the root bus number */
> +			if (pci_is_root_bus(bus) && reg == PCI_PRIMARY_BUS)
> +				pcie->root_bus_nr = *data & 0xff;
> +
> +			rcar_pci_write_reg(pcie->base, *data, PCICONF(index));
> +		}
> +
> +		return PCIBIOS_SUCCESSFUL;
> +	}
> +
> +	if (pcie->root_bus_nr < 0)
> +		return PCIBIOS_DEVICE_NOT_FOUND;
> +
> +	/* Clear errors */
> +	rcar_pci_write_reg(pcie->base, rcar_pci_read_reg(pcie->base, PCIEERRFR),
> +			   PCIEERRFR);
> +
> +	/* Set the PIO address */
> +	rcar_pci_write_reg(pcie->base, PCIE_CONF_BUS(bus->number) |
> +		PCIE_CONF_DEV(dev) | PCIE_CONF_FUNC(func) | reg, PCIECAR);
> +
> +	/* Enable the configuration access */
> +	if (bus->parent->number == pcie->root_bus_nr)
> +		rcar_pci_write_reg(pcie->base,
> +				   CONFIG_SEND_ENABLE | TYPE0, PCIECCTLR);
> +	else
> +		rcar_pci_write_reg(pcie->base,
> +				   CONFIG_SEND_ENABLE | TYPE1, PCIECCTLR);
> +
> +	/* Check for errors */
> +	if (rcar_pci_read_reg(pcie->base, PCIEERRFR) & UNSUPPORTED_REQUEST)
> +		return PCIBIOS_DEVICE_NOT_FOUND;
> +
> +	/* Check for master and target aborts */
> +	if (rcar_read_conf(pcie, RCONF(PCI_STATUS)) &
> +		(PCI_STATUS_REC_MASTER_ABORT | PCI_STATUS_REC_TARGET_ABORT))
> +		return PCIBIOS_DEVICE_NOT_FOUND;
> +
> +	if (access_type == RCAR_PCI_ACCESS_READ)
> +		*data = rcar_pci_read_reg(pcie->base, PCIECDR);
> +	else
> +		rcar_pci_write_reg(pcie->base, *data, PCIECDR);
> +
> +	/* Disable the configuration access */
> +	rcar_pci_write_reg(pcie->base, 0, PCIECCTLR);
> +
> +	return PCIBIOS_SUCCESSFUL;
> +}
> +
> +static int rcar_pcie_read_conf(struct pci_bus *bus, unsigned int devfn,
> +			       int where, int size, u32 *val)
> +{
> +	struct rcar_pcie *pcie = bus->sysdata;
> +	int ret;
> +
> +	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
> +				      bus, devfn, where, val);
> +	if (ret != PCIBIOS_SUCCESSFUL) {
> +		*val = 0xffffffff;
> +		return ret;
> +	}
> +
> +	if (size == 1)
> +		*val = (*val >> (BITS_PER_BYTE * (where & 3))) & 0xff;
> +	else if (size == 2)
> +		*val = (*val >> (BITS_PER_BYTE * (where & 2))) & 0xffff;
> +
> +	dev_dbg(&bus->dev, "pcie-config-read: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
> +		bus->number, devfn, where, size, *val);
> +
> +	return ret;
> +}
> +
> +/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
> +static int rcar_pcie_write_conf(struct pci_bus *bus, unsigned int devfn,
> +				int where, int size, u32 val)
> +{
> +	struct rcar_pcie *pcie = bus->sysdata;
> +	unsigned int shift;
> +	u32 data;
> +	int ret;
> +
> +	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
> +				      bus, devfn, where, &data);
> +	if (ret != PCIBIOS_SUCCESSFUL)
> +		return ret;
> +
> +	dev_dbg(&bus->dev, "pcie-config-write: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
> +		bus->number, devfn, where, size, val);
> +
> +	if (size == 1) {
> +		shift = BITS_PER_BYTE * (where & 3);
> +		data &= ~(0xff << shift);
> +		data |= ((val & 0xff) << shift);
> +	} else if (size == 2) {
> +		shift = BITS_PER_BYTE * (where & 2);
> +		data &= ~(0xffff << shift);
> +		data |= ((val & 0xffff) << shift);
> +	} else {
> +		data = val;
> +	}
> +
> +	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_WRITE,
> +				      bus, devfn, where, &data);
> +
> +	return ret;
> +}
> +
> +static struct pci_ops rcar_pcie_ops = {
> +	.read	= rcar_pcie_read_conf,
> +	.write	= rcar_pcie_write_conf,
> +};
> +
> +static int rcar_pcie_setup(struct list_head *resource, struct rcar_pcie *pci)
> +{
> +	struct resource_entry *win;
> +	int i = 0;
> +
> +	/* Setup PCI resources */
> +	resource_list_for_each_entry(win, &pci->resources) {
> +		struct resource *res = win->res;
> +
> +		if (!res->flags)
> +			continue;
> +
> +		switch (resource_type(res)) {
> +		case IORESOURCE_IO:
> +		case IORESOURCE_MEM:
> +			rcar_pcie_set_outbound(i, pci->base, res, true);
> +			i++;
> +			break;
> +		case IORESOURCE_BUS:
> +			pci->root_bus_nr = res->start;
> +			break;
> +		default:
> +			continue;
> +		}
> +
> +		pci_add_resource(resource, res);
> +	}
> +
> +	return 1;
> +}
> +
> +static void rcar_pcie_force_speedup(struct rcar_pcie *pcie)
> +{
> +	struct device *dev = pcie->dev;
> +	unsigned int timeout = 1000;
> +	u32 macsr;
> +
> +	if ((rcar_pci_read_reg(pcie->base, MACS2R) & LINK_SPEED) !=
> +		LINK_SPEED_5_0GTS)
> +		return;
> +
> +	if (rcar_pci_read_reg(pcie->base, MACCTLR) & SPEED_CHANGE) {
> +		dev_err(dev, "Speed change already in progress\n");
> +		return;
> +	}
> +
> +	macsr = rcar_pci_read_reg(pcie->base, MACSR);
> +	if ((macsr & LINK_SPEED) == LINK_SPEED_5_0GTS)
> +		goto done;
> +
> +	/* Set target link speed to 5.0 GT/s */
> +	rcar_rmw32(pcie->base, EXPCAP(12), PCI_EXP_LNKSTA_CLS,
> +		   PCI_EXP_LNKSTA_CLS_5_0GB);
> +
> +	/* Set speed change reason as intentional factor */
> +	rcar_rmw32(pcie->base, MACCGSPSETR, SPCNGRSN, 0);
> +
> +	/* Clear SPCHGFIN, SPCHGSUC, and SPCHGFAIL */
> +	if (macsr & (SPCHGFIN | SPCHGSUC | SPCHGFAIL))
> +		rcar_pci_write_reg(pcie->base, macsr, MACSR);
> +
> +	/* Start link speed change */
> +	rcar_rmw32(pcie->base, MACCTLR, SPEED_CHANGE, SPEED_CHANGE);
> +
> +	while (timeout--) {
> +		macsr = rcar_pci_read_reg(pcie->base, MACSR);
> +		if (macsr & SPCHGFIN) {
> +			/* Clear the interrupt bits */
> +			rcar_pci_write_reg(pcie->base, macsr, MACSR);
> +
> +			if (macsr & SPCHGFAIL)
> +				dev_err(dev, "Speed change failed\n");
> +
> +			goto done;
> +		}
> +
> +		usleep_range(1000, 2000);
> +	}
> +
> +	dev_err(dev, "Speed change timed out\n");
> +
> +done:
> +	dev_info(dev, "Current link speed is %s GT/s\n",
> +		 (macsr & LINK_SPEED) == LINK_SPEED_5_0GTS ? "5" : "2.5");
> +}
> +
> +static int rcar_pcie_enable(struct rcar_pcie *pcie)
> +{
> +	struct pci_host_bridge *bridge = pci_host_bridge_from_priv(pcie);
> +	struct device *dev = pcie->dev;
> +	struct pci_bus *bus, *child;
> +	int ret;
> +
> +	/* Try setting 5 GT/s link speed */
> +	rcar_pcie_force_speedup(pcie);
> +
> +	rcar_pcie_setup(&bridge->windows, pcie);
> +
> +	pci_add_flags(PCI_REASSIGN_ALL_BUS);
> +
> +	bridge->dev.parent = dev;
> +	bridge->sysdata = pcie;
> +	bridge->busnr = pcie->root_bus_nr;
> +	bridge->ops = &rcar_pcie_ops;
> +	bridge->map_irq = of_irq_parse_and_map_pci;
> +	bridge->swizzle_irq = pci_common_swizzle;
> +	if (IS_ENABLED(CONFIG_PCI_MSI))
> +		bridge->msi = &pcie->msi.chip;
> +
> +	ret = pci_scan_root_bus_bridge(bridge);
> +	if (ret < 0)
> +		return ret;
> +
> +	bus = bridge->bus;
> +
> +	pci_bus_size_bridges(bus);
> +	pci_bus_assign_resources(bus);
> +
> +	list_for_each_entry(child, &bus->children, node)
> +		pcie_bus_configure_settings(child);
> +
> +	pci_bus_add_devices(bus);
> +
> +	return 0;
> +}
> +
> +static int phy_wait_for_ack(struct rcar_pcie *pcie)
> +{
> +	struct device *dev = pcie->dev;
> +	unsigned int timeout = 100;
> +
> +	while (timeout--) {
> +		if (rcar_pci_read_reg(pcie->base, H1_PCIEPHYADRR) & PHY_ACK)
> +			return 0;
> +
> +		usleep_range(90, 110);
> +	}
> +
> +	dev_err(dev, "Access to PCIe phy timed out\n");
> +
> +	return -ETIMEDOUT;
> +}
> +
> +static void phy_write_reg(struct rcar_pcie *pcie,
> +			  unsigned int rate, u32 addr,
> +			  unsigned int lane, u32 data)
> +{
> +	u32 phyaddr;
> +
> +	phyaddr = WRITE_CMD |
> +		((rate & 1) << RATE_POS) |
> +		((lane & 0xf) << LANE_POS) |
> +		((addr & 0xff) << ADR_POS);
> +
> +	/* Set write data */
> +	rcar_pci_write_reg(pcie->base, data, H1_PCIEPHYDOUTR);
> +	rcar_pci_write_reg(pcie->base, phyaddr, H1_PCIEPHYADRR);
> +
> +	/* Ignore errors as they will be dealt with if the data link is down */
> +	phy_wait_for_ack(pcie);
> +
> +	/* Clear command */
> +	rcar_pci_write_reg(pcie->base, 0, H1_PCIEPHYDOUTR);
> +	rcar_pci_write_reg(pcie->base, 0, H1_PCIEPHYADRR);
> +
> +	/* Ignore errors as they will be dealt with if the data link is down */
> +	phy_wait_for_ack(pcie);
> +}
> +
> +static int rcar_pcie_hw_init(struct rcar_pcie *pcie)
> +{
> +	int err;
> +
> +	/* Begin initialization */
> +	rcar_pci_write_reg(pcie->base, 0, PCIETCTLR);
> +
> +	/* Set mode */
> +	rcar_pci_write_reg(pcie->base, 1, PCIEMSR);
> +
> +	err = rcar_pcie_wait_for_phyrdy(pcie->base);
> +	if (err)
> +		return err;
> +
> +	/*
> +	 * Initial header for port config space is type 1, set the device
> +	 * class to match. Hardware takes care of propagating the IDSETR
> +	 * settings, so there is no need to bother with a quirk.
> +	 */
> +	rcar_pci_write_reg(pcie->base, PCI_CLASS_BRIDGE_PCI << 16, IDSETR1);
> +
> +	/*
> +	 * Setup Secondary Bus Number & Subordinate Bus Number, even though
> +	 * they aren't used, to avoid bridge being detected as broken.
> +	 */
> +	rcar_rmw32(pcie->base, RCONF(PCI_SECONDARY_BUS), 0xff, 1);
> +	rcar_rmw32(pcie->base, RCONF(PCI_SUBORDINATE_BUS), 0xff, 1);
> +
> +	/* Initialize default capabilities. */
> +	rcar_rmw32(pcie->base, REXPCAP(0), 0xff, PCI_CAP_ID_EXP);
> +	rcar_rmw32(pcie->base, REXPCAP(PCI_EXP_FLAGS),
> +		   PCI_EXP_FLAGS_TYPE, PCI_EXP_TYPE_ROOT_PORT << 4);
> +	rcar_rmw32(pcie->base, RCONF(PCI_HEADER_TYPE), 0x7f,
> +		   PCI_HEADER_TYPE_BRIDGE);
> +
> +	/* Enable data link layer active state reporting */
> +	rcar_rmw32(pcie->base, REXPCAP(PCI_EXP_LNKCAP), PCI_EXP_LNKCAP_DLLLARC,
> +		   PCI_EXP_LNKCAP_DLLLARC);
> +
> +	/* Write out the physical slot number = 0 */
> +	rcar_rmw32(pcie->base, REXPCAP(PCI_EXP_SLTCAP), PCI_EXP_SLTCAP_PSN, 0);
> +
> +	/* Set the completion timer timeout to the maximum 50ms. */
> +	rcar_rmw32(pcie->base, TLCTLR + 1, 0x3f, 50);
> +
> +	/* Terminate list of capabilities (Next Capability Offset=0) */
> +	rcar_rmw32(pcie->base, RVCCAP(0), 0xfff00000, 0);
> +
> +	/* Enable MSI */
> +	if (IS_ENABLED(CONFIG_PCI_MSI))
> +		rcar_pci_write_reg(pcie->base, 0x801f0000, PCIEMSITXR);
> +
> +	rcar_rmw32(pcie->base, MACCTLR, MACCTLR_RESERVED, 0);
> +
> +	/* Finish initialization - establish a PCI Express link */
> +	rcar_pci_write_reg(pcie->base, CFINIT, PCIETCTLR);
> +
> +	/* This will timeout if we don't have a link. */
> +	err = rcar_pcie_wait_for_dl(pcie->base);
> +	if (err)
> +		return err;
> +
> +	/* Enable INTx interrupts */
> +	rcar_rmw32(pcie->base, PCIEINTXR, 0, 0xF << 8);
> +
> +	/* flush modifications */
> +	wmb();
> +
> +	return 0;
> +}
> +
> +static int rcar_pcie_phy_init_h1(struct rcar_pcie *pcie)
> +{
> +	/* Initialize the phy */
> +	phy_write_reg(pcie, 0, 0x42, 0x1, 0x0EC34191);
> +	phy_write_reg(pcie, 1, 0x42, 0x1, 0x0EC34180);
> +	phy_write_reg(pcie, 0, 0x43, 0x1, 0x00210188);
> +	phy_write_reg(pcie, 1, 0x43, 0x1, 0x00210188);
> +	phy_write_reg(pcie, 0, 0x44, 0x1, 0x015C0014);
> +	phy_write_reg(pcie, 1, 0x44, 0x1, 0x015C0014);
> +	phy_write_reg(pcie, 1, 0x4C, 0x1, 0x786174A0);
> +	phy_write_reg(pcie, 1, 0x4D, 0x1, 0x048000BB);
> +	phy_write_reg(pcie, 0, 0x51, 0x1, 0x079EC062);
> +	phy_write_reg(pcie, 0, 0x52, 0x1, 0x20000000);
> +	phy_write_reg(pcie, 1, 0x52, 0x1, 0x20000000);
> +	phy_write_reg(pcie, 1, 0x56, 0x1, 0x00003806);
> +
> +	phy_write_reg(pcie, 0, 0x60, 0x1, 0x004B03A5);
> +	phy_write_reg(pcie, 0, 0x64, 0x1, 0x3F0F1F0F);
> +	phy_write_reg(pcie, 0, 0x66, 0x1, 0x00008000);
> +
> +	return 0;
> +}
> +
> +static int rcar_pcie_phy_init_gen2(struct rcar_pcie *pcie)
> +{
> +	/*
> +	 * These settings come from the R-Car Series, 2nd Generation User's
> +	 * Manual, section 50.3.1 (2) Initialization of the physical layer.
> +	 */
> +	rcar_pci_write_reg(pcie->base, 0x000f0030, GEN2_PCIEPHYADDR);
> +	rcar_pci_write_reg(pcie->base, 0x00381203, GEN2_PCIEPHYDATA);
> +	rcar_pci_write_reg(pcie->base, 0x00000001, GEN2_PCIEPHYCTRL);
> +	rcar_pci_write_reg(pcie->base, 0x00000006, GEN2_PCIEPHYCTRL);
> +
> +	rcar_pci_write_reg(pcie->base, 0x000f0054, GEN2_PCIEPHYADDR);
> +	/* The following value is for DC connection, no termination resistor */
> +	rcar_pci_write_reg(pcie->base, 0x13802007, GEN2_PCIEPHYDATA);
> +	rcar_pci_write_reg(pcie->base, 0x00000001, GEN2_PCIEPHYCTRL);
> +	rcar_pci_write_reg(pcie->base, 0x00000006, GEN2_PCIEPHYCTRL);
> +
> +	return 0;
> +}
> +
> +static int rcar_pcie_phy_init_gen3(struct rcar_pcie *pcie)
> +{
> +	int err;
> +
> +	err = phy_init(pcie->phy);
> +	if (err)
> +		return err;
> +
> +	err = phy_power_on(pcie->phy);
> +	if (err)
> +		phy_exit(pcie->phy);
> +
> +	return err;
> +}
> +
> +static int rcar_msi_alloc(struct rcar_msi *chip)
> +{
> +	int msi;
> +
> +	mutex_lock(&chip->lock);
> +
> +	msi = find_first_zero_bit(chip->used, INT_PCI_MSI_NR);
> +	if (msi < INT_PCI_MSI_NR)
> +		set_bit(msi, chip->used);
> +	else
> +		msi = -ENOSPC;
> +
> +	mutex_unlock(&chip->lock);
> +
> +	return msi;
> +}
> +
> +static int rcar_msi_alloc_region(struct rcar_msi *chip, int no_irqs)
> +{
> +	int msi;
> +
> +	mutex_lock(&chip->lock);
> +	msi = bitmap_find_free_region(chip->used, INT_PCI_MSI_NR,
> +				      order_base_2(no_irqs));
> +	mutex_unlock(&chip->lock);
> +
> +	return msi;
> +}
> +
> +static void rcar_msi_free(struct rcar_msi *chip, unsigned long irq)
> +{
> +	mutex_lock(&chip->lock);
> +	clear_bit(irq, chip->used);
> +	mutex_unlock(&chip->lock);
> +}
> +
> +static irqreturn_t rcar_pcie_msi_irq(int irq, void *data)
> +{
> +	struct rcar_pcie *pcie = data;
> +	struct rcar_msi *msi = &pcie->msi;
> +	struct device *dev = pcie->dev;
> +	unsigned long reg;
> +
> +	reg = rcar_pci_read_reg(pcie->base, PCIEMSIFR);
> +
> +	/* MSI & INTx share an interrupt - we only handle MSI here */
> +	if (!reg)
> +		return IRQ_NONE;
> +
> +	while (reg) {
> +		unsigned int index = find_first_bit(&reg, 32);
> +		unsigned int msi_irq;
> +
> +		/* clear the interrupt */
> +		rcar_pci_write_reg(pcie->base, 1 << index, PCIEMSIFR);
> +
> +		msi_irq = irq_find_mapping(msi->domain, index);
> +		if (msi_irq) {
> +			if (test_bit(index, msi->used))
> +				generic_handle_irq(msi_irq);
> +			else
> +				dev_info(dev, "unhandled MSI\n");
> +		} else {
> +			/* Unknown MSI, just clear it */
> +			dev_dbg(dev, "unexpected MSI\n");
> +		}
> +
> +		/* see if there's any more pending in this vector */
> +		reg = rcar_pci_read_reg(pcie->base, PCIEMSIFR);
> +	}
> +
> +	return IRQ_HANDLED;
> +}
> +
> +static int rcar_msi_setup_irq(struct msi_controller *chip, struct pci_dev *pdev,
> +			      struct msi_desc *desc)
> +{
> +	struct rcar_msi *msi = to_rcar_msi(chip);
> +	struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
> +	struct msi_msg msg;
> +	unsigned int irq;
> +	int hwirq;
> +
> +	hwirq = rcar_msi_alloc(msi);
> +	if (hwirq < 0)
> +		return hwirq;
> +
> +	irq = irq_find_mapping(msi->domain, hwirq);
> +	if (!irq) {
> +		rcar_msi_free(msi, hwirq);
> +		return -EINVAL;
> +	}
> +
> +	irq_set_msi_desc(irq, desc);
> +
> +	msg.address_lo = rcar_pci_read_reg(pcie->base, PCIEMSIALR) & ~MSIFE;
> +	msg.address_hi = rcar_pci_read_reg(pcie->base, PCIEMSIAUR);
> +	msg.data = hwirq;
> +
> +	pci_write_msi_msg(irq, &msg);
> +
> +	return 0;
> +}
> +
> +static int rcar_msi_setup_irqs(struct msi_controller *chip,
> +			       struct pci_dev *pdev, int nvec, int type)
> +{
> +	struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
> +	struct rcar_msi *msi = to_rcar_msi(chip);
> +	struct msi_desc *desc;
> +	struct msi_msg msg;
> +	unsigned int irq;
> +	int hwirq;
> +	int i;
> +
> +	/* MSI-X interrupts are not supported */
> +	if (type == PCI_CAP_ID_MSIX)
> +		return -EINVAL;
> +
> +	WARN_ON(!list_is_singular(&pdev->dev.msi_list));
> +	desc = list_entry(pdev->dev.msi_list.next, struct msi_desc, list);
> +
> +	hwirq = rcar_msi_alloc_region(msi, nvec);
> +	if (hwirq < 0)
> +		return -ENOSPC;
> +
> +	irq = irq_find_mapping(msi->domain, hwirq);
> +	if (!irq)
> +		return -ENOSPC;
> +
> +	for (i = 0; i < nvec; i++) {
> +		/*
> +		 * irq_create_mapping() called from rcar_pcie_probe() pre-
> +		 * allocates descs,  so there is no need to allocate descs here.
> +		 * We can therefore assume that if irq_find_mapping() above
> +		 * returns non-zero, then the descs are also successfully
> +		 * allocated.
> +		 */
> +		if (irq_set_msi_desc_off(irq, i, desc)) {
> +			/* TODO: clear */
> +			return -EINVAL;
> +		}
> +	}
> +
> +	desc->nvec_used = nvec;
> +	desc->msi_attrib.multiple = order_base_2(nvec);
> +
> +	msg.address_lo = rcar_pci_read_reg(pcie->base, PCIEMSIALR) & ~MSIFE;
> +	msg.address_hi = rcar_pci_read_reg(pcie->base, PCIEMSIAUR);
> +	msg.data = hwirq;
> +
> +	pci_write_msi_msg(irq, &msg);
> +
> +	return 0;
> +}
> +
> +static void rcar_msi_teardown_irq(struct msi_controller *chip, unsigned int irq)
> +{
> +	struct irq_data *d = irq_get_irq_data(irq);
> +	struct rcar_msi *msi = to_rcar_msi(chip);
> +
> +	rcar_msi_free(msi, d->hwirq);
> +}
> +
> +static struct irq_chip rcar_msi_irq_chip = {
> +	.name = "R-Car PCIe MSI",
> +	.irq_enable = pci_msi_unmask_irq,
> +	.irq_disable = pci_msi_mask_irq,
> +	.irq_mask = pci_msi_mask_irq,
> +	.irq_unmask = pci_msi_unmask_irq,
> +};
> +
> +static int rcar_msi_map(struct irq_domain *domain, unsigned int irq,
> +			irq_hw_number_t hwirq)
> +{
> +	irq_set_chip_and_handler(irq, &rcar_msi_irq_chip, handle_simple_irq);
> +	irq_set_chip_data(irq, domain->host_data);
> +
> +	return 0;
> +}
> +
> +static const struct irq_domain_ops msi_domain_ops = {
> +	.map = rcar_msi_map,
> +};
> +
> +static void rcar_pcie_unmap_msi(struct rcar_pcie *pcie)
> +{
> +	struct rcar_msi *msi = &pcie->msi;
> +	int i, irq;
> +
> +	for (i = 0; i < INT_PCI_MSI_NR; i++) {
> +		irq = irq_find_mapping(msi->domain, i);
> +		if (irq > 0)
> +			irq_dispose_mapping(irq);
> +	}
> +
> +	irq_domain_remove(msi->domain);
> +}
> +
> +static int rcar_pcie_enable_msi(struct rcar_pcie *pcie)
> +{
> +	struct rcar_msi *msi = &pcie->msi;
> +	struct device *dev = pcie->dev;
> +	phys_addr_t base;
> +	int err, i;
> +
> +	mutex_init(&msi->lock);
> +
> +	msi->chip.dev = dev;
> +	msi->chip.setup_irq = rcar_msi_setup_irq;
> +	msi->chip.setup_irqs = rcar_msi_setup_irqs;
> +	msi->chip.teardown_irq = rcar_msi_teardown_irq;
> +
> +	msi->domain = irq_domain_add_linear(dev->of_node, INT_PCI_MSI_NR,
> +					    &msi_domain_ops, &msi->chip);
> +	if (!msi->domain) {
> +		dev_err(dev, "failed to create IRQ domain\n");
> +		return -ENOMEM;
> +	}
> +
> +	for (i = 0; i < INT_PCI_MSI_NR; i++)
> +		irq_create_mapping(msi->domain, i);
> +
> +	/* Two irqs are for MSI, but they are also used for non-MSI irqs */
> +	err = devm_request_irq(dev, msi->irq1, rcar_pcie_msi_irq,
> +			       IRQF_SHARED | IRQF_NO_THREAD,
> +			       rcar_msi_irq_chip.name, pcie);
> +	if (err < 0) {
> +		dev_err(dev, "failed to request IRQ: %d\n", err);
> +		goto err;
> +	}
> +
> +	err = devm_request_irq(dev, msi->irq2, rcar_pcie_msi_irq,
> +			       IRQF_SHARED | IRQF_NO_THREAD,
> +			       rcar_msi_irq_chip.name, pcie);
> +	if (err < 0) {
> +		dev_err(dev, "failed to request IRQ: %d\n", err);
> +		goto err;
> +	}
> +
> +	/* setup MSI data target */
> +	msi->pages = __get_free_pages(GFP_KERNEL, 0);
> +	if (!msi->pages) {
> +		err = -ENOMEM;
> +		goto err;
> +	}
> +	base = virt_to_phys((void *)msi->pages);
> +
> +	rcar_pci_write_reg(pcie->base, lower_32_bits(base) | MSIFE, PCIEMSIALR);
> +	rcar_pci_write_reg(pcie->base, upper_32_bits(base), PCIEMSIAUR);
> +
> +	/* enable all MSI interrupts */
> +	rcar_pci_write_reg(pcie->base, 0xffffffff, PCIEMSIIER);
> +
> +	return 0;
> +
> +err:
> +	rcar_pcie_unmap_msi(pcie);
> +	return err;
> +}
> +
> +static void rcar_pcie_teardown_msi(struct rcar_pcie *pcie)
> +{
> +	struct rcar_msi *msi = &pcie->msi;
> +
> +	/* Disable all MSI interrupts */
> +	rcar_pci_write_reg(pcie->base, 0, PCIEMSIIER);
> +
> +	/* Disable address decoding of the MSI interrupt, MSIFE */
> +	rcar_pci_write_reg(pcie->base, 0, PCIEMSIALR);
> +
> +	free_pages(msi->pages, 0);
> +
> +	rcar_pcie_unmap_msi(pcie);
> +}
> +
> +static int rcar_pcie_get_resources(struct rcar_pcie *pcie)
> +{
> +	struct device *dev = pcie->dev;
> +	struct resource res;
> +	int err, i;
> +
> +	pcie->phy = devm_phy_optional_get(dev, "pcie");
> +	if (IS_ERR(pcie->phy))
> +		return PTR_ERR(pcie->phy);
> +
> +	err = of_address_to_resource(dev->of_node, 0, &res);
> +	if (err)
> +		return err;
> +
> +	pcie->base = devm_ioremap_resource(dev, &res);
> +	if (IS_ERR(pcie->base))
> +		return PTR_ERR(pcie->base);
> +
> +	pcie->bus_clk = devm_clk_get(dev, "pcie_bus");
> +	if (IS_ERR(pcie->bus_clk)) {
> +		dev_err(dev, "cannot get pcie bus clock\n");
> +		return PTR_ERR(pcie->bus_clk);
> +	}
> +
> +	i = irq_of_parse_and_map(dev->of_node, 0);
> +	if (!i) {
> +		dev_err(dev, "cannot get platform resources for msi interrupt\n");
> +		err = -ENOENT;
> +		goto err_irq1;
> +	}
> +	pcie->msi.irq1 = i;
> +
> +	i = irq_of_parse_and_map(dev->of_node, 1);
> +	if (!i) {
> +		dev_err(dev, "cannot get platform resources for msi interrupt\n");
> +		err = -ENOENT;
> +		goto err_irq2;
> +	}
> +	pcie->msi.irq2 = i;
> +
> +	return 0;
> +
> +err_irq2:
> +	irq_dispose_mapping(pcie->msi.irq1);
> +err_irq1:
> +	return err;
> +}
> +
> +static int rcar_pcie_inbound_ranges(struct rcar_pcie *pcie,
> +				    struct of_pci_range *range,
> +				    int *index)
> +{
> +	u64 restype = range->flags;
> +	u64 cpu_addr = range->cpu_addr;
> +	u64 cpu_end = range->cpu_addr + range->size;
> +	u64 pci_addr = range->pci_addr;
> +	u32 flags = LAM_64BIT | LAR_ENABLE;
> +	u64 mask;
> +	u64 size;
> +	int idx = *index;
> +
> +	if (restype & IORESOURCE_PREFETCH)
> +		flags |= LAM_PREFETCH;
> +
> +	/*
> +	 * If the size of the range is larger than the alignment of the start
> +	 * address, we have to use multiple entries to perform the mapping.
> +	 */
> +	if (cpu_addr > 0) {
> +		unsigned long nr_zeros = __ffs64(cpu_addr);
> +		u64 alignment = 1ULL << nr_zeros;
> +
> +		size = min(range->size, alignment);
> +	} else {
> +		size = range->size;
> +	}
> +	/* Hardware supports max 4GiB inbound region */
> +	size = min(size, 1ULL << 32);
> +
> +	mask = roundup_pow_of_two(size) - 1;
> +	mask &= ~0xf;
> +
> +	while (cpu_addr < cpu_end) {
> +		rcar_pcie_set_inbound(pcie->base, cpu_addr, pci_addr,
> +				      mask | flags, idx, true);
> +		pci_addr += size;
> +		cpu_addr += size;
> +		idx += 2;
> +
> +		if (idx > MAX_NR_INBOUND_MAPS) {
> +			dev_err(pcie->dev, "Failed to map inbound regions!\n");
> +			return -EINVAL;
> +		}
> +	}
> +	*index = idx;
> +
> +	return 0;
> +}
> +
> +static int rcar_pcie_parse_map_dma_ranges(struct rcar_pcie *pcie,
> +					  struct device_node *np)
> +{
> +	struct of_pci_range_parser parser;
> +	struct of_pci_range range;
> +	int index = 0;
> +	int err;
> +
> +	if (of_pci_dma_range_parser_init(&parser, np))
> +		return -EINVAL;
> +
> +	/* Get the dma-ranges from DT */
> +	for_each_of_pci_range(&parser, &range) {
> +		u64 end = range.cpu_addr + range.size - 1;
> +
> +		dev_dbg(pcie->dev, "0x%08x 0x%016llx..0x%016llx -> 0x%016llx\n",
> +			range.flags, range.cpu_addr, end, range.pci_addr);
> +
> +		err = rcar_pcie_inbound_ranges(pcie, &range, &index);
> +		if (err)
> +			return err;
> +	}
> +
> +	return 0;
> +}
> +
> +static const struct of_device_id rcar_pcie_of_match[] = {
> +	{ .compatible = "renesas,pcie-r8a7779",
> +	  .data = rcar_pcie_phy_init_h1 },
> +	{ .compatible = "renesas,pcie-r8a7790",
> +	  .data = rcar_pcie_phy_init_gen2 },
> +	{ .compatible = "renesas,pcie-r8a7791",
> +	  .data = rcar_pcie_phy_init_gen2 },
> +	{ .compatible = "renesas,pcie-rcar-gen2",
> +	  .data = rcar_pcie_phy_init_gen2 },
> +	{ .compatible = "renesas,pcie-r8a7795",
> +	  .data = rcar_pcie_phy_init_gen3 },
> +	{ .compatible = "renesas,pcie-rcar-gen3",
> +	  .data = rcar_pcie_phy_init_gen3 },
> +	{},
> +};
> +
> +static int rcar_pcie_probe(struct platform_device *pdev)
> +{
> +	int (*phy_init_fn)(struct rcar_pcie *pcie);
> +	struct device *dev = &pdev->dev;
> +	struct pci_host_bridge *bridge;
> +	struct rcar_pcie *pcie;
> +	u32 data;
> +	int err;
> +
> +	bridge = pci_alloc_host_bridge(sizeof(*pcie));
> +	if (!bridge)
> +		return -ENOMEM;
> +
> +	pcie = pci_host_bridge_priv(bridge);
> +
> +	pcie->dev = dev;
> +	platform_set_drvdata(pdev, pcie);
> +
> +	err = pci_parse_request_of_pci_ranges(dev, &pcie->resources, NULL);
> +	if (err)
> +		goto err_free_bridge;
> +
> +	pm_runtime_enable(pcie->dev);
> +	err = pm_runtime_get_sync(pcie->dev);
> +	if (err < 0) {
> +		dev_err(pcie->dev, "pm_runtime_get_sync failed\n");
> +		goto err_pm_disable;
> +	}
> +
> +	err = rcar_pcie_get_resources(pcie);
> +	if (err < 0) {
> +		dev_err(dev, "failed to request resources: %d\n", err);
> +		goto err_pm_put;
> +	}
> +
> +	err = clk_prepare_enable(pcie->bus_clk);
> +	if (err) {
> +		dev_err(dev, "failed to enable bus clock: %d\n", err);
> +		goto err_unmap_msi_irqs;
> +	}
> +
> +	err = rcar_pcie_parse_map_dma_ranges(pcie, dev->of_node);
> +	if (err)
> +		goto err_clk_disable;
> +
> +	phy_init_fn = of_device_get_match_data(dev);
> +	err = phy_init_fn(pcie);
> +	if (err) {
> +		dev_err(dev, "failed to init PCIe PHY\n");
> +		goto err_clk_disable;
> +	}
> +
> +	/* Failure to get a link might just be that no cards are inserted */
> +	if (rcar_pcie_hw_init(pcie)) {
> +		dev_info(dev, "PCIe link down\n");
> +		err = -ENODEV;
> +		goto err_phy_shutdown;
> +	}
> +
> +	data = rcar_pci_read_reg(pcie->base, MACSR);
> +	dev_info(dev, "PCIe x%d: link up\n", (data >> 20) & 0x3f);
> +
> +	if (IS_ENABLED(CONFIG_PCI_MSI)) {
> +		err = rcar_pcie_enable_msi(pcie);
> +		if (err < 0) {
> +			dev_err(dev,
> +				"failed to enable MSI support: %d\n",
> +				err);
> +			goto err_phy_shutdown;
> +		}
> +	}
> +
> +	err = rcar_pcie_enable(pcie);
> +	if (err)
> +		goto err_msi_teardown;
> +
> +	return 0;
> +
> +err_msi_teardown:
> +	if (IS_ENABLED(CONFIG_PCI_MSI))
> +		rcar_pcie_teardown_msi(pcie);
> +
> +err_phy_shutdown:
> +	if (pcie->phy) {
> +		phy_power_off(pcie->phy);
> +		phy_exit(pcie->phy);
> +	}
> +
> +err_clk_disable:
> +	clk_disable_unprepare(pcie->bus_clk);
> +
> +err_unmap_msi_irqs:
> +	irq_dispose_mapping(pcie->msi.irq2);
> +	irq_dispose_mapping(pcie->msi.irq1);
> +
> +err_pm_put:
> +	pm_runtime_put(dev);
> +
> +err_pm_disable:
> +	pm_runtime_disable(dev);
> +	pci_free_resource_list(&pcie->resources);
> +
> +err_free_bridge:
> +	pci_free_host_bridge(bridge);
> +
> +	return err;
> +}
> +
> +static int rcar_pcie_resume_noirq(struct device *dev)
> +{
> +	struct rcar_pcie *pcie = dev_get_drvdata(dev);
> +
> +	if (rcar_pci_read_reg(pcie->base, PMSR) &&
> +	    !(rcar_pci_read_reg(pcie->base, PCIETCTLR) & DL_DOWN))
> +		return 0;
> +
> +	/* Re-establish the PCIe link */
> +	rcar_rmw32(pcie->base, MACCTLR, MACCTLR_RESERVED, 0);
> +	rcar_pci_write_reg(pcie->base, CFINIT, PCIETCTLR);
> +	return rcar_pcie_wait_for_dl(pcie->base);
> +}
> +
> +static const struct dev_pm_ops rcar_pcie_pm_ops = {
> +	.resume_noirq = rcar_pcie_resume_noirq,
> +};
> +
> +static struct platform_driver rcar_pcie_driver = {
> +	.driver = {
> +		.name = "rcar-pcie",
> +		.of_match_table = rcar_pcie_of_match,
> +		.pm = &rcar_pcie_pm_ops,
> +		.suppress_bind_attrs = true,
> +	},
> +	.probe = rcar_pcie_probe,
> +};
> +builtin_platform_driver(rcar_pcie_driver);
> diff --git a/drivers/pci/controller/pcie-rcar.c b/drivers/pci/controller/pcie-rcar.c
> index e45bb2a7bfa5..1008ae6a896c 100644
> --- a/drivers/pci/controller/pcie-rcar.c
> +++ b/drivers/pci/controller/pcie-rcar.c
> @@ -1,7 +1,7 @@
>  // SPDX-License-Identifier: GPL-2.0
>  /*
>   * PCIe driver for Renesas R-Car SoCs
> - *  Copyright (C) 2014 Renesas Electronics Europe Ltd
> + *  Copyright (C) 2014-2019 Renesas Electronics Europe Ltd
>   *
>   * Based on:
>   *  arch/sh/drivers/pci/pcie-sh7786.c
> @@ -11,534 +11,80 @@
>   * Author: Phil Edworthy <phil.edworthy@renesas.com>
>   */
>  
> -#include <linux/bitops.h>
> -#include <linux/clk.h>
>  #include <linux/delay.h>
> -#include <linux/interrupt.h>
> -#include <linux/irq.h>
> -#include <linux/irqdomain.h>
> -#include <linux/kernel.h>
> -#include <linux/init.h>
> -#include <linux/msi.h>
> -#include <linux/of_address.h>
> -#include <linux/of_irq.h>
> -#include <linux/of_pci.h>
> -#include <linux/of_platform.h>
>  #include <linux/pci.h>
> -#include <linux/phy/phy.h>
> -#include <linux/platform_device.h>
> -#include <linux/pm_runtime.h>
> -#include <linux/slab.h>
>  
> -#define PCIECAR			0x000010
> -#define PCIECCTLR		0x000018
> -#define  CONFIG_SEND_ENABLE	BIT(31)
> -#define  TYPE0			(0 << 8)
> -#define  TYPE1			BIT(8)
> -#define PCIECDR			0x000020
> -#define PCIEMSR			0x000028
> -#define PCIEINTXR		0x000400
> -#define PCIEPHYSR		0x0007f0
> -#define  PHYRDY			BIT(0)
> -#define PCIEMSITXR		0x000840
> +#include "pcie-rcar.h"
>  
> -/* Transfer control */
> -#define PCIETCTLR		0x02000
> -#define  DL_DOWN		BIT(3)
> -#define  CFINIT			BIT(0)
> -#define PCIETSTR		0x02004
> -#define  DATA_LINK_ACTIVE	BIT(0)
> -#define PCIEERRFR		0x02020
> -#define  UNSUPPORTED_REQUEST	BIT(4)
> -#define PCIEMSIFR		0x02044
> -#define PCIEMSIALR		0x02048
> -#define  MSIFE			BIT(0)
> -#define PCIEMSIAUR		0x0204c
> -#define PCIEMSIIER		0x02050
> -
> -/* root port address */
> -#define PCIEPRAR(x)		(0x02080 + ((x) * 0x4))
> -
> -/* local address reg & mask */
> -#define PCIELAR(x)		(0x02200 + ((x) * 0x20))
> -#define PCIELAMR(x)		(0x02208 + ((x) * 0x20))
> -#define  LAM_PREFETCH		BIT(3)
> -#define  LAM_64BIT		BIT(2)
> -#define  LAR_ENABLE		BIT(1)
> -
> -/* PCIe address reg & mask */
> -#define PCIEPALR(x)		(0x03400 + ((x) * 0x20))
> -#define PCIEPAUR(x)		(0x03404 + ((x) * 0x20))
> -#define PCIEPAMR(x)		(0x03408 + ((x) * 0x20))
> -#define PCIEPTCTLR(x)		(0x0340c + ((x) * 0x20))
> -#define  PAR_ENABLE		BIT(31)
> -#define  IO_SPACE		BIT(8)
> -
> -/* Configuration */
> -#define PCICONF(x)		(0x010000 + ((x) * 0x4))
> -#define PMCAP(x)		(0x010040 + ((x) * 0x4))
> -#define EXPCAP(x)		(0x010070 + ((x) * 0x4))
> -#define VCCAP(x)		(0x010100 + ((x) * 0x4))
> -
> -/* link layer */
> -#define IDSETR1			0x011004
> -#define TLCTLR			0x011048
> -#define MACSR			0x011054
> -#define  SPCHGFIN		BIT(4)
> -#define  SPCHGFAIL		BIT(6)
> -#define  SPCHGSUC		BIT(7)
> -#define  LINK_SPEED		(0xf << 16)
> -#define  LINK_SPEED_2_5GTS	(1 << 16)
> -#define  LINK_SPEED_5_0GTS	(2 << 16)
> -#define MACCTLR			0x011058
> -#define  MACCTLR_RESERVED	BIT(0)
> -#define  SPEED_CHANGE		BIT(24)
> -#define  SCRAMBLE_DISABLE	BIT(27)
> -#define PMSR			0x01105c
> -#define MACS2R			0x011078
> -#define MACCGSPSETR		0x011084
> -#define  SPCNGRSN		BIT(31)
> -
> -/* R-Car H1 PHY */
> -#define H1_PCIEPHYADRR		0x04000c
> -#define  WRITE_CMD		BIT(16)
> -#define  PHY_ACK		BIT(24)
> -#define  RATE_POS		12
> -#define  LANE_POS		8
> -#define  ADR_POS		0
> -#define H1_PCIEPHYDOUTR		0x040014
> -
> -/* R-Car Gen2 PHY */
> -#define GEN2_PCIEPHYADDR	0x780
> -#define GEN2_PCIEPHYDATA	0x784
> -#define GEN2_PCIEPHYCTRL	0x78c
> -
> -#define INT_PCI_MSI_NR		32
> -
> -#define RCONF(x)		(PCICONF(0) + (x))
> -#define RPMCAP(x)		(PMCAP(0) + (x))
> -#define REXPCAP(x)		(EXPCAP(0) + (x))
> -#define RVCCAP(x)		(VCCAP(0) + (x))
> -
> -#define PCIE_CONF_BUS(b)	(((b) & 0xff) << 24)
> -#define PCIE_CONF_DEV(d)	(((d) & 0x1f) << 19)
> -#define PCIE_CONF_FUNC(f)	(((f) & 0x7) << 16)
> -
> -#define RCAR_PCI_MAX_RESOURCES	4
> -#define MAX_NR_INBOUND_MAPS	6
> -
> -struct rcar_msi {
> -	DECLARE_BITMAP(used, INT_PCI_MSI_NR);
> -	struct irq_domain *domain;
> -	struct msi_controller chip;
> -	unsigned long pages;
> -	struct mutex lock;
> -	int irq1;
> -	int irq2;
> -};
> -
> -static inline struct rcar_msi *to_rcar_msi(struct msi_controller *chip)
> -{
> -	return container_of(chip, struct rcar_msi, chip);
> -}
> -
> -/* Structure representing the PCIe interface */
> -struct rcar_pcie {
> -	struct device		*dev;
> -	struct phy		*phy;
> -	void __iomem		*base;
> -	struct list_head	resources;
> -	int			root_bus_nr;
> -	struct clk		*bus_clk;
> -	struct			rcar_msi msi;
> -};
> -
> -static void rcar_pci_write_reg(struct rcar_pcie *pcie, u32 val,
> -			       unsigned int reg)
> +void rcar_pci_write_reg(void __iomem *base, u32 val, unsigned int reg)
>  {
> -	writel(val, pcie->base + reg);
> +	writel(val, base + reg);
>  }
>  
> -static u32 rcar_pci_read_reg(struct rcar_pcie *pcie, unsigned int reg)
> +u32 rcar_pci_read_reg(void __iomem *base, unsigned int reg)
>  {
> -	return readl(pcie->base + reg);
> +	return readl(base + reg);
>  }
>  
> -enum {
> -	RCAR_PCI_ACCESS_READ,
> -	RCAR_PCI_ACCESS_WRITE,
> -};
> -
> -static void rcar_rmw32(struct rcar_pcie *pcie, int where, u32 mask, u32 data)
> +void rcar_rmw32(void __iomem *base, int where, u32 mask, u32 data)
>  {
>  	unsigned int shift = BITS_PER_BYTE * (where & 3);
> -	u32 val = rcar_pci_read_reg(pcie, where & ~3);
> +	u32 val = rcar_pci_read_reg(base, where & ~3);
>  
>  	val &= ~(mask << shift);
>  	val |= data << shift;
> -	rcar_pci_write_reg(pcie, val, where & ~3);
> -}
> -
> -static u32 rcar_read_conf(struct rcar_pcie *pcie, int where)
> -{
> -	unsigned int shift = BITS_PER_BYTE * (where & 3);
> -	u32 val = rcar_pci_read_reg(pcie, where & ~3);
> -
> -	return val >> shift;
> +	rcar_pci_write_reg(base, val, where & ~3);
>  }
>  
> -/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
> -static int rcar_pcie_config_access(struct rcar_pcie *pcie,
> -		unsigned char access_type, struct pci_bus *bus,
> -		unsigned int devfn, int where, u32 *data)
> -{
> -	unsigned int dev, func, reg, index;
> -
> -	dev = PCI_SLOT(devfn);
> -	func = PCI_FUNC(devfn);
> -	reg = where & ~3;
> -	index = reg / 4;
> -
> -	/*
> -	 * While each channel has its own memory-mapped extended config
> -	 * space, it's generally only accessible when in endpoint mode.
> -	 * When in root complex mode, the controller is unable to target
> -	 * itself with either type 0 or type 1 accesses, and indeed, any
> -	 * controller initiated target transfer to its own config space
> -	 * result in a completer abort.
> -	 *
> -	 * Each channel effectively only supports a single device, but as
> -	 * the same channel <-> device access works for any PCI_SLOT()
> -	 * value, we cheat a bit here and bind the controller's config
> -	 * space to devfn 0 in order to enable self-enumeration. In this
> -	 * case the regular ECAR/ECDR path is sidelined and the mangled
> -	 * config access itself is initiated as an internal bus transaction.
> -	 */
> -	if (pci_is_root_bus(bus)) {
> -		if (dev != 0)
> -			return PCIBIOS_DEVICE_NOT_FOUND;
> -
> -		if (access_type == RCAR_PCI_ACCESS_READ) {
> -			*data = rcar_pci_read_reg(pcie, PCICONF(index));
> -		} else {
> -			/* Keep an eye out for changes to the root bus number */
> -			if (pci_is_root_bus(bus) && (reg == PCI_PRIMARY_BUS))
> -				pcie->root_bus_nr = *data & 0xff;
> -
> -			rcar_pci_write_reg(pcie, *data, PCICONF(index));
> -		}
> -
> -		return PCIBIOS_SUCCESSFUL;
> -	}
> -
> -	if (pcie->root_bus_nr < 0)
> -		return PCIBIOS_DEVICE_NOT_FOUND;
> -
> -	/* Clear errors */
> -	rcar_pci_write_reg(pcie, rcar_pci_read_reg(pcie, PCIEERRFR), PCIEERRFR);
> -
> -	/* Set the PIO address */
> -	rcar_pci_write_reg(pcie, PCIE_CONF_BUS(bus->number) |
> -		PCIE_CONF_DEV(dev) | PCIE_CONF_FUNC(func) | reg, PCIECAR);
> -
> -	/* Enable the configuration access */
> -	if (bus->parent->number == pcie->root_bus_nr)
> -		rcar_pci_write_reg(pcie, CONFIG_SEND_ENABLE | TYPE0, PCIECCTLR);
> -	else
> -		rcar_pci_write_reg(pcie, CONFIG_SEND_ENABLE | TYPE1, PCIECCTLR);
> -
> -	/* Check for errors */
> -	if (rcar_pci_read_reg(pcie, PCIEERRFR) & UNSUPPORTED_REQUEST)
> -		return PCIBIOS_DEVICE_NOT_FOUND;
> -
> -	/* Check for master and target aborts */
> -	if (rcar_read_conf(pcie, RCONF(PCI_STATUS)) &
> -		(PCI_STATUS_REC_MASTER_ABORT | PCI_STATUS_REC_TARGET_ABORT))
> -		return PCIBIOS_DEVICE_NOT_FOUND;
> -
> -	if (access_type == RCAR_PCI_ACCESS_READ)
> -		*data = rcar_pci_read_reg(pcie, PCIECDR);
> -	else
> -		rcar_pci_write_reg(pcie, *data, PCIECDR);
> -
> -	/* Disable the configuration access */
> -	rcar_pci_write_reg(pcie, 0, PCIECCTLR);
> -
> -	return PCIBIOS_SUCCESSFUL;
> -}
> -
> -static int rcar_pcie_read_conf(struct pci_bus *bus, unsigned int devfn,
> -			       int where, int size, u32 *val)
> -{
> -	struct rcar_pcie *pcie = bus->sysdata;
> -	int ret;
> -
> -	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
> -				      bus, devfn, where, val);
> -	if (ret != PCIBIOS_SUCCESSFUL) {
> -		*val = 0xffffffff;
> -		return ret;
> -	}
> -
> -	if (size == 1)
> -		*val = (*val >> (BITS_PER_BYTE * (where & 3))) & 0xff;
> -	else if (size == 2)
> -		*val = (*val >> (BITS_PER_BYTE * (where & 2))) & 0xffff;
> -
> -	dev_dbg(&bus->dev, "pcie-config-read: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
> -		bus->number, devfn, where, size, *val);
> -
> -	return ret;
> -}
> -
> -/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
> -static int rcar_pcie_write_conf(struct pci_bus *bus, unsigned int devfn,
> -				int where, int size, u32 val)
> -{
> -	struct rcar_pcie *pcie = bus->sysdata;
> -	unsigned int shift;
> -	u32 data;
> -	int ret;
> -
> -	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
> -				      bus, devfn, where, &data);
> -	if (ret != PCIBIOS_SUCCESSFUL)
> -		return ret;
> -
> -	dev_dbg(&bus->dev, "pcie-config-write: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
> -		bus->number, devfn, where, size, val);
> -
> -	if (size == 1) {
> -		shift = BITS_PER_BYTE * (where & 3);
> -		data &= ~(0xff << shift);
> -		data |= ((val & 0xff) << shift);
> -	} else if (size == 2) {
> -		shift = BITS_PER_BYTE * (where & 2);
> -		data &= ~(0xffff << shift);
> -		data |= ((val & 0xffff) << shift);
> -	} else
> -		data = val;
> -
> -	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_WRITE,
> -				      bus, devfn, where, &data);
> -
> -	return ret;
> -}
> -
> -static struct pci_ops rcar_pcie_ops = {
> -	.read	= rcar_pcie_read_conf,
> -	.write	= rcar_pcie_write_conf,
> -};
> -
> -static void rcar_pcie_setup_window(int win, struct rcar_pcie *pcie,
> -				   struct resource *res)
> +void rcar_pcie_set_outbound(int win, void __iomem *base,
> +			    struct resource *res, bool host)
>  {
>  	/* Setup PCIe address space mappings for each resource */
> -	resource_size_t size;
> -	resource_size_t res_start;
> -	u32 mask;
> -
> -	rcar_pci_write_reg(pcie, 0x00000000, PCIEPTCTLR(win));
> +	resource_size_t size = 0;
> +	resource_size_t res_start = 0;
> +	u32 mask = 0x0;
>  
> +	rcar_pci_write_reg(base, mask, PCIEPTCTLR(win));
>  	/*
>  	 * The PAMR mask is calculated in units of 128Bytes, which
>  	 * keeps things pretty simple.
>  	 */
>  	size = resource_size(res);
> -	mask = (roundup_pow_of_two(size) / SZ_128) - 1;
> -	rcar_pci_write_reg(pcie, mask << 7, PCIEPAMR(win));
> +	if (size > 128)
> +		mask = (roundup_pow_of_two(size) / SZ_128) - 1;
> +	rcar_pci_write_reg(base, mask << 7, PCIEPAMR(win));
>  
> -	if (res->flags & IORESOURCE_IO)
> -		res_start = pci_pio_to_address(res->start);
> -	else
> +	if (!host) {
>  		res_start = res->start;
> +	} else {
> +		if (res->flags & IORESOURCE_IO)
> +			res_start = pci_pio_to_address(res->start);
> +		else
> +			res_start = res->start;
> +	}
>  
> -	rcar_pci_write_reg(pcie, upper_32_bits(res_start), PCIEPAUR(win));
> -	rcar_pci_write_reg(pcie, lower_32_bits(res_start) & ~0x7F,
> +	rcar_pci_write_reg(base, upper_32_bits(res_start), PCIEPAUR(win));
> +	rcar_pci_write_reg(base, lower_32_bits(res_start) & ~0x7F,
>  			   PCIEPALR(win));
>  
> -	/* First resource is for IO */
> -	mask = PAR_ENABLE;
> +	mask = 0x0;
> +	if (res->start)
> +		mask = PAR_ENABLE;
> +
>  	if (res->flags & IORESOURCE_IO)
>  		mask |= IO_SPACE;
>  
> -	rcar_pci_write_reg(pcie, mask, PCIEPTCTLR(win));
> -}
> -
> -static int rcar_pcie_setup(struct list_head *resource, struct rcar_pcie *pci)
> -{
> -	struct resource_entry *win;
> -	int i = 0;
> -
> -	/* Setup PCI resources */
> -	resource_list_for_each_entry(win, &pci->resources) {
> -		struct resource *res = win->res;
> -
> -		if (!res->flags)
> -			continue;
> -
> -		switch (resource_type(res)) {
> -		case IORESOURCE_IO:
> -		case IORESOURCE_MEM:
> -			rcar_pcie_setup_window(i, pci, res);
> -			i++;
> -			break;
> -		case IORESOURCE_BUS:
> -			pci->root_bus_nr = res->start;
> -			break;
> -		default:
> -			continue;
> -		}
> -
> -		pci_add_resource(resource, res);
> -	}
> -
> -	return 1;
> -}
> -
> -static void rcar_pcie_force_speedup(struct rcar_pcie *pcie)
> -{
> -	struct device *dev = pcie->dev;
> -	unsigned int timeout = 1000;
> -	u32 macsr;
> -
> -	if ((rcar_pci_read_reg(pcie, MACS2R) & LINK_SPEED) != LINK_SPEED_5_0GTS)
> -		return;
> -
> -	if (rcar_pci_read_reg(pcie, MACCTLR) & SPEED_CHANGE) {
> -		dev_err(dev, "Speed change already in progress\n");
> -		return;
> -	}
> -
> -	macsr = rcar_pci_read_reg(pcie, MACSR);
> -	if ((macsr & LINK_SPEED) == LINK_SPEED_5_0GTS)
> -		goto done;
> -
> -	/* Set target link speed to 5.0 GT/s */
> -	rcar_rmw32(pcie, EXPCAP(12), PCI_EXP_LNKSTA_CLS,
> -		   PCI_EXP_LNKSTA_CLS_5_0GB);
> -
> -	/* Set speed change reason as intentional factor */
> -	rcar_rmw32(pcie, MACCGSPSETR, SPCNGRSN, 0);
> -
> -	/* Clear SPCHGFIN, SPCHGSUC, and SPCHGFAIL */
> -	if (macsr & (SPCHGFIN | SPCHGSUC | SPCHGFAIL))
> -		rcar_pci_write_reg(pcie, macsr, MACSR);
> -
> -	/* Start link speed change */
> -	rcar_rmw32(pcie, MACCTLR, SPEED_CHANGE, SPEED_CHANGE);
> -
> -	while (timeout--) {
> -		macsr = rcar_pci_read_reg(pcie, MACSR);
> -		if (macsr & SPCHGFIN) {
> -			/* Clear the interrupt bits */
> -			rcar_pci_write_reg(pcie, macsr, MACSR);
> -
> -			if (macsr & SPCHGFAIL)
> -				dev_err(dev, "Speed change failed\n");
> -
> -			goto done;
> -		}
> -
> -		msleep(1);
> -	}
> -
> -	dev_err(dev, "Speed change timed out\n");
> -
> -done:
> -	dev_info(dev, "Current link speed is %s GT/s\n",
> -		 (macsr & LINK_SPEED) == LINK_SPEED_5_0GTS ? "5" : "2.5");
> -}
> -
> -static int rcar_pcie_enable(struct rcar_pcie *pcie)
> -{
> -	struct device *dev = pcie->dev;
> -	struct pci_host_bridge *bridge = pci_host_bridge_from_priv(pcie);
> -	struct pci_bus *bus, *child;
> -	int ret;
> -
> -	/* Try setting 5 GT/s link speed */
> -	rcar_pcie_force_speedup(pcie);
> -
> -	rcar_pcie_setup(&bridge->windows, pcie);
> -
> -	pci_add_flags(PCI_REASSIGN_ALL_BUS);
> -
> -	bridge->dev.parent = dev;
> -	bridge->sysdata = pcie;
> -	bridge->busnr = pcie->root_bus_nr;
> -	bridge->ops = &rcar_pcie_ops;
> -	bridge->map_irq = of_irq_parse_and_map_pci;
> -	bridge->swizzle_irq = pci_common_swizzle;
> -	if (IS_ENABLED(CONFIG_PCI_MSI))
> -		bridge->msi = &pcie->msi.chip;
> -
> -	ret = pci_scan_root_bus_bridge(bridge);
> -	if (ret < 0)
> -		return ret;
> -
> -	bus = bridge->bus;
> -
> -	pci_bus_size_bridges(bus);
> -	pci_bus_assign_resources(bus);
> -
> -	list_for_each_entry(child, &bus->children, node)
> -		pcie_bus_configure_settings(child);
> -
> -	pci_bus_add_devices(bus);
> -
> -	return 0;
> -}
> -
> -static int phy_wait_for_ack(struct rcar_pcie *pcie)
> -{
> -	struct device *dev = pcie->dev;
> -	unsigned int timeout = 100;
> -
> -	while (timeout--) {
> -		if (rcar_pci_read_reg(pcie, H1_PCIEPHYADRR) & PHY_ACK)
> -			return 0;
> -
> -		udelay(100);
> -	}
> -
> -	dev_err(dev, "Access to PCIe phy timed out\n");
> -
> -	return -ETIMEDOUT;
> -}
> -
> -static void phy_write_reg(struct rcar_pcie *pcie,
> -			  unsigned int rate, u32 addr,
> -			  unsigned int lane, u32 data)
> -{
> -	u32 phyaddr;
> -
> -	phyaddr = WRITE_CMD |
> -		((rate & 1) << RATE_POS) |
> -		((lane & 0xf) << LANE_POS) |
> -		((addr & 0xff) << ADR_POS);
> -
> -	/* Set write data */
> -	rcar_pci_write_reg(pcie, data, H1_PCIEPHYDOUTR);
> -	rcar_pci_write_reg(pcie, phyaddr, H1_PCIEPHYADRR);
> -
> -	/* Ignore errors as they will be dealt with if the data link is down */
> -	phy_wait_for_ack(pcie);
> -
> -	/* Clear command */
> -	rcar_pci_write_reg(pcie, 0, H1_PCIEPHYDOUTR);
> -	rcar_pci_write_reg(pcie, 0, H1_PCIEPHYADRR);
> -
> -	/* Ignore errors as they will be dealt with if the data link is down */
> -	phy_wait_for_ack(pcie);
> +	rcar_pci_write_reg(base, mask, PCIEPTCTLR(win));
> +	/* flush modifications */
> +	wmb();
>  }
>  
> -static int rcar_pcie_wait_for_phyrdy(struct rcar_pcie *pcie)
> +int rcar_pcie_wait_for_phyrdy(void __iomem *base)
>  {
>  	unsigned int timeout = 10;
>  
>  	while (timeout--) {
> -		if (rcar_pci_read_reg(pcie, PCIEPHYSR) & PHYRDY)
> +		if (rcar_pci_read_reg(base, PCIEPHYSR) & PHYRDY)
>  			return 0;
>  
>  		msleep(5);
> @@ -547,12 +93,12 @@ static int rcar_pcie_wait_for_phyrdy(struct rcar_pcie *pcie)
>  	return -ETIMEDOUT;
>  }
>  
> -static int rcar_pcie_wait_for_dl(struct rcar_pcie *pcie)
> +int rcar_pcie_wait_for_dl(void __iomem *base)
>  {
>  	unsigned int timeout = 10000;
>  
>  	while (timeout--) {
> -		if ((rcar_pci_read_reg(pcie, PCIETSTR) & DATA_LINK_ACTIVE))
> +		if ((rcar_pci_read_reg(base, PCIETSTR) & DATA_LINK_ACTIVE))
>  			return 0;
>  
>  		udelay(5);
> @@ -562,698 +108,27 @@ static int rcar_pcie_wait_for_dl(struct rcar_pcie *pcie)
>  	return -ETIMEDOUT;
>  }
>  
> -static int rcar_pcie_hw_init(struct rcar_pcie *pcie)
> -{
> -	int err;
> -
> -	/* Begin initialization */
> -	rcar_pci_write_reg(pcie, 0, PCIETCTLR);
> -
> -	/* Set mode */
> -	rcar_pci_write_reg(pcie, 1, PCIEMSR);
> -
> -	err = rcar_pcie_wait_for_phyrdy(pcie);
> -	if (err)
> -		return err;
> -
> -	/*
> -	 * Initial header for port config space is type 1, set the device
> -	 * class to match. Hardware takes care of propagating the IDSETR
> -	 * settings, so there is no need to bother with a quirk.
> -	 */
> -	rcar_pci_write_reg(pcie, PCI_CLASS_BRIDGE_PCI << 16, IDSETR1);
> -
> -	/*
> -	 * Setup Secondary Bus Number & Subordinate Bus Number, even though
> -	 * they aren't used, to avoid bridge being detected as broken.
> -	 */
> -	rcar_rmw32(pcie, RCONF(PCI_SECONDARY_BUS), 0xff, 1);
> -	rcar_rmw32(pcie, RCONF(PCI_SUBORDINATE_BUS), 0xff, 1);
> -
> -	/* Initialize default capabilities. */
> -	rcar_rmw32(pcie, REXPCAP(0), 0xff, PCI_CAP_ID_EXP);
> -	rcar_rmw32(pcie, REXPCAP(PCI_EXP_FLAGS),
> -		PCI_EXP_FLAGS_TYPE, PCI_EXP_TYPE_ROOT_PORT << 4);
> -	rcar_rmw32(pcie, RCONF(PCI_HEADER_TYPE), 0x7f,
> -		PCI_HEADER_TYPE_BRIDGE);
> -
> -	/* Enable data link layer active state reporting */
> -	rcar_rmw32(pcie, REXPCAP(PCI_EXP_LNKCAP), PCI_EXP_LNKCAP_DLLLARC,
> -		PCI_EXP_LNKCAP_DLLLARC);
> -
> -	/* Write out the physical slot number = 0 */
> -	rcar_rmw32(pcie, REXPCAP(PCI_EXP_SLTCAP), PCI_EXP_SLTCAP_PSN, 0);
> -
> -	/* Set the completion timer timeout to the maximum 50ms. */
> -	rcar_rmw32(pcie, TLCTLR + 1, 0x3f, 50);
> -
> -	/* Terminate list of capabilities (Next Capability Offset=0) */
> -	rcar_rmw32(pcie, RVCCAP(0), 0xfff00000, 0);
> -
> -	/* Enable MSI */
> -	if (IS_ENABLED(CONFIG_PCI_MSI))
> -		rcar_pci_write_reg(pcie, 0x801f0000, PCIEMSITXR);
> -
> -	rcar_rmw32(pcie, MACCTLR, MACCTLR_RESERVED, 0);
> -
> -	/* Finish initialization - establish a PCI Express link */
> -	rcar_pci_write_reg(pcie, CFINIT, PCIETCTLR);
> -
> -	/* This will timeout if we don't have a link. */
> -	err = rcar_pcie_wait_for_dl(pcie);
> -	if (err)
> -		return err;
> -
> -	/* Enable INTx interrupts */
> -	rcar_rmw32(pcie, PCIEINTXR, 0, 0xF << 8);
> -
> -	wmb();
> -
> -	return 0;
> -}
> -
> -static int rcar_pcie_phy_init_h1(struct rcar_pcie *pcie)
> -{
> -	/* Initialize the phy */
> -	phy_write_reg(pcie, 0, 0x42, 0x1, 0x0EC34191);
> -	phy_write_reg(pcie, 1, 0x42, 0x1, 0x0EC34180);
> -	phy_write_reg(pcie, 0, 0x43, 0x1, 0x00210188);
> -	phy_write_reg(pcie, 1, 0x43, 0x1, 0x00210188);
> -	phy_write_reg(pcie, 0, 0x44, 0x1, 0x015C0014);
> -	phy_write_reg(pcie, 1, 0x44, 0x1, 0x015C0014);
> -	phy_write_reg(pcie, 1, 0x4C, 0x1, 0x786174A0);
> -	phy_write_reg(pcie, 1, 0x4D, 0x1, 0x048000BB);
> -	phy_write_reg(pcie, 0, 0x51, 0x1, 0x079EC062);
> -	phy_write_reg(pcie, 0, 0x52, 0x1, 0x20000000);
> -	phy_write_reg(pcie, 1, 0x52, 0x1, 0x20000000);
> -	phy_write_reg(pcie, 1, 0x56, 0x1, 0x00003806);
> -
> -	phy_write_reg(pcie, 0, 0x60, 0x1, 0x004B03A5);
> -	phy_write_reg(pcie, 0, 0x64, 0x1, 0x3F0F1F0F);
> -	phy_write_reg(pcie, 0, 0x66, 0x1, 0x00008000);
> -
> -	return 0;
> -}
> -
> -static int rcar_pcie_phy_init_gen2(struct rcar_pcie *pcie)
> -{
> -	/*
> -	 * These settings come from the R-Car Series, 2nd Generation User's
> -	 * Manual, section 50.3.1 (2) Initialization of the physical layer.
> -	 */
> -	rcar_pci_write_reg(pcie, 0x000f0030, GEN2_PCIEPHYADDR);
> -	rcar_pci_write_reg(pcie, 0x00381203, GEN2_PCIEPHYDATA);
> -	rcar_pci_write_reg(pcie, 0x00000001, GEN2_PCIEPHYCTRL);
> -	rcar_pci_write_reg(pcie, 0x00000006, GEN2_PCIEPHYCTRL);
> -
> -	rcar_pci_write_reg(pcie, 0x000f0054, GEN2_PCIEPHYADDR);
> -	/* The following value is for DC connection, no termination resistor */
> -	rcar_pci_write_reg(pcie, 0x13802007, GEN2_PCIEPHYDATA);
> -	rcar_pci_write_reg(pcie, 0x00000001, GEN2_PCIEPHYCTRL);
> -	rcar_pci_write_reg(pcie, 0x00000006, GEN2_PCIEPHYCTRL);
> -
> -	return 0;
> -}
> -
> -static int rcar_pcie_phy_init_gen3(struct rcar_pcie *pcie)
> -{
> -	int err;
> -
> -	err = phy_init(pcie->phy);
> -	if (err)
> -		return err;
> -
> -	err = phy_power_on(pcie->phy);
> -	if (err)
> -		phy_exit(pcie->phy);
> -
> -	return err;
> -}
> -
> -static int rcar_msi_alloc(struct rcar_msi *chip)
> -{
> -	int msi;
> -
> -	mutex_lock(&chip->lock);
> -
> -	msi = find_first_zero_bit(chip->used, INT_PCI_MSI_NR);
> -	if (msi < INT_PCI_MSI_NR)
> -		set_bit(msi, chip->used);
> -	else
> -		msi = -ENOSPC;
> -
> -	mutex_unlock(&chip->lock);
> -
> -	return msi;
> -}
> -
> -static int rcar_msi_alloc_region(struct rcar_msi *chip, int no_irqs)
> -{
> -	int msi;
> -
> -	mutex_lock(&chip->lock);
> -	msi = bitmap_find_free_region(chip->used, INT_PCI_MSI_NR,
> -				      order_base_2(no_irqs));
> -	mutex_unlock(&chip->lock);
> -
> -	return msi;
> -}
> -
> -static void rcar_msi_free(struct rcar_msi *chip, unsigned long irq)
> -{
> -	mutex_lock(&chip->lock);
> -	clear_bit(irq, chip->used);
> -	mutex_unlock(&chip->lock);
> -}
> -
> -static irqreturn_t rcar_pcie_msi_irq(int irq, void *data)
> -{
> -	struct rcar_pcie *pcie = data;
> -	struct rcar_msi *msi = &pcie->msi;
> -	struct device *dev = pcie->dev;
> -	unsigned long reg;
> -
> -	reg = rcar_pci_read_reg(pcie, PCIEMSIFR);
> -
> -	/* MSI & INTx share an interrupt - we only handle MSI here */
> -	if (!reg)
> -		return IRQ_NONE;
> -
> -	while (reg) {
> -		unsigned int index = find_first_bit(&reg, 32);
> -		unsigned int msi_irq;
> -
> -		/* clear the interrupt */
> -		rcar_pci_write_reg(pcie, 1 << index, PCIEMSIFR);
> -
> -		msi_irq = irq_find_mapping(msi->domain, index);
> -		if (msi_irq) {
> -			if (test_bit(index, msi->used))
> -				generic_handle_irq(msi_irq);
> -			else
> -				dev_info(dev, "unhandled MSI\n");
> -		} else {
> -			/* Unknown MSI, just clear it */
> -			dev_dbg(dev, "unexpected MSI\n");
> -		}
> -
> -		/* see if there's any more pending in this vector */
> -		reg = rcar_pci_read_reg(pcie, PCIEMSIFR);
> -	}
> -
> -	return IRQ_HANDLED;
> -}
> -
> -static int rcar_msi_setup_irq(struct msi_controller *chip, struct pci_dev *pdev,
> -			      struct msi_desc *desc)
> -{
> -	struct rcar_msi *msi = to_rcar_msi(chip);
> -	struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
> -	struct msi_msg msg;
> -	unsigned int irq;
> -	int hwirq;
> -
> -	hwirq = rcar_msi_alloc(msi);
> -	if (hwirq < 0)
> -		return hwirq;
> -
> -	irq = irq_find_mapping(msi->domain, hwirq);
> -	if (!irq) {
> -		rcar_msi_free(msi, hwirq);
> -		return -EINVAL;
> -	}
> -
> -	irq_set_msi_desc(irq, desc);
> -
> -	msg.address_lo = rcar_pci_read_reg(pcie, PCIEMSIALR) & ~MSIFE;
> -	msg.address_hi = rcar_pci_read_reg(pcie, PCIEMSIAUR);
> -	msg.data = hwirq;
> -
> -	pci_write_msi_msg(irq, &msg);
> -
> -	return 0;
> -}
> -
> -static int rcar_msi_setup_irqs(struct msi_controller *chip,
> -			       struct pci_dev *pdev, int nvec, int type)
> -{
> -	struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
> -	struct rcar_msi *msi = to_rcar_msi(chip);
> -	struct msi_desc *desc;
> -	struct msi_msg msg;
> -	unsigned int irq;
> -	int hwirq;
> -	int i;
> -
> -	/* MSI-X interrupts are not supported */
> -	if (type == PCI_CAP_ID_MSIX)
> -		return -EINVAL;
> -
> -	WARN_ON(!list_is_singular(&pdev->dev.msi_list));
> -	desc = list_entry(pdev->dev.msi_list.next, struct msi_desc, list);
> -
> -	hwirq = rcar_msi_alloc_region(msi, nvec);
> -	if (hwirq < 0)
> -		return -ENOSPC;
> -
> -	irq = irq_find_mapping(msi->domain, hwirq);
> -	if (!irq)
> -		return -ENOSPC;
> -
> -	for (i = 0; i < nvec; i++) {
> -		/*
> -		 * irq_create_mapping() called from rcar_pcie_probe() pre-
> -		 * allocates descs,  so there is no need to allocate descs here.
> -		 * We can therefore assume that if irq_find_mapping() above
> -		 * returns non-zero, then the descs are also successfully
> -		 * allocated.
> -		 */
> -		if (irq_set_msi_desc_off(irq, i, desc)) {
> -			/* TODO: clear */
> -			return -EINVAL;
> -		}
> -	}
> -
> -	desc->nvec_used = nvec;
> -	desc->msi_attrib.multiple = order_base_2(nvec);
> -
> -	msg.address_lo = rcar_pci_read_reg(pcie, PCIEMSIALR) & ~MSIFE;
> -	msg.address_hi = rcar_pci_read_reg(pcie, PCIEMSIAUR);
> -	msg.data = hwirq;
> -
> -	pci_write_msi_msg(irq, &msg);
> -
> -	return 0;
> -}
> -
> -static void rcar_msi_teardown_irq(struct msi_controller *chip, unsigned int irq)
> -{
> -	struct rcar_msi *msi = to_rcar_msi(chip);
> -	struct irq_data *d = irq_get_irq_data(irq);
> -
> -	rcar_msi_free(msi, d->hwirq);
> -}
> -
> -static struct irq_chip rcar_msi_irq_chip = {
> -	.name = "R-Car PCIe MSI",
> -	.irq_enable = pci_msi_unmask_irq,
> -	.irq_disable = pci_msi_mask_irq,
> -	.irq_mask = pci_msi_mask_irq,
> -	.irq_unmask = pci_msi_unmask_irq,
> -};
> -
> -static int rcar_msi_map(struct irq_domain *domain, unsigned int irq,
> -			irq_hw_number_t hwirq)
> -{
> -	irq_set_chip_and_handler(irq, &rcar_msi_irq_chip, handle_simple_irq);
> -	irq_set_chip_data(irq, domain->host_data);
> -
> -	return 0;
> -}
> -
> -static const struct irq_domain_ops msi_domain_ops = {
> -	.map = rcar_msi_map,
> -};
> -
> -static void rcar_pcie_unmap_msi(struct rcar_pcie *pcie)
> -{
> -	struct rcar_msi *msi = &pcie->msi;
> -	int i, irq;
> -
> -	for (i = 0; i < INT_PCI_MSI_NR; i++) {
> -		irq = irq_find_mapping(msi->domain, i);
> -		if (irq > 0)
> -			irq_dispose_mapping(irq);
> -	}
> -
> -	irq_domain_remove(msi->domain);
> -}
> -
> -static int rcar_pcie_enable_msi(struct rcar_pcie *pcie)
> -{
> -	struct device *dev = pcie->dev;
> -	struct rcar_msi *msi = &pcie->msi;
> -	phys_addr_t base;
> -	int err, i;
> -
> -	mutex_init(&msi->lock);
> -
> -	msi->chip.dev = dev;
> -	msi->chip.setup_irq = rcar_msi_setup_irq;
> -	msi->chip.setup_irqs = rcar_msi_setup_irqs;
> -	msi->chip.teardown_irq = rcar_msi_teardown_irq;
> -
> -	msi->domain = irq_domain_add_linear(dev->of_node, INT_PCI_MSI_NR,
> -					    &msi_domain_ops, &msi->chip);
> -	if (!msi->domain) {
> -		dev_err(dev, "failed to create IRQ domain\n");
> -		return -ENOMEM;
> -	}
> -
> -	for (i = 0; i < INT_PCI_MSI_NR; i++)
> -		irq_create_mapping(msi->domain, i);
> -
> -	/* Two irqs are for MSI, but they are also used for non-MSI irqs */
> -	err = devm_request_irq(dev, msi->irq1, rcar_pcie_msi_irq,
> -			       IRQF_SHARED | IRQF_NO_THREAD,
> -			       rcar_msi_irq_chip.name, pcie);
> -	if (err < 0) {
> -		dev_err(dev, "failed to request IRQ: %d\n", err);
> -		goto err;
> -	}
> -
> -	err = devm_request_irq(dev, msi->irq2, rcar_pcie_msi_irq,
> -			       IRQF_SHARED | IRQF_NO_THREAD,
> -			       rcar_msi_irq_chip.name, pcie);
> -	if (err < 0) {
> -		dev_err(dev, "failed to request IRQ: %d\n", err);
> -		goto err;
> -	}
> -
> -	/* setup MSI data target */
> -	msi->pages = __get_free_pages(GFP_KERNEL, 0);
> -	if (!msi->pages) {
> -		err = -ENOMEM;
> -		goto err;
> -	}
> -	base = virt_to_phys((void *)msi->pages);
> -
> -	rcar_pci_write_reg(pcie, lower_32_bits(base) | MSIFE, PCIEMSIALR);
> -	rcar_pci_write_reg(pcie, upper_32_bits(base), PCIEMSIAUR);
> -
> -	/* enable all MSI interrupts */
> -	rcar_pci_write_reg(pcie, 0xffffffff, PCIEMSIIER);
> -
> -	return 0;
> -
> -err:
> -	rcar_pcie_unmap_msi(pcie);
> -	return err;
> -}
> -
> -static void rcar_pcie_teardown_msi(struct rcar_pcie *pcie)
> -{
> -	struct rcar_msi *msi = &pcie->msi;
> -
> -	/* Disable all MSI interrupts */
> -	rcar_pci_write_reg(pcie, 0, PCIEMSIIER);
> -
> -	/* Disable address decoding of the MSI interrupt, MSIFE */
> -	rcar_pci_write_reg(pcie, 0, PCIEMSIALR);
> -
> -	free_pages(msi->pages, 0);
> -
> -	rcar_pcie_unmap_msi(pcie);
> -}
> -
> -static int rcar_pcie_get_resources(struct rcar_pcie *pcie)
> -{
> -	struct device *dev = pcie->dev;
> -	struct resource res;
> -	int err, i;
> -
> -	pcie->phy = devm_phy_optional_get(dev, "pcie");
> -	if (IS_ERR(pcie->phy))
> -		return PTR_ERR(pcie->phy);
> -
> -	err = of_address_to_resource(dev->of_node, 0, &res);
> -	if (err)
> -		return err;
> -
> -	pcie->base = devm_ioremap_resource(dev, &res);
> -	if (IS_ERR(pcie->base))
> -		return PTR_ERR(pcie->base);
> -
> -	pcie->bus_clk = devm_clk_get(dev, "pcie_bus");
> -	if (IS_ERR(pcie->bus_clk)) {
> -		dev_err(dev, "cannot get pcie bus clock\n");
> -		return PTR_ERR(pcie->bus_clk);
> -	}
> -
> -	i = irq_of_parse_and_map(dev->of_node, 0);
> -	if (!i) {
> -		dev_err(dev, "cannot get platform resources for msi interrupt\n");
> -		err = -ENOENT;
> -		goto err_irq1;
> -	}
> -	pcie->msi.irq1 = i;
> -
> -	i = irq_of_parse_and_map(dev->of_node, 1);
> -	if (!i) {
> -		dev_err(dev, "cannot get platform resources for msi interrupt\n");
> -		err = -ENOENT;
> -		goto err_irq2;
> -	}
> -	pcie->msi.irq2 = i;
> -
> -	return 0;
> -
> -err_irq2:
> -	irq_dispose_mapping(pcie->msi.irq1);
> -err_irq1:
> -	return err;
> -}
> -
> -static int rcar_pcie_inbound_ranges(struct rcar_pcie *pcie,
> -				    struct of_pci_range *range,
> -				    int *index)
> +void rcar_pcie_set_inbound(void __iomem *base,
> +			   u64 cpu_addr, u64 pci_addr,
> +			   u64 mask, int idx, bool host)
>  {
> -	u64 restype = range->flags;
> -	u64 cpu_addr = range->cpu_addr;
> -	u64 cpu_end = range->cpu_addr + range->size;
> -	u64 pci_addr = range->pci_addr;
> -	u32 flags = LAM_64BIT | LAR_ENABLE;
> -	u64 mask;
> -	u64 size;
> -	int idx = *index;
> -
> -	if (restype & IORESOURCE_PREFETCH)
> -		flags |= LAM_PREFETCH;
> -
>  	/*
> -	 * If the size of the range is larger than the alignment of the start
> -	 * address, we have to use multiple entries to perform the mapping.
> -	 */
> -	if (cpu_addr > 0) {
> -		unsigned long nr_zeros = __ffs64(cpu_addr);
> -		u64 alignment = 1ULL << nr_zeros;
> -
> -		size = min(range->size, alignment);
> -	} else {
> -		size = range->size;
> -	}
> -	/* Hardware supports max 4GiB inbound region */
> -	size = min(size, 1ULL << 32);
> -
> -	mask = roundup_pow_of_two(size) - 1;
> -	mask &= ~0xf;
> -
> -	while (cpu_addr < cpu_end) {
> -		/*
> -		 * Set up 64-bit inbound regions as the range parser doesn't
> -		 * distinguish between 32 and 64-bit types.
> -		 */
> -		rcar_pci_write_reg(pcie, lower_32_bits(pci_addr),
> +	* Set up 64-bit inbound regions as the range parser doesn't
> +	* distinguish between 32 and 64-bit types.
> +	*/
> +	if (host) {
> +		rcar_pci_write_reg(base, lower_32_bits(pci_addr),
>  				   PCIEPRAR(idx));
> -		rcar_pci_write_reg(pcie, lower_32_bits(cpu_addr), PCIELAR(idx));
> -		rcar_pci_write_reg(pcie, lower_32_bits(mask) | flags,
> -				   PCIELAMR(idx));
> -
> -		rcar_pci_write_reg(pcie, upper_32_bits(pci_addr),
> +		rcar_pci_write_reg(base, upper_32_bits(pci_addr),
>  				   PCIEPRAR(idx + 1));
> -		rcar_pci_write_reg(pcie, upper_32_bits(cpu_addr),
> -				   PCIELAR(idx + 1));
> -		rcar_pci_write_reg(pcie, 0, PCIELAMR(idx + 1));
> -
> -		pci_addr += size;
> -		cpu_addr += size;
> -		idx += 2;
> -
> -		if (idx > MAX_NR_INBOUND_MAPS) {
> -			dev_err(pcie->dev, "Failed to map inbound regions!\n");
> -			return -EINVAL;
> -		}
>  	}
> -	*index = idx;
> -
> -	return 0;
> -}
>  
> -static int rcar_pcie_parse_map_dma_ranges(struct rcar_pcie *pcie,
> -					  struct device_node *np)
> -{
> -	struct of_pci_range range;
> -	struct of_pci_range_parser parser;
> -	int index = 0;
> -	int err;
> -
> -	if (of_pci_dma_range_parser_init(&parser, np))
> -		return -EINVAL;
> -
> -	/* Get the dma-ranges from DT */
> -	for_each_of_pci_range(&parser, &range) {
> -		u64 end = range.cpu_addr + range.size - 1;
> +	rcar_pci_write_reg(base, lower_32_bits(cpu_addr), PCIELAR(idx));
> +	rcar_pci_write_reg(base, lower_32_bits(mask), PCIELAMR(idx));
>  
> -		dev_dbg(pcie->dev, "0x%08x 0x%016llx..0x%016llx -> 0x%016llx\n",
> -			range.flags, range.cpu_addr, end, range.pci_addr);
> -
> -		err = rcar_pcie_inbound_ranges(pcie, &range, &index);
> -		if (err)
> -			return err;
> -	}
> +	rcar_pci_write_reg(base, upper_32_bits(cpu_addr), PCIELAR(idx + 1));
> +	rcar_pci_write_reg(base, 0, PCIELAMR(idx + 1));
>  
> -	return 0;
> -}
> -
> -static const struct of_device_id rcar_pcie_of_match[] = {
> -	{ .compatible = "renesas,pcie-r8a7779",
> -	  .data = rcar_pcie_phy_init_h1 },
> -	{ .compatible = "renesas,pcie-r8a7790",
> -	  .data = rcar_pcie_phy_init_gen2 },
> -	{ .compatible = "renesas,pcie-r8a7791",
> -	  .data = rcar_pcie_phy_init_gen2 },
> -	{ .compatible = "renesas,pcie-rcar-gen2",
> -	  .data = rcar_pcie_phy_init_gen2 },
> -	{ .compatible = "renesas,pcie-r8a7795",
> -	  .data = rcar_pcie_phy_init_gen3 },
> -	{ .compatible = "renesas,pcie-rcar-gen3",
> -	  .data = rcar_pcie_phy_init_gen3 },
> -	{},
> -};
> -
> -static int rcar_pcie_probe(struct platform_device *pdev)
> -{
> -	struct device *dev = &pdev->dev;
> -	struct rcar_pcie *pcie;
> -	u32 data;
> -	int err;
> -	int (*phy_init_fn)(struct rcar_pcie *);
> -	struct pci_host_bridge *bridge;
> -
> -	bridge = pci_alloc_host_bridge(sizeof(*pcie));
> -	if (!bridge)
> -		return -ENOMEM;
> -
> -	pcie = pci_host_bridge_priv(bridge);
> -
> -	pcie->dev = dev;
> -	platform_set_drvdata(pdev, pcie);
> -
> -	err = pci_parse_request_of_pci_ranges(dev, &pcie->resources, NULL);
> -	if (err)
> -		goto err_free_bridge;
> -
> -	pm_runtime_enable(pcie->dev);
> -	err = pm_runtime_get_sync(pcie->dev);
> -	if (err < 0) {
> -		dev_err(pcie->dev, "pm_runtime_get_sync failed\n");
> -		goto err_pm_disable;
> -	}
> -
> -	err = rcar_pcie_get_resources(pcie);
> -	if (err < 0) {
> -		dev_err(dev, "failed to request resources: %d\n", err);
> -		goto err_pm_put;
> -	}
> -
> -	err = clk_prepare_enable(pcie->bus_clk);
> -	if (err) {
> -		dev_err(dev, "failed to enable bus clock: %d\n", err);
> -		goto err_unmap_msi_irqs;
> -	}
> -
> -	err = rcar_pcie_parse_map_dma_ranges(pcie, dev->of_node);
> -	if (err)
> -		goto err_clk_disable;
> -
> -	phy_init_fn = of_device_get_match_data(dev);
> -	err = phy_init_fn(pcie);
> -	if (err) {
> -		dev_err(dev, "failed to init PCIe PHY\n");
> -		goto err_clk_disable;
> -	}
> -
> -	/* Failure to get a link might just be that no cards are inserted */
> -	if (rcar_pcie_hw_init(pcie)) {
> -		dev_info(dev, "PCIe link down\n");
> -		err = -ENODEV;
> -		goto err_phy_shutdown;
> -	}
> -
> -	data = rcar_pci_read_reg(pcie, MACSR);
> -	dev_info(dev, "PCIe x%d: link up\n", (data >> 20) & 0x3f);
> -
> -	if (IS_ENABLED(CONFIG_PCI_MSI)) {
> -		err = rcar_pcie_enable_msi(pcie);
> -		if (err < 0) {
> -			dev_err(dev,
> -				"failed to enable MSI support: %d\n",
> -				err);
> -			goto err_phy_shutdown;
> -		}
> -	}
> -
> -	err = rcar_pcie_enable(pcie);
> -	if (err)
> -		goto err_msi_teardown;
> -
> -	return 0;
> -
> -err_msi_teardown:
> -	if (IS_ENABLED(CONFIG_PCI_MSI))
> -		rcar_pcie_teardown_msi(pcie);
> -
> -err_phy_shutdown:
> -	if (pcie->phy) {
> -		phy_power_off(pcie->phy);
> -		phy_exit(pcie->phy);
> -	}
> -
> -err_clk_disable:
> -	clk_disable_unprepare(pcie->bus_clk);
> -
> -err_unmap_msi_irqs:
> -	irq_dispose_mapping(pcie->msi.irq2);
> -	irq_dispose_mapping(pcie->msi.irq1);
> -
> -err_pm_put:
> -	pm_runtime_put(dev);
> -
> -err_pm_disable:
> -	pm_runtime_disable(dev);
> -	pci_free_resource_list(&pcie->resources);
> -
> -err_free_bridge:
> -	pci_free_host_bridge(bridge);
> -
> -	return err;
> -}
> -
> -static int rcar_pcie_resume_noirq(struct device *dev)
> -{
> -	struct rcar_pcie *pcie = dev_get_drvdata(dev);
> -
> -	if (rcar_pci_read_reg(pcie, PMSR) &&
> -	    !(rcar_pci_read_reg(pcie, PCIETCTLR) & DL_DOWN))
> -		return 0;
> -
> -	/* Re-establish the PCIe link */
> -	rcar_rmw32(pcie, MACCTLR, MACCTLR_RESERVED, 0);
> -	rcar_pci_write_reg(pcie, CFINIT, PCIETCTLR);
> -	return rcar_pcie_wait_for_dl(pcie);
> +	/* flush modifications */
> +	wmb();
>  }
> -
> -static const struct dev_pm_ops rcar_pcie_pm_ops = {
> -	.resume_noirq = rcar_pcie_resume_noirq,
> -};
> -
> -static struct platform_driver rcar_pcie_driver = {
> -	.driver = {
> -		.name = "rcar-pcie",
> -		.of_match_table = rcar_pcie_of_match,
> -		.pm = &rcar_pcie_pm_ops,
> -		.suppress_bind_attrs = true,
> -	},
> -	.probe = rcar_pcie_probe,
> -};
> -builtin_platform_driver(rcar_pcie_driver);
> diff --git a/drivers/pci/controller/pcie-rcar.h b/drivers/pci/controller/pcie-rcar.h
> new file mode 100644
> index 000000000000..502621d4e6ac
> --- /dev/null
> +++ b/drivers/pci/controller/pcie-rcar.h
> @@ -0,0 +1,129 @@
> +/* SPDX-License-Identifier: GPL-2.0 */
> +/*
> + * PCIe driver for Renesas R-Car SoCs
> + *  Copyright (C) 2014-2019 Renesas Electronics Europe Ltd
> + *
> + * Author: Phil Edworthy <phil.edworthy@renesas.com>
> + */
> +#ifndef _PCIE_RCAR_H
> +#define _PCIE_RCAR_H
> +
> +#define PCIECAR			0x000010
> +#define PCIECCTLR		0x000018
> +#define  CONFIG_SEND_ENABLE	BIT(31)
> +#define  TYPE0			(0 << 8)
> +#define  TYPE1			BIT(8)
> +#define PCIECDR			0x000020
> +#define PCIEMSR			0x000028
> +#define PCIEINTXR		0x000400
> +#define PCIEPHYSR		0x0007f0
> +#define  PHYRDY			BIT(0)
> +#define PCIEMSITXR		0x000840
> +
> +/* Transfer control */
> +#define PCIETCTLR		0x02000
> +#define  DL_DOWN		BIT(3)
> +#define  CFINIT			BIT(0)
> +#define PCIETSTR		0x02004
> +#define  DATA_LINK_ACTIVE	BIT(0)
> +#define PCIEERRFR		0x02020
> +#define  UNSUPPORTED_REQUEST	BIT(4)
> +#define PCIEMSIFR		0x02044
> +#define PCIEMSIALR		0x02048
> +#define  MSIFE			BIT(0)
> +#define PCIEMSIAUR		0x0204c
> +#define PCIEMSIIER		0x02050
> +
> +/* root port address */
> +#define PCIEPRAR(x)		(0x02080 + ((x) * 0x4))
> +
> +/* local address reg & mask */
> +#define PCIELAR(x)		(0x02200 + ((x) * 0x20))
> +#define PCIELAMR(x)		(0x02208 + ((x) * 0x20))
> +#define  LAM_PREFETCH		BIT(3)
> +#define  LAM_64BIT		BIT(2)
> +#define  LAR_ENABLE		BIT(1)
> +
> +/* PCIe address reg & mask */
> +#define PCIEPALR(x)		(0x03400 + ((x) * 0x20))
> +#define PCIEPAUR(x)		(0x03404 + ((x) * 0x20))
> +#define PCIEPAMR(x)		(0x03408 + ((x) * 0x20))
> +#define PCIEPTCTLR(x)		(0x0340c + ((x) * 0x20))
> +#define  PAR_ENABLE		BIT(31)
> +#define  IO_SPACE		BIT(8)
> +
> +/* Configuration */
> +#define PCICONF(x)		(0x010000 + ((x) * 0x4))
> +#define PMCAP(x)		(0x010040 + ((x) * 0x4))
> +#define EXPCAP(x)		(0x010070 + ((x) * 0x4))
> +#define VCCAP(x)		(0x010100 + ((x) * 0x4))
> +
> +/* link layer */
> +#define IDSETR0			0x011000
> +#define  DEVICE_ID_SHFIT	16
> +#define IDSETR1			0x011004
> +#define SUBIDSETR		0x011024
> +#define TLCTLR			0x011048
> +#define MACSR			0x011054
> +#define  SPCHGFIN		BIT(4)
> +#define  SPCHGFAIL		BIT(6)
> +#define  SPCHGSUC		BIT(7)
> +#define  LINK_SPEED		(0xf << 16)
> +#define  LINK_SPEED_2_5GTS	BIT(16)
> +#define  LINK_SPEED_5_0GTS	BIT(17)
> +#define MACCTLR			0x011058
> +#define  MACCTLR_RESERVED	BIT(0)
> +#define  SPEED_CHANGE		BIT(24)
> +#define  SCRAMBLE_DISABLE	BIT(27)
> +#define PMSR			0x01105c
> +#define MACS2R			0x011078
> +#define MACCGSPSETR		0x011084
> +#define  SPCNGRSN		BIT(31)
> +
> +/* R-Car H1 PHY */
> +#define H1_PCIEPHYADRR		0x04000c
> +#define  WRITE_CMD		BIT(16)
> +#define  PHY_ACK		BIT(24)
> +#define  RATE_POS		12
> +#define  LANE_POS		8
> +#define  ADR_POS		0
> +#define H1_PCIEPHYDOUTR		0x040014
> +
> +/* R-Car Gen2 PHY */
> +#define GEN2_PCIEPHYADDR	0x780
> +#define GEN2_PCIEPHYDATA	0x784
> +#define GEN2_PCIEPHYCTRL	0x78c
> +
> +#define INT_PCI_MSI_NR		32
> +
> +#define RCONF(x)		(PCICONF(0) + (x))
> +#define RPMCAP(x)		(PMCAP(0) + (x))
> +#define REXPCAP(x)		(EXPCAP(0) + (x))
> +#define RVCCAP(x)		(VCCAP(0) + (x))
> +
> +#define INTDIS_SHIFT		BIT(10)
> +#define ASTINTX_SHIFT		BIT(16)
> +
> +#define PCIE_CONF_BUS(b)	(((b) & 0xff) << 24)
> +#define PCIE_CONF_DEV(d)	(((d) & 0x1f) << 19)
> +#define PCIE_CONF_FUNC(f)	(((f) & 0x7) << 16)
> +
> +#define RCAR_PCI_MAX_RESOURCES	4
> +#define MAX_NR_INBOUND_MAPS	6
> +
> +enum {
> +	RCAR_PCI_ACCESS_READ,
> +	RCAR_PCI_ACCESS_WRITE,
> +};
> +
> +void rcar_pci_write_reg(void __iomem *base, u32 val, unsigned int reg);
> +u32 rcar_pci_read_reg(void __iomem *base, unsigned int reg);
> +void rcar_rmw32(void __iomem *base, int where, u32 mask, u32 data);
> +int rcar_pcie_wait_for_phyrdy(void __iomem *base);
> +int rcar_pcie_wait_for_dl(void __iomem *base);
> +void rcar_pcie_set_outbound(int win, void __iomem *base,
> +			    struct resource *res, bool host);
> +void rcar_pcie_set_inbound(void __iomem *base, u64 cpu_addr, u64 pci_addr,
> +			   u64 mask, int idx, bool host);
> +
> +#endif
> 

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 1/5] pci: pcie-rcar: preparation for adding endpoint support
@ 2019-11-27  4:55     ` Kishon Vijay Abraham I
  0 siblings, 0 replies; 60+ messages in thread
From: Kishon Vijay Abraham I @ 2019-11-27  4:55 UTC (permalink / raw)
  To: Lad Prabhakar, Bjorn Helgaas, Rob Herring, Mark Rutland,
	Geert Uytterhoeven, Magnus Damm, Marek Vasut, Yoshihiro Shimoda,
	linux-pci
  Cc: devicetree, Chris Paterson, Lorenzo Pieralisi, Arnd Bergmann,
	Catalin Marinas, linux-kernel, Lad, Prabhakar, linux-renesas-soc,
	Greg Kroah-Hartman, Andrew Murray, Will Deacon, linux-arm-kernel

Hi Prabhakar,

On 07/11/19 1:06 AM, Lad Prabhakar wrote:
> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> 
> this patch prepares for adding endpoint support to rcar controller,
> there are no functional changes with this patch, a common file is
> created so that it can be shared with endpoint driver.
> 
> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> ---
>  arch/arm64/configs/defconfig            |    2 +-
>  arch/arm64/configs/renesas_defconfig    |    2 +-
>  drivers/pci/controller/Kconfig          |    4 +-
>  drivers/pci/controller/Makefile         |    2 +-
>  drivers/pci/controller/pcie-rcar-host.c | 1058 +++++++++++++++++++
>  drivers/pci/controller/pcie-rcar.c      | 1231 +----------------------

Can you create this patch using "git format-patch -M", so that only whatever
changed appears here.

Thanks
Kishon

>  drivers/pci/controller/pcie-rcar.h      |  129 +++
>  7 files changed, 1245 insertions(+), 1183 deletions(-)
>  create mode 100644 drivers/pci/controller/pcie-rcar-host.c
>  create mode 100644 drivers/pci/controller/pcie-rcar.h
> 
> diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
> index 7d16f10a9251..f67ce0ccea1d 100644
> --- a/arch/arm64/configs/defconfig
> +++ b/arch/arm64/configs/defconfig
> @@ -186,7 +186,7 @@ CONFIG_HOTPLUG_PCI=y
>  CONFIG_HOTPLUG_PCI_ACPI=y
>  CONFIG_PCI_AARDVARK=y
>  CONFIG_PCI_TEGRA=y
> -CONFIG_PCIE_RCAR=y
> +CONFIG_PCIE_RCAR_HOST=y
>  CONFIG_PCI_HOST_GENERIC=y
>  CONFIG_PCI_XGENE=y
>  CONFIG_PCIE_ALTERA=y
> diff --git a/arch/arm64/configs/renesas_defconfig b/arch/arm64/configs/renesas_defconfig
> index f7411e8807d6..75fc418377c5 100644
> --- a/arch/arm64/configs/renesas_defconfig
> +++ b/arch/arm64/configs/renesas_defconfig
> @@ -91,7 +91,7 @@ CONFIG_PCIEPORTBUS=y
>  CONFIG_HOTPLUG_PCI_PCIE=y
>  CONFIG_PCI_IOV=y
>  CONFIG_HOTPLUG_PCI=y
> -CONFIG_PCIE_RCAR=y
> +CONFIG_PCIE_RCAR_HOST=y
>  CONFIG_PCI_HOST_GENERIC=y
>  CONFIG_DEVTMPFS=y
>  CONFIG_DEVTMPFS_MOUNT=y
> diff --git a/drivers/pci/controller/Kconfig b/drivers/pci/controller/Kconfig
> index 70e078238899..01c238e5ea1e 100644
> --- a/drivers/pci/controller/Kconfig
> +++ b/drivers/pci/controller/Kconfig
> @@ -82,12 +82,12 @@ config PCI_RCAR_GEN2
>  	  There are 3 internal PCI controllers available with a single
>  	  built-in EHCI/OHCI host controller present on each one.
>  
> -config PCIE_RCAR
> +config PCIE_RCAR_HOST
>  	bool "Renesas R-Car PCIe controller"
>  	depends on ARCH_RENESAS || COMPILE_TEST
>  	depends on PCI_MSI_IRQ_DOMAIN
>  	help
> -	  Say Y here if you want PCIe controller support on R-Car SoCs.
> +	  Say Y here if you want PCIe controller support on R-Car SoCs in host mode.
>  
>  config PCI_HOST_COMMON
>  	bool
> diff --git a/drivers/pci/controller/Makefile b/drivers/pci/controller/Makefile
> index a2a22c9d91af..3577902a0d40 100644
> --- a/drivers/pci/controller/Makefile
> +++ b/drivers/pci/controller/Makefile
> @@ -9,7 +9,7 @@ obj-$(CONFIG_PCI_MVEBU) += pci-mvebu.o
>  obj-$(CONFIG_PCI_AARDVARK) += pci-aardvark.o
>  obj-$(CONFIG_PCI_TEGRA) += pci-tegra.o
>  obj-$(CONFIG_PCI_RCAR_GEN2) += pci-rcar-gen2.o
> -obj-$(CONFIG_PCIE_RCAR) += pcie-rcar.o
> +obj-$(CONFIG_PCIE_RCAR_HOST) += pcie-rcar.o pcie-rcar-host.o
>  obj-$(CONFIG_PCI_HOST_COMMON) += pci-host-common.o
>  obj-$(CONFIG_PCI_HOST_GENERIC) += pci-host-generic.o
>  obj-$(CONFIG_PCIE_XILINX) += pcie-xilinx.o
> diff --git a/drivers/pci/controller/pcie-rcar-host.c b/drivers/pci/controller/pcie-rcar-host.c
> new file mode 100644
> index 000000000000..645ef3e41e27
> --- /dev/null
> +++ b/drivers/pci/controller/pcie-rcar-host.c
> @@ -0,0 +1,1058 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/*
> + * PCIe driver for Renesas R-Car SoCs
> + *  Copyright (C) 2014-2019 Renesas Electronics Europe Ltd
> + *
> + * Based on:
> + *  arch/sh/drivers/pci/pcie-sh7786.c
> + *  arch/sh/drivers/pci/ops-sh7786.c
> + *  Copyright (C) 2009 - 2011  Paul Mundt
> + *
> + * Author: Phil Edworthy <phil.edworthy@renesas.com>
> + */
> +
> +#include <linux/clk.h>
> +#include <linux/delay.h>
> +#include <linux/msi.h>
> +#include <linux/of_address.h>
> +#include <linux/of_irq.h>
> +#include <linux/of_pci.h>
> +#include <linux/of_platform.h>
> +#include <linux/pci.h>
> +#include <linux/phy/phy.h>
> +
> +#include "pcie-rcar.h"
> +
> +struct rcar_msi {
> +	DECLARE_BITMAP(used, INT_PCI_MSI_NR);
> +	struct irq_domain *domain;
> +	struct msi_controller chip;
> +	unsigned long pages;
> +	struct mutex lock; /* serializes msi */
> +	int irq1;
> +	int irq2;
> +};
> +
> +static inline struct rcar_msi *to_rcar_msi(struct msi_controller *chip)
> +{
> +	return container_of(chip, struct rcar_msi, chip);
> +}
> +
> +/* Structure representing the PCIe interface */
> +struct rcar_pcie {
> +	struct device		*dev;
> +	struct phy		*phy;
> +	void __iomem		*base;
> +	struct list_head	resources;
> +	int			root_bus_nr;
> +	struct clk		*bus_clk;
> +	struct			rcar_msi msi;
> +};
> +
> +static u32 rcar_read_conf(struct rcar_pcie *pcie, int where)
> +{
> +	unsigned int shift = BITS_PER_BYTE * (where & 3);
> +	u32 val = rcar_pci_read_reg(pcie->base, where & ~3);
> +
> +	return val >> shift;
> +}
> +
> +/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
> +static int
> +rcar_pcie_config_access(struct rcar_pcie *pcie,
> +			unsigned char access_type, struct pci_bus *bus,
> +			unsigned int devfn, int where, u32 *data)
> +{
> +	unsigned int dev, func, reg, index;
> +
> +	dev = PCI_SLOT(devfn);
> +	func = PCI_FUNC(devfn);
> +	reg = where & ~3;
> +	index = reg / 4;
> +
> +	/*
> +	 * While each channel has its own memory-mapped extended config
> +	 * space, it's generally only accessible when in endpoint mode.
> +	 * When in root complex mode, the controller is unable to target
> +	 * itself with either type 0 or type 1 accesses, and indeed, any
> +	 * controller initiated target transfer to its own config space
> +	 * result in a completer abort.
> +	 *
> +	 * Each channel effectively only supports a single device, but as
> +	 * the same channel <-> device access works for any PCI_SLOT()
> +	 * value, we cheat a bit here and bind the controller's config
> +	 * space to devfn 0 in order to enable self-enumeration. In this
> +	 * case the regular ECAR/ECDR path is sidelined and the mangled
> +	 * config access itself is initiated as an internal bus transaction.
> +	 */
> +	if (pci_is_root_bus(bus)) {
> +		if (dev != 0)
> +			return PCIBIOS_DEVICE_NOT_FOUND;
> +
> +		if (access_type == RCAR_PCI_ACCESS_READ) {
> +			*data = rcar_pci_read_reg(pcie->base, PCICONF(index));
> +		} else {
> +			/* Keep an eye out for changes to the root bus number */
> +			if (pci_is_root_bus(bus) && reg == PCI_PRIMARY_BUS)
> +				pcie->root_bus_nr = *data & 0xff;
> +
> +			rcar_pci_write_reg(pcie->base, *data, PCICONF(index));
> +		}
> +
> +		return PCIBIOS_SUCCESSFUL;
> +	}
> +
> +	if (pcie->root_bus_nr < 0)
> +		return PCIBIOS_DEVICE_NOT_FOUND;
> +
> +	/* Clear errors */
> +	rcar_pci_write_reg(pcie->base, rcar_pci_read_reg(pcie->base, PCIEERRFR),
> +			   PCIEERRFR);
> +
> +	/* Set the PIO address */
> +	rcar_pci_write_reg(pcie->base, PCIE_CONF_BUS(bus->number) |
> +		PCIE_CONF_DEV(dev) | PCIE_CONF_FUNC(func) | reg, PCIECAR);
> +
> +	/* Enable the configuration access */
> +	if (bus->parent->number == pcie->root_bus_nr)
> +		rcar_pci_write_reg(pcie->base,
> +				   CONFIG_SEND_ENABLE | TYPE0, PCIECCTLR);
> +	else
> +		rcar_pci_write_reg(pcie->base,
> +				   CONFIG_SEND_ENABLE | TYPE1, PCIECCTLR);
> +
> +	/* Check for errors */
> +	if (rcar_pci_read_reg(pcie->base, PCIEERRFR) & UNSUPPORTED_REQUEST)
> +		return PCIBIOS_DEVICE_NOT_FOUND;
> +
> +	/* Check for master and target aborts */
> +	if (rcar_read_conf(pcie, RCONF(PCI_STATUS)) &
> +		(PCI_STATUS_REC_MASTER_ABORT | PCI_STATUS_REC_TARGET_ABORT))
> +		return PCIBIOS_DEVICE_NOT_FOUND;
> +
> +	if (access_type == RCAR_PCI_ACCESS_READ)
> +		*data = rcar_pci_read_reg(pcie->base, PCIECDR);
> +	else
> +		rcar_pci_write_reg(pcie->base, *data, PCIECDR);
> +
> +	/* Disable the configuration access */
> +	rcar_pci_write_reg(pcie->base, 0, PCIECCTLR);
> +
> +	return PCIBIOS_SUCCESSFUL;
> +}
> +
> +static int rcar_pcie_read_conf(struct pci_bus *bus, unsigned int devfn,
> +			       int where, int size, u32 *val)
> +{
> +	struct rcar_pcie *pcie = bus->sysdata;
> +	int ret;
> +
> +	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
> +				      bus, devfn, where, val);
> +	if (ret != PCIBIOS_SUCCESSFUL) {
> +		*val = 0xffffffff;
> +		return ret;
> +	}
> +
> +	if (size == 1)
> +		*val = (*val >> (BITS_PER_BYTE * (where & 3))) & 0xff;
> +	else if (size == 2)
> +		*val = (*val >> (BITS_PER_BYTE * (where & 2))) & 0xffff;
> +
> +	dev_dbg(&bus->dev, "pcie-config-read: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
> +		bus->number, devfn, where, size, *val);
> +
> +	return ret;
> +}
> +
> +/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
> +static int rcar_pcie_write_conf(struct pci_bus *bus, unsigned int devfn,
> +				int where, int size, u32 val)
> +{
> +	struct rcar_pcie *pcie = bus->sysdata;
> +	unsigned int shift;
> +	u32 data;
> +	int ret;
> +
> +	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
> +				      bus, devfn, where, &data);
> +	if (ret != PCIBIOS_SUCCESSFUL)
> +		return ret;
> +
> +	dev_dbg(&bus->dev, "pcie-config-write: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
> +		bus->number, devfn, where, size, val);
> +
> +	if (size == 1) {
> +		shift = BITS_PER_BYTE * (where & 3);
> +		data &= ~(0xff << shift);
> +		data |= ((val & 0xff) << shift);
> +	} else if (size == 2) {
> +		shift = BITS_PER_BYTE * (where & 2);
> +		data &= ~(0xffff << shift);
> +		data |= ((val & 0xffff) << shift);
> +	} else {
> +		data = val;
> +	}
> +
> +	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_WRITE,
> +				      bus, devfn, where, &data);
> +
> +	return ret;
> +}
> +
> +static struct pci_ops rcar_pcie_ops = {
> +	.read	= rcar_pcie_read_conf,
> +	.write	= rcar_pcie_write_conf,
> +};
> +
> +static int rcar_pcie_setup(struct list_head *resource, struct rcar_pcie *pci)
> +{
> +	struct resource_entry *win;
> +	int i = 0;
> +
> +	/* Setup PCI resources */
> +	resource_list_for_each_entry(win, &pci->resources) {
> +		struct resource *res = win->res;
> +
> +		if (!res->flags)
> +			continue;
> +
> +		switch (resource_type(res)) {
> +		case IORESOURCE_IO:
> +		case IORESOURCE_MEM:
> +			rcar_pcie_set_outbound(i, pci->base, res, true);
> +			i++;
> +			break;
> +		case IORESOURCE_BUS:
> +			pci->root_bus_nr = res->start;
> +			break;
> +		default:
> +			continue;
> +		}
> +
> +		pci_add_resource(resource, res);
> +	}
> +
> +	return 1;
> +}
> +
> +static void rcar_pcie_force_speedup(struct rcar_pcie *pcie)
> +{
> +	struct device *dev = pcie->dev;
> +	unsigned int timeout = 1000;
> +	u32 macsr;
> +
> +	if ((rcar_pci_read_reg(pcie->base, MACS2R) & LINK_SPEED) !=
> +		LINK_SPEED_5_0GTS)
> +		return;
> +
> +	if (rcar_pci_read_reg(pcie->base, MACCTLR) & SPEED_CHANGE) {
> +		dev_err(dev, "Speed change already in progress\n");
> +		return;
> +	}
> +
> +	macsr = rcar_pci_read_reg(pcie->base, MACSR);
> +	if ((macsr & LINK_SPEED) == LINK_SPEED_5_0GTS)
> +		goto done;
> +
> +	/* Set target link speed to 5.0 GT/s */
> +	rcar_rmw32(pcie->base, EXPCAP(12), PCI_EXP_LNKSTA_CLS,
> +		   PCI_EXP_LNKSTA_CLS_5_0GB);
> +
> +	/* Set speed change reason as intentional factor */
> +	rcar_rmw32(pcie->base, MACCGSPSETR, SPCNGRSN, 0);
> +
> +	/* Clear SPCHGFIN, SPCHGSUC, and SPCHGFAIL */
> +	if (macsr & (SPCHGFIN | SPCHGSUC | SPCHGFAIL))
> +		rcar_pci_write_reg(pcie->base, macsr, MACSR);
> +
> +	/* Start link speed change */
> +	rcar_rmw32(pcie->base, MACCTLR, SPEED_CHANGE, SPEED_CHANGE);
> +
> +	while (timeout--) {
> +		macsr = rcar_pci_read_reg(pcie->base, MACSR);
> +		if (macsr & SPCHGFIN) {
> +			/* Clear the interrupt bits */
> +			rcar_pci_write_reg(pcie->base, macsr, MACSR);
> +
> +			if (macsr & SPCHGFAIL)
> +				dev_err(dev, "Speed change failed\n");
> +
> +			goto done;
> +		}
> +
> +		usleep_range(1000, 2000);
> +	}
> +
> +	dev_err(dev, "Speed change timed out\n");
> +
> +done:
> +	dev_info(dev, "Current link speed is %s GT/s\n",
> +		 (macsr & LINK_SPEED) == LINK_SPEED_5_0GTS ? "5" : "2.5");
> +}
> +
> +static int rcar_pcie_enable(struct rcar_pcie *pcie)
> +{
> +	struct pci_host_bridge *bridge = pci_host_bridge_from_priv(pcie);
> +	struct device *dev = pcie->dev;
> +	struct pci_bus *bus, *child;
> +	int ret;
> +
> +	/* Try setting 5 GT/s link speed */
> +	rcar_pcie_force_speedup(pcie);
> +
> +	rcar_pcie_setup(&bridge->windows, pcie);
> +
> +	pci_add_flags(PCI_REASSIGN_ALL_BUS);
> +
> +	bridge->dev.parent = dev;
> +	bridge->sysdata = pcie;
> +	bridge->busnr = pcie->root_bus_nr;
> +	bridge->ops = &rcar_pcie_ops;
> +	bridge->map_irq = of_irq_parse_and_map_pci;
> +	bridge->swizzle_irq = pci_common_swizzle;
> +	if (IS_ENABLED(CONFIG_PCI_MSI))
> +		bridge->msi = &pcie->msi.chip;
> +
> +	ret = pci_scan_root_bus_bridge(bridge);
> +	if (ret < 0)
> +		return ret;
> +
> +	bus = bridge->bus;
> +
> +	pci_bus_size_bridges(bus);
> +	pci_bus_assign_resources(bus);
> +
> +	list_for_each_entry(child, &bus->children, node)
> +		pcie_bus_configure_settings(child);
> +
> +	pci_bus_add_devices(bus);
> +
> +	return 0;
> +}
> +
> +static int phy_wait_for_ack(struct rcar_pcie *pcie)
> +{
> +	struct device *dev = pcie->dev;
> +	unsigned int timeout = 100;
> +
> +	while (timeout--) {
> +		if (rcar_pci_read_reg(pcie->base, H1_PCIEPHYADRR) & PHY_ACK)
> +			return 0;
> +
> +		usleep_range(90, 110);
> +	}
> +
> +	dev_err(dev, "Access to PCIe phy timed out\n");
> +
> +	return -ETIMEDOUT;
> +}
> +
> +static void phy_write_reg(struct rcar_pcie *pcie,
> +			  unsigned int rate, u32 addr,
> +			  unsigned int lane, u32 data)
> +{
> +	u32 phyaddr;
> +
> +	phyaddr = WRITE_CMD |
> +		((rate & 1) << RATE_POS) |
> +		((lane & 0xf) << LANE_POS) |
> +		((addr & 0xff) << ADR_POS);
> +
> +	/* Set write data */
> +	rcar_pci_write_reg(pcie->base, data, H1_PCIEPHYDOUTR);
> +	rcar_pci_write_reg(pcie->base, phyaddr, H1_PCIEPHYADRR);
> +
> +	/* Ignore errors as they will be dealt with if the data link is down */
> +	phy_wait_for_ack(pcie);
> +
> +	/* Clear command */
> +	rcar_pci_write_reg(pcie->base, 0, H1_PCIEPHYDOUTR);
> +	rcar_pci_write_reg(pcie->base, 0, H1_PCIEPHYADRR);
> +
> +	/* Ignore errors as they will be dealt with if the data link is down */
> +	phy_wait_for_ack(pcie);
> +}
> +
> +static int rcar_pcie_hw_init(struct rcar_pcie *pcie)
> +{
> +	int err;
> +
> +	/* Begin initialization */
> +	rcar_pci_write_reg(pcie->base, 0, PCIETCTLR);
> +
> +	/* Set mode */
> +	rcar_pci_write_reg(pcie->base, 1, PCIEMSR);
> +
> +	err = rcar_pcie_wait_for_phyrdy(pcie->base);
> +	if (err)
> +		return err;
> +
> +	/*
> +	 * Initial header for port config space is type 1, set the device
> +	 * class to match. Hardware takes care of propagating the IDSETR
> +	 * settings, so there is no need to bother with a quirk.
> +	 */
> +	rcar_pci_write_reg(pcie->base, PCI_CLASS_BRIDGE_PCI << 16, IDSETR1);
> +
> +	/*
> +	 * Setup Secondary Bus Number & Subordinate Bus Number, even though
> +	 * they aren't used, to avoid bridge being detected as broken.
> +	 */
> +	rcar_rmw32(pcie->base, RCONF(PCI_SECONDARY_BUS), 0xff, 1);
> +	rcar_rmw32(pcie->base, RCONF(PCI_SUBORDINATE_BUS), 0xff, 1);
> +
> +	/* Initialize default capabilities. */
> +	rcar_rmw32(pcie->base, REXPCAP(0), 0xff, PCI_CAP_ID_EXP);
> +	rcar_rmw32(pcie->base, REXPCAP(PCI_EXP_FLAGS),
> +		   PCI_EXP_FLAGS_TYPE, PCI_EXP_TYPE_ROOT_PORT << 4);
> +	rcar_rmw32(pcie->base, RCONF(PCI_HEADER_TYPE), 0x7f,
> +		   PCI_HEADER_TYPE_BRIDGE);
> +
> +	/* Enable data link layer active state reporting */
> +	rcar_rmw32(pcie->base, REXPCAP(PCI_EXP_LNKCAP), PCI_EXP_LNKCAP_DLLLARC,
> +		   PCI_EXP_LNKCAP_DLLLARC);
> +
> +	/* Write out the physical slot number = 0 */
> +	rcar_rmw32(pcie->base, REXPCAP(PCI_EXP_SLTCAP), PCI_EXP_SLTCAP_PSN, 0);
> +
> +	/* Set the completion timer timeout to the maximum 50ms. */
> +	rcar_rmw32(pcie->base, TLCTLR + 1, 0x3f, 50);
> +
> +	/* Terminate list of capabilities (Next Capability Offset=0) */
> +	rcar_rmw32(pcie->base, RVCCAP(0), 0xfff00000, 0);
> +
> +	/* Enable MSI */
> +	if (IS_ENABLED(CONFIG_PCI_MSI))
> +		rcar_pci_write_reg(pcie->base, 0x801f0000, PCIEMSITXR);
> +
> +	rcar_rmw32(pcie->base, MACCTLR, MACCTLR_RESERVED, 0);
> +
> +	/* Finish initialization - establish a PCI Express link */
> +	rcar_pci_write_reg(pcie->base, CFINIT, PCIETCTLR);
> +
> +	/* This will timeout if we don't have a link. */
> +	err = rcar_pcie_wait_for_dl(pcie->base);
> +	if (err)
> +		return err;
> +
> +	/* Enable INTx interrupts */
> +	rcar_rmw32(pcie->base, PCIEINTXR, 0, 0xF << 8);
> +
> +	/* flush modifications */
> +	wmb();
> +
> +	return 0;
> +}
> +
> +static int rcar_pcie_phy_init_h1(struct rcar_pcie *pcie)
> +{
> +	/* Initialize the phy */
> +	phy_write_reg(pcie, 0, 0x42, 0x1, 0x0EC34191);
> +	phy_write_reg(pcie, 1, 0x42, 0x1, 0x0EC34180);
> +	phy_write_reg(pcie, 0, 0x43, 0x1, 0x00210188);
> +	phy_write_reg(pcie, 1, 0x43, 0x1, 0x00210188);
> +	phy_write_reg(pcie, 0, 0x44, 0x1, 0x015C0014);
> +	phy_write_reg(pcie, 1, 0x44, 0x1, 0x015C0014);
> +	phy_write_reg(pcie, 1, 0x4C, 0x1, 0x786174A0);
> +	phy_write_reg(pcie, 1, 0x4D, 0x1, 0x048000BB);
> +	phy_write_reg(pcie, 0, 0x51, 0x1, 0x079EC062);
> +	phy_write_reg(pcie, 0, 0x52, 0x1, 0x20000000);
> +	phy_write_reg(pcie, 1, 0x52, 0x1, 0x20000000);
> +	phy_write_reg(pcie, 1, 0x56, 0x1, 0x00003806);
> +
> +	phy_write_reg(pcie, 0, 0x60, 0x1, 0x004B03A5);
> +	phy_write_reg(pcie, 0, 0x64, 0x1, 0x3F0F1F0F);
> +	phy_write_reg(pcie, 0, 0x66, 0x1, 0x00008000);
> +
> +	return 0;
> +}
> +
> +static int rcar_pcie_phy_init_gen2(struct rcar_pcie *pcie)
> +{
> +	/*
> +	 * These settings come from the R-Car Series, 2nd Generation User's
> +	 * Manual, section 50.3.1 (2) Initialization of the physical layer.
> +	 */
> +	rcar_pci_write_reg(pcie->base, 0x000f0030, GEN2_PCIEPHYADDR);
> +	rcar_pci_write_reg(pcie->base, 0x00381203, GEN2_PCIEPHYDATA);
> +	rcar_pci_write_reg(pcie->base, 0x00000001, GEN2_PCIEPHYCTRL);
> +	rcar_pci_write_reg(pcie->base, 0x00000006, GEN2_PCIEPHYCTRL);
> +
> +	rcar_pci_write_reg(pcie->base, 0x000f0054, GEN2_PCIEPHYADDR);
> +	/* The following value is for DC connection, no termination resistor */
> +	rcar_pci_write_reg(pcie->base, 0x13802007, GEN2_PCIEPHYDATA);
> +	rcar_pci_write_reg(pcie->base, 0x00000001, GEN2_PCIEPHYCTRL);
> +	rcar_pci_write_reg(pcie->base, 0x00000006, GEN2_PCIEPHYCTRL);
> +
> +	return 0;
> +}
> +
> +static int rcar_pcie_phy_init_gen3(struct rcar_pcie *pcie)
> +{
> +	int err;
> +
> +	err = phy_init(pcie->phy);
> +	if (err)
> +		return err;
> +
> +	err = phy_power_on(pcie->phy);
> +	if (err)
> +		phy_exit(pcie->phy);
> +
> +	return err;
> +}
> +
> +static int rcar_msi_alloc(struct rcar_msi *chip)
> +{
> +	int msi;
> +
> +	mutex_lock(&chip->lock);
> +
> +	msi = find_first_zero_bit(chip->used, INT_PCI_MSI_NR);
> +	if (msi < INT_PCI_MSI_NR)
> +		set_bit(msi, chip->used);
> +	else
> +		msi = -ENOSPC;
> +
> +	mutex_unlock(&chip->lock);
> +
> +	return msi;
> +}
> +
> +static int rcar_msi_alloc_region(struct rcar_msi *chip, int no_irqs)
> +{
> +	int msi;
> +
> +	mutex_lock(&chip->lock);
> +	msi = bitmap_find_free_region(chip->used, INT_PCI_MSI_NR,
> +				      order_base_2(no_irqs));
> +	mutex_unlock(&chip->lock);
> +
> +	return msi;
> +}
> +
> +static void rcar_msi_free(struct rcar_msi *chip, unsigned long irq)
> +{
> +	mutex_lock(&chip->lock);
> +	clear_bit(irq, chip->used);
> +	mutex_unlock(&chip->lock);
> +}
> +
> +static irqreturn_t rcar_pcie_msi_irq(int irq, void *data)
> +{
> +	struct rcar_pcie *pcie = data;
> +	struct rcar_msi *msi = &pcie->msi;
> +	struct device *dev = pcie->dev;
> +	unsigned long reg;
> +
> +	reg = rcar_pci_read_reg(pcie->base, PCIEMSIFR);
> +
> +	/* MSI & INTx share an interrupt - we only handle MSI here */
> +	if (!reg)
> +		return IRQ_NONE;
> +
> +	while (reg) {
> +		unsigned int index = find_first_bit(&reg, 32);
> +		unsigned int msi_irq;
> +
> +		/* clear the interrupt */
> +		rcar_pci_write_reg(pcie->base, 1 << index, PCIEMSIFR);
> +
> +		msi_irq = irq_find_mapping(msi->domain, index);
> +		if (msi_irq) {
> +			if (test_bit(index, msi->used))
> +				generic_handle_irq(msi_irq);
> +			else
> +				dev_info(dev, "unhandled MSI\n");
> +		} else {
> +			/* Unknown MSI, just clear it */
> +			dev_dbg(dev, "unexpected MSI\n");
> +		}
> +
> +		/* see if there's any more pending in this vector */
> +		reg = rcar_pci_read_reg(pcie->base, PCIEMSIFR);
> +	}
> +
> +	return IRQ_HANDLED;
> +}
> +
> +static int rcar_msi_setup_irq(struct msi_controller *chip, struct pci_dev *pdev,
> +			      struct msi_desc *desc)
> +{
> +	struct rcar_msi *msi = to_rcar_msi(chip);
> +	struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
> +	struct msi_msg msg;
> +	unsigned int irq;
> +	int hwirq;
> +
> +	hwirq = rcar_msi_alloc(msi);
> +	if (hwirq < 0)
> +		return hwirq;
> +
> +	irq = irq_find_mapping(msi->domain, hwirq);
> +	if (!irq) {
> +		rcar_msi_free(msi, hwirq);
> +		return -EINVAL;
> +	}
> +
> +	irq_set_msi_desc(irq, desc);
> +
> +	msg.address_lo = rcar_pci_read_reg(pcie->base, PCIEMSIALR) & ~MSIFE;
> +	msg.address_hi = rcar_pci_read_reg(pcie->base, PCIEMSIAUR);
> +	msg.data = hwirq;
> +
> +	pci_write_msi_msg(irq, &msg);
> +
> +	return 0;
> +}
> +
> +static int rcar_msi_setup_irqs(struct msi_controller *chip,
> +			       struct pci_dev *pdev, int nvec, int type)
> +{
> +	struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
> +	struct rcar_msi *msi = to_rcar_msi(chip);
> +	struct msi_desc *desc;
> +	struct msi_msg msg;
> +	unsigned int irq;
> +	int hwirq;
> +	int i;
> +
> +	/* MSI-X interrupts are not supported */
> +	if (type == PCI_CAP_ID_MSIX)
> +		return -EINVAL;
> +
> +	WARN_ON(!list_is_singular(&pdev->dev.msi_list));
> +	desc = list_entry(pdev->dev.msi_list.next, struct msi_desc, list);
> +
> +	hwirq = rcar_msi_alloc_region(msi, nvec);
> +	if (hwirq < 0)
> +		return -ENOSPC;
> +
> +	irq = irq_find_mapping(msi->domain, hwirq);
> +	if (!irq)
> +		return -ENOSPC;
> +
> +	for (i = 0; i < nvec; i++) {
> +		/*
> +		 * irq_create_mapping() called from rcar_pcie_probe() pre-
> +		 * allocates descs,  so there is no need to allocate descs here.
> +		 * We can therefore assume that if irq_find_mapping() above
> +		 * returns non-zero, then the descs are also successfully
> +		 * allocated.
> +		 */
> +		if (irq_set_msi_desc_off(irq, i, desc)) {
> +			/* TODO: clear */
> +			return -EINVAL;
> +		}
> +	}
> +
> +	desc->nvec_used = nvec;
> +	desc->msi_attrib.multiple = order_base_2(nvec);
> +
> +	msg.address_lo = rcar_pci_read_reg(pcie->base, PCIEMSIALR) & ~MSIFE;
> +	msg.address_hi = rcar_pci_read_reg(pcie->base, PCIEMSIAUR);
> +	msg.data = hwirq;
> +
> +	pci_write_msi_msg(irq, &msg);
> +
> +	return 0;
> +}
> +
> +static void rcar_msi_teardown_irq(struct msi_controller *chip, unsigned int irq)
> +{
> +	struct irq_data *d = irq_get_irq_data(irq);
> +	struct rcar_msi *msi = to_rcar_msi(chip);
> +
> +	rcar_msi_free(msi, d->hwirq);
> +}
> +
> +static struct irq_chip rcar_msi_irq_chip = {
> +	.name = "R-Car PCIe MSI",
> +	.irq_enable = pci_msi_unmask_irq,
> +	.irq_disable = pci_msi_mask_irq,
> +	.irq_mask = pci_msi_mask_irq,
> +	.irq_unmask = pci_msi_unmask_irq,
> +};
> +
> +static int rcar_msi_map(struct irq_domain *domain, unsigned int irq,
> +			irq_hw_number_t hwirq)
> +{
> +	irq_set_chip_and_handler(irq, &rcar_msi_irq_chip, handle_simple_irq);
> +	irq_set_chip_data(irq, domain->host_data);
> +
> +	return 0;
> +}
> +
> +static const struct irq_domain_ops msi_domain_ops = {
> +	.map = rcar_msi_map,
> +};
> +
> +static void rcar_pcie_unmap_msi(struct rcar_pcie *pcie)
> +{
> +	struct rcar_msi *msi = &pcie->msi;
> +	int i, irq;
> +
> +	for (i = 0; i < INT_PCI_MSI_NR; i++) {
> +		irq = irq_find_mapping(msi->domain, i);
> +		if (irq > 0)
> +			irq_dispose_mapping(irq);
> +	}
> +
> +	irq_domain_remove(msi->domain);
> +}
> +
> +static int rcar_pcie_enable_msi(struct rcar_pcie *pcie)
> +{
> +	struct rcar_msi *msi = &pcie->msi;
> +	struct device *dev = pcie->dev;
> +	phys_addr_t base;
> +	int err, i;
> +
> +	mutex_init(&msi->lock);
> +
> +	msi->chip.dev = dev;
> +	msi->chip.setup_irq = rcar_msi_setup_irq;
> +	msi->chip.setup_irqs = rcar_msi_setup_irqs;
> +	msi->chip.teardown_irq = rcar_msi_teardown_irq;
> +
> +	msi->domain = irq_domain_add_linear(dev->of_node, INT_PCI_MSI_NR,
> +					    &msi_domain_ops, &msi->chip);
> +	if (!msi->domain) {
> +		dev_err(dev, "failed to create IRQ domain\n");
> +		return -ENOMEM;
> +	}
> +
> +	for (i = 0; i < INT_PCI_MSI_NR; i++)
> +		irq_create_mapping(msi->domain, i);
> +
> +	/* Two irqs are for MSI, but they are also used for non-MSI irqs */
> +	err = devm_request_irq(dev, msi->irq1, rcar_pcie_msi_irq,
> +			       IRQF_SHARED | IRQF_NO_THREAD,
> +			       rcar_msi_irq_chip.name, pcie);
> +	if (err < 0) {
> +		dev_err(dev, "failed to request IRQ: %d\n", err);
> +		goto err;
> +	}
> +
> +	err = devm_request_irq(dev, msi->irq2, rcar_pcie_msi_irq,
> +			       IRQF_SHARED | IRQF_NO_THREAD,
> +			       rcar_msi_irq_chip.name, pcie);
> +	if (err < 0) {
> +		dev_err(dev, "failed to request IRQ: %d\n", err);
> +		goto err;
> +	}
> +
> +	/* setup MSI data target */
> +	msi->pages = __get_free_pages(GFP_KERNEL, 0);
> +	if (!msi->pages) {
> +		err = -ENOMEM;
> +		goto err;
> +	}
> +	base = virt_to_phys((void *)msi->pages);
> +
> +	rcar_pci_write_reg(pcie->base, lower_32_bits(base) | MSIFE, PCIEMSIALR);
> +	rcar_pci_write_reg(pcie->base, upper_32_bits(base), PCIEMSIAUR);
> +
> +	/* enable all MSI interrupts */
> +	rcar_pci_write_reg(pcie->base, 0xffffffff, PCIEMSIIER);
> +
> +	return 0;
> +
> +err:
> +	rcar_pcie_unmap_msi(pcie);
> +	return err;
> +}
> +
> +static void rcar_pcie_teardown_msi(struct rcar_pcie *pcie)
> +{
> +	struct rcar_msi *msi = &pcie->msi;
> +
> +	/* Disable all MSI interrupts */
> +	rcar_pci_write_reg(pcie->base, 0, PCIEMSIIER);
> +
> +	/* Disable address decoding of the MSI interrupt, MSIFE */
> +	rcar_pci_write_reg(pcie->base, 0, PCIEMSIALR);
> +
> +	free_pages(msi->pages, 0);
> +
> +	rcar_pcie_unmap_msi(pcie);
> +}
> +
> +static int rcar_pcie_get_resources(struct rcar_pcie *pcie)
> +{
> +	struct device *dev = pcie->dev;
> +	struct resource res;
> +	int err, i;
> +
> +	pcie->phy = devm_phy_optional_get(dev, "pcie");
> +	if (IS_ERR(pcie->phy))
> +		return PTR_ERR(pcie->phy);
> +
> +	err = of_address_to_resource(dev->of_node, 0, &res);
> +	if (err)
> +		return err;
> +
> +	pcie->base = devm_ioremap_resource(dev, &res);
> +	if (IS_ERR(pcie->base))
> +		return PTR_ERR(pcie->base);
> +
> +	pcie->bus_clk = devm_clk_get(dev, "pcie_bus");
> +	if (IS_ERR(pcie->bus_clk)) {
> +		dev_err(dev, "cannot get pcie bus clock\n");
> +		return PTR_ERR(pcie->bus_clk);
> +	}
> +
> +	i = irq_of_parse_and_map(dev->of_node, 0);
> +	if (!i) {
> +		dev_err(dev, "cannot get platform resources for msi interrupt\n");
> +		err = -ENOENT;
> +		goto err_irq1;
> +	}
> +	pcie->msi.irq1 = i;
> +
> +	i = irq_of_parse_and_map(dev->of_node, 1);
> +	if (!i) {
> +		dev_err(dev, "cannot get platform resources for msi interrupt\n");
> +		err = -ENOENT;
> +		goto err_irq2;
> +	}
> +	pcie->msi.irq2 = i;
> +
> +	return 0;
> +
> +err_irq2:
> +	irq_dispose_mapping(pcie->msi.irq1);
> +err_irq1:
> +	return err;
> +}
> +
> +static int rcar_pcie_inbound_ranges(struct rcar_pcie *pcie,
> +				    struct of_pci_range *range,
> +				    int *index)
> +{
> +	u64 restype = range->flags;
> +	u64 cpu_addr = range->cpu_addr;
> +	u64 cpu_end = range->cpu_addr + range->size;
> +	u64 pci_addr = range->pci_addr;
> +	u32 flags = LAM_64BIT | LAR_ENABLE;
> +	u64 mask;
> +	u64 size;
> +	int idx = *index;
> +
> +	if (restype & IORESOURCE_PREFETCH)
> +		flags |= LAM_PREFETCH;
> +
> +	/*
> +	 * If the size of the range is larger than the alignment of the start
> +	 * address, we have to use multiple entries to perform the mapping.
> +	 */
> +	if (cpu_addr > 0) {
> +		unsigned long nr_zeros = __ffs64(cpu_addr);
> +		u64 alignment = 1ULL << nr_zeros;
> +
> +		size = min(range->size, alignment);
> +	} else {
> +		size = range->size;
> +	}
> +	/* Hardware supports max 4GiB inbound region */
> +	size = min(size, 1ULL << 32);
> +
> +	mask = roundup_pow_of_two(size) - 1;
> +	mask &= ~0xf;
> +
> +	while (cpu_addr < cpu_end) {
> +		rcar_pcie_set_inbound(pcie->base, cpu_addr, pci_addr,
> +				      mask | flags, idx, true);
> +		pci_addr += size;
> +		cpu_addr += size;
> +		idx += 2;
> +
> +		if (idx > MAX_NR_INBOUND_MAPS) {
> +			dev_err(pcie->dev, "Failed to map inbound regions!\n");
> +			return -EINVAL;
> +		}
> +	}
> +	*index = idx;
> +
> +	return 0;
> +}
> +
> +static int rcar_pcie_parse_map_dma_ranges(struct rcar_pcie *pcie,
> +					  struct device_node *np)
> +{
> +	struct of_pci_range_parser parser;
> +	struct of_pci_range range;
> +	int index = 0;
> +	int err;
> +
> +	if (of_pci_dma_range_parser_init(&parser, np))
> +		return -EINVAL;
> +
> +	/* Get the dma-ranges from DT */
> +	for_each_of_pci_range(&parser, &range) {
> +		u64 end = range.cpu_addr + range.size - 1;
> +
> +		dev_dbg(pcie->dev, "0x%08x 0x%016llx..0x%016llx -> 0x%016llx\n",
> +			range.flags, range.cpu_addr, end, range.pci_addr);
> +
> +		err = rcar_pcie_inbound_ranges(pcie, &range, &index);
> +		if (err)
> +			return err;
> +	}
> +
> +	return 0;
> +}
> +
> +static const struct of_device_id rcar_pcie_of_match[] = {
> +	{ .compatible = "renesas,pcie-r8a7779",
> +	  .data = rcar_pcie_phy_init_h1 },
> +	{ .compatible = "renesas,pcie-r8a7790",
> +	  .data = rcar_pcie_phy_init_gen2 },
> +	{ .compatible = "renesas,pcie-r8a7791",
> +	  .data = rcar_pcie_phy_init_gen2 },
> +	{ .compatible = "renesas,pcie-rcar-gen2",
> +	  .data = rcar_pcie_phy_init_gen2 },
> +	{ .compatible = "renesas,pcie-r8a7795",
> +	  .data = rcar_pcie_phy_init_gen3 },
> +	{ .compatible = "renesas,pcie-rcar-gen3",
> +	  .data = rcar_pcie_phy_init_gen3 },
> +	{},
> +};
> +
> +static int rcar_pcie_probe(struct platform_device *pdev)
> +{
> +	int (*phy_init_fn)(struct rcar_pcie *pcie);
> +	struct device *dev = &pdev->dev;
> +	struct pci_host_bridge *bridge;
> +	struct rcar_pcie *pcie;
> +	u32 data;
> +	int err;
> +
> +	bridge = pci_alloc_host_bridge(sizeof(*pcie));
> +	if (!bridge)
> +		return -ENOMEM;
> +
> +	pcie = pci_host_bridge_priv(bridge);
> +
> +	pcie->dev = dev;
> +	platform_set_drvdata(pdev, pcie);
> +
> +	err = pci_parse_request_of_pci_ranges(dev, &pcie->resources, NULL);
> +	if (err)
> +		goto err_free_bridge;
> +
> +	pm_runtime_enable(pcie->dev);
> +	err = pm_runtime_get_sync(pcie->dev);
> +	if (err < 0) {
> +		dev_err(pcie->dev, "pm_runtime_get_sync failed\n");
> +		goto err_pm_disable;
> +	}
> +
> +	err = rcar_pcie_get_resources(pcie);
> +	if (err < 0) {
> +		dev_err(dev, "failed to request resources: %d\n", err);
> +		goto err_pm_put;
> +	}
> +
> +	err = clk_prepare_enable(pcie->bus_clk);
> +	if (err) {
> +		dev_err(dev, "failed to enable bus clock: %d\n", err);
> +		goto err_unmap_msi_irqs;
> +	}
> +
> +	err = rcar_pcie_parse_map_dma_ranges(pcie, dev->of_node);
> +	if (err)
> +		goto err_clk_disable;
> +
> +	phy_init_fn = of_device_get_match_data(dev);
> +	err = phy_init_fn(pcie);
> +	if (err) {
> +		dev_err(dev, "failed to init PCIe PHY\n");
> +		goto err_clk_disable;
> +	}
> +
> +	/* Failure to get a link might just be that no cards are inserted */
> +	if (rcar_pcie_hw_init(pcie)) {
> +		dev_info(dev, "PCIe link down\n");
> +		err = -ENODEV;
> +		goto err_phy_shutdown;
> +	}
> +
> +	data = rcar_pci_read_reg(pcie->base, MACSR);
> +	dev_info(dev, "PCIe x%d: link up\n", (data >> 20) & 0x3f);
> +
> +	if (IS_ENABLED(CONFIG_PCI_MSI)) {
> +		err = rcar_pcie_enable_msi(pcie);
> +		if (err < 0) {
> +			dev_err(dev,
> +				"failed to enable MSI support: %d\n",
> +				err);
> +			goto err_phy_shutdown;
> +		}
> +	}
> +
> +	err = rcar_pcie_enable(pcie);
> +	if (err)
> +		goto err_msi_teardown;
> +
> +	return 0;
> +
> +err_msi_teardown:
> +	if (IS_ENABLED(CONFIG_PCI_MSI))
> +		rcar_pcie_teardown_msi(pcie);
> +
> +err_phy_shutdown:
> +	if (pcie->phy) {
> +		phy_power_off(pcie->phy);
> +		phy_exit(pcie->phy);
> +	}
> +
> +err_clk_disable:
> +	clk_disable_unprepare(pcie->bus_clk);
> +
> +err_unmap_msi_irqs:
> +	irq_dispose_mapping(pcie->msi.irq2);
> +	irq_dispose_mapping(pcie->msi.irq1);
> +
> +err_pm_put:
> +	pm_runtime_put(dev);
> +
> +err_pm_disable:
> +	pm_runtime_disable(dev);
> +	pci_free_resource_list(&pcie->resources);
> +
> +err_free_bridge:
> +	pci_free_host_bridge(bridge);
> +
> +	return err;
> +}
> +
> +static int rcar_pcie_resume_noirq(struct device *dev)
> +{
> +	struct rcar_pcie *pcie = dev_get_drvdata(dev);
> +
> +	if (rcar_pci_read_reg(pcie->base, PMSR) &&
> +	    !(rcar_pci_read_reg(pcie->base, PCIETCTLR) & DL_DOWN))
> +		return 0;
> +
> +	/* Re-establish the PCIe link */
> +	rcar_rmw32(pcie->base, MACCTLR, MACCTLR_RESERVED, 0);
> +	rcar_pci_write_reg(pcie->base, CFINIT, PCIETCTLR);
> +	return rcar_pcie_wait_for_dl(pcie->base);
> +}
> +
> +static const struct dev_pm_ops rcar_pcie_pm_ops = {
> +	.resume_noirq = rcar_pcie_resume_noirq,
> +};
> +
> +static struct platform_driver rcar_pcie_driver = {
> +	.driver = {
> +		.name = "rcar-pcie",
> +		.of_match_table = rcar_pcie_of_match,
> +		.pm = &rcar_pcie_pm_ops,
> +		.suppress_bind_attrs = true,
> +	},
> +	.probe = rcar_pcie_probe,
> +};
> +builtin_platform_driver(rcar_pcie_driver);
> diff --git a/drivers/pci/controller/pcie-rcar.c b/drivers/pci/controller/pcie-rcar.c
> index e45bb2a7bfa5..1008ae6a896c 100644
> --- a/drivers/pci/controller/pcie-rcar.c
> +++ b/drivers/pci/controller/pcie-rcar.c
> @@ -1,7 +1,7 @@
>  // SPDX-License-Identifier: GPL-2.0
>  /*
>   * PCIe driver for Renesas R-Car SoCs
> - *  Copyright (C) 2014 Renesas Electronics Europe Ltd
> + *  Copyright (C) 2014-2019 Renesas Electronics Europe Ltd
>   *
>   * Based on:
>   *  arch/sh/drivers/pci/pcie-sh7786.c
> @@ -11,534 +11,80 @@
>   * Author: Phil Edworthy <phil.edworthy@renesas.com>
>   */
>  
> -#include <linux/bitops.h>
> -#include <linux/clk.h>
>  #include <linux/delay.h>
> -#include <linux/interrupt.h>
> -#include <linux/irq.h>
> -#include <linux/irqdomain.h>
> -#include <linux/kernel.h>
> -#include <linux/init.h>
> -#include <linux/msi.h>
> -#include <linux/of_address.h>
> -#include <linux/of_irq.h>
> -#include <linux/of_pci.h>
> -#include <linux/of_platform.h>
>  #include <linux/pci.h>
> -#include <linux/phy/phy.h>
> -#include <linux/platform_device.h>
> -#include <linux/pm_runtime.h>
> -#include <linux/slab.h>
>  
> -#define PCIECAR			0x000010
> -#define PCIECCTLR		0x000018
> -#define  CONFIG_SEND_ENABLE	BIT(31)
> -#define  TYPE0			(0 << 8)
> -#define  TYPE1			BIT(8)
> -#define PCIECDR			0x000020
> -#define PCIEMSR			0x000028
> -#define PCIEINTXR		0x000400
> -#define PCIEPHYSR		0x0007f0
> -#define  PHYRDY			BIT(0)
> -#define PCIEMSITXR		0x000840
> +#include "pcie-rcar.h"
>  
> -/* Transfer control */
> -#define PCIETCTLR		0x02000
> -#define  DL_DOWN		BIT(3)
> -#define  CFINIT			BIT(0)
> -#define PCIETSTR		0x02004
> -#define  DATA_LINK_ACTIVE	BIT(0)
> -#define PCIEERRFR		0x02020
> -#define  UNSUPPORTED_REQUEST	BIT(4)
> -#define PCIEMSIFR		0x02044
> -#define PCIEMSIALR		0x02048
> -#define  MSIFE			BIT(0)
> -#define PCIEMSIAUR		0x0204c
> -#define PCIEMSIIER		0x02050
> -
> -/* root port address */
> -#define PCIEPRAR(x)		(0x02080 + ((x) * 0x4))
> -
> -/* local address reg & mask */
> -#define PCIELAR(x)		(0x02200 + ((x) * 0x20))
> -#define PCIELAMR(x)		(0x02208 + ((x) * 0x20))
> -#define  LAM_PREFETCH		BIT(3)
> -#define  LAM_64BIT		BIT(2)
> -#define  LAR_ENABLE		BIT(1)
> -
> -/* PCIe address reg & mask */
> -#define PCIEPALR(x)		(0x03400 + ((x) * 0x20))
> -#define PCIEPAUR(x)		(0x03404 + ((x) * 0x20))
> -#define PCIEPAMR(x)		(0x03408 + ((x) * 0x20))
> -#define PCIEPTCTLR(x)		(0x0340c + ((x) * 0x20))
> -#define  PAR_ENABLE		BIT(31)
> -#define  IO_SPACE		BIT(8)
> -
> -/* Configuration */
> -#define PCICONF(x)		(0x010000 + ((x) * 0x4))
> -#define PMCAP(x)		(0x010040 + ((x) * 0x4))
> -#define EXPCAP(x)		(0x010070 + ((x) * 0x4))
> -#define VCCAP(x)		(0x010100 + ((x) * 0x4))
> -
> -/* link layer */
> -#define IDSETR1			0x011004
> -#define TLCTLR			0x011048
> -#define MACSR			0x011054
> -#define  SPCHGFIN		BIT(4)
> -#define  SPCHGFAIL		BIT(6)
> -#define  SPCHGSUC		BIT(7)
> -#define  LINK_SPEED		(0xf << 16)
> -#define  LINK_SPEED_2_5GTS	(1 << 16)
> -#define  LINK_SPEED_5_0GTS	(2 << 16)
> -#define MACCTLR			0x011058
> -#define  MACCTLR_RESERVED	BIT(0)
> -#define  SPEED_CHANGE		BIT(24)
> -#define  SCRAMBLE_DISABLE	BIT(27)
> -#define PMSR			0x01105c
> -#define MACS2R			0x011078
> -#define MACCGSPSETR		0x011084
> -#define  SPCNGRSN		BIT(31)
> -
> -/* R-Car H1 PHY */
> -#define H1_PCIEPHYADRR		0x04000c
> -#define  WRITE_CMD		BIT(16)
> -#define  PHY_ACK		BIT(24)
> -#define  RATE_POS		12
> -#define  LANE_POS		8
> -#define  ADR_POS		0
> -#define H1_PCIEPHYDOUTR		0x040014
> -
> -/* R-Car Gen2 PHY */
> -#define GEN2_PCIEPHYADDR	0x780
> -#define GEN2_PCIEPHYDATA	0x784
> -#define GEN2_PCIEPHYCTRL	0x78c
> -
> -#define INT_PCI_MSI_NR		32
> -
> -#define RCONF(x)		(PCICONF(0) + (x))
> -#define RPMCAP(x)		(PMCAP(0) + (x))
> -#define REXPCAP(x)		(EXPCAP(0) + (x))
> -#define RVCCAP(x)		(VCCAP(0) + (x))
> -
> -#define PCIE_CONF_BUS(b)	(((b) & 0xff) << 24)
> -#define PCIE_CONF_DEV(d)	(((d) & 0x1f) << 19)
> -#define PCIE_CONF_FUNC(f)	(((f) & 0x7) << 16)
> -
> -#define RCAR_PCI_MAX_RESOURCES	4
> -#define MAX_NR_INBOUND_MAPS	6
> -
> -struct rcar_msi {
> -	DECLARE_BITMAP(used, INT_PCI_MSI_NR);
> -	struct irq_domain *domain;
> -	struct msi_controller chip;
> -	unsigned long pages;
> -	struct mutex lock;
> -	int irq1;
> -	int irq2;
> -};
> -
> -static inline struct rcar_msi *to_rcar_msi(struct msi_controller *chip)
> -{
> -	return container_of(chip, struct rcar_msi, chip);
> -}
> -
> -/* Structure representing the PCIe interface */
> -struct rcar_pcie {
> -	struct device		*dev;
> -	struct phy		*phy;
> -	void __iomem		*base;
> -	struct list_head	resources;
> -	int			root_bus_nr;
> -	struct clk		*bus_clk;
> -	struct			rcar_msi msi;
> -};
> -
> -static void rcar_pci_write_reg(struct rcar_pcie *pcie, u32 val,
> -			       unsigned int reg)
> +void rcar_pci_write_reg(void __iomem *base, u32 val, unsigned int reg)
>  {
> -	writel(val, pcie->base + reg);
> +	writel(val, base + reg);
>  }
>  
> -static u32 rcar_pci_read_reg(struct rcar_pcie *pcie, unsigned int reg)
> +u32 rcar_pci_read_reg(void __iomem *base, unsigned int reg)
>  {
> -	return readl(pcie->base + reg);
> +	return readl(base + reg);
>  }
>  
> -enum {
> -	RCAR_PCI_ACCESS_READ,
> -	RCAR_PCI_ACCESS_WRITE,
> -};
> -
> -static void rcar_rmw32(struct rcar_pcie *pcie, int where, u32 mask, u32 data)
> +void rcar_rmw32(void __iomem *base, int where, u32 mask, u32 data)
>  {
>  	unsigned int shift = BITS_PER_BYTE * (where & 3);
> -	u32 val = rcar_pci_read_reg(pcie, where & ~3);
> +	u32 val = rcar_pci_read_reg(base, where & ~3);
>  
>  	val &= ~(mask << shift);
>  	val |= data << shift;
> -	rcar_pci_write_reg(pcie, val, where & ~3);
> -}
> -
> -static u32 rcar_read_conf(struct rcar_pcie *pcie, int where)
> -{
> -	unsigned int shift = BITS_PER_BYTE * (where & 3);
> -	u32 val = rcar_pci_read_reg(pcie, where & ~3);
> -
> -	return val >> shift;
> +	rcar_pci_write_reg(base, val, where & ~3);
>  }
>  
> -/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
> -static int rcar_pcie_config_access(struct rcar_pcie *pcie,
> -		unsigned char access_type, struct pci_bus *bus,
> -		unsigned int devfn, int where, u32 *data)
> -{
> -	unsigned int dev, func, reg, index;
> -
> -	dev = PCI_SLOT(devfn);
> -	func = PCI_FUNC(devfn);
> -	reg = where & ~3;
> -	index = reg / 4;
> -
> -	/*
> -	 * While each channel has its own memory-mapped extended config
> -	 * space, it's generally only accessible when in endpoint mode.
> -	 * When in root complex mode, the controller is unable to target
> -	 * itself with either type 0 or type 1 accesses, and indeed, any
> -	 * controller initiated target transfer to its own config space
> -	 * result in a completer abort.
> -	 *
> -	 * Each channel effectively only supports a single device, but as
> -	 * the same channel <-> device access works for any PCI_SLOT()
> -	 * value, we cheat a bit here and bind the controller's config
> -	 * space to devfn 0 in order to enable self-enumeration. In this
> -	 * case the regular ECAR/ECDR path is sidelined and the mangled
> -	 * config access itself is initiated as an internal bus transaction.
> -	 */
> -	if (pci_is_root_bus(bus)) {
> -		if (dev != 0)
> -			return PCIBIOS_DEVICE_NOT_FOUND;
> -
> -		if (access_type == RCAR_PCI_ACCESS_READ) {
> -			*data = rcar_pci_read_reg(pcie, PCICONF(index));
> -		} else {
> -			/* Keep an eye out for changes to the root bus number */
> -			if (pci_is_root_bus(bus) && (reg == PCI_PRIMARY_BUS))
> -				pcie->root_bus_nr = *data & 0xff;
> -
> -			rcar_pci_write_reg(pcie, *data, PCICONF(index));
> -		}
> -
> -		return PCIBIOS_SUCCESSFUL;
> -	}
> -
> -	if (pcie->root_bus_nr < 0)
> -		return PCIBIOS_DEVICE_NOT_FOUND;
> -
> -	/* Clear errors */
> -	rcar_pci_write_reg(pcie, rcar_pci_read_reg(pcie, PCIEERRFR), PCIEERRFR);
> -
> -	/* Set the PIO address */
> -	rcar_pci_write_reg(pcie, PCIE_CONF_BUS(bus->number) |
> -		PCIE_CONF_DEV(dev) | PCIE_CONF_FUNC(func) | reg, PCIECAR);
> -
> -	/* Enable the configuration access */
> -	if (bus->parent->number == pcie->root_bus_nr)
> -		rcar_pci_write_reg(pcie, CONFIG_SEND_ENABLE | TYPE0, PCIECCTLR);
> -	else
> -		rcar_pci_write_reg(pcie, CONFIG_SEND_ENABLE | TYPE1, PCIECCTLR);
> -
> -	/* Check for errors */
> -	if (rcar_pci_read_reg(pcie, PCIEERRFR) & UNSUPPORTED_REQUEST)
> -		return PCIBIOS_DEVICE_NOT_FOUND;
> -
> -	/* Check for master and target aborts */
> -	if (rcar_read_conf(pcie, RCONF(PCI_STATUS)) &
> -		(PCI_STATUS_REC_MASTER_ABORT | PCI_STATUS_REC_TARGET_ABORT))
> -		return PCIBIOS_DEVICE_NOT_FOUND;
> -
> -	if (access_type == RCAR_PCI_ACCESS_READ)
> -		*data = rcar_pci_read_reg(pcie, PCIECDR);
> -	else
> -		rcar_pci_write_reg(pcie, *data, PCIECDR);
> -
> -	/* Disable the configuration access */
> -	rcar_pci_write_reg(pcie, 0, PCIECCTLR);
> -
> -	return PCIBIOS_SUCCESSFUL;
> -}
> -
> -static int rcar_pcie_read_conf(struct pci_bus *bus, unsigned int devfn,
> -			       int where, int size, u32 *val)
> -{
> -	struct rcar_pcie *pcie = bus->sysdata;
> -	int ret;
> -
> -	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
> -				      bus, devfn, where, val);
> -	if (ret != PCIBIOS_SUCCESSFUL) {
> -		*val = 0xffffffff;
> -		return ret;
> -	}
> -
> -	if (size == 1)
> -		*val = (*val >> (BITS_PER_BYTE * (where & 3))) & 0xff;
> -	else if (size == 2)
> -		*val = (*val >> (BITS_PER_BYTE * (where & 2))) & 0xffff;
> -
> -	dev_dbg(&bus->dev, "pcie-config-read: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
> -		bus->number, devfn, where, size, *val);
> -
> -	return ret;
> -}
> -
> -/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
> -static int rcar_pcie_write_conf(struct pci_bus *bus, unsigned int devfn,
> -				int where, int size, u32 val)
> -{
> -	struct rcar_pcie *pcie = bus->sysdata;
> -	unsigned int shift;
> -	u32 data;
> -	int ret;
> -
> -	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
> -				      bus, devfn, where, &data);
> -	if (ret != PCIBIOS_SUCCESSFUL)
> -		return ret;
> -
> -	dev_dbg(&bus->dev, "pcie-config-write: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
> -		bus->number, devfn, where, size, val);
> -
> -	if (size == 1) {
> -		shift = BITS_PER_BYTE * (where & 3);
> -		data &= ~(0xff << shift);
> -		data |= ((val & 0xff) << shift);
> -	} else if (size == 2) {
> -		shift = BITS_PER_BYTE * (where & 2);
> -		data &= ~(0xffff << shift);
> -		data |= ((val & 0xffff) << shift);
> -	} else
> -		data = val;
> -
> -	ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_WRITE,
> -				      bus, devfn, where, &data);
> -
> -	return ret;
> -}
> -
> -static struct pci_ops rcar_pcie_ops = {
> -	.read	= rcar_pcie_read_conf,
> -	.write	= rcar_pcie_write_conf,
> -};
> -
> -static void rcar_pcie_setup_window(int win, struct rcar_pcie *pcie,
> -				   struct resource *res)
> +void rcar_pcie_set_outbound(int win, void __iomem *base,
> +			    struct resource *res, bool host)
>  {
>  	/* Setup PCIe address space mappings for each resource */
> -	resource_size_t size;
> -	resource_size_t res_start;
> -	u32 mask;
> -
> -	rcar_pci_write_reg(pcie, 0x00000000, PCIEPTCTLR(win));
> +	resource_size_t size = 0;
> +	resource_size_t res_start = 0;
> +	u32 mask = 0x0;
>  
> +	rcar_pci_write_reg(base, mask, PCIEPTCTLR(win));
>  	/*
>  	 * The PAMR mask is calculated in units of 128Bytes, which
>  	 * keeps things pretty simple.
>  	 */
>  	size = resource_size(res);
> -	mask = (roundup_pow_of_two(size) / SZ_128) - 1;
> -	rcar_pci_write_reg(pcie, mask << 7, PCIEPAMR(win));
> +	if (size > 128)
> +		mask = (roundup_pow_of_two(size) / SZ_128) - 1;
> +	rcar_pci_write_reg(base, mask << 7, PCIEPAMR(win));
>  
> -	if (res->flags & IORESOURCE_IO)
> -		res_start = pci_pio_to_address(res->start);
> -	else
> +	if (!host) {
>  		res_start = res->start;
> +	} else {
> +		if (res->flags & IORESOURCE_IO)
> +			res_start = pci_pio_to_address(res->start);
> +		else
> +			res_start = res->start;
> +	}
>  
> -	rcar_pci_write_reg(pcie, upper_32_bits(res_start), PCIEPAUR(win));
> -	rcar_pci_write_reg(pcie, lower_32_bits(res_start) & ~0x7F,
> +	rcar_pci_write_reg(base, upper_32_bits(res_start), PCIEPAUR(win));
> +	rcar_pci_write_reg(base, lower_32_bits(res_start) & ~0x7F,
>  			   PCIEPALR(win));
>  
> -	/* First resource is for IO */
> -	mask = PAR_ENABLE;
> +	mask = 0x0;
> +	if (res->start)
> +		mask = PAR_ENABLE;
> +
>  	if (res->flags & IORESOURCE_IO)
>  		mask |= IO_SPACE;
>  
> -	rcar_pci_write_reg(pcie, mask, PCIEPTCTLR(win));
> -}
> -
> -static int rcar_pcie_setup(struct list_head *resource, struct rcar_pcie *pci)
> -{
> -	struct resource_entry *win;
> -	int i = 0;
> -
> -	/* Setup PCI resources */
> -	resource_list_for_each_entry(win, &pci->resources) {
> -		struct resource *res = win->res;
> -
> -		if (!res->flags)
> -			continue;
> -
> -		switch (resource_type(res)) {
> -		case IORESOURCE_IO:
> -		case IORESOURCE_MEM:
> -			rcar_pcie_setup_window(i, pci, res);
> -			i++;
> -			break;
> -		case IORESOURCE_BUS:
> -			pci->root_bus_nr = res->start;
> -			break;
> -		default:
> -			continue;
> -		}
> -
> -		pci_add_resource(resource, res);
> -	}
> -
> -	return 1;
> -}
> -
> -static void rcar_pcie_force_speedup(struct rcar_pcie *pcie)
> -{
> -	struct device *dev = pcie->dev;
> -	unsigned int timeout = 1000;
> -	u32 macsr;
> -
> -	if ((rcar_pci_read_reg(pcie, MACS2R) & LINK_SPEED) != LINK_SPEED_5_0GTS)
> -		return;
> -
> -	if (rcar_pci_read_reg(pcie, MACCTLR) & SPEED_CHANGE) {
> -		dev_err(dev, "Speed change already in progress\n");
> -		return;
> -	}
> -
> -	macsr = rcar_pci_read_reg(pcie, MACSR);
> -	if ((macsr & LINK_SPEED) == LINK_SPEED_5_0GTS)
> -		goto done;
> -
> -	/* Set target link speed to 5.0 GT/s */
> -	rcar_rmw32(pcie, EXPCAP(12), PCI_EXP_LNKSTA_CLS,
> -		   PCI_EXP_LNKSTA_CLS_5_0GB);
> -
> -	/* Set speed change reason as intentional factor */
> -	rcar_rmw32(pcie, MACCGSPSETR, SPCNGRSN, 0);
> -
> -	/* Clear SPCHGFIN, SPCHGSUC, and SPCHGFAIL */
> -	if (macsr & (SPCHGFIN | SPCHGSUC | SPCHGFAIL))
> -		rcar_pci_write_reg(pcie, macsr, MACSR);
> -
> -	/* Start link speed change */
> -	rcar_rmw32(pcie, MACCTLR, SPEED_CHANGE, SPEED_CHANGE);
> -
> -	while (timeout--) {
> -		macsr = rcar_pci_read_reg(pcie, MACSR);
> -		if (macsr & SPCHGFIN) {
> -			/* Clear the interrupt bits */
> -			rcar_pci_write_reg(pcie, macsr, MACSR);
> -
> -			if (macsr & SPCHGFAIL)
> -				dev_err(dev, "Speed change failed\n");
> -
> -			goto done;
> -		}
> -
> -		msleep(1);
> -	}
> -
> -	dev_err(dev, "Speed change timed out\n");
> -
> -done:
> -	dev_info(dev, "Current link speed is %s GT/s\n",
> -		 (macsr & LINK_SPEED) == LINK_SPEED_5_0GTS ? "5" : "2.5");
> -}
> -
> -static int rcar_pcie_enable(struct rcar_pcie *pcie)
> -{
> -	struct device *dev = pcie->dev;
> -	struct pci_host_bridge *bridge = pci_host_bridge_from_priv(pcie);
> -	struct pci_bus *bus, *child;
> -	int ret;
> -
> -	/* Try setting 5 GT/s link speed */
> -	rcar_pcie_force_speedup(pcie);
> -
> -	rcar_pcie_setup(&bridge->windows, pcie);
> -
> -	pci_add_flags(PCI_REASSIGN_ALL_BUS);
> -
> -	bridge->dev.parent = dev;
> -	bridge->sysdata = pcie;
> -	bridge->busnr = pcie->root_bus_nr;
> -	bridge->ops = &rcar_pcie_ops;
> -	bridge->map_irq = of_irq_parse_and_map_pci;
> -	bridge->swizzle_irq = pci_common_swizzle;
> -	if (IS_ENABLED(CONFIG_PCI_MSI))
> -		bridge->msi = &pcie->msi.chip;
> -
> -	ret = pci_scan_root_bus_bridge(bridge);
> -	if (ret < 0)
> -		return ret;
> -
> -	bus = bridge->bus;
> -
> -	pci_bus_size_bridges(bus);
> -	pci_bus_assign_resources(bus);
> -
> -	list_for_each_entry(child, &bus->children, node)
> -		pcie_bus_configure_settings(child);
> -
> -	pci_bus_add_devices(bus);
> -
> -	return 0;
> -}
> -
> -static int phy_wait_for_ack(struct rcar_pcie *pcie)
> -{
> -	struct device *dev = pcie->dev;
> -	unsigned int timeout = 100;
> -
> -	while (timeout--) {
> -		if (rcar_pci_read_reg(pcie, H1_PCIEPHYADRR) & PHY_ACK)
> -			return 0;
> -
> -		udelay(100);
> -	}
> -
> -	dev_err(dev, "Access to PCIe phy timed out\n");
> -
> -	return -ETIMEDOUT;
> -}
> -
> -static void phy_write_reg(struct rcar_pcie *pcie,
> -			  unsigned int rate, u32 addr,
> -			  unsigned int lane, u32 data)
> -{
> -	u32 phyaddr;
> -
> -	phyaddr = WRITE_CMD |
> -		((rate & 1) << RATE_POS) |
> -		((lane & 0xf) << LANE_POS) |
> -		((addr & 0xff) << ADR_POS);
> -
> -	/* Set write data */
> -	rcar_pci_write_reg(pcie, data, H1_PCIEPHYDOUTR);
> -	rcar_pci_write_reg(pcie, phyaddr, H1_PCIEPHYADRR);
> -
> -	/* Ignore errors as they will be dealt with if the data link is down */
> -	phy_wait_for_ack(pcie);
> -
> -	/* Clear command */
> -	rcar_pci_write_reg(pcie, 0, H1_PCIEPHYDOUTR);
> -	rcar_pci_write_reg(pcie, 0, H1_PCIEPHYADRR);
> -
> -	/* Ignore errors as they will be dealt with if the data link is down */
> -	phy_wait_for_ack(pcie);
> +	rcar_pci_write_reg(base, mask, PCIEPTCTLR(win));
> +	/* flush modifications */
> +	wmb();
>  }
>  
> -static int rcar_pcie_wait_for_phyrdy(struct rcar_pcie *pcie)
> +int rcar_pcie_wait_for_phyrdy(void __iomem *base)
>  {
>  	unsigned int timeout = 10;
>  
>  	while (timeout--) {
> -		if (rcar_pci_read_reg(pcie, PCIEPHYSR) & PHYRDY)
> +		if (rcar_pci_read_reg(base, PCIEPHYSR) & PHYRDY)
>  			return 0;
>  
>  		msleep(5);
> @@ -547,12 +93,12 @@ static int rcar_pcie_wait_for_phyrdy(struct rcar_pcie *pcie)
>  	return -ETIMEDOUT;
>  }
>  
> -static int rcar_pcie_wait_for_dl(struct rcar_pcie *pcie)
> +int rcar_pcie_wait_for_dl(void __iomem *base)
>  {
>  	unsigned int timeout = 10000;
>  
>  	while (timeout--) {
> -		if ((rcar_pci_read_reg(pcie, PCIETSTR) & DATA_LINK_ACTIVE))
> +		if ((rcar_pci_read_reg(base, PCIETSTR) & DATA_LINK_ACTIVE))
>  			return 0;
>  
>  		udelay(5);
> @@ -562,698 +108,27 @@ static int rcar_pcie_wait_for_dl(struct rcar_pcie *pcie)
>  	return -ETIMEDOUT;
>  }
>  
> -static int rcar_pcie_hw_init(struct rcar_pcie *pcie)
> -{
> -	int err;
> -
> -	/* Begin initialization */
> -	rcar_pci_write_reg(pcie, 0, PCIETCTLR);
> -
> -	/* Set mode */
> -	rcar_pci_write_reg(pcie, 1, PCIEMSR);
> -
> -	err = rcar_pcie_wait_for_phyrdy(pcie);
> -	if (err)
> -		return err;
> -
> -	/*
> -	 * Initial header for port config space is type 1, set the device
> -	 * class to match. Hardware takes care of propagating the IDSETR
> -	 * settings, so there is no need to bother with a quirk.
> -	 */
> -	rcar_pci_write_reg(pcie, PCI_CLASS_BRIDGE_PCI << 16, IDSETR1);
> -
> -	/*
> -	 * Setup Secondary Bus Number & Subordinate Bus Number, even though
> -	 * they aren't used, to avoid bridge being detected as broken.
> -	 */
> -	rcar_rmw32(pcie, RCONF(PCI_SECONDARY_BUS), 0xff, 1);
> -	rcar_rmw32(pcie, RCONF(PCI_SUBORDINATE_BUS), 0xff, 1);
> -
> -	/* Initialize default capabilities. */
> -	rcar_rmw32(pcie, REXPCAP(0), 0xff, PCI_CAP_ID_EXP);
> -	rcar_rmw32(pcie, REXPCAP(PCI_EXP_FLAGS),
> -		PCI_EXP_FLAGS_TYPE, PCI_EXP_TYPE_ROOT_PORT << 4);
> -	rcar_rmw32(pcie, RCONF(PCI_HEADER_TYPE), 0x7f,
> -		PCI_HEADER_TYPE_BRIDGE);
> -
> -	/* Enable data link layer active state reporting */
> -	rcar_rmw32(pcie, REXPCAP(PCI_EXP_LNKCAP), PCI_EXP_LNKCAP_DLLLARC,
> -		PCI_EXP_LNKCAP_DLLLARC);
> -
> -	/* Write out the physical slot number = 0 */
> -	rcar_rmw32(pcie, REXPCAP(PCI_EXP_SLTCAP), PCI_EXP_SLTCAP_PSN, 0);
> -
> -	/* Set the completion timer timeout to the maximum 50ms. */
> -	rcar_rmw32(pcie, TLCTLR + 1, 0x3f, 50);
> -
> -	/* Terminate list of capabilities (Next Capability Offset=0) */
> -	rcar_rmw32(pcie, RVCCAP(0), 0xfff00000, 0);
> -
> -	/* Enable MSI */
> -	if (IS_ENABLED(CONFIG_PCI_MSI))
> -		rcar_pci_write_reg(pcie, 0x801f0000, PCIEMSITXR);
> -
> -	rcar_rmw32(pcie, MACCTLR, MACCTLR_RESERVED, 0);
> -
> -	/* Finish initialization - establish a PCI Express link */
> -	rcar_pci_write_reg(pcie, CFINIT, PCIETCTLR);
> -
> -	/* This will timeout if we don't have a link. */
> -	err = rcar_pcie_wait_for_dl(pcie);
> -	if (err)
> -		return err;
> -
> -	/* Enable INTx interrupts */
> -	rcar_rmw32(pcie, PCIEINTXR, 0, 0xF << 8);
> -
> -	wmb();
> -
> -	return 0;
> -}
> -
> -static int rcar_pcie_phy_init_h1(struct rcar_pcie *pcie)
> -{
> -	/* Initialize the phy */
> -	phy_write_reg(pcie, 0, 0x42, 0x1, 0x0EC34191);
> -	phy_write_reg(pcie, 1, 0x42, 0x1, 0x0EC34180);
> -	phy_write_reg(pcie, 0, 0x43, 0x1, 0x00210188);
> -	phy_write_reg(pcie, 1, 0x43, 0x1, 0x00210188);
> -	phy_write_reg(pcie, 0, 0x44, 0x1, 0x015C0014);
> -	phy_write_reg(pcie, 1, 0x44, 0x1, 0x015C0014);
> -	phy_write_reg(pcie, 1, 0x4C, 0x1, 0x786174A0);
> -	phy_write_reg(pcie, 1, 0x4D, 0x1, 0x048000BB);
> -	phy_write_reg(pcie, 0, 0x51, 0x1, 0x079EC062);
> -	phy_write_reg(pcie, 0, 0x52, 0x1, 0x20000000);
> -	phy_write_reg(pcie, 1, 0x52, 0x1, 0x20000000);
> -	phy_write_reg(pcie, 1, 0x56, 0x1, 0x00003806);
> -
> -	phy_write_reg(pcie, 0, 0x60, 0x1, 0x004B03A5);
> -	phy_write_reg(pcie, 0, 0x64, 0x1, 0x3F0F1F0F);
> -	phy_write_reg(pcie, 0, 0x66, 0x1, 0x00008000);
> -
> -	return 0;
> -}
> -
> -static int rcar_pcie_phy_init_gen2(struct rcar_pcie *pcie)
> -{
> -	/*
> -	 * These settings come from the R-Car Series, 2nd Generation User's
> -	 * Manual, section 50.3.1 (2) Initialization of the physical layer.
> -	 */
> -	rcar_pci_write_reg(pcie, 0x000f0030, GEN2_PCIEPHYADDR);
> -	rcar_pci_write_reg(pcie, 0x00381203, GEN2_PCIEPHYDATA);
> -	rcar_pci_write_reg(pcie, 0x00000001, GEN2_PCIEPHYCTRL);
> -	rcar_pci_write_reg(pcie, 0x00000006, GEN2_PCIEPHYCTRL);
> -
> -	rcar_pci_write_reg(pcie, 0x000f0054, GEN2_PCIEPHYADDR);
> -	/* The following value is for DC connection, no termination resistor */
> -	rcar_pci_write_reg(pcie, 0x13802007, GEN2_PCIEPHYDATA);
> -	rcar_pci_write_reg(pcie, 0x00000001, GEN2_PCIEPHYCTRL);
> -	rcar_pci_write_reg(pcie, 0x00000006, GEN2_PCIEPHYCTRL);
> -
> -	return 0;
> -}
> -
> -static int rcar_pcie_phy_init_gen3(struct rcar_pcie *pcie)
> -{
> -	int err;
> -
> -	err = phy_init(pcie->phy);
> -	if (err)
> -		return err;
> -
> -	err = phy_power_on(pcie->phy);
> -	if (err)
> -		phy_exit(pcie->phy);
> -
> -	return err;
> -}
> -
> -static int rcar_msi_alloc(struct rcar_msi *chip)
> -{
> -	int msi;
> -
> -	mutex_lock(&chip->lock);
> -
> -	msi = find_first_zero_bit(chip->used, INT_PCI_MSI_NR);
> -	if (msi < INT_PCI_MSI_NR)
> -		set_bit(msi, chip->used);
> -	else
> -		msi = -ENOSPC;
> -
> -	mutex_unlock(&chip->lock);
> -
> -	return msi;
> -}
> -
> -static int rcar_msi_alloc_region(struct rcar_msi *chip, int no_irqs)
> -{
> -	int msi;
> -
> -	mutex_lock(&chip->lock);
> -	msi = bitmap_find_free_region(chip->used, INT_PCI_MSI_NR,
> -				      order_base_2(no_irqs));
> -	mutex_unlock(&chip->lock);
> -
> -	return msi;
> -}
> -
> -static void rcar_msi_free(struct rcar_msi *chip, unsigned long irq)
> -{
> -	mutex_lock(&chip->lock);
> -	clear_bit(irq, chip->used);
> -	mutex_unlock(&chip->lock);
> -}
> -
> -static irqreturn_t rcar_pcie_msi_irq(int irq, void *data)
> -{
> -	struct rcar_pcie *pcie = data;
> -	struct rcar_msi *msi = &pcie->msi;
> -	struct device *dev = pcie->dev;
> -	unsigned long reg;
> -
> -	reg = rcar_pci_read_reg(pcie, PCIEMSIFR);
> -
> -	/* MSI & INTx share an interrupt - we only handle MSI here */
> -	if (!reg)
> -		return IRQ_NONE;
> -
> -	while (reg) {
> -		unsigned int index = find_first_bit(&reg, 32);
> -		unsigned int msi_irq;
> -
> -		/* clear the interrupt */
> -		rcar_pci_write_reg(pcie, 1 << index, PCIEMSIFR);
> -
> -		msi_irq = irq_find_mapping(msi->domain, index);
> -		if (msi_irq) {
> -			if (test_bit(index, msi->used))
> -				generic_handle_irq(msi_irq);
> -			else
> -				dev_info(dev, "unhandled MSI\n");
> -		} else {
> -			/* Unknown MSI, just clear it */
> -			dev_dbg(dev, "unexpected MSI\n");
> -		}
> -
> -		/* see if there's any more pending in this vector */
> -		reg = rcar_pci_read_reg(pcie, PCIEMSIFR);
> -	}
> -
> -	return IRQ_HANDLED;
> -}
> -
> -static int rcar_msi_setup_irq(struct msi_controller *chip, struct pci_dev *pdev,
> -			      struct msi_desc *desc)
> -{
> -	struct rcar_msi *msi = to_rcar_msi(chip);
> -	struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
> -	struct msi_msg msg;
> -	unsigned int irq;
> -	int hwirq;
> -
> -	hwirq = rcar_msi_alloc(msi);
> -	if (hwirq < 0)
> -		return hwirq;
> -
> -	irq = irq_find_mapping(msi->domain, hwirq);
> -	if (!irq) {
> -		rcar_msi_free(msi, hwirq);
> -		return -EINVAL;
> -	}
> -
> -	irq_set_msi_desc(irq, desc);
> -
> -	msg.address_lo = rcar_pci_read_reg(pcie, PCIEMSIALR) & ~MSIFE;
> -	msg.address_hi = rcar_pci_read_reg(pcie, PCIEMSIAUR);
> -	msg.data = hwirq;
> -
> -	pci_write_msi_msg(irq, &msg);
> -
> -	return 0;
> -}
> -
> -static int rcar_msi_setup_irqs(struct msi_controller *chip,
> -			       struct pci_dev *pdev, int nvec, int type)
> -{
> -	struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
> -	struct rcar_msi *msi = to_rcar_msi(chip);
> -	struct msi_desc *desc;
> -	struct msi_msg msg;
> -	unsigned int irq;
> -	int hwirq;
> -	int i;
> -
> -	/* MSI-X interrupts are not supported */
> -	if (type == PCI_CAP_ID_MSIX)
> -		return -EINVAL;
> -
> -	WARN_ON(!list_is_singular(&pdev->dev.msi_list));
> -	desc = list_entry(pdev->dev.msi_list.next, struct msi_desc, list);
> -
> -	hwirq = rcar_msi_alloc_region(msi, nvec);
> -	if (hwirq < 0)
> -		return -ENOSPC;
> -
> -	irq = irq_find_mapping(msi->domain, hwirq);
> -	if (!irq)
> -		return -ENOSPC;
> -
> -	for (i = 0; i < nvec; i++) {
> -		/*
> -		 * irq_create_mapping() called from rcar_pcie_probe() pre-
> -		 * allocates descs,  so there is no need to allocate descs here.
> -		 * We can therefore assume that if irq_find_mapping() above
> -		 * returns non-zero, then the descs are also successfully
> -		 * allocated.
> -		 */
> -		if (irq_set_msi_desc_off(irq, i, desc)) {
> -			/* TODO: clear */
> -			return -EINVAL;
> -		}
> -	}
> -
> -	desc->nvec_used = nvec;
> -	desc->msi_attrib.multiple = order_base_2(nvec);
> -
> -	msg.address_lo = rcar_pci_read_reg(pcie, PCIEMSIALR) & ~MSIFE;
> -	msg.address_hi = rcar_pci_read_reg(pcie, PCIEMSIAUR);
> -	msg.data = hwirq;
> -
> -	pci_write_msi_msg(irq, &msg);
> -
> -	return 0;
> -}
> -
> -static void rcar_msi_teardown_irq(struct msi_controller *chip, unsigned int irq)
> -{
> -	struct rcar_msi *msi = to_rcar_msi(chip);
> -	struct irq_data *d = irq_get_irq_data(irq);
> -
> -	rcar_msi_free(msi, d->hwirq);
> -}
> -
> -static struct irq_chip rcar_msi_irq_chip = {
> -	.name = "R-Car PCIe MSI",
> -	.irq_enable = pci_msi_unmask_irq,
> -	.irq_disable = pci_msi_mask_irq,
> -	.irq_mask = pci_msi_mask_irq,
> -	.irq_unmask = pci_msi_unmask_irq,
> -};
> -
> -static int rcar_msi_map(struct irq_domain *domain, unsigned int irq,
> -			irq_hw_number_t hwirq)
> -{
> -	irq_set_chip_and_handler(irq, &rcar_msi_irq_chip, handle_simple_irq);
> -	irq_set_chip_data(irq, domain->host_data);
> -
> -	return 0;
> -}
> -
> -static const struct irq_domain_ops msi_domain_ops = {
> -	.map = rcar_msi_map,
> -};
> -
> -static void rcar_pcie_unmap_msi(struct rcar_pcie *pcie)
> -{
> -	struct rcar_msi *msi = &pcie->msi;
> -	int i, irq;
> -
> -	for (i = 0; i < INT_PCI_MSI_NR; i++) {
> -		irq = irq_find_mapping(msi->domain, i);
> -		if (irq > 0)
> -			irq_dispose_mapping(irq);
> -	}
> -
> -	irq_domain_remove(msi->domain);
> -}
> -
> -static int rcar_pcie_enable_msi(struct rcar_pcie *pcie)
> -{
> -	struct device *dev = pcie->dev;
> -	struct rcar_msi *msi = &pcie->msi;
> -	phys_addr_t base;
> -	int err, i;
> -
> -	mutex_init(&msi->lock);
> -
> -	msi->chip.dev = dev;
> -	msi->chip.setup_irq = rcar_msi_setup_irq;
> -	msi->chip.setup_irqs = rcar_msi_setup_irqs;
> -	msi->chip.teardown_irq = rcar_msi_teardown_irq;
> -
> -	msi->domain = irq_domain_add_linear(dev->of_node, INT_PCI_MSI_NR,
> -					    &msi_domain_ops, &msi->chip);
> -	if (!msi->domain) {
> -		dev_err(dev, "failed to create IRQ domain\n");
> -		return -ENOMEM;
> -	}
> -
> -	for (i = 0; i < INT_PCI_MSI_NR; i++)
> -		irq_create_mapping(msi->domain, i);
> -
> -	/* Two irqs are for MSI, but they are also used for non-MSI irqs */
> -	err = devm_request_irq(dev, msi->irq1, rcar_pcie_msi_irq,
> -			       IRQF_SHARED | IRQF_NO_THREAD,
> -			       rcar_msi_irq_chip.name, pcie);
> -	if (err < 0) {
> -		dev_err(dev, "failed to request IRQ: %d\n", err);
> -		goto err;
> -	}
> -
> -	err = devm_request_irq(dev, msi->irq2, rcar_pcie_msi_irq,
> -			       IRQF_SHARED | IRQF_NO_THREAD,
> -			       rcar_msi_irq_chip.name, pcie);
> -	if (err < 0) {
> -		dev_err(dev, "failed to request IRQ: %d\n", err);
> -		goto err;
> -	}
> -
> -	/* setup MSI data target */
> -	msi->pages = __get_free_pages(GFP_KERNEL, 0);
> -	if (!msi->pages) {
> -		err = -ENOMEM;
> -		goto err;
> -	}
> -	base = virt_to_phys((void *)msi->pages);
> -
> -	rcar_pci_write_reg(pcie, lower_32_bits(base) | MSIFE, PCIEMSIALR);
> -	rcar_pci_write_reg(pcie, upper_32_bits(base), PCIEMSIAUR);
> -
> -	/* enable all MSI interrupts */
> -	rcar_pci_write_reg(pcie, 0xffffffff, PCIEMSIIER);
> -
> -	return 0;
> -
> -err:
> -	rcar_pcie_unmap_msi(pcie);
> -	return err;
> -}
> -
> -static void rcar_pcie_teardown_msi(struct rcar_pcie *pcie)
> -{
> -	struct rcar_msi *msi = &pcie->msi;
> -
> -	/* Disable all MSI interrupts */
> -	rcar_pci_write_reg(pcie, 0, PCIEMSIIER);
> -
> -	/* Disable address decoding of the MSI interrupt, MSIFE */
> -	rcar_pci_write_reg(pcie, 0, PCIEMSIALR);
> -
> -	free_pages(msi->pages, 0);
> -
> -	rcar_pcie_unmap_msi(pcie);
> -}
> -
> -static int rcar_pcie_get_resources(struct rcar_pcie *pcie)
> -{
> -	struct device *dev = pcie->dev;
> -	struct resource res;
> -	int err, i;
> -
> -	pcie->phy = devm_phy_optional_get(dev, "pcie");
> -	if (IS_ERR(pcie->phy))
> -		return PTR_ERR(pcie->phy);
> -
> -	err = of_address_to_resource(dev->of_node, 0, &res);
> -	if (err)
> -		return err;
> -
> -	pcie->base = devm_ioremap_resource(dev, &res);
> -	if (IS_ERR(pcie->base))
> -		return PTR_ERR(pcie->base);
> -
> -	pcie->bus_clk = devm_clk_get(dev, "pcie_bus");
> -	if (IS_ERR(pcie->bus_clk)) {
> -		dev_err(dev, "cannot get pcie bus clock\n");
> -		return PTR_ERR(pcie->bus_clk);
> -	}
> -
> -	i = irq_of_parse_and_map(dev->of_node, 0);
> -	if (!i) {
> -		dev_err(dev, "cannot get platform resources for msi interrupt\n");
> -		err = -ENOENT;
> -		goto err_irq1;
> -	}
> -	pcie->msi.irq1 = i;
> -
> -	i = irq_of_parse_and_map(dev->of_node, 1);
> -	if (!i) {
> -		dev_err(dev, "cannot get platform resources for msi interrupt\n");
> -		err = -ENOENT;
> -		goto err_irq2;
> -	}
> -	pcie->msi.irq2 = i;
> -
> -	return 0;
> -
> -err_irq2:
> -	irq_dispose_mapping(pcie->msi.irq1);
> -err_irq1:
> -	return err;
> -}
> -
> -static int rcar_pcie_inbound_ranges(struct rcar_pcie *pcie,
> -				    struct of_pci_range *range,
> -				    int *index)
> +void rcar_pcie_set_inbound(void __iomem *base,
> +			   u64 cpu_addr, u64 pci_addr,
> +			   u64 mask, int idx, bool host)
>  {
> -	u64 restype = range->flags;
> -	u64 cpu_addr = range->cpu_addr;
> -	u64 cpu_end = range->cpu_addr + range->size;
> -	u64 pci_addr = range->pci_addr;
> -	u32 flags = LAM_64BIT | LAR_ENABLE;
> -	u64 mask;
> -	u64 size;
> -	int idx = *index;
> -
> -	if (restype & IORESOURCE_PREFETCH)
> -		flags |= LAM_PREFETCH;
> -
>  	/*
> -	 * If the size of the range is larger than the alignment of the start
> -	 * address, we have to use multiple entries to perform the mapping.
> -	 */
> -	if (cpu_addr > 0) {
> -		unsigned long nr_zeros = __ffs64(cpu_addr);
> -		u64 alignment = 1ULL << nr_zeros;
> -
> -		size = min(range->size, alignment);
> -	} else {
> -		size = range->size;
> -	}
> -	/* Hardware supports max 4GiB inbound region */
> -	size = min(size, 1ULL << 32);
> -
> -	mask = roundup_pow_of_two(size) - 1;
> -	mask &= ~0xf;
> -
> -	while (cpu_addr < cpu_end) {
> -		/*
> -		 * Set up 64-bit inbound regions as the range parser doesn't
> -		 * distinguish between 32 and 64-bit types.
> -		 */
> -		rcar_pci_write_reg(pcie, lower_32_bits(pci_addr),
> +	* Set up 64-bit inbound regions as the range parser doesn't
> +	* distinguish between 32 and 64-bit types.
> +	*/
> +	if (host) {
> +		rcar_pci_write_reg(base, lower_32_bits(pci_addr),
>  				   PCIEPRAR(idx));
> -		rcar_pci_write_reg(pcie, lower_32_bits(cpu_addr), PCIELAR(idx));
> -		rcar_pci_write_reg(pcie, lower_32_bits(mask) | flags,
> -				   PCIELAMR(idx));
> -
> -		rcar_pci_write_reg(pcie, upper_32_bits(pci_addr),
> +		rcar_pci_write_reg(base, upper_32_bits(pci_addr),
>  				   PCIEPRAR(idx + 1));
> -		rcar_pci_write_reg(pcie, upper_32_bits(cpu_addr),
> -				   PCIELAR(idx + 1));
> -		rcar_pci_write_reg(pcie, 0, PCIELAMR(idx + 1));
> -
> -		pci_addr += size;
> -		cpu_addr += size;
> -		idx += 2;
> -
> -		if (idx > MAX_NR_INBOUND_MAPS) {
> -			dev_err(pcie->dev, "Failed to map inbound regions!\n");
> -			return -EINVAL;
> -		}
>  	}
> -	*index = idx;
> -
> -	return 0;
> -}
>  
> -static int rcar_pcie_parse_map_dma_ranges(struct rcar_pcie *pcie,
> -					  struct device_node *np)
> -{
> -	struct of_pci_range range;
> -	struct of_pci_range_parser parser;
> -	int index = 0;
> -	int err;
> -
> -	if (of_pci_dma_range_parser_init(&parser, np))
> -		return -EINVAL;
> -
> -	/* Get the dma-ranges from DT */
> -	for_each_of_pci_range(&parser, &range) {
> -		u64 end = range.cpu_addr + range.size - 1;
> +	rcar_pci_write_reg(base, lower_32_bits(cpu_addr), PCIELAR(idx));
> +	rcar_pci_write_reg(base, lower_32_bits(mask), PCIELAMR(idx));
>  
> -		dev_dbg(pcie->dev, "0x%08x 0x%016llx..0x%016llx -> 0x%016llx\n",
> -			range.flags, range.cpu_addr, end, range.pci_addr);
> -
> -		err = rcar_pcie_inbound_ranges(pcie, &range, &index);
> -		if (err)
> -			return err;
> -	}
> +	rcar_pci_write_reg(base, upper_32_bits(cpu_addr), PCIELAR(idx + 1));
> +	rcar_pci_write_reg(base, 0, PCIELAMR(idx + 1));
>  
> -	return 0;
> -}
> -
> -static const struct of_device_id rcar_pcie_of_match[] = {
> -	{ .compatible = "renesas,pcie-r8a7779",
> -	  .data = rcar_pcie_phy_init_h1 },
> -	{ .compatible = "renesas,pcie-r8a7790",
> -	  .data = rcar_pcie_phy_init_gen2 },
> -	{ .compatible = "renesas,pcie-r8a7791",
> -	  .data = rcar_pcie_phy_init_gen2 },
> -	{ .compatible = "renesas,pcie-rcar-gen2",
> -	  .data = rcar_pcie_phy_init_gen2 },
> -	{ .compatible = "renesas,pcie-r8a7795",
> -	  .data = rcar_pcie_phy_init_gen3 },
> -	{ .compatible = "renesas,pcie-rcar-gen3",
> -	  .data = rcar_pcie_phy_init_gen3 },
> -	{},
> -};
> -
> -static int rcar_pcie_probe(struct platform_device *pdev)
> -{
> -	struct device *dev = &pdev->dev;
> -	struct rcar_pcie *pcie;
> -	u32 data;
> -	int err;
> -	int (*phy_init_fn)(struct rcar_pcie *);
> -	struct pci_host_bridge *bridge;
> -
> -	bridge = pci_alloc_host_bridge(sizeof(*pcie));
> -	if (!bridge)
> -		return -ENOMEM;
> -
> -	pcie = pci_host_bridge_priv(bridge);
> -
> -	pcie->dev = dev;
> -	platform_set_drvdata(pdev, pcie);
> -
> -	err = pci_parse_request_of_pci_ranges(dev, &pcie->resources, NULL);
> -	if (err)
> -		goto err_free_bridge;
> -
> -	pm_runtime_enable(pcie->dev);
> -	err = pm_runtime_get_sync(pcie->dev);
> -	if (err < 0) {
> -		dev_err(pcie->dev, "pm_runtime_get_sync failed\n");
> -		goto err_pm_disable;
> -	}
> -
> -	err = rcar_pcie_get_resources(pcie);
> -	if (err < 0) {
> -		dev_err(dev, "failed to request resources: %d\n", err);
> -		goto err_pm_put;
> -	}
> -
> -	err = clk_prepare_enable(pcie->bus_clk);
> -	if (err) {
> -		dev_err(dev, "failed to enable bus clock: %d\n", err);
> -		goto err_unmap_msi_irqs;
> -	}
> -
> -	err = rcar_pcie_parse_map_dma_ranges(pcie, dev->of_node);
> -	if (err)
> -		goto err_clk_disable;
> -
> -	phy_init_fn = of_device_get_match_data(dev);
> -	err = phy_init_fn(pcie);
> -	if (err) {
> -		dev_err(dev, "failed to init PCIe PHY\n");
> -		goto err_clk_disable;
> -	}
> -
> -	/* Failure to get a link might just be that no cards are inserted */
> -	if (rcar_pcie_hw_init(pcie)) {
> -		dev_info(dev, "PCIe link down\n");
> -		err = -ENODEV;
> -		goto err_phy_shutdown;
> -	}
> -
> -	data = rcar_pci_read_reg(pcie, MACSR);
> -	dev_info(dev, "PCIe x%d: link up\n", (data >> 20) & 0x3f);
> -
> -	if (IS_ENABLED(CONFIG_PCI_MSI)) {
> -		err = rcar_pcie_enable_msi(pcie);
> -		if (err < 0) {
> -			dev_err(dev,
> -				"failed to enable MSI support: %d\n",
> -				err);
> -			goto err_phy_shutdown;
> -		}
> -	}
> -
> -	err = rcar_pcie_enable(pcie);
> -	if (err)
> -		goto err_msi_teardown;
> -
> -	return 0;
> -
> -err_msi_teardown:
> -	if (IS_ENABLED(CONFIG_PCI_MSI))
> -		rcar_pcie_teardown_msi(pcie);
> -
> -err_phy_shutdown:
> -	if (pcie->phy) {
> -		phy_power_off(pcie->phy);
> -		phy_exit(pcie->phy);
> -	}
> -
> -err_clk_disable:
> -	clk_disable_unprepare(pcie->bus_clk);
> -
> -err_unmap_msi_irqs:
> -	irq_dispose_mapping(pcie->msi.irq2);
> -	irq_dispose_mapping(pcie->msi.irq1);
> -
> -err_pm_put:
> -	pm_runtime_put(dev);
> -
> -err_pm_disable:
> -	pm_runtime_disable(dev);
> -	pci_free_resource_list(&pcie->resources);
> -
> -err_free_bridge:
> -	pci_free_host_bridge(bridge);
> -
> -	return err;
> -}
> -
> -static int rcar_pcie_resume_noirq(struct device *dev)
> -{
> -	struct rcar_pcie *pcie = dev_get_drvdata(dev);
> -
> -	if (rcar_pci_read_reg(pcie, PMSR) &&
> -	    !(rcar_pci_read_reg(pcie, PCIETCTLR) & DL_DOWN))
> -		return 0;
> -
> -	/* Re-establish the PCIe link */
> -	rcar_rmw32(pcie, MACCTLR, MACCTLR_RESERVED, 0);
> -	rcar_pci_write_reg(pcie, CFINIT, PCIETCTLR);
> -	return rcar_pcie_wait_for_dl(pcie);
> +	/* flush modifications */
> +	wmb();
>  }
> -
> -static const struct dev_pm_ops rcar_pcie_pm_ops = {
> -	.resume_noirq = rcar_pcie_resume_noirq,
> -};
> -
> -static struct platform_driver rcar_pcie_driver = {
> -	.driver = {
> -		.name = "rcar-pcie",
> -		.of_match_table = rcar_pcie_of_match,
> -		.pm = &rcar_pcie_pm_ops,
> -		.suppress_bind_attrs = true,
> -	},
> -	.probe = rcar_pcie_probe,
> -};
> -builtin_platform_driver(rcar_pcie_driver);
> diff --git a/drivers/pci/controller/pcie-rcar.h b/drivers/pci/controller/pcie-rcar.h
> new file mode 100644
> index 000000000000..502621d4e6ac
> --- /dev/null
> +++ b/drivers/pci/controller/pcie-rcar.h
> @@ -0,0 +1,129 @@
> +/* SPDX-License-Identifier: GPL-2.0 */
> +/*
> + * PCIe driver for Renesas R-Car SoCs
> + *  Copyright (C) 2014-2019 Renesas Electronics Europe Ltd
> + *
> + * Author: Phil Edworthy <phil.edworthy@renesas.com>
> + */
> +#ifndef _PCIE_RCAR_H
> +#define _PCIE_RCAR_H
> +
> +#define PCIECAR			0x000010
> +#define PCIECCTLR		0x000018
> +#define  CONFIG_SEND_ENABLE	BIT(31)
> +#define  TYPE0			(0 << 8)
> +#define  TYPE1			BIT(8)
> +#define PCIECDR			0x000020
> +#define PCIEMSR			0x000028
> +#define PCIEINTXR		0x000400
> +#define PCIEPHYSR		0x0007f0
> +#define  PHYRDY			BIT(0)
> +#define PCIEMSITXR		0x000840
> +
> +/* Transfer control */
> +#define PCIETCTLR		0x02000
> +#define  DL_DOWN		BIT(3)
> +#define  CFINIT			BIT(0)
> +#define PCIETSTR		0x02004
> +#define  DATA_LINK_ACTIVE	BIT(0)
> +#define PCIEERRFR		0x02020
> +#define  UNSUPPORTED_REQUEST	BIT(4)
> +#define PCIEMSIFR		0x02044
> +#define PCIEMSIALR		0x02048
> +#define  MSIFE			BIT(0)
> +#define PCIEMSIAUR		0x0204c
> +#define PCIEMSIIER		0x02050
> +
> +/* root port address */
> +#define PCIEPRAR(x)		(0x02080 + ((x) * 0x4))
> +
> +/* local address reg & mask */
> +#define PCIELAR(x)		(0x02200 + ((x) * 0x20))
> +#define PCIELAMR(x)		(0x02208 + ((x) * 0x20))
> +#define  LAM_PREFETCH		BIT(3)
> +#define  LAM_64BIT		BIT(2)
> +#define  LAR_ENABLE		BIT(1)
> +
> +/* PCIe address reg & mask */
> +#define PCIEPALR(x)		(0x03400 + ((x) * 0x20))
> +#define PCIEPAUR(x)		(0x03404 + ((x) * 0x20))
> +#define PCIEPAMR(x)		(0x03408 + ((x) * 0x20))
> +#define PCIEPTCTLR(x)		(0x0340c + ((x) * 0x20))
> +#define  PAR_ENABLE		BIT(31)
> +#define  IO_SPACE		BIT(8)
> +
> +/* Configuration */
> +#define PCICONF(x)		(0x010000 + ((x) * 0x4))
> +#define PMCAP(x)		(0x010040 + ((x) * 0x4))
> +#define EXPCAP(x)		(0x010070 + ((x) * 0x4))
> +#define VCCAP(x)		(0x010100 + ((x) * 0x4))
> +
> +/* link layer */
> +#define IDSETR0			0x011000
> +#define  DEVICE_ID_SHFIT	16
> +#define IDSETR1			0x011004
> +#define SUBIDSETR		0x011024
> +#define TLCTLR			0x011048
> +#define MACSR			0x011054
> +#define  SPCHGFIN		BIT(4)
> +#define  SPCHGFAIL		BIT(6)
> +#define  SPCHGSUC		BIT(7)
> +#define  LINK_SPEED		(0xf << 16)
> +#define  LINK_SPEED_2_5GTS	BIT(16)
> +#define  LINK_SPEED_5_0GTS	BIT(17)
> +#define MACCTLR			0x011058
> +#define  MACCTLR_RESERVED	BIT(0)
> +#define  SPEED_CHANGE		BIT(24)
> +#define  SCRAMBLE_DISABLE	BIT(27)
> +#define PMSR			0x01105c
> +#define MACS2R			0x011078
> +#define MACCGSPSETR		0x011084
> +#define  SPCNGRSN		BIT(31)
> +
> +/* R-Car H1 PHY */
> +#define H1_PCIEPHYADRR		0x04000c
> +#define  WRITE_CMD		BIT(16)
> +#define  PHY_ACK		BIT(24)
> +#define  RATE_POS		12
> +#define  LANE_POS		8
> +#define  ADR_POS		0
> +#define H1_PCIEPHYDOUTR		0x040014
> +
> +/* R-Car Gen2 PHY */
> +#define GEN2_PCIEPHYADDR	0x780
> +#define GEN2_PCIEPHYDATA	0x784
> +#define GEN2_PCIEPHYCTRL	0x78c
> +
> +#define INT_PCI_MSI_NR		32
> +
> +#define RCONF(x)		(PCICONF(0) + (x))
> +#define RPMCAP(x)		(PMCAP(0) + (x))
> +#define REXPCAP(x)		(EXPCAP(0) + (x))
> +#define RVCCAP(x)		(VCCAP(0) + (x))
> +
> +#define INTDIS_SHIFT		BIT(10)
> +#define ASTINTX_SHIFT		BIT(16)
> +
> +#define PCIE_CONF_BUS(b)	(((b) & 0xff) << 24)
> +#define PCIE_CONF_DEV(d)	(((d) & 0x1f) << 19)
> +#define PCIE_CONF_FUNC(f)	(((f) & 0x7) << 16)
> +
> +#define RCAR_PCI_MAX_RESOURCES	4
> +#define MAX_NR_INBOUND_MAPS	6
> +
> +enum {
> +	RCAR_PCI_ACCESS_READ,
> +	RCAR_PCI_ACCESS_WRITE,
> +};
> +
> +void rcar_pci_write_reg(void __iomem *base, u32 val, unsigned int reg);
> +u32 rcar_pci_read_reg(void __iomem *base, unsigned int reg);
> +void rcar_rmw32(void __iomem *base, int where, u32 mask, u32 data);
> +int rcar_pcie_wait_for_phyrdy(void __iomem *base);
> +int rcar_pcie_wait_for_dl(void __iomem *base);
> +void rcar_pcie_set_outbound(int win, void __iomem *base,
> +			    struct resource *res, bool host);
> +void rcar_pcie_set_inbound(void __iomem *base, u64 cpu_addr, u64 pci_addr,
> +			   u64 mask, int idx, bool host);
> +
> +#endif
> 

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 2/5] pci: endpoint: add support to handle multiple base for mapping outbound memory
  2019-11-06 19:36   ` Lad Prabhakar
  (?)
@ 2019-11-27  5:14     ` Kishon Vijay Abraham I
  -1 siblings, 0 replies; 60+ messages in thread
From: Kishon Vijay Abraham I @ 2019-11-27  5:14 UTC (permalink / raw)
  To: Lad Prabhakar, Bjorn Helgaas, Rob Herring, Mark Rutland,
	Geert Uytterhoeven, Magnus Damm, Marek Vasut, Yoshihiro Shimoda,
	linux-pci
  Cc: Catalin Marinas, Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray, devicetree, linux-kernel,
	linux-arm-kernel, linux-renesas-soc, Chris Paterson, Lad,
	Prabhakar, linux-rockchip, Shawn Lin, Heiko Stuebner, Tom Joseph,
	Jingoo Han, Gustavo Pimentel

Hi Prabhakar,

On 07/11/19 1:06 AM, Lad Prabhakar wrote:
> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> 
> rcar pcie controller has support to map multiple memory regions
> for mapping the outbound memory in local system, this feature
> inspires to add support for handling multiple memory bases in
> endpoint framework. In case of multiple memory regions only chunk
> or complete region can be mapped and this window needs to be
> passed to the controller driver.
> 
> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> Cc: <linux-rockchip@lists.infradead.org>
> Cc: Shawn Lin <shawn.lin@rock-chips.com>
> Cc: Heiko Stuebner <heiko@sntech.de>
> Cc: Tom Joseph <tjoseph@cadence.com>
> Cc: Jingoo Han <jingoohan1@gmail.com>
> Cc: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
> ---
>  .../pci/controller/dwc/pcie-designware-ep.c   |  30 ++-
>  drivers/pci/controller/pcie-cadence-ep.c      |  11 +-
>  drivers/pci/controller/pcie-rockchip-ep.c     |  13 +-
>  drivers/pci/endpoint/functions/pci-epf-test.c |  29 +--
>  drivers/pci/endpoint/pci-epc-core.c           |   7 +-
>  drivers/pci/endpoint/pci-epc-mem.c            | 189 ++++++++++++++----
>  include/linux/pci-epc.h                       |  43 ++--
>  7 files changed, 234 insertions(+), 88 deletions(-)
> 
> diff --git a/drivers/pci/controller/dwc/pcie-designware-ep.c b/drivers/pci/controller/dwc/pcie-designware-ep.c
> index 3dd2e2697294..8d23c20b9afd 100644
> --- a/drivers/pci/controller/dwc/pcie-designware-ep.c
> +++ b/drivers/pci/controller/dwc/pcie-designware-ep.c
> @@ -195,7 +195,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, u8 func_no,
>  }
>  
>  static int dw_pcie_ep_map_addr(struct pci_epc *epc, u8 func_no,
> -			       phys_addr_t addr,
> +			       phys_addr_t addr, int window,
>  			       u64 pci_addr, size_t size)
>  {
>  	int ret;
> @@ -367,6 +367,7 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
>  	unsigned int aligned_offset;
>  	u16 msg_ctrl, msg_data;
>  	u32 msg_addr_lower, msg_addr_upper, reg;
> +	int window = PCI_EPC_DEFAULT_WINDOW;
>  	u64 msg_addr;
>  	bool has_upper;
>  	int ret;
> @@ -390,11 +391,11 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
>  		reg = ep->msi_cap + PCI_MSI_DATA_32;
>  		msg_data = dw_pcie_readw_dbi(pci, reg);
>  	}
> -	aligned_offset = msg_addr_lower & (epc->mem->page_size - 1);
> +	aligned_offset = msg_addr_lower & (epc->mem[window]->page_size - 1);
>  	msg_addr = ((u64)msg_addr_upper) << 32 |
>  			(msg_addr_lower & ~aligned_offset);
> -	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> -				  epc->mem->page_size);
> +	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> +				  msg_addr, epc->mem[window]->page_size);
>  	if (ret)
>  		return ret;
>  
> @@ -416,6 +417,7 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>  	u32 reg, msg_data, vec_ctrl;
>  	u64 tbl_addr, msg_addr, reg_u64;
>  	void __iomem *msix_tbl;
> +	int window = PCI_EPC_DEFAULT_WINDOW;
>  	int ret;
>  
>  	reg = ep->msix_cap + PCI_MSIX_TABLE;
> @@ -452,8 +454,8 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>  		return -EPERM;
>  	}
>  
> -	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> -				  epc->mem->page_size);
> +	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> +				  msg_addr, epc->mem[window]->page_size);
>  	if (ret)
>  		return ret;
>  
> @@ -466,10 +468,11 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>  
>  void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
>  {
> +	int window = PCI_EPC_DEFAULT_WINDOW;
>  	struct pci_epc *epc = ep->epc;
>  
>  	pci_epc_mem_free_addr(epc, ep->msi_mem_phys, ep->msi_mem,
> -			      epc->mem->page_size);
> +			      epc->mem[window]->page_size);
>  
>  	pci_epc_mem_exit(epc);
>  }
> @@ -499,9 +502,12 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
>  	u32 reg;
>  	void *addr;
>  	u8 hdr_type;
> +	int window;
>  	unsigned int nbars;
>  	unsigned int offset;
>  	struct pci_epc *epc;
> +	size_t msi_page_size;
> +	struct pci_epc_mem_window mem_window;
>  	struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
>  	struct device *dev = pci->dev;
>  	struct device_node *np = dev->of_node;
> @@ -574,15 +580,17 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
>  	if (ret < 0)
>  		epc->max_functions = 1;
>  
> -	ret = __pci_epc_mem_init(epc, ep->phys_base, ep->addr_size,
> -				 ep->page_size);
> +	mem_window.phys_base = ep->phys_base;
> +	mem_window.size = ep->addr_size;
> +	ret = __pci_epc_mem_init(epc, &mem_window, 1, ep->page_size);
>  	if (ret < 0) {
>  		dev_err(dev, "Failed to initialize address space\n");
>  		return ret;
>  	}
>  
> -	ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys,
> -					     epc->mem->page_size);
> +	msi_page_size = epc->mem[PCI_EPC_DEFAULT_WINDOW]->page_size;
> +	ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys, &window,
> +					     msi_page_size);
>  	if (!ep->msi_mem) {
>  		dev_err(dev, "Failed to reserve memory for MSI/MSI-X\n");
>  		return -ENOMEM;
> diff --git a/drivers/pci/controller/pcie-cadence-ep.c b/drivers/pci/controller/pcie-cadence-ep.c
> index def7820cb824..7991b38a5350 100644
> --- a/drivers/pci/controller/pcie-cadence-ep.c
> +++ b/drivers/pci/controller/pcie-cadence-ep.c
> @@ -172,7 +172,7 @@ static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
>  }
>  
>  static int cdns_pcie_ep_map_addr(struct pci_epc *epc, u8 fn, phys_addr_t addr,
> -				 u64 pci_addr, size_t size)
> +				 int window, u64 pci_addr, size_t size)
>  {
>  	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>  	struct cdns_pcie *pcie = &ep->pcie;
> @@ -434,12 +434,14 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
>  {
>  	struct device *dev = &pdev->dev;
>  	struct device_node *np = dev->of_node;
> +	struct pci_epc_mem_window mem_window;
>  	struct cdns_pcie_ep *ep;
>  	struct cdns_pcie *pcie;
>  	struct pci_epc *epc;
>  	struct resource *res;
>  	int ret;
>  	int phy_count;
> +	int window;
>  
>  	ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
>  	if (!ep)
> @@ -502,15 +504,16 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
>  	if (of_property_read_u8(np, "max-functions", &epc->max_functions) < 0)
>  		epc->max_functions = 1;
>  
> -	ret = pci_epc_mem_init(epc, pcie->mem_res->start,
> -			       resource_size(pcie->mem_res));
> +	mem_window.phys_base = pcie->mem_res->start;
> +	mem_window.size = resource_size(pcie->mem_res);
> +	ret = pci_epc_mem_init(epc, &mem_window, 1);
>  	if (ret < 0) {
>  		dev_err(dev, "failed to initialize the memory space\n");
>  		goto err_init;
>  	}
>  
>  	ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> -						  SZ_128K);
> +						  &window, SZ_128K);
>  	if (!ep->irq_cpu_addr) {
>  		dev_err(dev, "failed to reserve memory space for MSI\n");
>  		ret = -ENOMEM;
> diff --git a/drivers/pci/controller/pcie-rockchip-ep.c b/drivers/pci/controller/pcie-rockchip-ep.c
> index d743b0a48988..d59e85c8d319 100644
> --- a/drivers/pci/controller/pcie-rockchip-ep.c
> +++ b/drivers/pci/controller/pcie-rockchip-ep.c
> @@ -256,8 +256,8 @@ static void rockchip_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
>  }
>  
>  static int rockchip_pcie_ep_map_addr(struct pci_epc *epc, u8 fn,
> -				     phys_addr_t addr, u64 pci_addr,
> -				     size_t size)
> +				     phys_addr_t addr, int window,
> +				     u64 pci_addr, size_t size)
>  {
>  	struct rockchip_pcie_ep *ep = epc_get_drvdata(epc);
>  	struct rockchip_pcie *pcie = &ep->rockchip;
> @@ -562,11 +562,13 @@ static const struct of_device_id rockchip_pcie_ep_of_match[] = {
>  
>  static int rockchip_pcie_ep_probe(struct platform_device *pdev)
>  {
> +	struct pci_epc_mem_window mem_window;
>  	struct device *dev = &pdev->dev;
>  	struct rockchip_pcie_ep *ep;
>  	struct rockchip_pcie *rockchip;
>  	struct pci_epc *epc;
>  	size_t max_regions;
> +	int window;
>  	int err;
>  
>  	ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> @@ -614,15 +616,16 @@ static int rockchip_pcie_ep_probe(struct platform_device *pdev)
>  	/* Only enable function 0 by default */
>  	rockchip_pcie_write(rockchip, BIT(0), PCIE_CORE_PHY_FUNC_CFG);
>  
> -	err = pci_epc_mem_init(epc, rockchip->mem_res->start,
> -			       resource_size(rockchip->mem_res));
> +	mem_window.phys_base = rockchip->mem_res->start;
> +	mem_window.size = resource_size(rockchip->mem_res);
> +	err = pci_epc_mem_init(epc, &mem_window, 1);
>  	if (err < 0) {
>  		dev_err(dev, "failed to initialize the memory space\n");
>  		goto err_uninit_port;
>  	}
>  
>  	ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> -						  SZ_128K);
> +						  &window, SZ_128K);
>  	if (!ep->irq_cpu_addr) {
>  		dev_err(dev, "failed to reserve memory space for MSI\n");
>  		err = -ENOMEM;
> diff --git a/drivers/pci/endpoint/functions/pci-epf-test.c b/drivers/pci/endpoint/functions/pci-epf-test.c
> index 5d74f81ddfe4..475228011703 100644
> --- a/drivers/pci/endpoint/functions/pci-epf-test.c
> +++ b/drivers/pci/endpoint/functions/pci-epf-test.c
> @@ -84,8 +84,10 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>  	struct pci_epc *epc = epf->epc;
>  	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>  	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> +	int window;
>  
> -	src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr, reg->size);
> +	src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr,
> +					  &window, reg->size);
>  	if (!src_addr) {
>  		dev_err(dev, "Failed to allocate source address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
> @@ -93,15 +95,16 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>  		goto err;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, reg->src_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, window,
> +			       reg->src_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map source address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
>  		goto err_src_addr;
>  	}
>  
> -	dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr, reg->size);
> +	dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr,
> +					  &window, reg->size);
>  	if (!dst_addr) {
>  		dev_err(dev, "Failed to allocate destination address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> @@ -109,8 +112,8 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>  		goto err_src_map_addr;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, reg->dst_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, window,
> +			       reg->dst_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map destination address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> @@ -146,8 +149,9 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
>  	struct pci_epc *epc = epf->epc;
>  	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>  	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> +	int window;
>  
> -	src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> +	src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
>  	if (!src_addr) {
>  		dev_err(dev, "Failed to allocate address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
> @@ -155,8 +159,8 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
>  		goto err;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->src_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> +			       reg->src_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
> @@ -193,13 +197,14 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
>  	void __iomem *dst_addr;
>  	void *buf;
>  	phys_addr_t phys_addr;
> +	int window;
>  	struct pci_epf *epf = epf_test->epf;
>  	struct device *dev = &epf->dev;
>  	struct pci_epc *epc = epf->epc;
>  	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>  	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
>  
> -	dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> +	dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
>  	if (!dst_addr) {
>  		dev_err(dev, "Failed to allocate address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> @@ -207,8 +212,8 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
>  		goto err;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->dst_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> +			       reg->dst_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> diff --git a/drivers/pci/endpoint/pci-epc-core.c b/drivers/pci/endpoint/pci-epc-core.c
> index 2091508c1620..289c266c2d90 100644
> --- a/drivers/pci/endpoint/pci-epc-core.c
> +++ b/drivers/pci/endpoint/pci-epc-core.c
> @@ -358,13 +358,15 @@ EXPORT_SYMBOL_GPL(pci_epc_unmap_addr);
>   * @epc: the EPC device on which address is allocated
>   * @func_no: the endpoint function number in the EPC device
>   * @phys_addr: physical address of the local system
> + * @window: index to the window region where PCI address will be mapped
>   * @pci_addr: PCI address to which the physical address should be mapped
>   * @size: the size of the allocation
>   *
>   * Invoke to map CPU address with PCI address.
>   */
>  int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> -		     phys_addr_t phys_addr, u64 pci_addr, size_t size)
> +		     phys_addr_t phys_addr, int window,
> +		     u64 pci_addr, size_t size)
>  {
>  	int ret;
>  	unsigned long flags;
> @@ -376,7 +378,8 @@ int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
>  		return 0;
>  
>  	spin_lock_irqsave(&epc->lock, flags);
> -	ret = epc->ops->map_addr(epc, func_no, phys_addr, pci_addr, size);
> +	ret = epc->ops->map_addr(epc, func_no, phys_addr,
> +				 window, pci_addr, size);
>  	spin_unlock_irqrestore(&epc->lock, flags);
>  
>  	return ret;
> diff --git a/drivers/pci/endpoint/pci-epc-mem.c b/drivers/pci/endpoint/pci-epc-mem.c
> index d2b174ce15de..c955f2c97944 100644
> --- a/drivers/pci/endpoint/pci-epc-mem.c
> +++ b/drivers/pci/endpoint/pci-epc-mem.c
> @@ -39,56 +39,77 @@ static int pci_epc_mem_get_order(struct pci_epc_mem *mem, size_t size)
>   * __pci_epc_mem_init() - initialize the pci_epc_mem structure
>   * @epc: the EPC device that invoked pci_epc_mem_init
>   * @phys_base: the physical address of the base
> - * @size: the size of the address space
> + * @num_windows: number of windows device supports

struct pci_epc_mem_window is missing here.
>   * @page_size: size of each page
>   *
>   * Invoke to initialize the pci_epc_mem structure used by the
>   * endpoint functions to allocate mapped PCI address.
>   */
> -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_base, size_t size,
> -		       size_t page_size)
> +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *windows,
> +		       int num_windows, size_t page_size)
>  {
> -	int ret;
> -	struct pci_epc_mem *mem;
> -	unsigned long *bitmap;
> +	struct pci_epc_mem *mem = NULL;
> +	unsigned long *bitmap = NULL;
>  	unsigned int page_shift;
> -	int pages;
>  	int bitmap_size;
> +	int pages;
> +	int ret;
> +	int i;
> +
> +	epc->mem_windows = 0;
> +
> +	if (!windows)
> +		return -EINVAL;
> +
> +	if (num_windows <= 0)
> +		return -EINVAL;
>  
>  	if (page_size < PAGE_SIZE)
>  		page_size = PAGE_SIZE;
>  
>  	page_shift = ilog2(page_size);
> -	pages = size >> page_shift;
> -	bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
>  
> -	mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> -	if (!mem) {
> -		ret = -ENOMEM;
> -		goto err;
> -	}
> +	epc->mem = kcalloc(num_windows, sizeof(*mem), GFP_KERNEL);
> +	if (!epc->mem)
> +		return -EINVAL;
>  
> -	bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> -	if (!bitmap) {
> -		ret = -ENOMEM;
> -		goto err_mem;
> -	}
> +	for (i = 0; i < num_windows; i++) {
> +		pages = windows[i].phys_base >> page_shift;
> +		bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
>  
> -	mem->bitmap = bitmap;
> -	mem->phys_base = phys_base;
> -	mem->page_size = page_size;
> -	mem->pages = pages;
> -	mem->size = size;
> +		mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> +		if (!mem) {
> +			ret = -ENOMEM;
> +			goto err_mem;
> +		}
>  
> -	epc->mem = mem;
> +		bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> +		if (!bitmap) {
> +			ret = -ENOMEM;
> +			goto err_mem;
> +		}
> +
> +		mem->bitmap = bitmap;
> +		mem->window.phys_base = windows[i].phys_base;
> +		mem->page_size = page_size;
> +		mem->pages = pages;
> +		mem->window.size = windows[i].size;
> +		mem->window.map_size = 0;
> +
> +		epc->mem[i] = mem;
> +	}
> +	epc->mem_windows = num_windows;
>  
>  	return 0;
>  
>  err_mem:
> -	kfree(mem);
> +	for (; i >= 0; i--) {
> +		kfree(mem->bitmap);
> +		kfree(epc->mem[i]);
> +	}
> +	kfree(epc->mem);
>  
> -err:
> -return ret;
> +	return ret;
>  }
>  EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
>  
> @@ -101,48 +122,126 @@ EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
>   */
>  void pci_epc_mem_exit(struct pci_epc *epc)
>  {
> -	struct pci_epc_mem *mem = epc->mem;
> +	struct pci_epc_mem *mem;
> +	int i;
> +
> +	if (!epc->mem_windows)
> +		return;
> +
> +	for (i = 0; i <= epc->mem_windows; i--) {
> +		mem = epc->mem[i];
> +		kfree(mem->bitmap);
> +		kfree(epc->mem[i]);
> +	}
> +	kfree(epc->mem);
>  
>  	epc->mem = NULL;
> -	kfree(mem->bitmap);
> -	kfree(mem);
> +	epc->mem_windows = 0;
>  }
>  EXPORT_SYMBOL_GPL(pci_epc_mem_exit);
>  
> +static int pci_epc_find_best_fit_window(struct pci_epc *epc, size_t size)
> +{
> +	size_t window_least_size = 0;
> +	int best_fit_window = -1;
> +	struct pci_epc_mem *mem;
> +	size_t actual_size;
> +	int i;
> +
> +	for (i = 0; i < epc->mem_windows; i++) {
> +		mem = epc->mem[i];
> +
> +		/* if chunk from this region is already used skip it */
> +		if (mem->window.map_size)
> +			continue;
> +
> +		actual_size = ALIGN(size, mem->page_size);
> +
> +		if (best_fit_window == -1) {
> +			best_fit_window = i;
> +			window_least_size = mem->window.size;
> +		} else {
> +			if (actual_size <= mem->window.size &&
> +			    mem->window.size < window_least_size) {
> +				best_fit_window = i;
> +				window_least_size = mem->window.size;
> +			}
> +		}
> +	}
> +
> +	return best_fit_window;
> +}
> +
>  /**
>   * pci_epc_mem_alloc_addr() - allocate memory address from EPC addr space
>   * @epc: the EPC device on which memory has to be allocated
>   * @phys_addr: populate the allocated physical address here
> + * @window: populate the window here which will be used to map PCI address
>   * @size: the size of the address space that has to be allocated
>   *
>   * Invoke to allocate memory address from the EPC address space. This
>   * is usually done to map the remote RC address into the local system.
>   */
>  void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
> -				     phys_addr_t *phys_addr, size_t size)
> +				     phys_addr_t *phys_addr,
> +				     int *window, size_t size)
>  {
> +	int best_fit = PCI_EPC_DEFAULT_WINDOW;
> +	void __iomem *virt_addr = NULL;
> +	struct pci_epc_mem *mem;
> +	unsigned int page_shift;
>  	int pageno;
> -	void __iomem *virt_addr;
> -	struct pci_epc_mem *mem = epc->mem;
> -	unsigned int page_shift = ilog2(mem->page_size);
>  	int order;
>  
> +	if (epc->mem_windows <= 0)
> +		return NULL;
> +
> +	if (epc->mem_windows > 1) {
> +		best_fit = pci_epc_find_best_fit_window(epc, size);
> +		if (best_fit < 0)
> +			return NULL;
> +	}
> +
> +	mem = epc->mem[best_fit];
>  	size = ALIGN(size, mem->page_size);
> +	if (size > (mem->window.size - mem->window.map_size))
> +		return NULL;

Assume I have two mem regions, first region is of size 128MB and the second
region is of size 4GB. If there is a allocation request for 4GB, will the
allocation succeed?

Thanks
Kishon

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 2/5] pci: endpoint: add support to handle multiple base for mapping outbound memory
@ 2019-11-27  5:14     ` Kishon Vijay Abraham I
  0 siblings, 0 replies; 60+ messages in thread
From: Kishon Vijay Abraham I @ 2019-11-27  5:14 UTC (permalink / raw)
  To: Lad Prabhakar, Bjorn Helgaas, Rob Herring, Mark Rutland,
	Geert Uytterhoeven, Magnus Damm, Marek Vasut, Yoshihiro Shimoda,
	linux-pci
  Cc: devicetree, Chris Paterson, Lorenzo Pieralisi, Heiko Stuebner,
	Arnd Bergmann, Jingoo Han, Catalin Marinas, Shawn Lin,
	linux-kernel, Lad, Prabhakar, linux-renesas-soc, linux-rockchip,
	Tom Joseph, Greg Kroah-Hartman, Andrew Murray, Gustavo Pimentel,
	Will Deacon, linux-arm-kernel

Hi Prabhakar,

On 07/11/19 1:06 AM, Lad Prabhakar wrote:
> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> 
> rcar pcie controller has support to map multiple memory regions
> for mapping the outbound memory in local system, this feature
> inspires to add support for handling multiple memory bases in
> endpoint framework. In case of multiple memory regions only chunk
> or complete region can be mapped and this window needs to be
> passed to the controller driver.
> 
> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> Cc: <linux-rockchip@lists.infradead.org>
> Cc: Shawn Lin <shawn.lin@rock-chips.com>
> Cc: Heiko Stuebner <heiko@sntech.de>
> Cc: Tom Joseph <tjoseph@cadence.com>
> Cc: Jingoo Han <jingoohan1@gmail.com>
> Cc: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
> ---
>  .../pci/controller/dwc/pcie-designware-ep.c   |  30 ++-
>  drivers/pci/controller/pcie-cadence-ep.c      |  11 +-
>  drivers/pci/controller/pcie-rockchip-ep.c     |  13 +-
>  drivers/pci/endpoint/functions/pci-epf-test.c |  29 +--
>  drivers/pci/endpoint/pci-epc-core.c           |   7 +-
>  drivers/pci/endpoint/pci-epc-mem.c            | 189 ++++++++++++++----
>  include/linux/pci-epc.h                       |  43 ++--
>  7 files changed, 234 insertions(+), 88 deletions(-)
> 
> diff --git a/drivers/pci/controller/dwc/pcie-designware-ep.c b/drivers/pci/controller/dwc/pcie-designware-ep.c
> index 3dd2e2697294..8d23c20b9afd 100644
> --- a/drivers/pci/controller/dwc/pcie-designware-ep.c
> +++ b/drivers/pci/controller/dwc/pcie-designware-ep.c
> @@ -195,7 +195,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, u8 func_no,
>  }
>  
>  static int dw_pcie_ep_map_addr(struct pci_epc *epc, u8 func_no,
> -			       phys_addr_t addr,
> +			       phys_addr_t addr, int window,
>  			       u64 pci_addr, size_t size)
>  {
>  	int ret;
> @@ -367,6 +367,7 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
>  	unsigned int aligned_offset;
>  	u16 msg_ctrl, msg_data;
>  	u32 msg_addr_lower, msg_addr_upper, reg;
> +	int window = PCI_EPC_DEFAULT_WINDOW;
>  	u64 msg_addr;
>  	bool has_upper;
>  	int ret;
> @@ -390,11 +391,11 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
>  		reg = ep->msi_cap + PCI_MSI_DATA_32;
>  		msg_data = dw_pcie_readw_dbi(pci, reg);
>  	}
> -	aligned_offset = msg_addr_lower & (epc->mem->page_size - 1);
> +	aligned_offset = msg_addr_lower & (epc->mem[window]->page_size - 1);
>  	msg_addr = ((u64)msg_addr_upper) << 32 |
>  			(msg_addr_lower & ~aligned_offset);
> -	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> -				  epc->mem->page_size);
> +	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> +				  msg_addr, epc->mem[window]->page_size);
>  	if (ret)
>  		return ret;
>  
> @@ -416,6 +417,7 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>  	u32 reg, msg_data, vec_ctrl;
>  	u64 tbl_addr, msg_addr, reg_u64;
>  	void __iomem *msix_tbl;
> +	int window = PCI_EPC_DEFAULT_WINDOW;
>  	int ret;
>  
>  	reg = ep->msix_cap + PCI_MSIX_TABLE;
> @@ -452,8 +454,8 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>  		return -EPERM;
>  	}
>  
> -	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> -				  epc->mem->page_size);
> +	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> +				  msg_addr, epc->mem[window]->page_size);
>  	if (ret)
>  		return ret;
>  
> @@ -466,10 +468,11 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>  
>  void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
>  {
> +	int window = PCI_EPC_DEFAULT_WINDOW;
>  	struct pci_epc *epc = ep->epc;
>  
>  	pci_epc_mem_free_addr(epc, ep->msi_mem_phys, ep->msi_mem,
> -			      epc->mem->page_size);
> +			      epc->mem[window]->page_size);
>  
>  	pci_epc_mem_exit(epc);
>  }
> @@ -499,9 +502,12 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
>  	u32 reg;
>  	void *addr;
>  	u8 hdr_type;
> +	int window;
>  	unsigned int nbars;
>  	unsigned int offset;
>  	struct pci_epc *epc;
> +	size_t msi_page_size;
> +	struct pci_epc_mem_window mem_window;
>  	struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
>  	struct device *dev = pci->dev;
>  	struct device_node *np = dev->of_node;
> @@ -574,15 +580,17 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
>  	if (ret < 0)
>  		epc->max_functions = 1;
>  
> -	ret = __pci_epc_mem_init(epc, ep->phys_base, ep->addr_size,
> -				 ep->page_size);
> +	mem_window.phys_base = ep->phys_base;
> +	mem_window.size = ep->addr_size;
> +	ret = __pci_epc_mem_init(epc, &mem_window, 1, ep->page_size);
>  	if (ret < 0) {
>  		dev_err(dev, "Failed to initialize address space\n");
>  		return ret;
>  	}
>  
> -	ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys,
> -					     epc->mem->page_size);
> +	msi_page_size = epc->mem[PCI_EPC_DEFAULT_WINDOW]->page_size;
> +	ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys, &window,
> +					     msi_page_size);
>  	if (!ep->msi_mem) {
>  		dev_err(dev, "Failed to reserve memory for MSI/MSI-X\n");
>  		return -ENOMEM;
> diff --git a/drivers/pci/controller/pcie-cadence-ep.c b/drivers/pci/controller/pcie-cadence-ep.c
> index def7820cb824..7991b38a5350 100644
> --- a/drivers/pci/controller/pcie-cadence-ep.c
> +++ b/drivers/pci/controller/pcie-cadence-ep.c
> @@ -172,7 +172,7 @@ static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
>  }
>  
>  static int cdns_pcie_ep_map_addr(struct pci_epc *epc, u8 fn, phys_addr_t addr,
> -				 u64 pci_addr, size_t size)
> +				 int window, u64 pci_addr, size_t size)
>  {
>  	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>  	struct cdns_pcie *pcie = &ep->pcie;
> @@ -434,12 +434,14 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
>  {
>  	struct device *dev = &pdev->dev;
>  	struct device_node *np = dev->of_node;
> +	struct pci_epc_mem_window mem_window;
>  	struct cdns_pcie_ep *ep;
>  	struct cdns_pcie *pcie;
>  	struct pci_epc *epc;
>  	struct resource *res;
>  	int ret;
>  	int phy_count;
> +	int window;
>  
>  	ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
>  	if (!ep)
> @@ -502,15 +504,16 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
>  	if (of_property_read_u8(np, "max-functions", &epc->max_functions) < 0)
>  		epc->max_functions = 1;
>  
> -	ret = pci_epc_mem_init(epc, pcie->mem_res->start,
> -			       resource_size(pcie->mem_res));
> +	mem_window.phys_base = pcie->mem_res->start;
> +	mem_window.size = resource_size(pcie->mem_res);
> +	ret = pci_epc_mem_init(epc, &mem_window, 1);
>  	if (ret < 0) {
>  		dev_err(dev, "failed to initialize the memory space\n");
>  		goto err_init;
>  	}
>  
>  	ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> -						  SZ_128K);
> +						  &window, SZ_128K);
>  	if (!ep->irq_cpu_addr) {
>  		dev_err(dev, "failed to reserve memory space for MSI\n");
>  		ret = -ENOMEM;
> diff --git a/drivers/pci/controller/pcie-rockchip-ep.c b/drivers/pci/controller/pcie-rockchip-ep.c
> index d743b0a48988..d59e85c8d319 100644
> --- a/drivers/pci/controller/pcie-rockchip-ep.c
> +++ b/drivers/pci/controller/pcie-rockchip-ep.c
> @@ -256,8 +256,8 @@ static void rockchip_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
>  }
>  
>  static int rockchip_pcie_ep_map_addr(struct pci_epc *epc, u8 fn,
> -				     phys_addr_t addr, u64 pci_addr,
> -				     size_t size)
> +				     phys_addr_t addr, int window,
> +				     u64 pci_addr, size_t size)
>  {
>  	struct rockchip_pcie_ep *ep = epc_get_drvdata(epc);
>  	struct rockchip_pcie *pcie = &ep->rockchip;
> @@ -562,11 +562,13 @@ static const struct of_device_id rockchip_pcie_ep_of_match[] = {
>  
>  static int rockchip_pcie_ep_probe(struct platform_device *pdev)
>  {
> +	struct pci_epc_mem_window mem_window;
>  	struct device *dev = &pdev->dev;
>  	struct rockchip_pcie_ep *ep;
>  	struct rockchip_pcie *rockchip;
>  	struct pci_epc *epc;
>  	size_t max_regions;
> +	int window;
>  	int err;
>  
>  	ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> @@ -614,15 +616,16 @@ static int rockchip_pcie_ep_probe(struct platform_device *pdev)
>  	/* Only enable function 0 by default */
>  	rockchip_pcie_write(rockchip, BIT(0), PCIE_CORE_PHY_FUNC_CFG);
>  
> -	err = pci_epc_mem_init(epc, rockchip->mem_res->start,
> -			       resource_size(rockchip->mem_res));
> +	mem_window.phys_base = rockchip->mem_res->start;
> +	mem_window.size = resource_size(rockchip->mem_res);
> +	err = pci_epc_mem_init(epc, &mem_window, 1);
>  	if (err < 0) {
>  		dev_err(dev, "failed to initialize the memory space\n");
>  		goto err_uninit_port;
>  	}
>  
>  	ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> -						  SZ_128K);
> +						  &window, SZ_128K);
>  	if (!ep->irq_cpu_addr) {
>  		dev_err(dev, "failed to reserve memory space for MSI\n");
>  		err = -ENOMEM;
> diff --git a/drivers/pci/endpoint/functions/pci-epf-test.c b/drivers/pci/endpoint/functions/pci-epf-test.c
> index 5d74f81ddfe4..475228011703 100644
> --- a/drivers/pci/endpoint/functions/pci-epf-test.c
> +++ b/drivers/pci/endpoint/functions/pci-epf-test.c
> @@ -84,8 +84,10 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>  	struct pci_epc *epc = epf->epc;
>  	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>  	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> +	int window;
>  
> -	src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr, reg->size);
> +	src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr,
> +					  &window, reg->size);
>  	if (!src_addr) {
>  		dev_err(dev, "Failed to allocate source address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
> @@ -93,15 +95,16 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>  		goto err;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, reg->src_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, window,
> +			       reg->src_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map source address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
>  		goto err_src_addr;
>  	}
>  
> -	dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr, reg->size);
> +	dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr,
> +					  &window, reg->size);
>  	if (!dst_addr) {
>  		dev_err(dev, "Failed to allocate destination address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> @@ -109,8 +112,8 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>  		goto err_src_map_addr;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, reg->dst_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, window,
> +			       reg->dst_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map destination address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> @@ -146,8 +149,9 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
>  	struct pci_epc *epc = epf->epc;
>  	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>  	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> +	int window;
>  
> -	src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> +	src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
>  	if (!src_addr) {
>  		dev_err(dev, "Failed to allocate address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
> @@ -155,8 +159,8 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
>  		goto err;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->src_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> +			       reg->src_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
> @@ -193,13 +197,14 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
>  	void __iomem *dst_addr;
>  	void *buf;
>  	phys_addr_t phys_addr;
> +	int window;
>  	struct pci_epf *epf = epf_test->epf;
>  	struct device *dev = &epf->dev;
>  	struct pci_epc *epc = epf->epc;
>  	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>  	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
>  
> -	dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> +	dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
>  	if (!dst_addr) {
>  		dev_err(dev, "Failed to allocate address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> @@ -207,8 +212,8 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
>  		goto err;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->dst_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> +			       reg->dst_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> diff --git a/drivers/pci/endpoint/pci-epc-core.c b/drivers/pci/endpoint/pci-epc-core.c
> index 2091508c1620..289c266c2d90 100644
> --- a/drivers/pci/endpoint/pci-epc-core.c
> +++ b/drivers/pci/endpoint/pci-epc-core.c
> @@ -358,13 +358,15 @@ EXPORT_SYMBOL_GPL(pci_epc_unmap_addr);
>   * @epc: the EPC device on which address is allocated
>   * @func_no: the endpoint function number in the EPC device
>   * @phys_addr: physical address of the local system
> + * @window: index to the window region where PCI address will be mapped
>   * @pci_addr: PCI address to which the physical address should be mapped
>   * @size: the size of the allocation
>   *
>   * Invoke to map CPU address with PCI address.
>   */
>  int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> -		     phys_addr_t phys_addr, u64 pci_addr, size_t size)
> +		     phys_addr_t phys_addr, int window,
> +		     u64 pci_addr, size_t size)
>  {
>  	int ret;
>  	unsigned long flags;
> @@ -376,7 +378,8 @@ int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
>  		return 0;
>  
>  	spin_lock_irqsave(&epc->lock, flags);
> -	ret = epc->ops->map_addr(epc, func_no, phys_addr, pci_addr, size);
> +	ret = epc->ops->map_addr(epc, func_no, phys_addr,
> +				 window, pci_addr, size);
>  	spin_unlock_irqrestore(&epc->lock, flags);
>  
>  	return ret;
> diff --git a/drivers/pci/endpoint/pci-epc-mem.c b/drivers/pci/endpoint/pci-epc-mem.c
> index d2b174ce15de..c955f2c97944 100644
> --- a/drivers/pci/endpoint/pci-epc-mem.c
> +++ b/drivers/pci/endpoint/pci-epc-mem.c
> @@ -39,56 +39,77 @@ static int pci_epc_mem_get_order(struct pci_epc_mem *mem, size_t size)
>   * __pci_epc_mem_init() - initialize the pci_epc_mem structure
>   * @epc: the EPC device that invoked pci_epc_mem_init
>   * @phys_base: the physical address of the base
> - * @size: the size of the address space
> + * @num_windows: number of windows device supports

struct pci_epc_mem_window is missing here.
>   * @page_size: size of each page
>   *
>   * Invoke to initialize the pci_epc_mem structure used by the
>   * endpoint functions to allocate mapped PCI address.
>   */
> -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_base, size_t size,
> -		       size_t page_size)
> +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *windows,
> +		       int num_windows, size_t page_size)
>  {
> -	int ret;
> -	struct pci_epc_mem *mem;
> -	unsigned long *bitmap;
> +	struct pci_epc_mem *mem = NULL;
> +	unsigned long *bitmap = NULL;
>  	unsigned int page_shift;
> -	int pages;
>  	int bitmap_size;
> +	int pages;
> +	int ret;
> +	int i;
> +
> +	epc->mem_windows = 0;
> +
> +	if (!windows)
> +		return -EINVAL;
> +
> +	if (num_windows <= 0)
> +		return -EINVAL;
>  
>  	if (page_size < PAGE_SIZE)
>  		page_size = PAGE_SIZE;
>  
>  	page_shift = ilog2(page_size);
> -	pages = size >> page_shift;
> -	bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
>  
> -	mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> -	if (!mem) {
> -		ret = -ENOMEM;
> -		goto err;
> -	}
> +	epc->mem = kcalloc(num_windows, sizeof(*mem), GFP_KERNEL);
> +	if (!epc->mem)
> +		return -EINVAL;
>  
> -	bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> -	if (!bitmap) {
> -		ret = -ENOMEM;
> -		goto err_mem;
> -	}
> +	for (i = 0; i < num_windows; i++) {
> +		pages = windows[i].phys_base >> page_shift;
> +		bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
>  
> -	mem->bitmap = bitmap;
> -	mem->phys_base = phys_base;
> -	mem->page_size = page_size;
> -	mem->pages = pages;
> -	mem->size = size;
> +		mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> +		if (!mem) {
> +			ret = -ENOMEM;
> +			goto err_mem;
> +		}
>  
> -	epc->mem = mem;
> +		bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> +		if (!bitmap) {
> +			ret = -ENOMEM;
> +			goto err_mem;
> +		}
> +
> +		mem->bitmap = bitmap;
> +		mem->window.phys_base = windows[i].phys_base;
> +		mem->page_size = page_size;
> +		mem->pages = pages;
> +		mem->window.size = windows[i].size;
> +		mem->window.map_size = 0;
> +
> +		epc->mem[i] = mem;
> +	}
> +	epc->mem_windows = num_windows;
>  
>  	return 0;
>  
>  err_mem:
> -	kfree(mem);
> +	for (; i >= 0; i--) {
> +		kfree(mem->bitmap);
> +		kfree(epc->mem[i]);
> +	}
> +	kfree(epc->mem);
>  
> -err:
> -return ret;
> +	return ret;
>  }
>  EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
>  
> @@ -101,48 +122,126 @@ EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
>   */
>  void pci_epc_mem_exit(struct pci_epc *epc)
>  {
> -	struct pci_epc_mem *mem = epc->mem;
> +	struct pci_epc_mem *mem;
> +	int i;
> +
> +	if (!epc->mem_windows)
> +		return;
> +
> +	for (i = 0; i <= epc->mem_windows; i--) {
> +		mem = epc->mem[i];
> +		kfree(mem->bitmap);
> +		kfree(epc->mem[i]);
> +	}
> +	kfree(epc->mem);
>  
>  	epc->mem = NULL;
> -	kfree(mem->bitmap);
> -	kfree(mem);
> +	epc->mem_windows = 0;
>  }
>  EXPORT_SYMBOL_GPL(pci_epc_mem_exit);
>  
> +static int pci_epc_find_best_fit_window(struct pci_epc *epc, size_t size)
> +{
> +	size_t window_least_size = 0;
> +	int best_fit_window = -1;
> +	struct pci_epc_mem *mem;
> +	size_t actual_size;
> +	int i;
> +
> +	for (i = 0; i < epc->mem_windows; i++) {
> +		mem = epc->mem[i];
> +
> +		/* if chunk from this region is already used skip it */
> +		if (mem->window.map_size)
> +			continue;
> +
> +		actual_size = ALIGN(size, mem->page_size);
> +
> +		if (best_fit_window == -1) {
> +			best_fit_window = i;
> +			window_least_size = mem->window.size;
> +		} else {
> +			if (actual_size <= mem->window.size &&
> +			    mem->window.size < window_least_size) {
> +				best_fit_window = i;
> +				window_least_size = mem->window.size;
> +			}
> +		}
> +	}
> +
> +	return best_fit_window;
> +}
> +
>  /**
>   * pci_epc_mem_alloc_addr() - allocate memory address from EPC addr space
>   * @epc: the EPC device on which memory has to be allocated
>   * @phys_addr: populate the allocated physical address here
> + * @window: populate the window here which will be used to map PCI address
>   * @size: the size of the address space that has to be allocated
>   *
>   * Invoke to allocate memory address from the EPC address space. This
>   * is usually done to map the remote RC address into the local system.
>   */
>  void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
> -				     phys_addr_t *phys_addr, size_t size)
> +				     phys_addr_t *phys_addr,
> +				     int *window, size_t size)
>  {
> +	int best_fit = PCI_EPC_DEFAULT_WINDOW;
> +	void __iomem *virt_addr = NULL;
> +	struct pci_epc_mem *mem;
> +	unsigned int page_shift;
>  	int pageno;
> -	void __iomem *virt_addr;
> -	struct pci_epc_mem *mem = epc->mem;
> -	unsigned int page_shift = ilog2(mem->page_size);
>  	int order;
>  
> +	if (epc->mem_windows <= 0)
> +		return NULL;
> +
> +	if (epc->mem_windows > 1) {
> +		best_fit = pci_epc_find_best_fit_window(epc, size);
> +		if (best_fit < 0)
> +			return NULL;
> +	}
> +
> +	mem = epc->mem[best_fit];
>  	size = ALIGN(size, mem->page_size);
> +	if (size > (mem->window.size - mem->window.map_size))
> +		return NULL;

Assume I have two mem regions, first region is of size 128MB and the second
region is of size 4GB. If there is a allocation request for 4GB, will the
allocation succeed?

Thanks
Kishon

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 2/5] pci: endpoint: add support to handle multiple base for mapping outbound memory
@ 2019-11-27  5:14     ` Kishon Vijay Abraham I
  0 siblings, 0 replies; 60+ messages in thread
From: Kishon Vijay Abraham I @ 2019-11-27  5:14 UTC (permalink / raw)
  To: Lad Prabhakar, Bjorn Helgaas, Rob Herring, Mark Rutland,
	Geert Uytterhoeven, Magnus Damm, Marek Vasut, Yoshihiro Shimoda,
	linux-pci
  Cc: devicetree, Chris Paterson, Lorenzo Pieralisi, Heiko Stuebner,
	Arnd Bergmann, Jingoo Han, Catalin Marinas, Shawn Lin,
	linux-kernel, Lad, Prabhakar, linux-renesas-soc, linux-rockchip,
	Tom Joseph, Greg Kroah-Hartman, Andrew Murray, Gustavo Pimentel,
	Will Deacon, linux-arm-kernel

Hi Prabhakar,

On 07/11/19 1:06 AM, Lad Prabhakar wrote:
> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> 
> rcar pcie controller has support to map multiple memory regions
> for mapping the outbound memory in local system, this feature
> inspires to add support for handling multiple memory bases in
> endpoint framework. In case of multiple memory regions only chunk
> or complete region can be mapped and this window needs to be
> passed to the controller driver.
> 
> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> Cc: <linux-rockchip@lists.infradead.org>
> Cc: Shawn Lin <shawn.lin@rock-chips.com>
> Cc: Heiko Stuebner <heiko@sntech.de>
> Cc: Tom Joseph <tjoseph@cadence.com>
> Cc: Jingoo Han <jingoohan1@gmail.com>
> Cc: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
> ---
>  .../pci/controller/dwc/pcie-designware-ep.c   |  30 ++-
>  drivers/pci/controller/pcie-cadence-ep.c      |  11 +-
>  drivers/pci/controller/pcie-rockchip-ep.c     |  13 +-
>  drivers/pci/endpoint/functions/pci-epf-test.c |  29 +--
>  drivers/pci/endpoint/pci-epc-core.c           |   7 +-
>  drivers/pci/endpoint/pci-epc-mem.c            | 189 ++++++++++++++----
>  include/linux/pci-epc.h                       |  43 ++--
>  7 files changed, 234 insertions(+), 88 deletions(-)
> 
> diff --git a/drivers/pci/controller/dwc/pcie-designware-ep.c b/drivers/pci/controller/dwc/pcie-designware-ep.c
> index 3dd2e2697294..8d23c20b9afd 100644
> --- a/drivers/pci/controller/dwc/pcie-designware-ep.c
> +++ b/drivers/pci/controller/dwc/pcie-designware-ep.c
> @@ -195,7 +195,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, u8 func_no,
>  }
>  
>  static int dw_pcie_ep_map_addr(struct pci_epc *epc, u8 func_no,
> -			       phys_addr_t addr,
> +			       phys_addr_t addr, int window,
>  			       u64 pci_addr, size_t size)
>  {
>  	int ret;
> @@ -367,6 +367,7 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
>  	unsigned int aligned_offset;
>  	u16 msg_ctrl, msg_data;
>  	u32 msg_addr_lower, msg_addr_upper, reg;
> +	int window = PCI_EPC_DEFAULT_WINDOW;
>  	u64 msg_addr;
>  	bool has_upper;
>  	int ret;
> @@ -390,11 +391,11 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
>  		reg = ep->msi_cap + PCI_MSI_DATA_32;
>  		msg_data = dw_pcie_readw_dbi(pci, reg);
>  	}
> -	aligned_offset = msg_addr_lower & (epc->mem->page_size - 1);
> +	aligned_offset = msg_addr_lower & (epc->mem[window]->page_size - 1);
>  	msg_addr = ((u64)msg_addr_upper) << 32 |
>  			(msg_addr_lower & ~aligned_offset);
> -	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> -				  epc->mem->page_size);
> +	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> +				  msg_addr, epc->mem[window]->page_size);
>  	if (ret)
>  		return ret;
>  
> @@ -416,6 +417,7 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>  	u32 reg, msg_data, vec_ctrl;
>  	u64 tbl_addr, msg_addr, reg_u64;
>  	void __iomem *msix_tbl;
> +	int window = PCI_EPC_DEFAULT_WINDOW;
>  	int ret;
>  
>  	reg = ep->msix_cap + PCI_MSIX_TABLE;
> @@ -452,8 +454,8 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>  		return -EPERM;
>  	}
>  
> -	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> -				  epc->mem->page_size);
> +	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> +				  msg_addr, epc->mem[window]->page_size);
>  	if (ret)
>  		return ret;
>  
> @@ -466,10 +468,11 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>  
>  void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
>  {
> +	int window = PCI_EPC_DEFAULT_WINDOW;
>  	struct pci_epc *epc = ep->epc;
>  
>  	pci_epc_mem_free_addr(epc, ep->msi_mem_phys, ep->msi_mem,
> -			      epc->mem->page_size);
> +			      epc->mem[window]->page_size);
>  
>  	pci_epc_mem_exit(epc);
>  }
> @@ -499,9 +502,12 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
>  	u32 reg;
>  	void *addr;
>  	u8 hdr_type;
> +	int window;
>  	unsigned int nbars;
>  	unsigned int offset;
>  	struct pci_epc *epc;
> +	size_t msi_page_size;
> +	struct pci_epc_mem_window mem_window;
>  	struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
>  	struct device *dev = pci->dev;
>  	struct device_node *np = dev->of_node;
> @@ -574,15 +580,17 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
>  	if (ret < 0)
>  		epc->max_functions = 1;
>  
> -	ret = __pci_epc_mem_init(epc, ep->phys_base, ep->addr_size,
> -				 ep->page_size);
> +	mem_window.phys_base = ep->phys_base;
> +	mem_window.size = ep->addr_size;
> +	ret = __pci_epc_mem_init(epc, &mem_window, 1, ep->page_size);
>  	if (ret < 0) {
>  		dev_err(dev, "Failed to initialize address space\n");
>  		return ret;
>  	}
>  
> -	ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys,
> -					     epc->mem->page_size);
> +	msi_page_size = epc->mem[PCI_EPC_DEFAULT_WINDOW]->page_size;
> +	ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys, &window,
> +					     msi_page_size);
>  	if (!ep->msi_mem) {
>  		dev_err(dev, "Failed to reserve memory for MSI/MSI-X\n");
>  		return -ENOMEM;
> diff --git a/drivers/pci/controller/pcie-cadence-ep.c b/drivers/pci/controller/pcie-cadence-ep.c
> index def7820cb824..7991b38a5350 100644
> --- a/drivers/pci/controller/pcie-cadence-ep.c
> +++ b/drivers/pci/controller/pcie-cadence-ep.c
> @@ -172,7 +172,7 @@ static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
>  }
>  
>  static int cdns_pcie_ep_map_addr(struct pci_epc *epc, u8 fn, phys_addr_t addr,
> -				 u64 pci_addr, size_t size)
> +				 int window, u64 pci_addr, size_t size)
>  {
>  	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>  	struct cdns_pcie *pcie = &ep->pcie;
> @@ -434,12 +434,14 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
>  {
>  	struct device *dev = &pdev->dev;
>  	struct device_node *np = dev->of_node;
> +	struct pci_epc_mem_window mem_window;
>  	struct cdns_pcie_ep *ep;
>  	struct cdns_pcie *pcie;
>  	struct pci_epc *epc;
>  	struct resource *res;
>  	int ret;
>  	int phy_count;
> +	int window;
>  
>  	ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
>  	if (!ep)
> @@ -502,15 +504,16 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
>  	if (of_property_read_u8(np, "max-functions", &epc->max_functions) < 0)
>  		epc->max_functions = 1;
>  
> -	ret = pci_epc_mem_init(epc, pcie->mem_res->start,
> -			       resource_size(pcie->mem_res));
> +	mem_window.phys_base = pcie->mem_res->start;
> +	mem_window.size = resource_size(pcie->mem_res);
> +	ret = pci_epc_mem_init(epc, &mem_window, 1);
>  	if (ret < 0) {
>  		dev_err(dev, "failed to initialize the memory space\n");
>  		goto err_init;
>  	}
>  
>  	ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> -						  SZ_128K);
> +						  &window, SZ_128K);
>  	if (!ep->irq_cpu_addr) {
>  		dev_err(dev, "failed to reserve memory space for MSI\n");
>  		ret = -ENOMEM;
> diff --git a/drivers/pci/controller/pcie-rockchip-ep.c b/drivers/pci/controller/pcie-rockchip-ep.c
> index d743b0a48988..d59e85c8d319 100644
> --- a/drivers/pci/controller/pcie-rockchip-ep.c
> +++ b/drivers/pci/controller/pcie-rockchip-ep.c
> @@ -256,8 +256,8 @@ static void rockchip_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
>  }
>  
>  static int rockchip_pcie_ep_map_addr(struct pci_epc *epc, u8 fn,
> -				     phys_addr_t addr, u64 pci_addr,
> -				     size_t size)
> +				     phys_addr_t addr, int window,
> +				     u64 pci_addr, size_t size)
>  {
>  	struct rockchip_pcie_ep *ep = epc_get_drvdata(epc);
>  	struct rockchip_pcie *pcie = &ep->rockchip;
> @@ -562,11 +562,13 @@ static const struct of_device_id rockchip_pcie_ep_of_match[] = {
>  
>  static int rockchip_pcie_ep_probe(struct platform_device *pdev)
>  {
> +	struct pci_epc_mem_window mem_window;
>  	struct device *dev = &pdev->dev;
>  	struct rockchip_pcie_ep *ep;
>  	struct rockchip_pcie *rockchip;
>  	struct pci_epc *epc;
>  	size_t max_regions;
> +	int window;
>  	int err;
>  
>  	ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> @@ -614,15 +616,16 @@ static int rockchip_pcie_ep_probe(struct platform_device *pdev)
>  	/* Only enable function 0 by default */
>  	rockchip_pcie_write(rockchip, BIT(0), PCIE_CORE_PHY_FUNC_CFG);
>  
> -	err = pci_epc_mem_init(epc, rockchip->mem_res->start,
> -			       resource_size(rockchip->mem_res));
> +	mem_window.phys_base = rockchip->mem_res->start;
> +	mem_window.size = resource_size(rockchip->mem_res);
> +	err = pci_epc_mem_init(epc, &mem_window, 1);
>  	if (err < 0) {
>  		dev_err(dev, "failed to initialize the memory space\n");
>  		goto err_uninit_port;
>  	}
>  
>  	ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> -						  SZ_128K);
> +						  &window, SZ_128K);
>  	if (!ep->irq_cpu_addr) {
>  		dev_err(dev, "failed to reserve memory space for MSI\n");
>  		err = -ENOMEM;
> diff --git a/drivers/pci/endpoint/functions/pci-epf-test.c b/drivers/pci/endpoint/functions/pci-epf-test.c
> index 5d74f81ddfe4..475228011703 100644
> --- a/drivers/pci/endpoint/functions/pci-epf-test.c
> +++ b/drivers/pci/endpoint/functions/pci-epf-test.c
> @@ -84,8 +84,10 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>  	struct pci_epc *epc = epf->epc;
>  	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>  	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> +	int window;
>  
> -	src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr, reg->size);
> +	src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr,
> +					  &window, reg->size);
>  	if (!src_addr) {
>  		dev_err(dev, "Failed to allocate source address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
> @@ -93,15 +95,16 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>  		goto err;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, reg->src_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, window,
> +			       reg->src_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map source address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
>  		goto err_src_addr;
>  	}
>  
> -	dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr, reg->size);
> +	dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr,
> +					  &window, reg->size);
>  	if (!dst_addr) {
>  		dev_err(dev, "Failed to allocate destination address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> @@ -109,8 +112,8 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>  		goto err_src_map_addr;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, reg->dst_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, window,
> +			       reg->dst_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map destination address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> @@ -146,8 +149,9 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
>  	struct pci_epc *epc = epf->epc;
>  	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>  	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> +	int window;
>  
> -	src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> +	src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
>  	if (!src_addr) {
>  		dev_err(dev, "Failed to allocate address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
> @@ -155,8 +159,8 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
>  		goto err;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->src_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> +			       reg->src_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
> @@ -193,13 +197,14 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
>  	void __iomem *dst_addr;
>  	void *buf;
>  	phys_addr_t phys_addr;
> +	int window;
>  	struct pci_epf *epf = epf_test->epf;
>  	struct device *dev = &epf->dev;
>  	struct pci_epc *epc = epf->epc;
>  	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>  	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
>  
> -	dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> +	dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
>  	if (!dst_addr) {
>  		dev_err(dev, "Failed to allocate address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> @@ -207,8 +212,8 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
>  		goto err;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->dst_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> +			       reg->dst_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> diff --git a/drivers/pci/endpoint/pci-epc-core.c b/drivers/pci/endpoint/pci-epc-core.c
> index 2091508c1620..289c266c2d90 100644
> --- a/drivers/pci/endpoint/pci-epc-core.c
> +++ b/drivers/pci/endpoint/pci-epc-core.c
> @@ -358,13 +358,15 @@ EXPORT_SYMBOL_GPL(pci_epc_unmap_addr);
>   * @epc: the EPC device on which address is allocated
>   * @func_no: the endpoint function number in the EPC device
>   * @phys_addr: physical address of the local system
> + * @window: index to the window region where PCI address will be mapped
>   * @pci_addr: PCI address to which the physical address should be mapped
>   * @size: the size of the allocation
>   *
>   * Invoke to map CPU address with PCI address.
>   */
>  int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> -		     phys_addr_t phys_addr, u64 pci_addr, size_t size)
> +		     phys_addr_t phys_addr, int window,
> +		     u64 pci_addr, size_t size)
>  {
>  	int ret;
>  	unsigned long flags;
> @@ -376,7 +378,8 @@ int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
>  		return 0;
>  
>  	spin_lock_irqsave(&epc->lock, flags);
> -	ret = epc->ops->map_addr(epc, func_no, phys_addr, pci_addr, size);
> +	ret = epc->ops->map_addr(epc, func_no, phys_addr,
> +				 window, pci_addr, size);
>  	spin_unlock_irqrestore(&epc->lock, flags);
>  
>  	return ret;
> diff --git a/drivers/pci/endpoint/pci-epc-mem.c b/drivers/pci/endpoint/pci-epc-mem.c
> index d2b174ce15de..c955f2c97944 100644
> --- a/drivers/pci/endpoint/pci-epc-mem.c
> +++ b/drivers/pci/endpoint/pci-epc-mem.c
> @@ -39,56 +39,77 @@ static int pci_epc_mem_get_order(struct pci_epc_mem *mem, size_t size)
>   * __pci_epc_mem_init() - initialize the pci_epc_mem structure
>   * @epc: the EPC device that invoked pci_epc_mem_init
>   * @phys_base: the physical address of the base
> - * @size: the size of the address space
> + * @num_windows: number of windows device supports

struct pci_epc_mem_window is missing here.
>   * @page_size: size of each page
>   *
>   * Invoke to initialize the pci_epc_mem structure used by the
>   * endpoint functions to allocate mapped PCI address.
>   */
> -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_base, size_t size,
> -		       size_t page_size)
> +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *windows,
> +		       int num_windows, size_t page_size)
>  {
> -	int ret;
> -	struct pci_epc_mem *mem;
> -	unsigned long *bitmap;
> +	struct pci_epc_mem *mem = NULL;
> +	unsigned long *bitmap = NULL;
>  	unsigned int page_shift;
> -	int pages;
>  	int bitmap_size;
> +	int pages;
> +	int ret;
> +	int i;
> +
> +	epc->mem_windows = 0;
> +
> +	if (!windows)
> +		return -EINVAL;
> +
> +	if (num_windows <= 0)
> +		return -EINVAL;
>  
>  	if (page_size < PAGE_SIZE)
>  		page_size = PAGE_SIZE;
>  
>  	page_shift = ilog2(page_size);
> -	pages = size >> page_shift;
> -	bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
>  
> -	mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> -	if (!mem) {
> -		ret = -ENOMEM;
> -		goto err;
> -	}
> +	epc->mem = kcalloc(num_windows, sizeof(*mem), GFP_KERNEL);
> +	if (!epc->mem)
> +		return -EINVAL;
>  
> -	bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> -	if (!bitmap) {
> -		ret = -ENOMEM;
> -		goto err_mem;
> -	}
> +	for (i = 0; i < num_windows; i++) {
> +		pages = windows[i].phys_base >> page_shift;
> +		bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
>  
> -	mem->bitmap = bitmap;
> -	mem->phys_base = phys_base;
> -	mem->page_size = page_size;
> -	mem->pages = pages;
> -	mem->size = size;
> +		mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> +		if (!mem) {
> +			ret = -ENOMEM;
> +			goto err_mem;
> +		}
>  
> -	epc->mem = mem;
> +		bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> +		if (!bitmap) {
> +			ret = -ENOMEM;
> +			goto err_mem;
> +		}
> +
> +		mem->bitmap = bitmap;
> +		mem->window.phys_base = windows[i].phys_base;
> +		mem->page_size = page_size;
> +		mem->pages = pages;
> +		mem->window.size = windows[i].size;
> +		mem->window.map_size = 0;
> +
> +		epc->mem[i] = mem;
> +	}
> +	epc->mem_windows = num_windows;
>  
>  	return 0;
>  
>  err_mem:
> -	kfree(mem);
> +	for (; i >= 0; i--) {
> +		kfree(mem->bitmap);
> +		kfree(epc->mem[i]);
> +	}
> +	kfree(epc->mem);
>  
> -err:
> -return ret;
> +	return ret;
>  }
>  EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
>  
> @@ -101,48 +122,126 @@ EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
>   */
>  void pci_epc_mem_exit(struct pci_epc *epc)
>  {
> -	struct pci_epc_mem *mem = epc->mem;
> +	struct pci_epc_mem *mem;
> +	int i;
> +
> +	if (!epc->mem_windows)
> +		return;
> +
> +	for (i = 0; i <= epc->mem_windows; i--) {
> +		mem = epc->mem[i];
> +		kfree(mem->bitmap);
> +		kfree(epc->mem[i]);
> +	}
> +	kfree(epc->mem);
>  
>  	epc->mem = NULL;
> -	kfree(mem->bitmap);
> -	kfree(mem);
> +	epc->mem_windows = 0;
>  }
>  EXPORT_SYMBOL_GPL(pci_epc_mem_exit);
>  
> +static int pci_epc_find_best_fit_window(struct pci_epc *epc, size_t size)
> +{
> +	size_t window_least_size = 0;
> +	int best_fit_window = -1;
> +	struct pci_epc_mem *mem;
> +	size_t actual_size;
> +	int i;
> +
> +	for (i = 0; i < epc->mem_windows; i++) {
> +		mem = epc->mem[i];
> +
> +		/* if chunk from this region is already used skip it */
> +		if (mem->window.map_size)
> +			continue;
> +
> +		actual_size = ALIGN(size, mem->page_size);
> +
> +		if (best_fit_window == -1) {
> +			best_fit_window = i;
> +			window_least_size = mem->window.size;
> +		} else {
> +			if (actual_size <= mem->window.size &&
> +			    mem->window.size < window_least_size) {
> +				best_fit_window = i;
> +				window_least_size = mem->window.size;
> +			}
> +		}
> +	}
> +
> +	return best_fit_window;
> +}
> +
>  /**
>   * pci_epc_mem_alloc_addr() - allocate memory address from EPC addr space
>   * @epc: the EPC device on which memory has to be allocated
>   * @phys_addr: populate the allocated physical address here
> + * @window: populate the window here which will be used to map PCI address
>   * @size: the size of the address space that has to be allocated
>   *
>   * Invoke to allocate memory address from the EPC address space. This
>   * is usually done to map the remote RC address into the local system.
>   */
>  void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
> -				     phys_addr_t *phys_addr, size_t size)
> +				     phys_addr_t *phys_addr,
> +				     int *window, size_t size)
>  {
> +	int best_fit = PCI_EPC_DEFAULT_WINDOW;
> +	void __iomem *virt_addr = NULL;
> +	struct pci_epc_mem *mem;
> +	unsigned int page_shift;
>  	int pageno;
> -	void __iomem *virt_addr;
> -	struct pci_epc_mem *mem = epc->mem;
> -	unsigned int page_shift = ilog2(mem->page_size);
>  	int order;
>  
> +	if (epc->mem_windows <= 0)
> +		return NULL;
> +
> +	if (epc->mem_windows > 1) {
> +		best_fit = pci_epc_find_best_fit_window(epc, size);
> +		if (best_fit < 0)
> +			return NULL;
> +	}
> +
> +	mem = epc->mem[best_fit];
>  	size = ALIGN(size, mem->page_size);
> +	if (size > (mem->window.size - mem->window.map_size))
> +		return NULL;

Assume I have two mem regions, first region is of size 128MB and the second
region is of size 4GB. If there is a allocation request for 4GB, will the
allocation succeed?

Thanks
Kishon

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
  2019-11-13  4:08           ` Rob Herring
@ 2019-11-27  5:44             ` Kishon Vijay Abraham I
  -1 siblings, 0 replies; 60+ messages in thread
From: Kishon Vijay Abraham I @ 2019-11-27  5:44 UTC (permalink / raw)
  To: Rob Herring, Geert Uytterhoeven
  Cc: Lad, Prabhakar, Bjorn Helgaas, Mark Rutland, Magnus Damm,
	Marek Vasut, Yoshihiro Shimoda, linux-pci, Catalin Marinas,
	Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Linux Kernel Mailing List, Linux ARM, Linux-Renesas,
	Chris Paterson, Lad, Prabhakar

Hi,

On 13/11/19 9:38 AM, Rob Herring wrote:
> On Thu, Nov 07, 2019 at 09:08:35PM +0100, Geert Uytterhoeven wrote:
>> Hi Prabhakar,
>>
>> On Thu, Nov 7, 2019 at 10:26 AM Lad, Prabhakar
>> <prabhakar.csengg@gmail.com> wrote:
>>> On Thu, Nov 7, 2019 at 8:44 AM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
>>>> On Wed, Nov 6, 2019 at 8:36 PM Lad Prabhakar <prabhakar.csengg@gmail.com> wrote:
>>>>> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
>>>>>
>>>>> This patch adds the bindings for the R-Car PCIe endpoint driver.
>>>>>
>>>>> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
>>>>
>>>> Thanks for your patch!
>>>>
>>>>> --- /dev/null
>>>>> +++ b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
>>>>> @@ -0,0 +1,43 @@
>>>>> +* Renesas R-Car PCIe Endpoint Controller DT description
>>>>> +
>>>>> +Required properties:
>>>>> +           "renesas,pcie-ep-r8a774c0" for the R8A774C0 SoC;
>>>>> +           "renesas,pcie-ep-rcar-gen3" for a generic R-Car Gen3 or
>>>>> +                                    RZ/G2 compatible device.
>>>>
>>>> Unless I'm missing something, this is for the exact same hardware block as
>>>> Documentation/devicetree/bindings/pci/rcar-pci.txt?
>>>> So shouldn't you amend those bindings, instead of adding new compatible
>>>> values?
>>>> Please remember that DT describes hardware, not software policy.
>>>> So IMHO choosing between host and endpoint is purely a configuration
>>>> issue, and could be indicated by the presence or lack of some DT properties.
>>>> E.g. host mode requires both "bus-range" and "device_type" properties,
>>>> so their absence could indicate endpoint mode.
>>>>
>>> yes its the same hardware block as described in the rcar-pci.txt, I
>>> did think about amending it
>>> but  it might turn out to be bit messy,
>>>
>>> required properties host ======required properties Endpoint
>>> ====================||==================
>>> 1: reg                                || reg
>>> 2:bus-range                      || reg names
>>> 3: device_type                  || resets
>>> 4: ranges                          || clocks
>>> 5: dma-ranges                  || clock-names
>>> 6: interrupts                      ||
>>> 7: interrupt-cells               ||
>>> 8: interrupt-map-mask     ||
>>> 9: clocks                          ||
>>> 10: clock-names             ||
>>
>> We have a similar situation with SPI, where a controller can operate in
>> master or slave mode, based on the absence or presence of the
>> "spi-slave" DT property.
>>
>>> and if I go ahead with the same compatible string that would mean to
>>> add support for endpoint
>>> mode in the host driver itself. I did follow the examples of
>>
>> You can still have two separate drivers, binding against the same
>> compatible value.  Just let the .probe() function return -ENODEV if it
>> discovers (by looking at DT properties) if the node is configured for
>> the other mode.
>> Which brings us to my next questions: is there any code that could be
>> shared between the drivers for the two modes?
>>
>>> rockchip/cadence/designware where
>>> its the same hardware block but has two different binding files one
>>> for host mode and other for
>>> endpoint mode.
>>
>> Having two separate DT binding documents sounds fine to me, if unifying
>> them makes things too complex.
>> However, I think they should use the same compatible value, because the
>> hardware block is the same, but just used in a different mode.
>>
>> Rob/Mark: Any input from the DT maintainers?
> 
> Separate files makes sense because different modes will want to 
> include different common schemas. We've generally been doing different 
> compatibles too which makes validating the node has the right set of 
> properties easier.
>  
>>>>> +- reg: Five register ranges as listed in the reg-names property
>>>>> +- reg-names: Must include the following names
>>>>> +       - "apb-base"
>>>>> +       - "memory0"
>>>>> +       - "memory1"
>>>>> +       - "memory2"
>>>>> +       - "memory3"
>>>>
>>>> What is the purpose of the last 4 regions?
>>>> Can they be chosen by the driver, at runtime?
>>>>
>>> no the driver cannot choose them at runtime, as these are the only
>>> PCIE memory(0/1/2/3) ranges
>>> in the AXI address space where host memory can be mapped.
>>
>> Are they fixed by the PCIe hardware, i.e. could they be looked up by the
>> driver based on the compatible value?
> 
> That would be strange for a memory range.
> 
> Sounds like like 'ranges' though I'm not sure if 'ranges' for an EP 
> makes sense or what that should look like.

These are similar to "memory node" with multiple address, size pairs. I'm
thinking if these should be added as a subnode within PCIe EP controller device
tree node?

Thanks
Kishon

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
@ 2019-11-27  5:44             ` Kishon Vijay Abraham I
  0 siblings, 0 replies; 60+ messages in thread
From: Kishon Vijay Abraham I @ 2019-11-27  5:44 UTC (permalink / raw)
  To: Rob Herring, Geert Uytterhoeven
  Cc: Mark Rutland,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Chris Paterson, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, linux-pci, Yoshihiro Shimoda, Magnus Damm,
	Linux Kernel Mailing List, Lad, Prabhakar, Linux-Renesas, Lad,
	Prabhakar, Catalin Marinas, Bjorn Helgaas, Andrew Murray,
	Will Deacon, Linux ARM, Marek Vasut

Hi,

On 13/11/19 9:38 AM, Rob Herring wrote:
> On Thu, Nov 07, 2019 at 09:08:35PM +0100, Geert Uytterhoeven wrote:
>> Hi Prabhakar,
>>
>> On Thu, Nov 7, 2019 at 10:26 AM Lad, Prabhakar
>> <prabhakar.csengg@gmail.com> wrote:
>>> On Thu, Nov 7, 2019 at 8:44 AM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
>>>> On Wed, Nov 6, 2019 at 8:36 PM Lad Prabhakar <prabhakar.csengg@gmail.com> wrote:
>>>>> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
>>>>>
>>>>> This patch adds the bindings for the R-Car PCIe endpoint driver.
>>>>>
>>>>> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
>>>>
>>>> Thanks for your patch!
>>>>
>>>>> --- /dev/null
>>>>> +++ b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
>>>>> @@ -0,0 +1,43 @@
>>>>> +* Renesas R-Car PCIe Endpoint Controller DT description
>>>>> +
>>>>> +Required properties:
>>>>> +           "renesas,pcie-ep-r8a774c0" for the R8A774C0 SoC;
>>>>> +           "renesas,pcie-ep-rcar-gen3" for a generic R-Car Gen3 or
>>>>> +                                    RZ/G2 compatible device.
>>>>
>>>> Unless I'm missing something, this is for the exact same hardware block as
>>>> Documentation/devicetree/bindings/pci/rcar-pci.txt?
>>>> So shouldn't you amend those bindings, instead of adding new compatible
>>>> values?
>>>> Please remember that DT describes hardware, not software policy.
>>>> So IMHO choosing between host and endpoint is purely a configuration
>>>> issue, and could be indicated by the presence or lack of some DT properties.
>>>> E.g. host mode requires both "bus-range" and "device_type" properties,
>>>> so their absence could indicate endpoint mode.
>>>>
>>> yes its the same hardware block as described in the rcar-pci.txt, I
>>> did think about amending it
>>> but  it might turn out to be bit messy,
>>>
>>> required properties host ======required properties Endpoint
>>> ====================||==================
>>> 1: reg                                || reg
>>> 2:bus-range                      || reg names
>>> 3: device_type                  || resets
>>> 4: ranges                          || clocks
>>> 5: dma-ranges                  || clock-names
>>> 6: interrupts                      ||
>>> 7: interrupt-cells               ||
>>> 8: interrupt-map-mask     ||
>>> 9: clocks                          ||
>>> 10: clock-names             ||
>>
>> We have a similar situation with SPI, where a controller can operate in
>> master or slave mode, based on the absence or presence of the
>> "spi-slave" DT property.
>>
>>> and if I go ahead with the same compatible string that would mean to
>>> add support for endpoint
>>> mode in the host driver itself. I did follow the examples of
>>
>> You can still have two separate drivers, binding against the same
>> compatible value.  Just let the .probe() function return -ENODEV if it
>> discovers (by looking at DT properties) if the node is configured for
>> the other mode.
>> Which brings us to my next questions: is there any code that could be
>> shared between the drivers for the two modes?
>>
>>> rockchip/cadence/designware where
>>> its the same hardware block but has two different binding files one
>>> for host mode and other for
>>> endpoint mode.
>>
>> Having two separate DT binding documents sounds fine to me, if unifying
>> them makes things too complex.
>> However, I think they should use the same compatible value, because the
>> hardware block is the same, but just used in a different mode.
>>
>> Rob/Mark: Any input from the DT maintainers?
> 
> Separate files makes sense because different modes will want to 
> include different common schemas. We've generally been doing different 
> compatibles too which makes validating the node has the right set of 
> properties easier.
>  
>>>>> +- reg: Five register ranges as listed in the reg-names property
>>>>> +- reg-names: Must include the following names
>>>>> +       - "apb-base"
>>>>> +       - "memory0"
>>>>> +       - "memory1"
>>>>> +       - "memory2"
>>>>> +       - "memory3"
>>>>
>>>> What is the purpose of the last 4 regions?
>>>> Can they be chosen by the driver, at runtime?
>>>>
>>> no the driver cannot choose them at runtime, as these are the only
>>> PCIE memory(0/1/2/3) ranges
>>> in the AXI address space where host memory can be mapped.
>>
>> Are they fixed by the PCIe hardware, i.e. could they be looked up by the
>> driver based on the compatible value?
> 
> That would be strange for a memory range.
> 
> Sounds like like 'ranges' though I'm not sure if 'ranges' for an EP 
> makes sense or what that should look like.

These are similar to "memory node" with multiple address, size pairs. I'm
thinking if these should be added as a subnode within PCIe EP controller device
tree node?

Thanks
Kishon

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 5/5] misc: pci_endpoint_test: add device-id for RZ/G2 pcie controller
  2019-11-06 19:36   ` Lad Prabhakar
@ 2019-11-27  5:45     ` Kishon Vijay Abraham I
  -1 siblings, 0 replies; 60+ messages in thread
From: Kishon Vijay Abraham I @ 2019-11-27  5:45 UTC (permalink / raw)
  To: Lad Prabhakar, Bjorn Helgaas, Rob Herring, Mark Rutland,
	Geert Uytterhoeven, Magnus Damm, Marek Vasut, Yoshihiro Shimoda,
	linux-pci
  Cc: Catalin Marinas, Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray, devicetree, linux-kernel,
	linux-arm-kernel, linux-renesas-soc, Chris Paterson, Lad,
	Prabhakar



On 07/11/19 1:06 AM, Lad Prabhakar wrote:
> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>

Please add a simple commit log here.

Thanks
Kishon

> 
> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> ---
>  drivers/misc/pci_endpoint_test.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/drivers/misc/pci_endpoint_test.c b/drivers/misc/pci_endpoint_test.c
> index a5e317073d95..346b03b6e4b9 100644
> --- a/drivers/misc/pci_endpoint_test.c
> +++ b/drivers/misc/pci_endpoint_test.c
> @@ -66,6 +66,8 @@
>  
>  #define PCI_DEVICE_ID_TI_AM654			0xb00c
>  
> +#define PCI_DEVICE_ID_RENESAS_RZG2		0x002d
> +
>  #define is_am654_pci_dev(pdev)		\
>  		((pdev)->device == PCI_DEVICE_ID_TI_AM654)
>  
> @@ -797,6 +799,7 @@ static const struct pci_device_id pci_endpoint_test_tbl[] = {
>  	{ PCI_DEVICE(PCI_VENDOR_ID_TI, PCI_DEVICE_ID_TI_AM654),
>  	  .driver_data = (kernel_ulong_t)&am654_data
>  	},
> +	{ PCI_DEVICE(PCI_VENDOR_ID_RENESAS, PCI_DEVICE_ID_RENESAS_RZG2) },
>  	{ }
>  };
>  MODULE_DEVICE_TABLE(pci, pci_endpoint_test_tbl);
> 

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 5/5] misc: pci_endpoint_test: add device-id for RZ/G2 pcie controller
@ 2019-11-27  5:45     ` Kishon Vijay Abraham I
  0 siblings, 0 replies; 60+ messages in thread
From: Kishon Vijay Abraham I @ 2019-11-27  5:45 UTC (permalink / raw)
  To: Lad Prabhakar, Bjorn Helgaas, Rob Herring, Mark Rutland,
	Geert Uytterhoeven, Magnus Damm, Marek Vasut, Yoshihiro Shimoda,
	linux-pci
  Cc: devicetree, Chris Paterson, Lorenzo Pieralisi, Arnd Bergmann,
	Catalin Marinas, linux-kernel, Lad, Prabhakar, linux-renesas-soc,
	Greg Kroah-Hartman, Andrew Murray, Will Deacon, linux-arm-kernel



On 07/11/19 1:06 AM, Lad Prabhakar wrote:
> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>

Please add a simple commit log here.

Thanks
Kishon

> 
> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> ---
>  drivers/misc/pci_endpoint_test.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/drivers/misc/pci_endpoint_test.c b/drivers/misc/pci_endpoint_test.c
> index a5e317073d95..346b03b6e4b9 100644
> --- a/drivers/misc/pci_endpoint_test.c
> +++ b/drivers/misc/pci_endpoint_test.c
> @@ -66,6 +66,8 @@
>  
>  #define PCI_DEVICE_ID_TI_AM654			0xb00c
>  
> +#define PCI_DEVICE_ID_RENESAS_RZG2		0x002d
> +
>  #define is_am654_pci_dev(pdev)		\
>  		((pdev)->device == PCI_DEVICE_ID_TI_AM654)
>  
> @@ -797,6 +799,7 @@ static const struct pci_device_id pci_endpoint_test_tbl[] = {
>  	{ PCI_DEVICE(PCI_VENDOR_ID_TI, PCI_DEVICE_ID_TI_AM654),
>  	  .driver_data = (kernel_ulong_t)&am654_data
>  	},
> +	{ PCI_DEVICE(PCI_VENDOR_ID_RENESAS, PCI_DEVICE_ID_RENESAS_RZG2) },
>  	{ }
>  };
>  MODULE_DEVICE_TABLE(pci, pci_endpoint_test_tbl);
> 

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 2/5] pci: endpoint: add support to handle multiple base for mapping outbound memory
@ 2019-11-27  6:05     ` Kishon Vijay Abraham I
  0 siblings, 0 replies; 60+ messages in thread
From: Kishon Vijay Abraham I @ 2019-11-27  6:05 UTC (permalink / raw)
  To: Lad Prabhakar, Bjorn Helgaas, Rob Herring, Mark Rutland,
	Geert Uytterhoeven, Magnus Damm, Marek Vasut, Yoshihiro Shimoda,
	linux-pci
  Cc: Catalin Marinas, Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray, devicetree, linux-kernel,
	linux-arm-kernel, linux-renesas-soc, Chris Paterson, Lad,
	Prabhakar, linux-rockchip, Shawn Lin, Heiko Stuebner, Tom Joseph,
	Jingoo Han, Gustavo Pimentel

Hi Prabhakar,

On 07/11/19 1:06 AM, Lad Prabhakar wrote:
> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> 
> rcar pcie controller has support to map multiple memory regions
> for mapping the outbound memory in local system, this feature
> inspires to add support for handling multiple memory bases in
> endpoint framework. In case of multiple memory regions only chunk
> or complete region can be mapped and this window needs to be
> passed to the controller driver.

Wouldn't this result in lot of unused address and exhausting the memory regions
quickly? I think we could have a "type" associated with each of these regions
and smaller allocations could use one "type" and larger allocation could use a
different "type"? For instance we allocate memory regions for MSI interrupt and
for data transfer. We could use smaller allocations for MSI and larger
allocations for data transfers.

We could also extend the "type" for certain special requirements. For instance
I've seen platforms which have dedicated address space for high priority data
transfers (uses higher virtual channel numbers) and dedicated address space for
low priority data transfers (uses lower virtual channel numbers).

Here too we could give a special flag or type for higher priority address
regions and different type for lower priority address. While allocating, the
function drivers should provide a flag specifying the region where memory
should be allocated.

Can you add flags in struct pci_epc_mem_window and see how that comes out?

Thanks
Kishon

> 
> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> Cc: <linux-rockchip@lists.infradead.org>
> Cc: Shawn Lin <shawn.lin@rock-chips.com>
> Cc: Heiko Stuebner <heiko@sntech.de>
> Cc: Tom Joseph <tjoseph@cadence.com>
> Cc: Jingoo Han <jingoohan1@gmail.com>
> Cc: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
> ---
>  .../pci/controller/dwc/pcie-designware-ep.c   |  30 ++-
>  drivers/pci/controller/pcie-cadence-ep.c      |  11 +-
>  drivers/pci/controller/pcie-rockchip-ep.c     |  13 +-
>  drivers/pci/endpoint/functions/pci-epf-test.c |  29 +--
>  drivers/pci/endpoint/pci-epc-core.c           |   7 +-
>  drivers/pci/endpoint/pci-epc-mem.c            | 189 ++++++++++++++----
>  include/linux/pci-epc.h                       |  43 ++--
>  7 files changed, 234 insertions(+), 88 deletions(-)
> 
> diff --git a/drivers/pci/controller/dwc/pcie-designware-ep.c b/drivers/pci/controller/dwc/pcie-designware-ep.c
> index 3dd2e2697294..8d23c20b9afd 100644
> --- a/drivers/pci/controller/dwc/pcie-designware-ep.c
> +++ b/drivers/pci/controller/dwc/pcie-designware-ep.c
> @@ -195,7 +195,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, u8 func_no,
>  }
>  
>  static int dw_pcie_ep_map_addr(struct pci_epc *epc, u8 func_no,
> -			       phys_addr_t addr,
> +			       phys_addr_t addr, int window,
>  			       u64 pci_addr, size_t size)
>  {
>  	int ret;
> @@ -367,6 +367,7 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
>  	unsigned int aligned_offset;
>  	u16 msg_ctrl, msg_data;
>  	u32 msg_addr_lower, msg_addr_upper, reg;
> +	int window = PCI_EPC_DEFAULT_WINDOW;
>  	u64 msg_addr;
>  	bool has_upper;
>  	int ret;
> @@ -390,11 +391,11 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
>  		reg = ep->msi_cap + PCI_MSI_DATA_32;
>  		msg_data = dw_pcie_readw_dbi(pci, reg);
>  	}
> -	aligned_offset = msg_addr_lower & (epc->mem->page_size - 1);
> +	aligned_offset = msg_addr_lower & (epc->mem[window]->page_size - 1);
>  	msg_addr = ((u64)msg_addr_upper) << 32 |
>  			(msg_addr_lower & ~aligned_offset);
> -	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> -				  epc->mem->page_size);
> +	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> +				  msg_addr, epc->mem[window]->page_size);
>  	if (ret)
>  		return ret;
>  
> @@ -416,6 +417,7 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>  	u32 reg, msg_data, vec_ctrl;
>  	u64 tbl_addr, msg_addr, reg_u64;
>  	void __iomem *msix_tbl;
> +	int window = PCI_EPC_DEFAULT_WINDOW;
>  	int ret;
>  
>  	reg = ep->msix_cap + PCI_MSIX_TABLE;
> @@ -452,8 +454,8 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>  		return -EPERM;
>  	}
>  
> -	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> -				  epc->mem->page_size);
> +	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> +				  msg_addr, epc->mem[window]->page_size);
>  	if (ret)
>  		return ret;
>  
> @@ -466,10 +468,11 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>  
>  void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
>  {
> +	int window = PCI_EPC_DEFAULT_WINDOW;
>  	struct pci_epc *epc = ep->epc;
>  
>  	pci_epc_mem_free_addr(epc, ep->msi_mem_phys, ep->msi_mem,
> -			      epc->mem->page_size);
> +			      epc->mem[window]->page_size);
>  
>  	pci_epc_mem_exit(epc);
>  }
> @@ -499,9 +502,12 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
>  	u32 reg;
>  	void *addr;
>  	u8 hdr_type;
> +	int window;
>  	unsigned int nbars;
>  	unsigned int offset;
>  	struct pci_epc *epc;
> +	size_t msi_page_size;
> +	struct pci_epc_mem_window mem_window;
>  	struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
>  	struct device *dev = pci->dev;
>  	struct device_node *np = dev->of_node;
> @@ -574,15 +580,17 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
>  	if (ret < 0)
>  		epc->max_functions = 1;
>  
> -	ret = __pci_epc_mem_init(epc, ep->phys_base, ep->addr_size,
> -				 ep->page_size);
> +	mem_window.phys_base = ep->phys_base;
> +	mem_window.size = ep->addr_size;
> +	ret = __pci_epc_mem_init(epc, &mem_window, 1, ep->page_size);
>  	if (ret < 0) {
>  		dev_err(dev, "Failed to initialize address space\n");
>  		return ret;
>  	}
>  
> -	ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys,
> -					     epc->mem->page_size);
> +	msi_page_size = epc->mem[PCI_EPC_DEFAULT_WINDOW]->page_size;
> +	ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys, &window,
> +					     msi_page_size);
>  	if (!ep->msi_mem) {
>  		dev_err(dev, "Failed to reserve memory for MSI/MSI-X\n");
>  		return -ENOMEM;
> diff --git a/drivers/pci/controller/pcie-cadence-ep.c b/drivers/pci/controller/pcie-cadence-ep.c
> index def7820cb824..7991b38a5350 100644
> --- a/drivers/pci/controller/pcie-cadence-ep.c
> +++ b/drivers/pci/controller/pcie-cadence-ep.c
> @@ -172,7 +172,7 @@ static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
>  }
>  
>  static int cdns_pcie_ep_map_addr(struct pci_epc *epc, u8 fn, phys_addr_t addr,
> -				 u64 pci_addr, size_t size)
> +				 int window, u64 pci_addr, size_t size)
>  {
>  	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>  	struct cdns_pcie *pcie = &ep->pcie;
> @@ -434,12 +434,14 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
>  {
>  	struct device *dev = &pdev->dev;
>  	struct device_node *np = dev->of_node;
> +	struct pci_epc_mem_window mem_window;
>  	struct cdns_pcie_ep *ep;
>  	struct cdns_pcie *pcie;
>  	struct pci_epc *epc;
>  	struct resource *res;
>  	int ret;
>  	int phy_count;
> +	int window;
>  
>  	ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
>  	if (!ep)
> @@ -502,15 +504,16 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
>  	if (of_property_read_u8(np, "max-functions", &epc->max_functions) < 0)
>  		epc->max_functions = 1;
>  
> -	ret = pci_epc_mem_init(epc, pcie->mem_res->start,
> -			       resource_size(pcie->mem_res));
> +	mem_window.phys_base = pcie->mem_res->start;
> +	mem_window.size = resource_size(pcie->mem_res);
> +	ret = pci_epc_mem_init(epc, &mem_window, 1);
>  	if (ret < 0) {
>  		dev_err(dev, "failed to initialize the memory space\n");
>  		goto err_init;
>  	}
>  
>  	ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> -						  SZ_128K);
> +						  &window, SZ_128K);
>  	if (!ep->irq_cpu_addr) {
>  		dev_err(dev, "failed to reserve memory space for MSI\n");
>  		ret = -ENOMEM;
> diff --git a/drivers/pci/controller/pcie-rockchip-ep.c b/drivers/pci/controller/pcie-rockchip-ep.c
> index d743b0a48988..d59e85c8d319 100644
> --- a/drivers/pci/controller/pcie-rockchip-ep.c
> +++ b/drivers/pci/controller/pcie-rockchip-ep.c
> @@ -256,8 +256,8 @@ static void rockchip_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
>  }
>  
>  static int rockchip_pcie_ep_map_addr(struct pci_epc *epc, u8 fn,
> -				     phys_addr_t addr, u64 pci_addr,
> -				     size_t size)
> +				     phys_addr_t addr, int window,
> +				     u64 pci_addr, size_t size)
>  {
>  	struct rockchip_pcie_ep *ep = epc_get_drvdata(epc);
>  	struct rockchip_pcie *pcie = &ep->rockchip;
> @@ -562,11 +562,13 @@ static const struct of_device_id rockchip_pcie_ep_of_match[] = {
>  
>  static int rockchip_pcie_ep_probe(struct platform_device *pdev)
>  {
> +	struct pci_epc_mem_window mem_window;
>  	struct device *dev = &pdev->dev;
>  	struct rockchip_pcie_ep *ep;
>  	struct rockchip_pcie *rockchip;
>  	struct pci_epc *epc;
>  	size_t max_regions;
> +	int window;
>  	int err;
>  
>  	ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> @@ -614,15 +616,16 @@ static int rockchip_pcie_ep_probe(struct platform_device *pdev)
>  	/* Only enable function 0 by default */
>  	rockchip_pcie_write(rockchip, BIT(0), PCIE_CORE_PHY_FUNC_CFG);
>  
> -	err = pci_epc_mem_init(epc, rockchip->mem_res->start,
> -			       resource_size(rockchip->mem_res));
> +	mem_window.phys_base = rockchip->mem_res->start;
> +	mem_window.size = resource_size(rockchip->mem_res);
> +	err = pci_epc_mem_init(epc, &mem_window, 1);
>  	if (err < 0) {
>  		dev_err(dev, "failed to initialize the memory space\n");
>  		goto err_uninit_port;
>  	}
>  
>  	ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> -						  SZ_128K);
> +						  &window, SZ_128K);
>  	if (!ep->irq_cpu_addr) {
>  		dev_err(dev, "failed to reserve memory space for MSI\n");
>  		err = -ENOMEM;
> diff --git a/drivers/pci/endpoint/functions/pci-epf-test.c b/drivers/pci/endpoint/functions/pci-epf-test.c
> index 5d74f81ddfe4..475228011703 100644
> --- a/drivers/pci/endpoint/functions/pci-epf-test.c
> +++ b/drivers/pci/endpoint/functions/pci-epf-test.c
> @@ -84,8 +84,10 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>  	struct pci_epc *epc = epf->epc;
>  	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>  	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> +	int window;
>  
> -	src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr, reg->size);
> +	src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr,
> +					  &window, reg->size);
>  	if (!src_addr) {
>  		dev_err(dev, "Failed to allocate source address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
> @@ -93,15 +95,16 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>  		goto err;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, reg->src_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, window,
> +			       reg->src_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map source address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
>  		goto err_src_addr;
>  	}
>  
> -	dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr, reg->size);
> +	dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr,
> +					  &window, reg->size);
>  	if (!dst_addr) {
>  		dev_err(dev, "Failed to allocate destination address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> @@ -109,8 +112,8 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>  		goto err_src_map_addr;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, reg->dst_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, window,
> +			       reg->dst_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map destination address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> @@ -146,8 +149,9 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
>  	struct pci_epc *epc = epf->epc;
>  	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>  	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> +	int window;
>  
> -	src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> +	src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
>  	if (!src_addr) {
>  		dev_err(dev, "Failed to allocate address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
> @@ -155,8 +159,8 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
>  		goto err;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->src_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> +			       reg->src_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
> @@ -193,13 +197,14 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
>  	void __iomem *dst_addr;
>  	void *buf;
>  	phys_addr_t phys_addr;
> +	int window;
>  	struct pci_epf *epf = epf_test->epf;
>  	struct device *dev = &epf->dev;
>  	struct pci_epc *epc = epf->epc;
>  	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>  	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
>  
> -	dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> +	dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
>  	if (!dst_addr) {
>  		dev_err(dev, "Failed to allocate address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> @@ -207,8 +212,8 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
>  		goto err;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->dst_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> +			       reg->dst_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> diff --git a/drivers/pci/endpoint/pci-epc-core.c b/drivers/pci/endpoint/pci-epc-core.c
> index 2091508c1620..289c266c2d90 100644
> --- a/drivers/pci/endpoint/pci-epc-core.c
> +++ b/drivers/pci/endpoint/pci-epc-core.c
> @@ -358,13 +358,15 @@ EXPORT_SYMBOL_GPL(pci_epc_unmap_addr);
>   * @epc: the EPC device on which address is allocated
>   * @func_no: the endpoint function number in the EPC device
>   * @phys_addr: physical address of the local system
> + * @window: index to the window region where PCI address will be mapped
>   * @pci_addr: PCI address to which the physical address should be mapped
>   * @size: the size of the allocation
>   *
>   * Invoke to map CPU address with PCI address.
>   */
>  int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> -		     phys_addr_t phys_addr, u64 pci_addr, size_t size)
> +		     phys_addr_t phys_addr, int window,
> +		     u64 pci_addr, size_t size)
>  {
>  	int ret;
>  	unsigned long flags;
> @@ -376,7 +378,8 @@ int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
>  		return 0;
>  
>  	spin_lock_irqsave(&epc->lock, flags);
> -	ret = epc->ops->map_addr(epc, func_no, phys_addr, pci_addr, size);
> +	ret = epc->ops->map_addr(epc, func_no, phys_addr,
> +				 window, pci_addr, size);
>  	spin_unlock_irqrestore(&epc->lock, flags);
>  
>  	return ret;
> diff --git a/drivers/pci/endpoint/pci-epc-mem.c b/drivers/pci/endpoint/pci-epc-mem.c
> index d2b174ce15de..c955f2c97944 100644
> --- a/drivers/pci/endpoint/pci-epc-mem.c
> +++ b/drivers/pci/endpoint/pci-epc-mem.c
> @@ -39,56 +39,77 @@ static int pci_epc_mem_get_order(struct pci_epc_mem *mem, size_t size)
>   * __pci_epc_mem_init() - initialize the pci_epc_mem structure
>   * @epc: the EPC device that invoked pci_epc_mem_init
>   * @phys_base: the physical address of the base
> - * @size: the size of the address space
> + * @num_windows: number of windows device supports
>   * @page_size: size of each page
>   *
>   * Invoke to initialize the pci_epc_mem structure used by the
>   * endpoint functions to allocate mapped PCI address.
>   */
> -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_base, size_t size,
> -		       size_t page_size)
> +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *windows,
> +		       int num_windows, size_t page_size)
>  {
> -	int ret;
> -	struct pci_epc_mem *mem;
> -	unsigned long *bitmap;
> +	struct pci_epc_mem *mem = NULL;
> +	unsigned long *bitmap = NULL;
>  	unsigned int page_shift;
> -	int pages;
>  	int bitmap_size;
> +	int pages;
> +	int ret;
> +	int i;
> +
> +	epc->mem_windows = 0;
> +
> +	if (!windows)
> +		return -EINVAL;
> +
> +	if (num_windows <= 0)
> +		return -EINVAL;
>  
>  	if (page_size < PAGE_SIZE)
>  		page_size = PAGE_SIZE;
>  
>  	page_shift = ilog2(page_size);
> -	pages = size >> page_shift;
> -	bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
>  
> -	mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> -	if (!mem) {
> -		ret = -ENOMEM;
> -		goto err;
> -	}
> +	epc->mem = kcalloc(num_windows, sizeof(*mem), GFP_KERNEL);
> +	if (!epc->mem)
> +		return -EINVAL;
>  
> -	bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> -	if (!bitmap) {
> -		ret = -ENOMEM;
> -		goto err_mem;
> -	}
> +	for (i = 0; i < num_windows; i++) {
> +		pages = windows[i].phys_base >> page_shift;
> +		bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
>  
> -	mem->bitmap = bitmap;
> -	mem->phys_base = phys_base;
> -	mem->page_size = page_size;
> -	mem->pages = pages;
> -	mem->size = size;
> +		mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> +		if (!mem) {
> +			ret = -ENOMEM;
> +			goto err_mem;
> +		}
>  
> -	epc->mem = mem;
> +		bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> +		if (!bitmap) {
> +			ret = -ENOMEM;
> +			goto err_mem;
> +		}
> +
> +		mem->bitmap = bitmap;
> +		mem->window.phys_base = windows[i].phys_base;
> +		mem->page_size = page_size;
> +		mem->pages = pages;
> +		mem->window.size = windows[i].size;
> +		mem->window.map_size = 0;
> +
> +		epc->mem[i] = mem;
> +	}
> +	epc->mem_windows = num_windows;
>  
>  	return 0;
>  
>  err_mem:
> -	kfree(mem);
> +	for (; i >= 0; i--) {
> +		kfree(mem->bitmap);
> +		kfree(epc->mem[i]);
> +	}
> +	kfree(epc->mem);
>  
> -err:
> -return ret;
> +	return ret;
>  }
>  EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
>  
> @@ -101,48 +122,126 @@ EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
>   */
>  void pci_epc_mem_exit(struct pci_epc *epc)
>  {
> -	struct pci_epc_mem *mem = epc->mem;
> +	struct pci_epc_mem *mem;
> +	int i;
> +
> +	if (!epc->mem_windows)
> +		return;
> +
> +	for (i = 0; i <= epc->mem_windows; i--) {
> +		mem = epc->mem[i];
> +		kfree(mem->bitmap);
> +		kfree(epc->mem[i]);
> +	}
> +	kfree(epc->mem);
>  
>  	epc->mem = NULL;
> -	kfree(mem->bitmap);
> -	kfree(mem);
> +	epc->mem_windows = 0;
>  }
>  EXPORT_SYMBOL_GPL(pci_epc_mem_exit);
>  
> +static int pci_epc_find_best_fit_window(struct pci_epc *epc, size_t size)
> +{
> +	size_t window_least_size = 0;
> +	int best_fit_window = -1;
> +	struct pci_epc_mem *mem;
> +	size_t actual_size;
> +	int i;
> +
> +	for (i = 0; i < epc->mem_windows; i++) {
> +		mem = epc->mem[i];
> +
> +		/* if chunk from this region is already used skip it */
> +		if (mem->window.map_size)
> +			continue;
> +
> +		actual_size = ALIGN(size, mem->page_size);
> +
> +		if (best_fit_window == -1) {
> +			best_fit_window = i;
> +			window_least_size = mem->window.size;
> +		} else {
> +			if (actual_size <= mem->window.size &&
> +			    mem->window.size < window_least_size) {
> +				best_fit_window = i;
> +				window_least_size = mem->window.size;
> +			}
> +		}
> +	}
> +
> +	return best_fit_window;
> +}
> +
>  /**
>   * pci_epc_mem_alloc_addr() - allocate memory address from EPC addr space
>   * @epc: the EPC device on which memory has to be allocated
>   * @phys_addr: populate the allocated physical address here
> + * @window: populate the window here which will be used to map PCI address
>   * @size: the size of the address space that has to be allocated
>   *
>   * Invoke to allocate memory address from the EPC address space. This
>   * is usually done to map the remote RC address into the local system.
>   */
>  void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
> -				     phys_addr_t *phys_addr, size_t size)
> +				     phys_addr_t *phys_addr,
> +				     int *window, size_t size)
>  {
> +	int best_fit = PCI_EPC_DEFAULT_WINDOW;
> +	void __iomem *virt_addr = NULL;
> +	struct pci_epc_mem *mem;
> +	unsigned int page_shift;
>  	int pageno;
> -	void __iomem *virt_addr;
> -	struct pci_epc_mem *mem = epc->mem;
> -	unsigned int page_shift = ilog2(mem->page_size);
>  	int order;
>  
> +	if (epc->mem_windows <= 0)
> +		return NULL;
> +
> +	if (epc->mem_windows > 1) {
> +		best_fit = pci_epc_find_best_fit_window(epc, size);
> +		if (best_fit < 0)
> +			return NULL;
> +	}
> +
> +	mem = epc->mem[best_fit];
>  	size = ALIGN(size, mem->page_size);
> +	if (size > (mem->window.size - mem->window.map_size))
> +		return NULL;
> +	page_shift = ilog2(mem->page_size);
>  	order = pci_epc_mem_get_order(mem, size);
>  
>  	pageno = bitmap_find_free_region(mem->bitmap, mem->pages, order);
>  	if (pageno < 0)
>  		return NULL;
>  
> -	*phys_addr = mem->phys_base + ((phys_addr_t)pageno << page_shift);
> +	*phys_addr = mem->window.phys_base + (pageno << page_shift);
>  	virt_addr = ioremap(*phys_addr, size);
> -	if (!virt_addr)
> +	if (!virt_addr) {
>  		bitmap_release_region(mem->bitmap, pageno, order);
> +	} else {
> +		mem->window.map_size += size;
> +		*window = best_fit;
> +	}
>  
>  	return virt_addr;
>  }
>  EXPORT_SYMBOL_GPL(pci_epc_mem_alloc_addr);
>  
> +static int pci_epc_get_matching_window(struct pci_epc *epc,
> +				       phys_addr_t phys_addr)
> +{
> +	struct pci_epc_mem *mem;
> +	int i;
> +
> +	for (i = 0; i < epc->mem_windows; i++) {
> +		mem = epc->mem[i];
> +
> +		if (mem->window.phys_base == phys_addr)
> +			return i;
> +	}
> +
> +	return -EINVAL;
> +}
> +
>  /**
>   * pci_epc_mem_free_addr() - free the allocated memory address
>   * @epc: the EPC device on which memory was allocated
> @@ -155,16 +254,26 @@ EXPORT_SYMBOL_GPL(pci_epc_mem_alloc_addr);
>  void pci_epc_mem_free_addr(struct pci_epc *epc, phys_addr_t phys_addr,
>  			   void __iomem *virt_addr, size_t size)
>  {
> +	struct pci_epc_mem *mem;
> +	unsigned int page_shift;
> +	int window = 0;
>  	int pageno;
> -	struct pci_epc_mem *mem = epc->mem;
> -	unsigned int page_shift = ilog2(mem->page_size);
>  	int order;
>  
> +	if (epc->mem_windows > 1) {
> +		window = pci_epc_get_matching_window(epc, phys_addr);
> +		if (window < 0)
> +			return;
> +	}
> +
> +	mem = epc->mem[window];
> +	page_shift = ilog2(mem->page_size);
>  	iounmap(virt_addr);
> -	pageno = (phys_addr - mem->phys_base) >> page_shift;
> +	pageno = (phys_addr - mem->window.phys_base) >> page_shift;
>  	size = ALIGN(size, mem->page_size);
>  	order = pci_epc_mem_get_order(mem, size);
>  	bitmap_release_region(mem->bitmap, pageno, order);
> +	mem->window.map_size -= size;
>  }
>  EXPORT_SYMBOL_GPL(pci_epc_mem_free_addr);
>  
> diff --git a/include/linux/pci-epc.h b/include/linux/pci-epc.h
> index 56f1846b9d39..5e86d22dd5c5 100644
> --- a/include/linux/pci-epc.h
> +++ b/include/linux/pci-epc.h
> @@ -48,7 +48,8 @@ struct pci_epc_ops {
>  	void	(*clear_bar)(struct pci_epc *epc, u8 func_no,
>  			     struct pci_epf_bar *epf_bar);
>  	int	(*map_addr)(struct pci_epc *epc, u8 func_no,
> -			    phys_addr_t addr, u64 pci_addr, size_t size);
> +			    phys_addr_t addr, int window,
> +			    u64 pci_addr, size_t size);
>  	void	(*unmap_addr)(struct pci_epc *epc, u8 func_no,
>  			      phys_addr_t addr);
>  	int	(*set_msi)(struct pci_epc *epc, u8 func_no, u8 interrupts);
> @@ -64,17 +65,28 @@ struct pci_epc_ops {
>  	struct module *owner;
>  };
>  
> +#define PCI_EPC_DEFAULT_WINDOW		0
> +
> +/**
> + * struct pci_epc_mem_window - address window of the endpoint controller
> + * @phys_base: physical base address of the PCI address window
> + * @size: the size of the PCI address window
> + */
> +struct pci_epc_mem_window {
> +	phys_addr_t	phys_base;
> +	size_t		size;
> +	size_t		map_size;
> +};
> +
>  /**
>   * struct pci_epc_mem - address space of the endpoint controller
> - * @phys_base: physical base address of the PCI address space
> - * @size: the size of the PCI address space
> + * @window: address window of the endpoint controller
>   * @bitmap: bitmap to manage the PCI address space
> - * @pages: number of bits representing the address region
>   * @page_size: size of each page
> + * @pages: number of bits representing the address region
>   */
>  struct pci_epc_mem {
> -	phys_addr_t	phys_base;
> -	size_t		size;
> +	struct pci_epc_mem_window window;
>  	unsigned long	*bitmap;
>  	size_t		page_size;
>  	int		pages;
> @@ -85,7 +97,8 @@ struct pci_epc_mem {
>   * @dev: PCI EPC device
>   * @pci_epf: list of endpoint functions present in this EPC device
>   * @ops: function pointers for performing endpoint operations
> - * @mem: address space of the endpoint controller
> + * @mem: array of address space of the endpoint controller
> + * @mem_windows: number of windows supported by device
>   * @max_functions: max number of functions that can be configured in this EPC
>   * @group: configfs group representing the PCI EPC device
>   * @lock: spinlock to protect pci_epc ops
> @@ -94,7 +107,8 @@ struct pci_epc {
>  	struct device			dev;
>  	struct list_head		pci_epf;
>  	const struct pci_epc_ops	*ops;
> -	struct pci_epc_mem		*mem;
> +	struct pci_epc_mem		**mem;
> +	int				mem_windows;
>  	u8				max_functions;
>  	struct config_group		*group;
>  	/* spinlock to protect against concurrent access of EP controller */
> @@ -128,8 +142,8 @@ struct pci_epc_features {
>  #define devm_pci_epc_create(dev, ops)    \
>  		__devm_pci_epc_create((dev), (ops), THIS_MODULE)
>  
> -#define pci_epc_mem_init(epc, phys_addr, size)	\
> -		__pci_epc_mem_init((epc), (phys_addr), (size), PAGE_SIZE)
> +#define pci_epc_mem_init(epc, windows, num_windows)	\
> +		__pci_epc_mem_init((epc), windows, num_windows, PAGE_SIZE)
>  
>  static inline void epc_set_drvdata(struct pci_epc *epc, void *data)
>  {
> @@ -159,7 +173,7 @@ int pci_epc_set_bar(struct pci_epc *epc, u8 func_no,
>  void pci_epc_clear_bar(struct pci_epc *epc, u8 func_no,
>  		       struct pci_epf_bar *epf_bar);
>  int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> -		     phys_addr_t phys_addr,
> +		     phys_addr_t phys_addr, int window,
>  		     u64 pci_addr, size_t size);
>  void pci_epc_unmap_addr(struct pci_epc *epc, u8 func_no,
>  			phys_addr_t phys_addr);
> @@ -178,11 +192,12 @@ unsigned int pci_epc_get_first_free_bar(const struct pci_epc_features
>  struct pci_epc *pci_epc_get(const char *epc_name);
>  void pci_epc_put(struct pci_epc *epc);
>  
> -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_addr, size_t size,
> -		       size_t page_size);
> +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *window,
> +		       int num_windows, size_t page_size);
>  void pci_epc_mem_exit(struct pci_epc *epc);
>  void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
> -				     phys_addr_t *phys_addr, size_t size);
> +				     phys_addr_t *phys_addr,
> +				     int *window, size_t size);
>  void pci_epc_mem_free_addr(struct pci_epc *epc, phys_addr_t phys_addr,
>  			   void __iomem *virt_addr, size_t size);
>  #endif /* __LINUX_PCI_EPC_H */
> 

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 2/5] pci: endpoint: add support to handle multiple base for mapping outbound memory
@ 2019-11-27  6:05     ` Kishon Vijay Abraham I
  0 siblings, 0 replies; 60+ messages in thread
From: Kishon Vijay Abraham I @ 2019-11-27  6:05 UTC (permalink / raw)
  To: Lad Prabhakar, Bjorn Helgaas, Rob Herring, Mark Rutland,
	Geert Uytterhoeven, Magnus Damm, Marek Vasut, Yoshihiro Shimoda,
	linux-pci-u79uwXL29TY76Z2rM5mHXA
  Cc: devicetree-u79uwXL29TY76Z2rM5mHXA, Chris Paterson,
	Lorenzo Pieralisi, Heiko Stuebner, Arnd Bergmann, Jingoo Han,
	Catalin Marinas, Shawn Lin, linux-kernel-u79uwXL29TY76Z2rM5mHXA,
	Lad, Prabhakar, linux-renesas-soc-u79uwXL29TY76Z2rM5mHXA,
	linux-rockchip-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r, Tom Joseph,
	Greg Kroah-Hartman, Andrew Murray, Gustavo Pimentel, Will Deacon,
	linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r

Hi Prabhakar,

On 07/11/19 1:06 AM, Lad Prabhakar wrote:
> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj-kTT6dE0pTRh9uiUsa/gSgQ@public.gmane.org>
> 
> rcar pcie controller has support to map multiple memory regions
> for mapping the outbound memory in local system, this feature
> inspires to add support for handling multiple memory bases in
> endpoint framework. In case of multiple memory regions only chunk
> or complete region can be mapped and this window needs to be
> passed to the controller driver.

Wouldn't this result in lot of unused address and exhausting the memory regions
quickly? I think we could have a "type" associated with each of these regions
and smaller allocations could use one "type" and larger allocation could use a
different "type"? For instance we allocate memory regions for MSI interrupt and
for data transfer. We could use smaller allocations for MSI and larger
allocations for data transfers.

We could also extend the "type" for certain special requirements. For instance
I've seen platforms which have dedicated address space for high priority data
transfers (uses higher virtual channel numbers) and dedicated address space for
low priority data transfers (uses lower virtual channel numbers).

Here too we could give a special flag or type for higher priority address
regions and different type for lower priority address. While allocating, the
function drivers should provide a flag specifying the region where memory
should be allocated.

Can you add flags in struct pci_epc_mem_window and see how that comes out?

Thanks
Kishon

> 
> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj-kTT6dE0pTRh9uiUsa/gSgQ@public.gmane.org>
> Cc: <linux-rockchip-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org>
> Cc: Shawn Lin <shawn.lin-TNX95d0MmH7DzftRWevZcw@public.gmane.org>
> Cc: Heiko Stuebner <heiko-4mtYJXux2i+zQB+pC5nmwQ@public.gmane.org>
> Cc: Tom Joseph <tjoseph-vna1KIf7WgpBDgjK7y7TUQ@public.gmane.org>
> Cc: Jingoo Han <jingoohan1-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
> Cc: Gustavo Pimentel <gustavo.pimentel-HKixBCOQz3hWk0Htik3J/w@public.gmane.org>
> ---
>  .../pci/controller/dwc/pcie-designware-ep.c   |  30 ++-
>  drivers/pci/controller/pcie-cadence-ep.c      |  11 +-
>  drivers/pci/controller/pcie-rockchip-ep.c     |  13 +-
>  drivers/pci/endpoint/functions/pci-epf-test.c |  29 +--
>  drivers/pci/endpoint/pci-epc-core.c           |   7 +-
>  drivers/pci/endpoint/pci-epc-mem.c            | 189 ++++++++++++++----
>  include/linux/pci-epc.h                       |  43 ++--
>  7 files changed, 234 insertions(+), 88 deletions(-)
> 
> diff --git a/drivers/pci/controller/dwc/pcie-designware-ep.c b/drivers/pci/controller/dwc/pcie-designware-ep.c
> index 3dd2e2697294..8d23c20b9afd 100644
> --- a/drivers/pci/controller/dwc/pcie-designware-ep.c
> +++ b/drivers/pci/controller/dwc/pcie-designware-ep.c
> @@ -195,7 +195,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, u8 func_no,
>  }
>  
>  static int dw_pcie_ep_map_addr(struct pci_epc *epc, u8 func_no,
> -			       phys_addr_t addr,
> +			       phys_addr_t addr, int window,
>  			       u64 pci_addr, size_t size)
>  {
>  	int ret;
> @@ -367,6 +367,7 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
>  	unsigned int aligned_offset;
>  	u16 msg_ctrl, msg_data;
>  	u32 msg_addr_lower, msg_addr_upper, reg;
> +	int window = PCI_EPC_DEFAULT_WINDOW;
>  	u64 msg_addr;
>  	bool has_upper;
>  	int ret;
> @@ -390,11 +391,11 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
>  		reg = ep->msi_cap + PCI_MSI_DATA_32;
>  		msg_data = dw_pcie_readw_dbi(pci, reg);
>  	}
> -	aligned_offset = msg_addr_lower & (epc->mem->page_size - 1);
> +	aligned_offset = msg_addr_lower & (epc->mem[window]->page_size - 1);
>  	msg_addr = ((u64)msg_addr_upper) << 32 |
>  			(msg_addr_lower & ~aligned_offset);
> -	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> -				  epc->mem->page_size);
> +	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> +				  msg_addr, epc->mem[window]->page_size);
>  	if (ret)
>  		return ret;
>  
> @@ -416,6 +417,7 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>  	u32 reg, msg_data, vec_ctrl;
>  	u64 tbl_addr, msg_addr, reg_u64;
>  	void __iomem *msix_tbl;
> +	int window = PCI_EPC_DEFAULT_WINDOW;
>  	int ret;
>  
>  	reg = ep->msix_cap + PCI_MSIX_TABLE;
> @@ -452,8 +454,8 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>  		return -EPERM;
>  	}
>  
> -	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> -				  epc->mem->page_size);
> +	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> +				  msg_addr, epc->mem[window]->page_size);
>  	if (ret)
>  		return ret;
>  
> @@ -466,10 +468,11 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>  
>  void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
>  {
> +	int window = PCI_EPC_DEFAULT_WINDOW;
>  	struct pci_epc *epc = ep->epc;
>  
>  	pci_epc_mem_free_addr(epc, ep->msi_mem_phys, ep->msi_mem,
> -			      epc->mem->page_size);
> +			      epc->mem[window]->page_size);
>  
>  	pci_epc_mem_exit(epc);
>  }
> @@ -499,9 +502,12 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
>  	u32 reg;
>  	void *addr;
>  	u8 hdr_type;
> +	int window;
>  	unsigned int nbars;
>  	unsigned int offset;
>  	struct pci_epc *epc;
> +	size_t msi_page_size;
> +	struct pci_epc_mem_window mem_window;
>  	struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
>  	struct device *dev = pci->dev;
>  	struct device_node *np = dev->of_node;
> @@ -574,15 +580,17 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
>  	if (ret < 0)
>  		epc->max_functions = 1;
>  
> -	ret = __pci_epc_mem_init(epc, ep->phys_base, ep->addr_size,
> -				 ep->page_size);
> +	mem_window.phys_base = ep->phys_base;
> +	mem_window.size = ep->addr_size;
> +	ret = __pci_epc_mem_init(epc, &mem_window, 1, ep->page_size);
>  	if (ret < 0) {
>  		dev_err(dev, "Failed to initialize address space\n");
>  		return ret;
>  	}
>  
> -	ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys,
> -					     epc->mem->page_size);
> +	msi_page_size = epc->mem[PCI_EPC_DEFAULT_WINDOW]->page_size;
> +	ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys, &window,
> +					     msi_page_size);
>  	if (!ep->msi_mem) {
>  		dev_err(dev, "Failed to reserve memory for MSI/MSI-X\n");
>  		return -ENOMEM;
> diff --git a/drivers/pci/controller/pcie-cadence-ep.c b/drivers/pci/controller/pcie-cadence-ep.c
> index def7820cb824..7991b38a5350 100644
> --- a/drivers/pci/controller/pcie-cadence-ep.c
> +++ b/drivers/pci/controller/pcie-cadence-ep.c
> @@ -172,7 +172,7 @@ static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
>  }
>  
>  static int cdns_pcie_ep_map_addr(struct pci_epc *epc, u8 fn, phys_addr_t addr,
> -				 u64 pci_addr, size_t size)
> +				 int window, u64 pci_addr, size_t size)
>  {
>  	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>  	struct cdns_pcie *pcie = &ep->pcie;
> @@ -434,12 +434,14 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
>  {
>  	struct device *dev = &pdev->dev;
>  	struct device_node *np = dev->of_node;
> +	struct pci_epc_mem_window mem_window;
>  	struct cdns_pcie_ep *ep;
>  	struct cdns_pcie *pcie;
>  	struct pci_epc *epc;
>  	struct resource *res;
>  	int ret;
>  	int phy_count;
> +	int window;
>  
>  	ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
>  	if (!ep)
> @@ -502,15 +504,16 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
>  	if (of_property_read_u8(np, "max-functions", &epc->max_functions) < 0)
>  		epc->max_functions = 1;
>  
> -	ret = pci_epc_mem_init(epc, pcie->mem_res->start,
> -			       resource_size(pcie->mem_res));
> +	mem_window.phys_base = pcie->mem_res->start;
> +	mem_window.size = resource_size(pcie->mem_res);
> +	ret = pci_epc_mem_init(epc, &mem_window, 1);
>  	if (ret < 0) {
>  		dev_err(dev, "failed to initialize the memory space\n");
>  		goto err_init;
>  	}
>  
>  	ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> -						  SZ_128K);
> +						  &window, SZ_128K);
>  	if (!ep->irq_cpu_addr) {
>  		dev_err(dev, "failed to reserve memory space for MSI\n");
>  		ret = -ENOMEM;
> diff --git a/drivers/pci/controller/pcie-rockchip-ep.c b/drivers/pci/controller/pcie-rockchip-ep.c
> index d743b0a48988..d59e85c8d319 100644
> --- a/drivers/pci/controller/pcie-rockchip-ep.c
> +++ b/drivers/pci/controller/pcie-rockchip-ep.c
> @@ -256,8 +256,8 @@ static void rockchip_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
>  }
>  
>  static int rockchip_pcie_ep_map_addr(struct pci_epc *epc, u8 fn,
> -				     phys_addr_t addr, u64 pci_addr,
> -				     size_t size)
> +				     phys_addr_t addr, int window,
> +				     u64 pci_addr, size_t size)
>  {
>  	struct rockchip_pcie_ep *ep = epc_get_drvdata(epc);
>  	struct rockchip_pcie *pcie = &ep->rockchip;
> @@ -562,11 +562,13 @@ static const struct of_device_id rockchip_pcie_ep_of_match[] = {
>  
>  static int rockchip_pcie_ep_probe(struct platform_device *pdev)
>  {
> +	struct pci_epc_mem_window mem_window;
>  	struct device *dev = &pdev->dev;
>  	struct rockchip_pcie_ep *ep;
>  	struct rockchip_pcie *rockchip;
>  	struct pci_epc *epc;
>  	size_t max_regions;
> +	int window;
>  	int err;
>  
>  	ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> @@ -614,15 +616,16 @@ static int rockchip_pcie_ep_probe(struct platform_device *pdev)
>  	/* Only enable function 0 by default */
>  	rockchip_pcie_write(rockchip, BIT(0), PCIE_CORE_PHY_FUNC_CFG);
>  
> -	err = pci_epc_mem_init(epc, rockchip->mem_res->start,
> -			       resource_size(rockchip->mem_res));
> +	mem_window.phys_base = rockchip->mem_res->start;
> +	mem_window.size = resource_size(rockchip->mem_res);
> +	err = pci_epc_mem_init(epc, &mem_window, 1);
>  	if (err < 0) {
>  		dev_err(dev, "failed to initialize the memory space\n");
>  		goto err_uninit_port;
>  	}
>  
>  	ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> -						  SZ_128K);
> +						  &window, SZ_128K);
>  	if (!ep->irq_cpu_addr) {
>  		dev_err(dev, "failed to reserve memory space for MSI\n");
>  		err = -ENOMEM;
> diff --git a/drivers/pci/endpoint/functions/pci-epf-test.c b/drivers/pci/endpoint/functions/pci-epf-test.c
> index 5d74f81ddfe4..475228011703 100644
> --- a/drivers/pci/endpoint/functions/pci-epf-test.c
> +++ b/drivers/pci/endpoint/functions/pci-epf-test.c
> @@ -84,8 +84,10 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>  	struct pci_epc *epc = epf->epc;
>  	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>  	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> +	int window;
>  
> -	src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr, reg->size);
> +	src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr,
> +					  &window, reg->size);
>  	if (!src_addr) {
>  		dev_err(dev, "Failed to allocate source address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
> @@ -93,15 +95,16 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>  		goto err;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, reg->src_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, window,
> +			       reg->src_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map source address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
>  		goto err_src_addr;
>  	}
>  
> -	dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr, reg->size);
> +	dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr,
> +					  &window, reg->size);
>  	if (!dst_addr) {
>  		dev_err(dev, "Failed to allocate destination address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> @@ -109,8 +112,8 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>  		goto err_src_map_addr;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, reg->dst_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, window,
> +			       reg->dst_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map destination address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> @@ -146,8 +149,9 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
>  	struct pci_epc *epc = epf->epc;
>  	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>  	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> +	int window;
>  
> -	src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> +	src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
>  	if (!src_addr) {
>  		dev_err(dev, "Failed to allocate address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
> @@ -155,8 +159,8 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
>  		goto err;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->src_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> +			       reg->src_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
> @@ -193,13 +197,14 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
>  	void __iomem *dst_addr;
>  	void *buf;
>  	phys_addr_t phys_addr;
> +	int window;
>  	struct pci_epf *epf = epf_test->epf;
>  	struct device *dev = &epf->dev;
>  	struct pci_epc *epc = epf->epc;
>  	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>  	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
>  
> -	dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> +	dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
>  	if (!dst_addr) {
>  		dev_err(dev, "Failed to allocate address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> @@ -207,8 +212,8 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
>  		goto err;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->dst_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> +			       reg->dst_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> diff --git a/drivers/pci/endpoint/pci-epc-core.c b/drivers/pci/endpoint/pci-epc-core.c
> index 2091508c1620..289c266c2d90 100644
> --- a/drivers/pci/endpoint/pci-epc-core.c
> +++ b/drivers/pci/endpoint/pci-epc-core.c
> @@ -358,13 +358,15 @@ EXPORT_SYMBOL_GPL(pci_epc_unmap_addr);
>   * @epc: the EPC device on which address is allocated
>   * @func_no: the endpoint function number in the EPC device
>   * @phys_addr: physical address of the local system
> + * @window: index to the window region where PCI address will be mapped
>   * @pci_addr: PCI address to which the physical address should be mapped
>   * @size: the size of the allocation
>   *
>   * Invoke to map CPU address with PCI address.
>   */
>  int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> -		     phys_addr_t phys_addr, u64 pci_addr, size_t size)
> +		     phys_addr_t phys_addr, int window,
> +		     u64 pci_addr, size_t size)
>  {
>  	int ret;
>  	unsigned long flags;
> @@ -376,7 +378,8 @@ int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
>  		return 0;
>  
>  	spin_lock_irqsave(&epc->lock, flags);
> -	ret = epc->ops->map_addr(epc, func_no, phys_addr, pci_addr, size);
> +	ret = epc->ops->map_addr(epc, func_no, phys_addr,
> +				 window, pci_addr, size);
>  	spin_unlock_irqrestore(&epc->lock, flags);
>  
>  	return ret;
> diff --git a/drivers/pci/endpoint/pci-epc-mem.c b/drivers/pci/endpoint/pci-epc-mem.c
> index d2b174ce15de..c955f2c97944 100644
> --- a/drivers/pci/endpoint/pci-epc-mem.c
> +++ b/drivers/pci/endpoint/pci-epc-mem.c
> @@ -39,56 +39,77 @@ static int pci_epc_mem_get_order(struct pci_epc_mem *mem, size_t size)
>   * __pci_epc_mem_init() - initialize the pci_epc_mem structure
>   * @epc: the EPC device that invoked pci_epc_mem_init
>   * @phys_base: the physical address of the base
> - * @size: the size of the address space
> + * @num_windows: number of windows device supports
>   * @page_size: size of each page
>   *
>   * Invoke to initialize the pci_epc_mem structure used by the
>   * endpoint functions to allocate mapped PCI address.
>   */
> -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_base, size_t size,
> -		       size_t page_size)
> +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *windows,
> +		       int num_windows, size_t page_size)
>  {
> -	int ret;
> -	struct pci_epc_mem *mem;
> -	unsigned long *bitmap;
> +	struct pci_epc_mem *mem = NULL;
> +	unsigned long *bitmap = NULL;
>  	unsigned int page_shift;
> -	int pages;
>  	int bitmap_size;
> +	int pages;
> +	int ret;
> +	int i;
> +
> +	epc->mem_windows = 0;
> +
> +	if (!windows)
> +		return -EINVAL;
> +
> +	if (num_windows <= 0)
> +		return -EINVAL;
>  
>  	if (page_size < PAGE_SIZE)
>  		page_size = PAGE_SIZE;
>  
>  	page_shift = ilog2(page_size);
> -	pages = size >> page_shift;
> -	bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
>  
> -	mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> -	if (!mem) {
> -		ret = -ENOMEM;
> -		goto err;
> -	}
> +	epc->mem = kcalloc(num_windows, sizeof(*mem), GFP_KERNEL);
> +	if (!epc->mem)
> +		return -EINVAL;
>  
> -	bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> -	if (!bitmap) {
> -		ret = -ENOMEM;
> -		goto err_mem;
> -	}
> +	for (i = 0; i < num_windows; i++) {
> +		pages = windows[i].phys_base >> page_shift;
> +		bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
>  
> -	mem->bitmap = bitmap;
> -	mem->phys_base = phys_base;
> -	mem->page_size = page_size;
> -	mem->pages = pages;
> -	mem->size = size;
> +		mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> +		if (!mem) {
> +			ret = -ENOMEM;
> +			goto err_mem;
> +		}
>  
> -	epc->mem = mem;
> +		bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> +		if (!bitmap) {
> +			ret = -ENOMEM;
> +			goto err_mem;
> +		}
> +
> +		mem->bitmap = bitmap;
> +		mem->window.phys_base = windows[i].phys_base;
> +		mem->page_size = page_size;
> +		mem->pages = pages;
> +		mem->window.size = windows[i].size;
> +		mem->window.map_size = 0;
> +
> +		epc->mem[i] = mem;
> +	}
> +	epc->mem_windows = num_windows;
>  
>  	return 0;
>  
>  err_mem:
> -	kfree(mem);
> +	for (; i >= 0; i--) {
> +		kfree(mem->bitmap);
> +		kfree(epc->mem[i]);
> +	}
> +	kfree(epc->mem);
>  
> -err:
> -return ret;
> +	return ret;
>  }
>  EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
>  
> @@ -101,48 +122,126 @@ EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
>   */
>  void pci_epc_mem_exit(struct pci_epc *epc)
>  {
> -	struct pci_epc_mem *mem = epc->mem;
> +	struct pci_epc_mem *mem;
> +	int i;
> +
> +	if (!epc->mem_windows)
> +		return;
> +
> +	for (i = 0; i <= epc->mem_windows; i--) {
> +		mem = epc->mem[i];
> +		kfree(mem->bitmap);
> +		kfree(epc->mem[i]);
> +	}
> +	kfree(epc->mem);
>  
>  	epc->mem = NULL;
> -	kfree(mem->bitmap);
> -	kfree(mem);
> +	epc->mem_windows = 0;
>  }
>  EXPORT_SYMBOL_GPL(pci_epc_mem_exit);
>  
> +static int pci_epc_find_best_fit_window(struct pci_epc *epc, size_t size)
> +{
> +	size_t window_least_size = 0;
> +	int best_fit_window = -1;
> +	struct pci_epc_mem *mem;
> +	size_t actual_size;
> +	int i;
> +
> +	for (i = 0; i < epc->mem_windows; i++) {
> +		mem = epc->mem[i];
> +
> +		/* if chunk from this region is already used skip it */
> +		if (mem->window.map_size)
> +			continue;
> +
> +		actual_size = ALIGN(size, mem->page_size);
> +
> +		if (best_fit_window == -1) {
> +			best_fit_window = i;
> +			window_least_size = mem->window.size;
> +		} else {
> +			if (actual_size <= mem->window.size &&
> +			    mem->window.size < window_least_size) {
> +				best_fit_window = i;
> +				window_least_size = mem->window.size;
> +			}
> +		}
> +	}
> +
> +	return best_fit_window;
> +}
> +
>  /**
>   * pci_epc_mem_alloc_addr() - allocate memory address from EPC addr space
>   * @epc: the EPC device on which memory has to be allocated
>   * @phys_addr: populate the allocated physical address here
> + * @window: populate the window here which will be used to map PCI address
>   * @size: the size of the address space that has to be allocated
>   *
>   * Invoke to allocate memory address from the EPC address space. This
>   * is usually done to map the remote RC address into the local system.
>   */
>  void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
> -				     phys_addr_t *phys_addr, size_t size)
> +				     phys_addr_t *phys_addr,
> +				     int *window, size_t size)
>  {
> +	int best_fit = PCI_EPC_DEFAULT_WINDOW;
> +	void __iomem *virt_addr = NULL;
> +	struct pci_epc_mem *mem;
> +	unsigned int page_shift;
>  	int pageno;
> -	void __iomem *virt_addr;
> -	struct pci_epc_mem *mem = epc->mem;
> -	unsigned int page_shift = ilog2(mem->page_size);
>  	int order;
>  
> +	if (epc->mem_windows <= 0)
> +		return NULL;
> +
> +	if (epc->mem_windows > 1) {
> +		best_fit = pci_epc_find_best_fit_window(epc, size);
> +		if (best_fit < 0)
> +			return NULL;
> +	}
> +
> +	mem = epc->mem[best_fit];
>  	size = ALIGN(size, mem->page_size);
> +	if (size > (mem->window.size - mem->window.map_size))
> +		return NULL;
> +	page_shift = ilog2(mem->page_size);
>  	order = pci_epc_mem_get_order(mem, size);
>  
>  	pageno = bitmap_find_free_region(mem->bitmap, mem->pages, order);
>  	if (pageno < 0)
>  		return NULL;
>  
> -	*phys_addr = mem->phys_base + ((phys_addr_t)pageno << page_shift);
> +	*phys_addr = mem->window.phys_base + (pageno << page_shift);
>  	virt_addr = ioremap(*phys_addr, size);
> -	if (!virt_addr)
> +	if (!virt_addr) {
>  		bitmap_release_region(mem->bitmap, pageno, order);
> +	} else {
> +		mem->window.map_size += size;
> +		*window = best_fit;
> +	}
>  
>  	return virt_addr;
>  }
>  EXPORT_SYMBOL_GPL(pci_epc_mem_alloc_addr);
>  
> +static int pci_epc_get_matching_window(struct pci_epc *epc,
> +				       phys_addr_t phys_addr)
> +{
> +	struct pci_epc_mem *mem;
> +	int i;
> +
> +	for (i = 0; i < epc->mem_windows; i++) {
> +		mem = epc->mem[i];
> +
> +		if (mem->window.phys_base == phys_addr)
> +			return i;
> +	}
> +
> +	return -EINVAL;
> +}
> +
>  /**
>   * pci_epc_mem_free_addr() - free the allocated memory address
>   * @epc: the EPC device on which memory was allocated
> @@ -155,16 +254,26 @@ EXPORT_SYMBOL_GPL(pci_epc_mem_alloc_addr);
>  void pci_epc_mem_free_addr(struct pci_epc *epc, phys_addr_t phys_addr,
>  			   void __iomem *virt_addr, size_t size)
>  {
> +	struct pci_epc_mem *mem;
> +	unsigned int page_shift;
> +	int window = 0;
>  	int pageno;
> -	struct pci_epc_mem *mem = epc->mem;
> -	unsigned int page_shift = ilog2(mem->page_size);
>  	int order;
>  
> +	if (epc->mem_windows > 1) {
> +		window = pci_epc_get_matching_window(epc, phys_addr);
> +		if (window < 0)
> +			return;
> +	}
> +
> +	mem = epc->mem[window];
> +	page_shift = ilog2(mem->page_size);
>  	iounmap(virt_addr);
> -	pageno = (phys_addr - mem->phys_base) >> page_shift;
> +	pageno = (phys_addr - mem->window.phys_base) >> page_shift;
>  	size = ALIGN(size, mem->page_size);
>  	order = pci_epc_mem_get_order(mem, size);
>  	bitmap_release_region(mem->bitmap, pageno, order);
> +	mem->window.map_size -= size;
>  }
>  EXPORT_SYMBOL_GPL(pci_epc_mem_free_addr);
>  
> diff --git a/include/linux/pci-epc.h b/include/linux/pci-epc.h
> index 56f1846b9d39..5e86d22dd5c5 100644
> --- a/include/linux/pci-epc.h
> +++ b/include/linux/pci-epc.h
> @@ -48,7 +48,8 @@ struct pci_epc_ops {
>  	void	(*clear_bar)(struct pci_epc *epc, u8 func_no,
>  			     struct pci_epf_bar *epf_bar);
>  	int	(*map_addr)(struct pci_epc *epc, u8 func_no,
> -			    phys_addr_t addr, u64 pci_addr, size_t size);
> +			    phys_addr_t addr, int window,
> +			    u64 pci_addr, size_t size);
>  	void	(*unmap_addr)(struct pci_epc *epc, u8 func_no,
>  			      phys_addr_t addr);
>  	int	(*set_msi)(struct pci_epc *epc, u8 func_no, u8 interrupts);
> @@ -64,17 +65,28 @@ struct pci_epc_ops {
>  	struct module *owner;
>  };
>  
> +#define PCI_EPC_DEFAULT_WINDOW		0
> +
> +/**
> + * struct pci_epc_mem_window - address window of the endpoint controller
> + * @phys_base: physical base address of the PCI address window
> + * @size: the size of the PCI address window
> + */
> +struct pci_epc_mem_window {
> +	phys_addr_t	phys_base;
> +	size_t		size;
> +	size_t		map_size;
> +};
> +
>  /**
>   * struct pci_epc_mem - address space of the endpoint controller
> - * @phys_base: physical base address of the PCI address space
> - * @size: the size of the PCI address space
> + * @window: address window of the endpoint controller
>   * @bitmap: bitmap to manage the PCI address space
> - * @pages: number of bits representing the address region
>   * @page_size: size of each page
> + * @pages: number of bits representing the address region
>   */
>  struct pci_epc_mem {
> -	phys_addr_t	phys_base;
> -	size_t		size;
> +	struct pci_epc_mem_window window;
>  	unsigned long	*bitmap;
>  	size_t		page_size;
>  	int		pages;
> @@ -85,7 +97,8 @@ struct pci_epc_mem {
>   * @dev: PCI EPC device
>   * @pci_epf: list of endpoint functions present in this EPC device
>   * @ops: function pointers for performing endpoint operations
> - * @mem: address space of the endpoint controller
> + * @mem: array of address space of the endpoint controller
> + * @mem_windows: number of windows supported by device
>   * @max_functions: max number of functions that can be configured in this EPC
>   * @group: configfs group representing the PCI EPC device
>   * @lock: spinlock to protect pci_epc ops
> @@ -94,7 +107,8 @@ struct pci_epc {
>  	struct device			dev;
>  	struct list_head		pci_epf;
>  	const struct pci_epc_ops	*ops;
> -	struct pci_epc_mem		*mem;
> +	struct pci_epc_mem		**mem;
> +	int				mem_windows;
>  	u8				max_functions;
>  	struct config_group		*group;
>  	/* spinlock to protect against concurrent access of EP controller */
> @@ -128,8 +142,8 @@ struct pci_epc_features {
>  #define devm_pci_epc_create(dev, ops)    \
>  		__devm_pci_epc_create((dev), (ops), THIS_MODULE)
>  
> -#define pci_epc_mem_init(epc, phys_addr, size)	\
> -		__pci_epc_mem_init((epc), (phys_addr), (size), PAGE_SIZE)
> +#define pci_epc_mem_init(epc, windows, num_windows)	\
> +		__pci_epc_mem_init((epc), windows, num_windows, PAGE_SIZE)
>  
>  static inline void epc_set_drvdata(struct pci_epc *epc, void *data)
>  {
> @@ -159,7 +173,7 @@ int pci_epc_set_bar(struct pci_epc *epc, u8 func_no,
>  void pci_epc_clear_bar(struct pci_epc *epc, u8 func_no,
>  		       struct pci_epf_bar *epf_bar);
>  int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> -		     phys_addr_t phys_addr,
> +		     phys_addr_t phys_addr, int window,
>  		     u64 pci_addr, size_t size);
>  void pci_epc_unmap_addr(struct pci_epc *epc, u8 func_no,
>  			phys_addr_t phys_addr);
> @@ -178,11 +192,12 @@ unsigned int pci_epc_get_first_free_bar(const struct pci_epc_features
>  struct pci_epc *pci_epc_get(const char *epc_name);
>  void pci_epc_put(struct pci_epc *epc);
>  
> -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_addr, size_t size,
> -		       size_t page_size);
> +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *window,
> +		       int num_windows, size_t page_size);
>  void pci_epc_mem_exit(struct pci_epc *epc);
>  void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
> -				     phys_addr_t *phys_addr, size_t size);
> +				     phys_addr_t *phys_addr,
> +				     int *window, size_t size);
>  void pci_epc_mem_free_addr(struct pci_epc *epc, phys_addr_t phys_addr,
>  			   void __iomem *virt_addr, size_t size);
>  #endif /* __LINUX_PCI_EPC_H */
> 

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 2/5] pci: endpoint: add support to handle multiple base for mapping outbound memory
@ 2019-11-27  6:05     ` Kishon Vijay Abraham I
  0 siblings, 0 replies; 60+ messages in thread
From: Kishon Vijay Abraham I @ 2019-11-27  6:05 UTC (permalink / raw)
  To: Lad Prabhakar, Bjorn Helgaas, Rob Herring, Mark Rutland,
	Geert Uytterhoeven, Magnus Damm, Marek Vasut, Yoshihiro Shimoda,
	linux-pci
  Cc: devicetree, Chris Paterson, Lorenzo Pieralisi, Heiko Stuebner,
	Arnd Bergmann, Jingoo Han, Catalin Marinas, Shawn Lin,
	linux-kernel, Lad, Prabhakar, linux-renesas-soc, linux-rockchip,
	Tom Joseph, Greg Kroah-Hartman, Andrew Murray, Gustavo Pimentel,
	Will Deacon, linux-arm-kernel

Hi Prabhakar,

On 07/11/19 1:06 AM, Lad Prabhakar wrote:
> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> 
> rcar pcie controller has support to map multiple memory regions
> for mapping the outbound memory in local system, this feature
> inspires to add support for handling multiple memory bases in
> endpoint framework. In case of multiple memory regions only chunk
> or complete region can be mapped and this window needs to be
> passed to the controller driver.

Wouldn't this result in lot of unused address and exhausting the memory regions
quickly? I think we could have a "type" associated with each of these regions
and smaller allocations could use one "type" and larger allocation could use a
different "type"? For instance we allocate memory regions for MSI interrupt and
for data transfer. We could use smaller allocations for MSI and larger
allocations for data transfers.

We could also extend the "type" for certain special requirements. For instance
I've seen platforms which have dedicated address space for high priority data
transfers (uses higher virtual channel numbers) and dedicated address space for
low priority data transfers (uses lower virtual channel numbers).

Here too we could give a special flag or type for higher priority address
regions and different type for lower priority address. While allocating, the
function drivers should provide a flag specifying the region where memory
should be allocated.

Can you add flags in struct pci_epc_mem_window and see how that comes out?

Thanks
Kishon

> 
> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> Cc: <linux-rockchip@lists.infradead.org>
> Cc: Shawn Lin <shawn.lin@rock-chips.com>
> Cc: Heiko Stuebner <heiko@sntech.de>
> Cc: Tom Joseph <tjoseph@cadence.com>
> Cc: Jingoo Han <jingoohan1@gmail.com>
> Cc: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
> ---
>  .../pci/controller/dwc/pcie-designware-ep.c   |  30 ++-
>  drivers/pci/controller/pcie-cadence-ep.c      |  11 +-
>  drivers/pci/controller/pcie-rockchip-ep.c     |  13 +-
>  drivers/pci/endpoint/functions/pci-epf-test.c |  29 +--
>  drivers/pci/endpoint/pci-epc-core.c           |   7 +-
>  drivers/pci/endpoint/pci-epc-mem.c            | 189 ++++++++++++++----
>  include/linux/pci-epc.h                       |  43 ++--
>  7 files changed, 234 insertions(+), 88 deletions(-)
> 
> diff --git a/drivers/pci/controller/dwc/pcie-designware-ep.c b/drivers/pci/controller/dwc/pcie-designware-ep.c
> index 3dd2e2697294..8d23c20b9afd 100644
> --- a/drivers/pci/controller/dwc/pcie-designware-ep.c
> +++ b/drivers/pci/controller/dwc/pcie-designware-ep.c
> @@ -195,7 +195,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, u8 func_no,
>  }
>  
>  static int dw_pcie_ep_map_addr(struct pci_epc *epc, u8 func_no,
> -			       phys_addr_t addr,
> +			       phys_addr_t addr, int window,
>  			       u64 pci_addr, size_t size)
>  {
>  	int ret;
> @@ -367,6 +367,7 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
>  	unsigned int aligned_offset;
>  	u16 msg_ctrl, msg_data;
>  	u32 msg_addr_lower, msg_addr_upper, reg;
> +	int window = PCI_EPC_DEFAULT_WINDOW;
>  	u64 msg_addr;
>  	bool has_upper;
>  	int ret;
> @@ -390,11 +391,11 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
>  		reg = ep->msi_cap + PCI_MSI_DATA_32;
>  		msg_data = dw_pcie_readw_dbi(pci, reg);
>  	}
> -	aligned_offset = msg_addr_lower & (epc->mem->page_size - 1);
> +	aligned_offset = msg_addr_lower & (epc->mem[window]->page_size - 1);
>  	msg_addr = ((u64)msg_addr_upper) << 32 |
>  			(msg_addr_lower & ~aligned_offset);
> -	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> -				  epc->mem->page_size);
> +	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> +				  msg_addr, epc->mem[window]->page_size);
>  	if (ret)
>  		return ret;
>  
> @@ -416,6 +417,7 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>  	u32 reg, msg_data, vec_ctrl;
>  	u64 tbl_addr, msg_addr, reg_u64;
>  	void __iomem *msix_tbl;
> +	int window = PCI_EPC_DEFAULT_WINDOW;
>  	int ret;
>  
>  	reg = ep->msix_cap + PCI_MSIX_TABLE;
> @@ -452,8 +454,8 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>  		return -EPERM;
>  	}
>  
> -	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> -				  epc->mem->page_size);
> +	ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> +				  msg_addr, epc->mem[window]->page_size);
>  	if (ret)
>  		return ret;
>  
> @@ -466,10 +468,11 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>  
>  void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
>  {
> +	int window = PCI_EPC_DEFAULT_WINDOW;
>  	struct pci_epc *epc = ep->epc;
>  
>  	pci_epc_mem_free_addr(epc, ep->msi_mem_phys, ep->msi_mem,
> -			      epc->mem->page_size);
> +			      epc->mem[window]->page_size);
>  
>  	pci_epc_mem_exit(epc);
>  }
> @@ -499,9 +502,12 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
>  	u32 reg;
>  	void *addr;
>  	u8 hdr_type;
> +	int window;
>  	unsigned int nbars;
>  	unsigned int offset;
>  	struct pci_epc *epc;
> +	size_t msi_page_size;
> +	struct pci_epc_mem_window mem_window;
>  	struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
>  	struct device *dev = pci->dev;
>  	struct device_node *np = dev->of_node;
> @@ -574,15 +580,17 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
>  	if (ret < 0)
>  		epc->max_functions = 1;
>  
> -	ret = __pci_epc_mem_init(epc, ep->phys_base, ep->addr_size,
> -				 ep->page_size);
> +	mem_window.phys_base = ep->phys_base;
> +	mem_window.size = ep->addr_size;
> +	ret = __pci_epc_mem_init(epc, &mem_window, 1, ep->page_size);
>  	if (ret < 0) {
>  		dev_err(dev, "Failed to initialize address space\n");
>  		return ret;
>  	}
>  
> -	ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys,
> -					     epc->mem->page_size);
> +	msi_page_size = epc->mem[PCI_EPC_DEFAULT_WINDOW]->page_size;
> +	ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys, &window,
> +					     msi_page_size);
>  	if (!ep->msi_mem) {
>  		dev_err(dev, "Failed to reserve memory for MSI/MSI-X\n");
>  		return -ENOMEM;
> diff --git a/drivers/pci/controller/pcie-cadence-ep.c b/drivers/pci/controller/pcie-cadence-ep.c
> index def7820cb824..7991b38a5350 100644
> --- a/drivers/pci/controller/pcie-cadence-ep.c
> +++ b/drivers/pci/controller/pcie-cadence-ep.c
> @@ -172,7 +172,7 @@ static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
>  }
>  
>  static int cdns_pcie_ep_map_addr(struct pci_epc *epc, u8 fn, phys_addr_t addr,
> -				 u64 pci_addr, size_t size)
> +				 int window, u64 pci_addr, size_t size)
>  {
>  	struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>  	struct cdns_pcie *pcie = &ep->pcie;
> @@ -434,12 +434,14 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
>  {
>  	struct device *dev = &pdev->dev;
>  	struct device_node *np = dev->of_node;
> +	struct pci_epc_mem_window mem_window;
>  	struct cdns_pcie_ep *ep;
>  	struct cdns_pcie *pcie;
>  	struct pci_epc *epc;
>  	struct resource *res;
>  	int ret;
>  	int phy_count;
> +	int window;
>  
>  	ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
>  	if (!ep)
> @@ -502,15 +504,16 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
>  	if (of_property_read_u8(np, "max-functions", &epc->max_functions) < 0)
>  		epc->max_functions = 1;
>  
> -	ret = pci_epc_mem_init(epc, pcie->mem_res->start,
> -			       resource_size(pcie->mem_res));
> +	mem_window.phys_base = pcie->mem_res->start;
> +	mem_window.size = resource_size(pcie->mem_res);
> +	ret = pci_epc_mem_init(epc, &mem_window, 1);
>  	if (ret < 0) {
>  		dev_err(dev, "failed to initialize the memory space\n");
>  		goto err_init;
>  	}
>  
>  	ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> -						  SZ_128K);
> +						  &window, SZ_128K);
>  	if (!ep->irq_cpu_addr) {
>  		dev_err(dev, "failed to reserve memory space for MSI\n");
>  		ret = -ENOMEM;
> diff --git a/drivers/pci/controller/pcie-rockchip-ep.c b/drivers/pci/controller/pcie-rockchip-ep.c
> index d743b0a48988..d59e85c8d319 100644
> --- a/drivers/pci/controller/pcie-rockchip-ep.c
> +++ b/drivers/pci/controller/pcie-rockchip-ep.c
> @@ -256,8 +256,8 @@ static void rockchip_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
>  }
>  
>  static int rockchip_pcie_ep_map_addr(struct pci_epc *epc, u8 fn,
> -				     phys_addr_t addr, u64 pci_addr,
> -				     size_t size)
> +				     phys_addr_t addr, int window,
> +				     u64 pci_addr, size_t size)
>  {
>  	struct rockchip_pcie_ep *ep = epc_get_drvdata(epc);
>  	struct rockchip_pcie *pcie = &ep->rockchip;
> @@ -562,11 +562,13 @@ static const struct of_device_id rockchip_pcie_ep_of_match[] = {
>  
>  static int rockchip_pcie_ep_probe(struct platform_device *pdev)
>  {
> +	struct pci_epc_mem_window mem_window;
>  	struct device *dev = &pdev->dev;
>  	struct rockchip_pcie_ep *ep;
>  	struct rockchip_pcie *rockchip;
>  	struct pci_epc *epc;
>  	size_t max_regions;
> +	int window;
>  	int err;
>  
>  	ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> @@ -614,15 +616,16 @@ static int rockchip_pcie_ep_probe(struct platform_device *pdev)
>  	/* Only enable function 0 by default */
>  	rockchip_pcie_write(rockchip, BIT(0), PCIE_CORE_PHY_FUNC_CFG);
>  
> -	err = pci_epc_mem_init(epc, rockchip->mem_res->start,
> -			       resource_size(rockchip->mem_res));
> +	mem_window.phys_base = rockchip->mem_res->start;
> +	mem_window.size = resource_size(rockchip->mem_res);
> +	err = pci_epc_mem_init(epc, &mem_window, 1);
>  	if (err < 0) {
>  		dev_err(dev, "failed to initialize the memory space\n");
>  		goto err_uninit_port;
>  	}
>  
>  	ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> -						  SZ_128K);
> +						  &window, SZ_128K);
>  	if (!ep->irq_cpu_addr) {
>  		dev_err(dev, "failed to reserve memory space for MSI\n");
>  		err = -ENOMEM;
> diff --git a/drivers/pci/endpoint/functions/pci-epf-test.c b/drivers/pci/endpoint/functions/pci-epf-test.c
> index 5d74f81ddfe4..475228011703 100644
> --- a/drivers/pci/endpoint/functions/pci-epf-test.c
> +++ b/drivers/pci/endpoint/functions/pci-epf-test.c
> @@ -84,8 +84,10 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>  	struct pci_epc *epc = epf->epc;
>  	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>  	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> +	int window;
>  
> -	src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr, reg->size);
> +	src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr,
> +					  &window, reg->size);
>  	if (!src_addr) {
>  		dev_err(dev, "Failed to allocate source address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
> @@ -93,15 +95,16 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>  		goto err;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, reg->src_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, window,
> +			       reg->src_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map source address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
>  		goto err_src_addr;
>  	}
>  
> -	dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr, reg->size);
> +	dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr,
> +					  &window, reg->size);
>  	if (!dst_addr) {
>  		dev_err(dev, "Failed to allocate destination address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> @@ -109,8 +112,8 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>  		goto err_src_map_addr;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, reg->dst_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, window,
> +			       reg->dst_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map destination address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> @@ -146,8 +149,9 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
>  	struct pci_epc *epc = epf->epc;
>  	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>  	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> +	int window;
>  
> -	src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> +	src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
>  	if (!src_addr) {
>  		dev_err(dev, "Failed to allocate address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
> @@ -155,8 +159,8 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
>  		goto err;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->src_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> +			       reg->src_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map address\n");
>  		reg->status = STATUS_SRC_ADDR_INVALID;
> @@ -193,13 +197,14 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
>  	void __iomem *dst_addr;
>  	void *buf;
>  	phys_addr_t phys_addr;
> +	int window;
>  	struct pci_epf *epf = epf_test->epf;
>  	struct device *dev = &epf->dev;
>  	struct pci_epc *epc = epf->epc;
>  	enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>  	struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
>  
> -	dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> +	dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
>  	if (!dst_addr) {
>  		dev_err(dev, "Failed to allocate address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> @@ -207,8 +212,8 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
>  		goto err;
>  	}
>  
> -	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->dst_addr,
> -			       reg->size);
> +	ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> +			       reg->dst_addr, reg->size);
>  	if (ret) {
>  		dev_err(dev, "Failed to map address\n");
>  		reg->status = STATUS_DST_ADDR_INVALID;
> diff --git a/drivers/pci/endpoint/pci-epc-core.c b/drivers/pci/endpoint/pci-epc-core.c
> index 2091508c1620..289c266c2d90 100644
> --- a/drivers/pci/endpoint/pci-epc-core.c
> +++ b/drivers/pci/endpoint/pci-epc-core.c
> @@ -358,13 +358,15 @@ EXPORT_SYMBOL_GPL(pci_epc_unmap_addr);
>   * @epc: the EPC device on which address is allocated
>   * @func_no: the endpoint function number in the EPC device
>   * @phys_addr: physical address of the local system
> + * @window: index to the window region where PCI address will be mapped
>   * @pci_addr: PCI address to which the physical address should be mapped
>   * @size: the size of the allocation
>   *
>   * Invoke to map CPU address with PCI address.
>   */
>  int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> -		     phys_addr_t phys_addr, u64 pci_addr, size_t size)
> +		     phys_addr_t phys_addr, int window,
> +		     u64 pci_addr, size_t size)
>  {
>  	int ret;
>  	unsigned long flags;
> @@ -376,7 +378,8 @@ int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
>  		return 0;
>  
>  	spin_lock_irqsave(&epc->lock, flags);
> -	ret = epc->ops->map_addr(epc, func_no, phys_addr, pci_addr, size);
> +	ret = epc->ops->map_addr(epc, func_no, phys_addr,
> +				 window, pci_addr, size);
>  	spin_unlock_irqrestore(&epc->lock, flags);
>  
>  	return ret;
> diff --git a/drivers/pci/endpoint/pci-epc-mem.c b/drivers/pci/endpoint/pci-epc-mem.c
> index d2b174ce15de..c955f2c97944 100644
> --- a/drivers/pci/endpoint/pci-epc-mem.c
> +++ b/drivers/pci/endpoint/pci-epc-mem.c
> @@ -39,56 +39,77 @@ static int pci_epc_mem_get_order(struct pci_epc_mem *mem, size_t size)
>   * __pci_epc_mem_init() - initialize the pci_epc_mem structure
>   * @epc: the EPC device that invoked pci_epc_mem_init
>   * @phys_base: the physical address of the base
> - * @size: the size of the address space
> + * @num_windows: number of windows device supports
>   * @page_size: size of each page
>   *
>   * Invoke to initialize the pci_epc_mem structure used by the
>   * endpoint functions to allocate mapped PCI address.
>   */
> -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_base, size_t size,
> -		       size_t page_size)
> +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *windows,
> +		       int num_windows, size_t page_size)
>  {
> -	int ret;
> -	struct pci_epc_mem *mem;
> -	unsigned long *bitmap;
> +	struct pci_epc_mem *mem = NULL;
> +	unsigned long *bitmap = NULL;
>  	unsigned int page_shift;
> -	int pages;
>  	int bitmap_size;
> +	int pages;
> +	int ret;
> +	int i;
> +
> +	epc->mem_windows = 0;
> +
> +	if (!windows)
> +		return -EINVAL;
> +
> +	if (num_windows <= 0)
> +		return -EINVAL;
>  
>  	if (page_size < PAGE_SIZE)
>  		page_size = PAGE_SIZE;
>  
>  	page_shift = ilog2(page_size);
> -	pages = size >> page_shift;
> -	bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
>  
> -	mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> -	if (!mem) {
> -		ret = -ENOMEM;
> -		goto err;
> -	}
> +	epc->mem = kcalloc(num_windows, sizeof(*mem), GFP_KERNEL);
> +	if (!epc->mem)
> +		return -EINVAL;
>  
> -	bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> -	if (!bitmap) {
> -		ret = -ENOMEM;
> -		goto err_mem;
> -	}
> +	for (i = 0; i < num_windows; i++) {
> +		pages = windows[i].phys_base >> page_shift;
> +		bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
>  
> -	mem->bitmap = bitmap;
> -	mem->phys_base = phys_base;
> -	mem->page_size = page_size;
> -	mem->pages = pages;
> -	mem->size = size;
> +		mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> +		if (!mem) {
> +			ret = -ENOMEM;
> +			goto err_mem;
> +		}
>  
> -	epc->mem = mem;
> +		bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> +		if (!bitmap) {
> +			ret = -ENOMEM;
> +			goto err_mem;
> +		}
> +
> +		mem->bitmap = bitmap;
> +		mem->window.phys_base = windows[i].phys_base;
> +		mem->page_size = page_size;
> +		mem->pages = pages;
> +		mem->window.size = windows[i].size;
> +		mem->window.map_size = 0;
> +
> +		epc->mem[i] = mem;
> +	}
> +	epc->mem_windows = num_windows;
>  
>  	return 0;
>  
>  err_mem:
> -	kfree(mem);
> +	for (; i >= 0; i--) {
> +		kfree(mem->bitmap);
> +		kfree(epc->mem[i]);
> +	}
> +	kfree(epc->mem);
>  
> -err:
> -return ret;
> +	return ret;
>  }
>  EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
>  
> @@ -101,48 +122,126 @@ EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
>   */
>  void pci_epc_mem_exit(struct pci_epc *epc)
>  {
> -	struct pci_epc_mem *mem = epc->mem;
> +	struct pci_epc_mem *mem;
> +	int i;
> +
> +	if (!epc->mem_windows)
> +		return;
> +
> +	for (i = 0; i <= epc->mem_windows; i--) {
> +		mem = epc->mem[i];
> +		kfree(mem->bitmap);
> +		kfree(epc->mem[i]);
> +	}
> +	kfree(epc->mem);
>  
>  	epc->mem = NULL;
> -	kfree(mem->bitmap);
> -	kfree(mem);
> +	epc->mem_windows = 0;
>  }
>  EXPORT_SYMBOL_GPL(pci_epc_mem_exit);
>  
> +static int pci_epc_find_best_fit_window(struct pci_epc *epc, size_t size)
> +{
> +	size_t window_least_size = 0;
> +	int best_fit_window = -1;
> +	struct pci_epc_mem *mem;
> +	size_t actual_size;
> +	int i;
> +
> +	for (i = 0; i < epc->mem_windows; i++) {
> +		mem = epc->mem[i];
> +
> +		/* if chunk from this region is already used skip it */
> +		if (mem->window.map_size)
> +			continue;
> +
> +		actual_size = ALIGN(size, mem->page_size);
> +
> +		if (best_fit_window == -1) {
> +			best_fit_window = i;
> +			window_least_size = mem->window.size;
> +		} else {
> +			if (actual_size <= mem->window.size &&
> +			    mem->window.size < window_least_size) {
> +				best_fit_window = i;
> +				window_least_size = mem->window.size;
> +			}
> +		}
> +	}
> +
> +	return best_fit_window;
> +}
> +
>  /**
>   * pci_epc_mem_alloc_addr() - allocate memory address from EPC addr space
>   * @epc: the EPC device on which memory has to be allocated
>   * @phys_addr: populate the allocated physical address here
> + * @window: populate the window here which will be used to map PCI address
>   * @size: the size of the address space that has to be allocated
>   *
>   * Invoke to allocate memory address from the EPC address space. This
>   * is usually done to map the remote RC address into the local system.
>   */
>  void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
> -				     phys_addr_t *phys_addr, size_t size)
> +				     phys_addr_t *phys_addr,
> +				     int *window, size_t size)
>  {
> +	int best_fit = PCI_EPC_DEFAULT_WINDOW;
> +	void __iomem *virt_addr = NULL;
> +	struct pci_epc_mem *mem;
> +	unsigned int page_shift;
>  	int pageno;
> -	void __iomem *virt_addr;
> -	struct pci_epc_mem *mem = epc->mem;
> -	unsigned int page_shift = ilog2(mem->page_size);
>  	int order;
>  
> +	if (epc->mem_windows <= 0)
> +		return NULL;
> +
> +	if (epc->mem_windows > 1) {
> +		best_fit = pci_epc_find_best_fit_window(epc, size);
> +		if (best_fit < 0)
> +			return NULL;
> +	}
> +
> +	mem = epc->mem[best_fit];
>  	size = ALIGN(size, mem->page_size);
> +	if (size > (mem->window.size - mem->window.map_size))
> +		return NULL;
> +	page_shift = ilog2(mem->page_size);
>  	order = pci_epc_mem_get_order(mem, size);
>  
>  	pageno = bitmap_find_free_region(mem->bitmap, mem->pages, order);
>  	if (pageno < 0)
>  		return NULL;
>  
> -	*phys_addr = mem->phys_base + ((phys_addr_t)pageno << page_shift);
> +	*phys_addr = mem->window.phys_base + (pageno << page_shift);
>  	virt_addr = ioremap(*phys_addr, size);
> -	if (!virt_addr)
> +	if (!virt_addr) {
>  		bitmap_release_region(mem->bitmap, pageno, order);
> +	} else {
> +		mem->window.map_size += size;
> +		*window = best_fit;
> +	}
>  
>  	return virt_addr;
>  }
>  EXPORT_SYMBOL_GPL(pci_epc_mem_alloc_addr);
>  
> +static int pci_epc_get_matching_window(struct pci_epc *epc,
> +				       phys_addr_t phys_addr)
> +{
> +	struct pci_epc_mem *mem;
> +	int i;
> +
> +	for (i = 0; i < epc->mem_windows; i++) {
> +		mem = epc->mem[i];
> +
> +		if (mem->window.phys_base == phys_addr)
> +			return i;
> +	}
> +
> +	return -EINVAL;
> +}
> +
>  /**
>   * pci_epc_mem_free_addr() - free the allocated memory address
>   * @epc: the EPC device on which memory was allocated
> @@ -155,16 +254,26 @@ EXPORT_SYMBOL_GPL(pci_epc_mem_alloc_addr);
>  void pci_epc_mem_free_addr(struct pci_epc *epc, phys_addr_t phys_addr,
>  			   void __iomem *virt_addr, size_t size)
>  {
> +	struct pci_epc_mem *mem;
> +	unsigned int page_shift;
> +	int window = 0;
>  	int pageno;
> -	struct pci_epc_mem *mem = epc->mem;
> -	unsigned int page_shift = ilog2(mem->page_size);
>  	int order;
>  
> +	if (epc->mem_windows > 1) {
> +		window = pci_epc_get_matching_window(epc, phys_addr);
> +		if (window < 0)
> +			return;
> +	}
> +
> +	mem = epc->mem[window];
> +	page_shift = ilog2(mem->page_size);
>  	iounmap(virt_addr);
> -	pageno = (phys_addr - mem->phys_base) >> page_shift;
> +	pageno = (phys_addr - mem->window.phys_base) >> page_shift;
>  	size = ALIGN(size, mem->page_size);
>  	order = pci_epc_mem_get_order(mem, size);
>  	bitmap_release_region(mem->bitmap, pageno, order);
> +	mem->window.map_size -= size;
>  }
>  EXPORT_SYMBOL_GPL(pci_epc_mem_free_addr);
>  
> diff --git a/include/linux/pci-epc.h b/include/linux/pci-epc.h
> index 56f1846b9d39..5e86d22dd5c5 100644
> --- a/include/linux/pci-epc.h
> +++ b/include/linux/pci-epc.h
> @@ -48,7 +48,8 @@ struct pci_epc_ops {
>  	void	(*clear_bar)(struct pci_epc *epc, u8 func_no,
>  			     struct pci_epf_bar *epf_bar);
>  	int	(*map_addr)(struct pci_epc *epc, u8 func_no,
> -			    phys_addr_t addr, u64 pci_addr, size_t size);
> +			    phys_addr_t addr, int window,
> +			    u64 pci_addr, size_t size);
>  	void	(*unmap_addr)(struct pci_epc *epc, u8 func_no,
>  			      phys_addr_t addr);
>  	int	(*set_msi)(struct pci_epc *epc, u8 func_no, u8 interrupts);
> @@ -64,17 +65,28 @@ struct pci_epc_ops {
>  	struct module *owner;
>  };
>  
> +#define PCI_EPC_DEFAULT_WINDOW		0
> +
> +/**
> + * struct pci_epc_mem_window - address window of the endpoint controller
> + * @phys_base: physical base address of the PCI address window
> + * @size: the size of the PCI address window
> + */
> +struct pci_epc_mem_window {
> +	phys_addr_t	phys_base;
> +	size_t		size;
> +	size_t		map_size;
> +};
> +
>  /**
>   * struct pci_epc_mem - address space of the endpoint controller
> - * @phys_base: physical base address of the PCI address space
> - * @size: the size of the PCI address space
> + * @window: address window of the endpoint controller
>   * @bitmap: bitmap to manage the PCI address space
> - * @pages: number of bits representing the address region
>   * @page_size: size of each page
> + * @pages: number of bits representing the address region
>   */
>  struct pci_epc_mem {
> -	phys_addr_t	phys_base;
> -	size_t		size;
> +	struct pci_epc_mem_window window;
>  	unsigned long	*bitmap;
>  	size_t		page_size;
>  	int		pages;
> @@ -85,7 +97,8 @@ struct pci_epc_mem {
>   * @dev: PCI EPC device
>   * @pci_epf: list of endpoint functions present in this EPC device
>   * @ops: function pointers for performing endpoint operations
> - * @mem: address space of the endpoint controller
> + * @mem: array of address space of the endpoint controller
> + * @mem_windows: number of windows supported by device
>   * @max_functions: max number of functions that can be configured in this EPC
>   * @group: configfs group representing the PCI EPC device
>   * @lock: spinlock to protect pci_epc ops
> @@ -94,7 +107,8 @@ struct pci_epc {
>  	struct device			dev;
>  	struct list_head		pci_epf;
>  	const struct pci_epc_ops	*ops;
> -	struct pci_epc_mem		*mem;
> +	struct pci_epc_mem		**mem;
> +	int				mem_windows;
>  	u8				max_functions;
>  	struct config_group		*group;
>  	/* spinlock to protect against concurrent access of EP controller */
> @@ -128,8 +142,8 @@ struct pci_epc_features {
>  #define devm_pci_epc_create(dev, ops)    \
>  		__devm_pci_epc_create((dev), (ops), THIS_MODULE)
>  
> -#define pci_epc_mem_init(epc, phys_addr, size)	\
> -		__pci_epc_mem_init((epc), (phys_addr), (size), PAGE_SIZE)
> +#define pci_epc_mem_init(epc, windows, num_windows)	\
> +		__pci_epc_mem_init((epc), windows, num_windows, PAGE_SIZE)
>  
>  static inline void epc_set_drvdata(struct pci_epc *epc, void *data)
>  {
> @@ -159,7 +173,7 @@ int pci_epc_set_bar(struct pci_epc *epc, u8 func_no,
>  void pci_epc_clear_bar(struct pci_epc *epc, u8 func_no,
>  		       struct pci_epf_bar *epf_bar);
>  int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> -		     phys_addr_t phys_addr,
> +		     phys_addr_t phys_addr, int window,
>  		     u64 pci_addr, size_t size);
>  void pci_epc_unmap_addr(struct pci_epc *epc, u8 func_no,
>  			phys_addr_t phys_addr);
> @@ -178,11 +192,12 @@ unsigned int pci_epc_get_first_free_bar(const struct pci_epc_features
>  struct pci_epc *pci_epc_get(const char *epc_name);
>  void pci_epc_put(struct pci_epc *epc);
>  
> -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_addr, size_t size,
> -		       size_t page_size);
> +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *window,
> +		       int num_windows, size_t page_size);
>  void pci_epc_mem_exit(struct pci_epc *epc);
>  void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
> -				     phys_addr_t *phys_addr, size_t size);
> +				     phys_addr_t *phys_addr,
> +				     int *window, size_t size);
>  void pci_epc_mem_free_addr(struct pci_epc *epc, phys_addr_t phys_addr,
>  			   void __iomem *virt_addr, size_t size);
>  #endif /* __LINUX_PCI_EPC_H */
> 

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 0/5] Add support for PCIe controller to work in endpoint mode on R-Car SoCs.
  2019-11-26 14:33   ` Lad, Prabhakar
@ 2019-11-27  6:07     ` Kishon Vijay Abraham I
  -1 siblings, 0 replies; 60+ messages in thread
From: Kishon Vijay Abraham I @ 2019-11-27  6:07 UTC (permalink / raw)
  To: Lad, Prabhakar, Bjorn Helgaas, Rob Herring, Mark Rutland,
	Geert Uytterhoeven, Magnus Damm, Marek Vasut, Yoshihiro Shimoda,
	linux-pci
  Cc: Catalin Marinas, Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS, LKML,
	LAK, Linux-Renesas, Chris Paterson, Lad, Prabhakar

Prabhakar,

On 26/11/19 8:03 PM, Lad, Prabhakar wrote:
> Hi Bjorn/Kishon,
> 
> Gentle ping for this patch set.

Reviewed them now. Sorry for the delay.

Thanks
Kishon

> 
> On Wed, Nov 6, 2019 at 7:36 PM Lad Prabhakar <prabhakar.csengg@gmail.com> wrote:
>>
>> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
>>
>> This patch series adds support for PCIe controller on rcar to work in endpoint mode,
>> this also extends the epf framework to handle multiple outbound regions.
>>
> Cheers,
> --Prabhakar Lad
> 
> 
>> Note:
>> The cadence/rockchip/designware endpoint drivers are build tested only.
>>
>> Endpoint configuration:
>> $ cd /sys/kernel/config/pci_ep
>> $ mkdir functions/pci_epf_test/func1
>> $ echo 0x1912 > functions/pci_epf_test/func1/vendorid
>> $ echo 0x002d > functions/pci_epf_test/func1/deviceid
>> $ echo 32 > functions/pci_epf_test/func1/msi_interrupts
>> $ ln -s functions/pci_epf_test/func1/ controllers/fe000000.pcie_ep/
>> $ echo 1 > controllers/fe000000.pcie_ep/start
>>
>> Host side results:
>> lspci output:
>> 01:00.0 Unassigned class [ff00]: Renesas Technology Corp. Device 002d
>>         Flags: bus master, fast devsel, latency 0, IRQ 103
>>         Memory at fe200200 (64-bit, non-prefetchable) [size=128]
>>         Memory at fe200000 (64-bit, non-prefetchable) [size=256]
>>         Memory at fe200100 (64-bit, non-prefetchable) [size=256]
>>         Capabilities: [40] Power Management version 3
>>         Capabilities: [50] MSI: Enable- Count=1/1 Maskable+ 64bit+
>>         Capabilities: [70] Express Endpoint, MSI 00
>>         Capabilities: [100] Virtual Channel
>>         Kernel driver in use: pci-endpoint-test
>>
>> pcitest results:
>>
>> BAR tests
>>
>> BAR0:           OKAY
>> BAR1:           NOT OKAY
>> BAR2:           OKAY
>> BAR3:           NOT OKAY
>> BAR4:           OKAY
>> BAR5:           NOT OKAY
>>
>> Interrupt tests
>>
>> SET IRQ TYPE TO LEGACY:         OKAY
>> LEGACY IRQ:     OKAY
>> Read Tests
>>
>> SET IRQ TYPE TO LEGACY:         OKAY
>> READ (      1 bytes):           OKAY
>> READ (   1024 bytes):           OKAY
>> READ (   1025 bytes):           OKAY
>> READ (1024000 bytes):           OKAY
>> READ (1024001 bytes):           OKAY
>>
>> Write Tests
>>
>> WRITE (      1 bytes):          OKAY
>> WRITE (   1024 bytes):          OKAY
>> WRITE (   1025 bytes):          OKAY
>> WRITE (1024000 bytes):          OKAY
>> WRITE (1024001 bytes):          OKAY
>>
>> Copy Tests
>>
>> COPY (      1 bytes):           OKAY
>> COPY (   1024 bytes):           OKAY
>> COPY (   1025 bytes):           OKAY
>> COPY (1024000 bytes):           OKAY
>> COPY (1024001 bytes):           OKAY
>>
>> BAR tests for 1/3/5 fail because its configured to use 64bit bars
>>
>> Lad, Prabhakar (5):
>>   pci: pcie-rcar: preparation for adding endpoint support
>>   pci: endpoint: add support to handle multiple base for mapping
>>     outbound memory
>>   PCI: rcar: Add R-Car PCIe endpoint device tree bindings
>>   pci: rcar: add support for rcar pcie controller in endpoint mode
>>   misc: pci_endpoint_test: add device-id for RZ/G2 pcie controller
>>
>>  .../devicetree/bindings/pci/rcar-pci-ep.txt   |   43 +
>>  arch/arm64/configs/defconfig                  |    2 +-
>>  arch/arm64/configs/renesas_defconfig          |    2 +-
>>  drivers/misc/pci_endpoint_test.c              |    3 +
>>  drivers/pci/controller/Kconfig                |   11 +-
>>  drivers/pci/controller/Makefile               |    3 +-
>>  .../pci/controller/dwc/pcie-designware-ep.c   |   30 +-
>>  drivers/pci/controller/pcie-cadence-ep.c      |   11 +-
>>  drivers/pci/controller/pcie-rcar-ep.c         |  483 +++++++
>>  drivers/pci/controller/pcie-rcar-host.c       | 1058 ++++++++++++++
>>  drivers/pci/controller/pcie-rcar.c            | 1231 +----------------
>>  drivers/pci/controller/pcie-rcar.h            |  129 ++
>>  drivers/pci/controller/pcie-rockchip-ep.c     |   13 +-
>>  drivers/pci/endpoint/functions/pci-epf-test.c |   29 +-
>>  drivers/pci/endpoint/pci-epc-core.c           |    7 +-
>>  drivers/pci/endpoint/pci-epc-mem.c            |  189 ++-
>>  include/linux/pci-epc.h                       |   43 +-
>>  17 files changed, 2016 insertions(+), 1271 deletions(-)
>>  create mode 100644 Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
>>  create mode 100644 drivers/pci/controller/pcie-rcar-ep.c
>>  create mode 100644 drivers/pci/controller/pcie-rcar-host.c
>>  create mode 100644 drivers/pci/controller/pcie-rcar.h
>>
>> --
>> 2.20.1
>>

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 0/5] Add support for PCIe controller to work in endpoint mode on R-Car SoCs.
@ 2019-11-27  6:07     ` Kishon Vijay Abraham I
  0 siblings, 0 replies; 60+ messages in thread
From: Kishon Vijay Abraham I @ 2019-11-27  6:07 UTC (permalink / raw)
  To: Lad, Prabhakar, Bjorn Helgaas, Rob Herring, Mark Rutland,
	Geert Uytterhoeven, Magnus Damm, Marek Vasut, Yoshihiro Shimoda,
	linux-pci
  Cc: open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Chris Paterson, Lorenzo Pieralisi, Arnd Bergmann,
	Catalin Marinas, LKML, Lad, Prabhakar, Linux-Renesas,
	Greg Kroah-Hartman, Andrew Murray, Will Deacon, LAK

Prabhakar,

On 26/11/19 8:03 PM, Lad, Prabhakar wrote:
> Hi Bjorn/Kishon,
> 
> Gentle ping for this patch set.

Reviewed them now. Sorry for the delay.

Thanks
Kishon

> 
> On Wed, Nov 6, 2019 at 7:36 PM Lad Prabhakar <prabhakar.csengg@gmail.com> wrote:
>>
>> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
>>
>> This patch series adds support for PCIe controller on rcar to work in endpoint mode,
>> this also extends the epf framework to handle multiple outbound regions.
>>
> Cheers,
> --Prabhakar Lad
> 
> 
>> Note:
>> The cadence/rockchip/designware endpoint drivers are build tested only.
>>
>> Endpoint configuration:
>> $ cd /sys/kernel/config/pci_ep
>> $ mkdir functions/pci_epf_test/func1
>> $ echo 0x1912 > functions/pci_epf_test/func1/vendorid
>> $ echo 0x002d > functions/pci_epf_test/func1/deviceid
>> $ echo 32 > functions/pci_epf_test/func1/msi_interrupts
>> $ ln -s functions/pci_epf_test/func1/ controllers/fe000000.pcie_ep/
>> $ echo 1 > controllers/fe000000.pcie_ep/start
>>
>> Host side results:
>> lspci output:
>> 01:00.0 Unassigned class [ff00]: Renesas Technology Corp. Device 002d
>>         Flags: bus master, fast devsel, latency 0, IRQ 103
>>         Memory at fe200200 (64-bit, non-prefetchable) [size=128]
>>         Memory at fe200000 (64-bit, non-prefetchable) [size=256]
>>         Memory at fe200100 (64-bit, non-prefetchable) [size=256]
>>         Capabilities: [40] Power Management version 3
>>         Capabilities: [50] MSI: Enable- Count=1/1 Maskable+ 64bit+
>>         Capabilities: [70] Express Endpoint, MSI 00
>>         Capabilities: [100] Virtual Channel
>>         Kernel driver in use: pci-endpoint-test
>>
>> pcitest results:
>>
>> BAR tests
>>
>> BAR0:           OKAY
>> BAR1:           NOT OKAY
>> BAR2:           OKAY
>> BAR3:           NOT OKAY
>> BAR4:           OKAY
>> BAR5:           NOT OKAY
>>
>> Interrupt tests
>>
>> SET IRQ TYPE TO LEGACY:         OKAY
>> LEGACY IRQ:     OKAY
>> Read Tests
>>
>> SET IRQ TYPE TO LEGACY:         OKAY
>> READ (      1 bytes):           OKAY
>> READ (   1024 bytes):           OKAY
>> READ (   1025 bytes):           OKAY
>> READ (1024000 bytes):           OKAY
>> READ (1024001 bytes):           OKAY
>>
>> Write Tests
>>
>> WRITE (      1 bytes):          OKAY
>> WRITE (   1024 bytes):          OKAY
>> WRITE (   1025 bytes):          OKAY
>> WRITE (1024000 bytes):          OKAY
>> WRITE (1024001 bytes):          OKAY
>>
>> Copy Tests
>>
>> COPY (      1 bytes):           OKAY
>> COPY (   1024 bytes):           OKAY
>> COPY (   1025 bytes):           OKAY
>> COPY (1024000 bytes):           OKAY
>> COPY (1024001 bytes):           OKAY
>>
>> BAR tests for 1/3/5 fail because its configured to use 64bit bars
>>
>> Lad, Prabhakar (5):
>>   pci: pcie-rcar: preparation for adding endpoint support
>>   pci: endpoint: add support to handle multiple base for mapping
>>     outbound memory
>>   PCI: rcar: Add R-Car PCIe endpoint device tree bindings
>>   pci: rcar: add support for rcar pcie controller in endpoint mode
>>   misc: pci_endpoint_test: add device-id for RZ/G2 pcie controller
>>
>>  .../devicetree/bindings/pci/rcar-pci-ep.txt   |   43 +
>>  arch/arm64/configs/defconfig                  |    2 +-
>>  arch/arm64/configs/renesas_defconfig          |    2 +-
>>  drivers/misc/pci_endpoint_test.c              |    3 +
>>  drivers/pci/controller/Kconfig                |   11 +-
>>  drivers/pci/controller/Makefile               |    3 +-
>>  .../pci/controller/dwc/pcie-designware-ep.c   |   30 +-
>>  drivers/pci/controller/pcie-cadence-ep.c      |   11 +-
>>  drivers/pci/controller/pcie-rcar-ep.c         |  483 +++++++
>>  drivers/pci/controller/pcie-rcar-host.c       | 1058 ++++++++++++++
>>  drivers/pci/controller/pcie-rcar.c            | 1231 +----------------
>>  drivers/pci/controller/pcie-rcar.h            |  129 ++
>>  drivers/pci/controller/pcie-rockchip-ep.c     |   13 +-
>>  drivers/pci/endpoint/functions/pci-epf-test.c |   29 +-
>>  drivers/pci/endpoint/pci-epc-core.c           |    7 +-
>>  drivers/pci/endpoint/pci-epc-mem.c            |  189 ++-
>>  include/linux/pci-epc.h                       |   43 +-
>>  17 files changed, 2016 insertions(+), 1271 deletions(-)
>>  create mode 100644 Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
>>  create mode 100644 drivers/pci/controller/pcie-rcar-ep.c
>>  create mode 100644 drivers/pci/controller/pcie-rcar-host.c
>>  create mode 100644 drivers/pci/controller/pcie-rcar.h
>>
>> --
>> 2.20.1
>>

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 1/5] pci: pcie-rcar: preparation for adding endpoint support
  2019-11-27  4:55     ` Kishon Vijay Abraham I
@ 2019-11-27 20:51       ` Lad, Prabhakar
  -1 siblings, 0 replies; 60+ messages in thread
From: Lad, Prabhakar @ 2019-11-27 20:51 UTC (permalink / raw)
  To: Kishon Vijay Abraham I
  Cc: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Marek Vasut, Yoshihiro Shimoda, linux-pci,
	Catalin Marinas, Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS, LKML,
	LAK, Linux-Renesas, Chris Paterson, Lad, Prabhakar

Hi Kishon,

On Wed, Nov 27, 2019 at 4:56 AM Kishon Vijay Abraham I <kishon@ti.com> wrote:
>
> Hi Prabhakar,
>
> On 07/11/19 1:06 AM, Lad Prabhakar wrote:
> > From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >
> > this patch prepares for adding endpoint support to rcar controller,
> > there are no functional changes with this patch, a common file is
> > created so that it can be shared with endpoint driver.
> >
> > Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > ---
> >  arch/arm64/configs/defconfig            |    2 +-
> >  arch/arm64/configs/renesas_defconfig    |    2 +-
> >  drivers/pci/controller/Kconfig          |    4 +-
> >  drivers/pci/controller/Makefile         |    2 +-
> >  drivers/pci/controller/pcie-rcar-host.c | 1058 +++++++++++++++++++
> >  drivers/pci/controller/pcie-rcar.c      | 1231 +----------------------
>
> Can you create this patch using "git format-patch -M", so that only whatever
> changed appears here.
>
sure will do that in the next version.

Cheers,
--Prabhakar

> Thanks
> Kishon
>
> >  drivers/pci/controller/pcie-rcar.h      |  129 +++
> >  7 files changed, 1245 insertions(+), 1183 deletions(-)
> >  create mode 100644 drivers/pci/controller/pcie-rcar-host.c
> >  create mode 100644 drivers/pci/controller/pcie-rcar.h
> >
> > diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
> > index 7d16f10a9251..f67ce0ccea1d 100644
> > --- a/arch/arm64/configs/defconfig
> > +++ b/arch/arm64/configs/defconfig
> > @@ -186,7 +186,7 @@ CONFIG_HOTPLUG_PCI=y
> >  CONFIG_HOTPLUG_PCI_ACPI=y
> >  CONFIG_PCI_AARDVARK=y
> >  CONFIG_PCI_TEGRA=y
> > -CONFIG_PCIE_RCAR=y
> > +CONFIG_PCIE_RCAR_HOST=y
> >  CONFIG_PCI_HOST_GENERIC=y
> >  CONFIG_PCI_XGENE=y
> >  CONFIG_PCIE_ALTERA=y
> > diff --git a/arch/arm64/configs/renesas_defconfig b/arch/arm64/configs/renesas_defconfig
> > index f7411e8807d6..75fc418377c5 100644
> > --- a/arch/arm64/configs/renesas_defconfig
> > +++ b/arch/arm64/configs/renesas_defconfig
> > @@ -91,7 +91,7 @@ CONFIG_PCIEPORTBUS=y
> >  CONFIG_HOTPLUG_PCI_PCIE=y
> >  CONFIG_PCI_IOV=y
> >  CONFIG_HOTPLUG_PCI=y
> > -CONFIG_PCIE_RCAR=y
> > +CONFIG_PCIE_RCAR_HOST=y
> >  CONFIG_PCI_HOST_GENERIC=y
> >  CONFIG_DEVTMPFS=y
> >  CONFIG_DEVTMPFS_MOUNT=y
> > diff --git a/drivers/pci/controller/Kconfig b/drivers/pci/controller/Kconfig
> > index 70e078238899..01c238e5ea1e 100644
> > --- a/drivers/pci/controller/Kconfig
> > +++ b/drivers/pci/controller/Kconfig
> > @@ -82,12 +82,12 @@ config PCI_RCAR_GEN2
> >         There are 3 internal PCI controllers available with a single
> >         built-in EHCI/OHCI host controller present on each one.
> >
> > -config PCIE_RCAR
> > +config PCIE_RCAR_HOST
> >       bool "Renesas R-Car PCIe controller"
> >       depends on ARCH_RENESAS || COMPILE_TEST
> >       depends on PCI_MSI_IRQ_DOMAIN
> >       help
> > -       Say Y here if you want PCIe controller support on R-Car SoCs.
> > +       Say Y here if you want PCIe controller support on R-Car SoCs in host mode.
> >
> >  config PCI_HOST_COMMON
> >       bool
> > diff --git a/drivers/pci/controller/Makefile b/drivers/pci/controller/Makefile
> > index a2a22c9d91af..3577902a0d40 100644
> > --- a/drivers/pci/controller/Makefile
> > +++ b/drivers/pci/controller/Makefile
> > @@ -9,7 +9,7 @@ obj-$(CONFIG_PCI_MVEBU) += pci-mvebu.o
> >  obj-$(CONFIG_PCI_AARDVARK) += pci-aardvark.o
> >  obj-$(CONFIG_PCI_TEGRA) += pci-tegra.o
> >  obj-$(CONFIG_PCI_RCAR_GEN2) += pci-rcar-gen2.o
> > -obj-$(CONFIG_PCIE_RCAR) += pcie-rcar.o
> > +obj-$(CONFIG_PCIE_RCAR_HOST) += pcie-rcar.o pcie-rcar-host.o
> >  obj-$(CONFIG_PCI_HOST_COMMON) += pci-host-common.o
> >  obj-$(CONFIG_PCI_HOST_GENERIC) += pci-host-generic.o
> >  obj-$(CONFIG_PCIE_XILINX) += pcie-xilinx.o
> > diff --git a/drivers/pci/controller/pcie-rcar-host.c b/drivers/pci/controller/pcie-rcar-host.c
> > new file mode 100644
> > index 000000000000..645ef3e41e27
> > --- /dev/null
> > +++ b/drivers/pci/controller/pcie-rcar-host.c
> > @@ -0,0 +1,1058 @@
> > +// SPDX-License-Identifier: GPL-2.0
> > +/*
> > + * PCIe driver for Renesas R-Car SoCs
> > + *  Copyright (C) 2014-2019 Renesas Electronics Europe Ltd
> > + *
> > + * Based on:
> > + *  arch/sh/drivers/pci/pcie-sh7786.c
> > + *  arch/sh/drivers/pci/ops-sh7786.c
> > + *  Copyright (C) 2009 - 2011  Paul Mundt
> > + *
> > + * Author: Phil Edworthy <phil.edworthy@renesas.com>
> > + */
> > +
> > +#include <linux/clk.h>
> > +#include <linux/delay.h>
> > +#include <linux/msi.h>
> > +#include <linux/of_address.h>
> > +#include <linux/of_irq.h>
> > +#include <linux/of_pci.h>
> > +#include <linux/of_platform.h>
> > +#include <linux/pci.h>
> > +#include <linux/phy/phy.h>
> > +
> > +#include "pcie-rcar.h"
> > +
> > +struct rcar_msi {
> > +     DECLARE_BITMAP(used, INT_PCI_MSI_NR);
> > +     struct irq_domain *domain;
> > +     struct msi_controller chip;
> > +     unsigned long pages;
> > +     struct mutex lock; /* serializes msi */
> > +     int irq1;
> > +     int irq2;
> > +};
> > +
> > +static inline struct rcar_msi *to_rcar_msi(struct msi_controller *chip)
> > +{
> > +     return container_of(chip, struct rcar_msi, chip);
> > +}
> > +
> > +/* Structure representing the PCIe interface */
> > +struct rcar_pcie {
> > +     struct device           *dev;
> > +     struct phy              *phy;
> > +     void __iomem            *base;
> > +     struct list_head        resources;
> > +     int                     root_bus_nr;
> > +     struct clk              *bus_clk;
> > +     struct                  rcar_msi msi;
> > +};
> > +
> > +static u32 rcar_read_conf(struct rcar_pcie *pcie, int where)
> > +{
> > +     unsigned int shift = BITS_PER_BYTE * (where & 3);
> > +     u32 val = rcar_pci_read_reg(pcie->base, where & ~3);
> > +
> > +     return val >> shift;
> > +}
> > +
> > +/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
> > +static int
> > +rcar_pcie_config_access(struct rcar_pcie *pcie,
> > +                     unsigned char access_type, struct pci_bus *bus,
> > +                     unsigned int devfn, int where, u32 *data)
> > +{
> > +     unsigned int dev, func, reg, index;
> > +
> > +     dev = PCI_SLOT(devfn);
> > +     func = PCI_FUNC(devfn);
> > +     reg = where & ~3;
> > +     index = reg / 4;
> > +
> > +     /*
> > +      * While each channel has its own memory-mapped extended config
> > +      * space, it's generally only accessible when in endpoint mode.
> > +      * When in root complex mode, the controller is unable to target
> > +      * itself with either type 0 or type 1 accesses, and indeed, any
> > +      * controller initiated target transfer to its own config space
> > +      * result in a completer abort.
> > +      *
> > +      * Each channel effectively only supports a single device, but as
> > +      * the same channel <-> device access works for any PCI_SLOT()
> > +      * value, we cheat a bit here and bind the controller's config
> > +      * space to devfn 0 in order to enable self-enumeration. In this
> > +      * case the regular ECAR/ECDR path is sidelined and the mangled
> > +      * config access itself is initiated as an internal bus transaction.
> > +      */
> > +     if (pci_is_root_bus(bus)) {
> > +             if (dev != 0)
> > +                     return PCIBIOS_DEVICE_NOT_FOUND;
> > +
> > +             if (access_type == RCAR_PCI_ACCESS_READ) {
> > +                     *data = rcar_pci_read_reg(pcie->base, PCICONF(index));
> > +             } else {
> > +                     /* Keep an eye out for changes to the root bus number */
> > +                     if (pci_is_root_bus(bus) && reg == PCI_PRIMARY_BUS)
> > +                             pcie->root_bus_nr = *data & 0xff;
> > +
> > +                     rcar_pci_write_reg(pcie->base, *data, PCICONF(index));
> > +             }
> > +
> > +             return PCIBIOS_SUCCESSFUL;
> > +     }
> > +
> > +     if (pcie->root_bus_nr < 0)
> > +             return PCIBIOS_DEVICE_NOT_FOUND;
> > +
> > +     /* Clear errors */
> > +     rcar_pci_write_reg(pcie->base, rcar_pci_read_reg(pcie->base, PCIEERRFR),
> > +                        PCIEERRFR);
> > +
> > +     /* Set the PIO address */
> > +     rcar_pci_write_reg(pcie->base, PCIE_CONF_BUS(bus->number) |
> > +             PCIE_CONF_DEV(dev) | PCIE_CONF_FUNC(func) | reg, PCIECAR);
> > +
> > +     /* Enable the configuration access */
> > +     if (bus->parent->number == pcie->root_bus_nr)
> > +             rcar_pci_write_reg(pcie->base,
> > +                                CONFIG_SEND_ENABLE | TYPE0, PCIECCTLR);
> > +     else
> > +             rcar_pci_write_reg(pcie->base,
> > +                                CONFIG_SEND_ENABLE | TYPE1, PCIECCTLR);
> > +
> > +     /* Check for errors */
> > +     if (rcar_pci_read_reg(pcie->base, PCIEERRFR) & UNSUPPORTED_REQUEST)
> > +             return PCIBIOS_DEVICE_NOT_FOUND;
> > +
> > +     /* Check for master and target aborts */
> > +     if (rcar_read_conf(pcie, RCONF(PCI_STATUS)) &
> > +             (PCI_STATUS_REC_MASTER_ABORT | PCI_STATUS_REC_TARGET_ABORT))
> > +             return PCIBIOS_DEVICE_NOT_FOUND;
> > +
> > +     if (access_type == RCAR_PCI_ACCESS_READ)
> > +             *data = rcar_pci_read_reg(pcie->base, PCIECDR);
> > +     else
> > +             rcar_pci_write_reg(pcie->base, *data, PCIECDR);
> > +
> > +     /* Disable the configuration access */
> > +     rcar_pci_write_reg(pcie->base, 0, PCIECCTLR);
> > +
> > +     return PCIBIOS_SUCCESSFUL;
> > +}
> > +
> > +static int rcar_pcie_read_conf(struct pci_bus *bus, unsigned int devfn,
> > +                            int where, int size, u32 *val)
> > +{
> > +     struct rcar_pcie *pcie = bus->sysdata;
> > +     int ret;
> > +
> > +     ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
> > +                                   bus, devfn, where, val);
> > +     if (ret != PCIBIOS_SUCCESSFUL) {
> > +             *val = 0xffffffff;
> > +             return ret;
> > +     }
> > +
> > +     if (size == 1)
> > +             *val = (*val >> (BITS_PER_BYTE * (where & 3))) & 0xff;
> > +     else if (size == 2)
> > +             *val = (*val >> (BITS_PER_BYTE * (where & 2))) & 0xffff;
> > +
> > +     dev_dbg(&bus->dev, "pcie-config-read: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
> > +             bus->number, devfn, where, size, *val);
> > +
> > +     return ret;
> > +}
> > +
> > +/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
> > +static int rcar_pcie_write_conf(struct pci_bus *bus, unsigned int devfn,
> > +                             int where, int size, u32 val)
> > +{
> > +     struct rcar_pcie *pcie = bus->sysdata;
> > +     unsigned int shift;
> > +     u32 data;
> > +     int ret;
> > +
> > +     ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
> > +                                   bus, devfn, where, &data);
> > +     if (ret != PCIBIOS_SUCCESSFUL)
> > +             return ret;
> > +
> > +     dev_dbg(&bus->dev, "pcie-config-write: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
> > +             bus->number, devfn, where, size, val);
> > +
> > +     if (size == 1) {
> > +             shift = BITS_PER_BYTE * (where & 3);
> > +             data &= ~(0xff << shift);
> > +             data |= ((val & 0xff) << shift);
> > +     } else if (size == 2) {
> > +             shift = BITS_PER_BYTE * (where & 2);
> > +             data &= ~(0xffff << shift);
> > +             data |= ((val & 0xffff) << shift);
> > +     } else {
> > +             data = val;
> > +     }
> > +
> > +     ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_WRITE,
> > +                                   bus, devfn, where, &data);
> > +
> > +     return ret;
> > +}
> > +
> > +static struct pci_ops rcar_pcie_ops = {
> > +     .read   = rcar_pcie_read_conf,
> > +     .write  = rcar_pcie_write_conf,
> > +};
> > +
> > +static int rcar_pcie_setup(struct list_head *resource, struct rcar_pcie *pci)
> > +{
> > +     struct resource_entry *win;
> > +     int i = 0;
> > +
> > +     /* Setup PCI resources */
> > +     resource_list_for_each_entry(win, &pci->resources) {
> > +             struct resource *res = win->res;
> > +
> > +             if (!res->flags)
> > +                     continue;
> > +
> > +             switch (resource_type(res)) {
> > +             case IORESOURCE_IO:
> > +             case IORESOURCE_MEM:
> > +                     rcar_pcie_set_outbound(i, pci->base, res, true);
> > +                     i++;
> > +                     break;
> > +             case IORESOURCE_BUS:
> > +                     pci->root_bus_nr = res->start;
> > +                     break;
> > +             default:
> > +                     continue;
> > +             }
> > +
> > +             pci_add_resource(resource, res);
> > +     }
> > +
> > +     return 1;
> > +}
> > +
> > +static void rcar_pcie_force_speedup(struct rcar_pcie *pcie)
> > +{
> > +     struct device *dev = pcie->dev;
> > +     unsigned int timeout = 1000;
> > +     u32 macsr;
> > +
> > +     if ((rcar_pci_read_reg(pcie->base, MACS2R) & LINK_SPEED) !=
> > +             LINK_SPEED_5_0GTS)
> > +             return;
> > +
> > +     if (rcar_pci_read_reg(pcie->base, MACCTLR) & SPEED_CHANGE) {
> > +             dev_err(dev, "Speed change already in progress\n");
> > +             return;
> > +     }
> > +
> > +     macsr = rcar_pci_read_reg(pcie->base, MACSR);
> > +     if ((macsr & LINK_SPEED) == LINK_SPEED_5_0GTS)
> > +             goto done;
> > +
> > +     /* Set target link speed to 5.0 GT/s */
> > +     rcar_rmw32(pcie->base, EXPCAP(12), PCI_EXP_LNKSTA_CLS,
> > +                PCI_EXP_LNKSTA_CLS_5_0GB);
> > +
> > +     /* Set speed change reason as intentional factor */
> > +     rcar_rmw32(pcie->base, MACCGSPSETR, SPCNGRSN, 0);
> > +
> > +     /* Clear SPCHGFIN, SPCHGSUC, and SPCHGFAIL */
> > +     if (macsr & (SPCHGFIN | SPCHGSUC | SPCHGFAIL))
> > +             rcar_pci_write_reg(pcie->base, macsr, MACSR);
> > +
> > +     /* Start link speed change */
> > +     rcar_rmw32(pcie->base, MACCTLR, SPEED_CHANGE, SPEED_CHANGE);
> > +
> > +     while (timeout--) {
> > +             macsr = rcar_pci_read_reg(pcie->base, MACSR);
> > +             if (macsr & SPCHGFIN) {
> > +                     /* Clear the interrupt bits */
> > +                     rcar_pci_write_reg(pcie->base, macsr, MACSR);
> > +
> > +                     if (macsr & SPCHGFAIL)
> > +                             dev_err(dev, "Speed change failed\n");
> > +
> > +                     goto done;
> > +             }
> > +
> > +             usleep_range(1000, 2000);
> > +     }
> > +
> > +     dev_err(dev, "Speed change timed out\n");
> > +
> > +done:
> > +     dev_info(dev, "Current link speed is %s GT/s\n",
> > +              (macsr & LINK_SPEED) == LINK_SPEED_5_0GTS ? "5" : "2.5");
> > +}
> > +
> > +static int rcar_pcie_enable(struct rcar_pcie *pcie)
> > +{
> > +     struct pci_host_bridge *bridge = pci_host_bridge_from_priv(pcie);
> > +     struct device *dev = pcie->dev;
> > +     struct pci_bus *bus, *child;
> > +     int ret;
> > +
> > +     /* Try setting 5 GT/s link speed */
> > +     rcar_pcie_force_speedup(pcie);
> > +
> > +     rcar_pcie_setup(&bridge->windows, pcie);
> > +
> > +     pci_add_flags(PCI_REASSIGN_ALL_BUS);
> > +
> > +     bridge->dev.parent = dev;
> > +     bridge->sysdata = pcie;
> > +     bridge->busnr = pcie->root_bus_nr;
> > +     bridge->ops = &rcar_pcie_ops;
> > +     bridge->map_irq = of_irq_parse_and_map_pci;
> > +     bridge->swizzle_irq = pci_common_swizzle;
> > +     if (IS_ENABLED(CONFIG_PCI_MSI))
> > +             bridge->msi = &pcie->msi.chip;
> > +
> > +     ret = pci_scan_root_bus_bridge(bridge);
> > +     if (ret < 0)
> > +             return ret;
> > +
> > +     bus = bridge->bus;
> > +
> > +     pci_bus_size_bridges(bus);
> > +     pci_bus_assign_resources(bus);
> > +
> > +     list_for_each_entry(child, &bus->children, node)
> > +             pcie_bus_configure_settings(child);
> > +
> > +     pci_bus_add_devices(bus);
> > +
> > +     return 0;
> > +}
> > +
> > +static int phy_wait_for_ack(struct rcar_pcie *pcie)
> > +{
> > +     struct device *dev = pcie->dev;
> > +     unsigned int timeout = 100;
> > +
> > +     while (timeout--) {
> > +             if (rcar_pci_read_reg(pcie->base, H1_PCIEPHYADRR) & PHY_ACK)
> > +                     return 0;
> > +
> > +             usleep_range(90, 110);
> > +     }
> > +
> > +     dev_err(dev, "Access to PCIe phy timed out\n");
> > +
> > +     return -ETIMEDOUT;
> > +}
> > +
> > +static void phy_write_reg(struct rcar_pcie *pcie,
> > +                       unsigned int rate, u32 addr,
> > +                       unsigned int lane, u32 data)
> > +{
> > +     u32 phyaddr;
> > +
> > +     phyaddr = WRITE_CMD |
> > +             ((rate & 1) << RATE_POS) |
> > +             ((lane & 0xf) << LANE_POS) |
> > +             ((addr & 0xff) << ADR_POS);
> > +
> > +     /* Set write data */
> > +     rcar_pci_write_reg(pcie->base, data, H1_PCIEPHYDOUTR);
> > +     rcar_pci_write_reg(pcie->base, phyaddr, H1_PCIEPHYADRR);
> > +
> > +     /* Ignore errors as they will be dealt with if the data link is down */
> > +     phy_wait_for_ack(pcie);
> > +
> > +     /* Clear command */
> > +     rcar_pci_write_reg(pcie->base, 0, H1_PCIEPHYDOUTR);
> > +     rcar_pci_write_reg(pcie->base, 0, H1_PCIEPHYADRR);
> > +
> > +     /* Ignore errors as they will be dealt with if the data link is down */
> > +     phy_wait_for_ack(pcie);
> > +}
> > +
> > +static int rcar_pcie_hw_init(struct rcar_pcie *pcie)
> > +{
> > +     int err;
> > +
> > +     /* Begin initialization */
> > +     rcar_pci_write_reg(pcie->base, 0, PCIETCTLR);
> > +
> > +     /* Set mode */
> > +     rcar_pci_write_reg(pcie->base, 1, PCIEMSR);
> > +
> > +     err = rcar_pcie_wait_for_phyrdy(pcie->base);
> > +     if (err)
> > +             return err;
> > +
> > +     /*
> > +      * Initial header for port config space is type 1, set the device
> > +      * class to match. Hardware takes care of propagating the IDSETR
> > +      * settings, so there is no need to bother with a quirk.
> > +      */
> > +     rcar_pci_write_reg(pcie->base, PCI_CLASS_BRIDGE_PCI << 16, IDSETR1);
> > +
> > +     /*
> > +      * Setup Secondary Bus Number & Subordinate Bus Number, even though
> > +      * they aren't used, to avoid bridge being detected as broken.
> > +      */
> > +     rcar_rmw32(pcie->base, RCONF(PCI_SECONDARY_BUS), 0xff, 1);
> > +     rcar_rmw32(pcie->base, RCONF(PCI_SUBORDINATE_BUS), 0xff, 1);
> > +
> > +     /* Initialize default capabilities. */
> > +     rcar_rmw32(pcie->base, REXPCAP(0), 0xff, PCI_CAP_ID_EXP);
> > +     rcar_rmw32(pcie->base, REXPCAP(PCI_EXP_FLAGS),
> > +                PCI_EXP_FLAGS_TYPE, PCI_EXP_TYPE_ROOT_PORT << 4);
> > +     rcar_rmw32(pcie->base, RCONF(PCI_HEADER_TYPE), 0x7f,
> > +                PCI_HEADER_TYPE_BRIDGE);
> > +
> > +     /* Enable data link layer active state reporting */
> > +     rcar_rmw32(pcie->base, REXPCAP(PCI_EXP_LNKCAP), PCI_EXP_LNKCAP_DLLLARC,
> > +                PCI_EXP_LNKCAP_DLLLARC);
> > +
> > +     /* Write out the physical slot number = 0 */
> > +     rcar_rmw32(pcie->base, REXPCAP(PCI_EXP_SLTCAP), PCI_EXP_SLTCAP_PSN, 0);
> > +
> > +     /* Set the completion timer timeout to the maximum 50ms. */
> > +     rcar_rmw32(pcie->base, TLCTLR + 1, 0x3f, 50);
> > +
> > +     /* Terminate list of capabilities (Next Capability Offset=0) */
> > +     rcar_rmw32(pcie->base, RVCCAP(0), 0xfff00000, 0);
> > +
> > +     /* Enable MSI */
> > +     if (IS_ENABLED(CONFIG_PCI_MSI))
> > +             rcar_pci_write_reg(pcie->base, 0x801f0000, PCIEMSITXR);
> > +
> > +     rcar_rmw32(pcie->base, MACCTLR, MACCTLR_RESERVED, 0);
> > +
> > +     /* Finish initialization - establish a PCI Express link */
> > +     rcar_pci_write_reg(pcie->base, CFINIT, PCIETCTLR);
> > +
> > +     /* This will timeout if we don't have a link. */
> > +     err = rcar_pcie_wait_for_dl(pcie->base);
> > +     if (err)
> > +             return err;
> > +
> > +     /* Enable INTx interrupts */
> > +     rcar_rmw32(pcie->base, PCIEINTXR, 0, 0xF << 8);
> > +
> > +     /* flush modifications */
> > +     wmb();
> > +
> > +     return 0;
> > +}
> > +
> > +static int rcar_pcie_phy_init_h1(struct rcar_pcie *pcie)
> > +{
> > +     /* Initialize the phy */
> > +     phy_write_reg(pcie, 0, 0x42, 0x1, 0x0EC34191);
> > +     phy_write_reg(pcie, 1, 0x42, 0x1, 0x0EC34180);
> > +     phy_write_reg(pcie, 0, 0x43, 0x1, 0x00210188);
> > +     phy_write_reg(pcie, 1, 0x43, 0x1, 0x00210188);
> > +     phy_write_reg(pcie, 0, 0x44, 0x1, 0x015C0014);
> > +     phy_write_reg(pcie, 1, 0x44, 0x1, 0x015C0014);
> > +     phy_write_reg(pcie, 1, 0x4C, 0x1, 0x786174A0);
> > +     phy_write_reg(pcie, 1, 0x4D, 0x1, 0x048000BB);
> > +     phy_write_reg(pcie, 0, 0x51, 0x1, 0x079EC062);
> > +     phy_write_reg(pcie, 0, 0x52, 0x1, 0x20000000);
> > +     phy_write_reg(pcie, 1, 0x52, 0x1, 0x20000000);
> > +     phy_write_reg(pcie, 1, 0x56, 0x1, 0x00003806);
> > +
> > +     phy_write_reg(pcie, 0, 0x60, 0x1, 0x004B03A5);
> > +     phy_write_reg(pcie, 0, 0x64, 0x1, 0x3F0F1F0F);
> > +     phy_write_reg(pcie, 0, 0x66, 0x1, 0x00008000);
> > +
> > +     return 0;
> > +}
> > +
> > +static int rcar_pcie_phy_init_gen2(struct rcar_pcie *pcie)
> > +{
> > +     /*
> > +      * These settings come from the R-Car Series, 2nd Generation User's
> > +      * Manual, section 50.3.1 (2) Initialization of the physical layer.
> > +      */
> > +     rcar_pci_write_reg(pcie->base, 0x000f0030, GEN2_PCIEPHYADDR);
> > +     rcar_pci_write_reg(pcie->base, 0x00381203, GEN2_PCIEPHYDATA);
> > +     rcar_pci_write_reg(pcie->base, 0x00000001, GEN2_PCIEPHYCTRL);
> > +     rcar_pci_write_reg(pcie->base, 0x00000006, GEN2_PCIEPHYCTRL);
> > +
> > +     rcar_pci_write_reg(pcie->base, 0x000f0054, GEN2_PCIEPHYADDR);
> > +     /* The following value is for DC connection, no termination resistor */
> > +     rcar_pci_write_reg(pcie->base, 0x13802007, GEN2_PCIEPHYDATA);
> > +     rcar_pci_write_reg(pcie->base, 0x00000001, GEN2_PCIEPHYCTRL);
> > +     rcar_pci_write_reg(pcie->base, 0x00000006, GEN2_PCIEPHYCTRL);
> > +
> > +     return 0;
> > +}
> > +
> > +static int rcar_pcie_phy_init_gen3(struct rcar_pcie *pcie)
> > +{
> > +     int err;
> > +
> > +     err = phy_init(pcie->phy);
> > +     if (err)
> > +             return err;
> > +
> > +     err = phy_power_on(pcie->phy);
> > +     if (err)
> > +             phy_exit(pcie->phy);
> > +
> > +     return err;
> > +}
> > +
> > +static int rcar_msi_alloc(struct rcar_msi *chip)
> > +{
> > +     int msi;
> > +
> > +     mutex_lock(&chip->lock);
> > +
> > +     msi = find_first_zero_bit(chip->used, INT_PCI_MSI_NR);
> > +     if (msi < INT_PCI_MSI_NR)
> > +             set_bit(msi, chip->used);
> > +     else
> > +             msi = -ENOSPC;
> > +
> > +     mutex_unlock(&chip->lock);
> > +
> > +     return msi;
> > +}
> > +
> > +static int rcar_msi_alloc_region(struct rcar_msi *chip, int no_irqs)
> > +{
> > +     int msi;
> > +
> > +     mutex_lock(&chip->lock);
> > +     msi = bitmap_find_free_region(chip->used, INT_PCI_MSI_NR,
> > +                                   order_base_2(no_irqs));
> > +     mutex_unlock(&chip->lock);
> > +
> > +     return msi;
> > +}
> > +
> > +static void rcar_msi_free(struct rcar_msi *chip, unsigned long irq)
> > +{
> > +     mutex_lock(&chip->lock);
> > +     clear_bit(irq, chip->used);
> > +     mutex_unlock(&chip->lock);
> > +}
> > +
> > +static irqreturn_t rcar_pcie_msi_irq(int irq, void *data)
> > +{
> > +     struct rcar_pcie *pcie = data;
> > +     struct rcar_msi *msi = &pcie->msi;
> > +     struct device *dev = pcie->dev;
> > +     unsigned long reg;
> > +
> > +     reg = rcar_pci_read_reg(pcie->base, PCIEMSIFR);
> > +
> > +     /* MSI & INTx share an interrupt - we only handle MSI here */
> > +     if (!reg)
> > +             return IRQ_NONE;
> > +
> > +     while (reg) {
> > +             unsigned int index = find_first_bit(&reg, 32);
> > +             unsigned int msi_irq;
> > +
> > +             /* clear the interrupt */
> > +             rcar_pci_write_reg(pcie->base, 1 << index, PCIEMSIFR);
> > +
> > +             msi_irq = irq_find_mapping(msi->domain, index);
> > +             if (msi_irq) {
> > +                     if (test_bit(index, msi->used))
> > +                             generic_handle_irq(msi_irq);
> > +                     else
> > +                             dev_info(dev, "unhandled MSI\n");
> > +             } else {
> > +                     /* Unknown MSI, just clear it */
> > +                     dev_dbg(dev, "unexpected MSI\n");
> > +             }
> > +
> > +             /* see if there's any more pending in this vector */
> > +             reg = rcar_pci_read_reg(pcie->base, PCIEMSIFR);
> > +     }
> > +
> > +     return IRQ_HANDLED;
> > +}
> > +
> > +static int rcar_msi_setup_irq(struct msi_controller *chip, struct pci_dev *pdev,
> > +                           struct msi_desc *desc)
> > +{
> > +     struct rcar_msi *msi = to_rcar_msi(chip);
> > +     struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
> > +     struct msi_msg msg;
> > +     unsigned int irq;
> > +     int hwirq;
> > +
> > +     hwirq = rcar_msi_alloc(msi);
> > +     if (hwirq < 0)
> > +             return hwirq;
> > +
> > +     irq = irq_find_mapping(msi->domain, hwirq);
> > +     if (!irq) {
> > +             rcar_msi_free(msi, hwirq);
> > +             return -EINVAL;
> > +     }
> > +
> > +     irq_set_msi_desc(irq, desc);
> > +
> > +     msg.address_lo = rcar_pci_read_reg(pcie->base, PCIEMSIALR) & ~MSIFE;
> > +     msg.address_hi = rcar_pci_read_reg(pcie->base, PCIEMSIAUR);
> > +     msg.data = hwirq;
> > +
> > +     pci_write_msi_msg(irq, &msg);
> > +
> > +     return 0;
> > +}
> > +
> > +static int rcar_msi_setup_irqs(struct msi_controller *chip,
> > +                            struct pci_dev *pdev, int nvec, int type)
> > +{
> > +     struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
> > +     struct rcar_msi *msi = to_rcar_msi(chip);
> > +     struct msi_desc *desc;
> > +     struct msi_msg msg;
> > +     unsigned int irq;
> > +     int hwirq;
> > +     int i;
> > +
> > +     /* MSI-X interrupts are not supported */
> > +     if (type == PCI_CAP_ID_MSIX)
> > +             return -EINVAL;
> > +
> > +     WARN_ON(!list_is_singular(&pdev->dev.msi_list));
> > +     desc = list_entry(pdev->dev.msi_list.next, struct msi_desc, list);
> > +
> > +     hwirq = rcar_msi_alloc_region(msi, nvec);
> > +     if (hwirq < 0)
> > +             return -ENOSPC;
> > +
> > +     irq = irq_find_mapping(msi->domain, hwirq);
> > +     if (!irq)
> > +             return -ENOSPC;
> > +
> > +     for (i = 0; i < nvec; i++) {
> > +             /*
> > +              * irq_create_mapping() called from rcar_pcie_probe() pre-
> > +              * allocates descs,  so there is no need to allocate descs here.
> > +              * We can therefore assume that if irq_find_mapping() above
> > +              * returns non-zero, then the descs are also successfully
> > +              * allocated.
> > +              */
> > +             if (irq_set_msi_desc_off(irq, i, desc)) {
> > +                     /* TODO: clear */
> > +                     return -EINVAL;
> > +             }
> > +     }
> > +
> > +     desc->nvec_used = nvec;
> > +     desc->msi_attrib.multiple = order_base_2(nvec);
> > +
> > +     msg.address_lo = rcar_pci_read_reg(pcie->base, PCIEMSIALR) & ~MSIFE;
> > +     msg.address_hi = rcar_pci_read_reg(pcie->base, PCIEMSIAUR);
> > +     msg.data = hwirq;
> > +
> > +     pci_write_msi_msg(irq, &msg);
> > +
> > +     return 0;
> > +}
> > +
> > +static void rcar_msi_teardown_irq(struct msi_controller *chip, unsigned int irq)
> > +{
> > +     struct irq_data *d = irq_get_irq_data(irq);
> > +     struct rcar_msi *msi = to_rcar_msi(chip);
> > +
> > +     rcar_msi_free(msi, d->hwirq);
> > +}
> > +
> > +static struct irq_chip rcar_msi_irq_chip = {
> > +     .name = "R-Car PCIe MSI",
> > +     .irq_enable = pci_msi_unmask_irq,
> > +     .irq_disable = pci_msi_mask_irq,
> > +     .irq_mask = pci_msi_mask_irq,
> > +     .irq_unmask = pci_msi_unmask_irq,
> > +};
> > +
> > +static int rcar_msi_map(struct irq_domain *domain, unsigned int irq,
> > +                     irq_hw_number_t hwirq)
> > +{
> > +     irq_set_chip_and_handler(irq, &rcar_msi_irq_chip, handle_simple_irq);
> > +     irq_set_chip_data(irq, domain->host_data);
> > +
> > +     return 0;
> > +}
> > +
> > +static const struct irq_domain_ops msi_domain_ops = {
> > +     .map = rcar_msi_map,
> > +};
> > +
> > +static void rcar_pcie_unmap_msi(struct rcar_pcie *pcie)
> > +{
> > +     struct rcar_msi *msi = &pcie->msi;
> > +     int i, irq;
> > +
> > +     for (i = 0; i < INT_PCI_MSI_NR; i++) {
> > +             irq = irq_find_mapping(msi->domain, i);
> > +             if (irq > 0)
> > +                     irq_dispose_mapping(irq);
> > +     }
> > +
> > +     irq_domain_remove(msi->domain);
> > +}
> > +
> > +static int rcar_pcie_enable_msi(struct rcar_pcie *pcie)
> > +{
> > +     struct rcar_msi *msi = &pcie->msi;
> > +     struct device *dev = pcie->dev;
> > +     phys_addr_t base;
> > +     int err, i;
> > +
> > +     mutex_init(&msi->lock);
> > +
> > +     msi->chip.dev = dev;
> > +     msi->chip.setup_irq = rcar_msi_setup_irq;
> > +     msi->chip.setup_irqs = rcar_msi_setup_irqs;
> > +     msi->chip.teardown_irq = rcar_msi_teardown_irq;
> > +
> > +     msi->domain = irq_domain_add_linear(dev->of_node, INT_PCI_MSI_NR,
> > +                                         &msi_domain_ops, &msi->chip);
> > +     if (!msi->domain) {
> > +             dev_err(dev, "failed to create IRQ domain\n");
> > +             return -ENOMEM;
> > +     }
> > +
> > +     for (i = 0; i < INT_PCI_MSI_NR; i++)
> > +             irq_create_mapping(msi->domain, i);
> > +
> > +     /* Two irqs are for MSI, but they are also used for non-MSI irqs */
> > +     err = devm_request_irq(dev, msi->irq1, rcar_pcie_msi_irq,
> > +                            IRQF_SHARED | IRQF_NO_THREAD,
> > +                            rcar_msi_irq_chip.name, pcie);
> > +     if (err < 0) {
> > +             dev_err(dev, "failed to request IRQ: %d\n", err);
> > +             goto err;
> > +     }
> > +
> > +     err = devm_request_irq(dev, msi->irq2, rcar_pcie_msi_irq,
> > +                            IRQF_SHARED | IRQF_NO_THREAD,
> > +                            rcar_msi_irq_chip.name, pcie);
> > +     if (err < 0) {
> > +             dev_err(dev, "failed to request IRQ: %d\n", err);
> > +             goto err;
> > +     }
> > +
> > +     /* setup MSI data target */
> > +     msi->pages = __get_free_pages(GFP_KERNEL, 0);
> > +     if (!msi->pages) {
> > +             err = -ENOMEM;
> > +             goto err;
> > +     }
> > +     base = virt_to_phys((void *)msi->pages);
> > +
> > +     rcar_pci_write_reg(pcie->base, lower_32_bits(base) | MSIFE, PCIEMSIALR);
> > +     rcar_pci_write_reg(pcie->base, upper_32_bits(base), PCIEMSIAUR);
> > +
> > +     /* enable all MSI interrupts */
> > +     rcar_pci_write_reg(pcie->base, 0xffffffff, PCIEMSIIER);
> > +
> > +     return 0;
> > +
> > +err:
> > +     rcar_pcie_unmap_msi(pcie);
> > +     return err;
> > +}
> > +
> > +static void rcar_pcie_teardown_msi(struct rcar_pcie *pcie)
> > +{
> > +     struct rcar_msi *msi = &pcie->msi;
> > +
> > +     /* Disable all MSI interrupts */
> > +     rcar_pci_write_reg(pcie->base, 0, PCIEMSIIER);
> > +
> > +     /* Disable address decoding of the MSI interrupt, MSIFE */
> > +     rcar_pci_write_reg(pcie->base, 0, PCIEMSIALR);
> > +
> > +     free_pages(msi->pages, 0);
> > +
> > +     rcar_pcie_unmap_msi(pcie);
> > +}
> > +
> > +static int rcar_pcie_get_resources(struct rcar_pcie *pcie)
> > +{
> > +     struct device *dev = pcie->dev;
> > +     struct resource res;
> > +     int err, i;
> > +
> > +     pcie->phy = devm_phy_optional_get(dev, "pcie");
> > +     if (IS_ERR(pcie->phy))
> > +             return PTR_ERR(pcie->phy);
> > +
> > +     err = of_address_to_resource(dev->of_node, 0, &res);
> > +     if (err)
> > +             return err;
> > +
> > +     pcie->base = devm_ioremap_resource(dev, &res);
> > +     if (IS_ERR(pcie->base))
> > +             return PTR_ERR(pcie->base);
> > +
> > +     pcie->bus_clk = devm_clk_get(dev, "pcie_bus");
> > +     if (IS_ERR(pcie->bus_clk)) {
> > +             dev_err(dev, "cannot get pcie bus clock\n");
> > +             return PTR_ERR(pcie->bus_clk);
> > +     }
> > +
> > +     i = irq_of_parse_and_map(dev->of_node, 0);
> > +     if (!i) {
> > +             dev_err(dev, "cannot get platform resources for msi interrupt\n");
> > +             err = -ENOENT;
> > +             goto err_irq1;
> > +     }
> > +     pcie->msi.irq1 = i;
> > +
> > +     i = irq_of_parse_and_map(dev->of_node, 1);
> > +     if (!i) {
> > +             dev_err(dev, "cannot get platform resources for msi interrupt\n");
> > +             err = -ENOENT;
> > +             goto err_irq2;
> > +     }
> > +     pcie->msi.irq2 = i;
> > +
> > +     return 0;
> > +
> > +err_irq2:
> > +     irq_dispose_mapping(pcie->msi.irq1);
> > +err_irq1:
> > +     return err;
> > +}
> > +
> > +static int rcar_pcie_inbound_ranges(struct rcar_pcie *pcie,
> > +                                 struct of_pci_range *range,
> > +                                 int *index)
> > +{
> > +     u64 restype = range->flags;
> > +     u64 cpu_addr = range->cpu_addr;
> > +     u64 cpu_end = range->cpu_addr + range->size;
> > +     u64 pci_addr = range->pci_addr;
> > +     u32 flags = LAM_64BIT | LAR_ENABLE;
> > +     u64 mask;
> > +     u64 size;
> > +     int idx = *index;
> > +
> > +     if (restype & IORESOURCE_PREFETCH)
> > +             flags |= LAM_PREFETCH;
> > +
> > +     /*
> > +      * If the size of the range is larger than the alignment of the start
> > +      * address, we have to use multiple entries to perform the mapping.
> > +      */
> > +     if (cpu_addr > 0) {
> > +             unsigned long nr_zeros = __ffs64(cpu_addr);
> > +             u64 alignment = 1ULL << nr_zeros;
> > +
> > +             size = min(range->size, alignment);
> > +     } else {
> > +             size = range->size;
> > +     }
> > +     /* Hardware supports max 4GiB inbound region */
> > +     size = min(size, 1ULL << 32);
> > +
> > +     mask = roundup_pow_of_two(size) - 1;
> > +     mask &= ~0xf;
> > +
> > +     while (cpu_addr < cpu_end) {
> > +             rcar_pcie_set_inbound(pcie->base, cpu_addr, pci_addr,
> > +                                   mask | flags, idx, true);
> > +             pci_addr += size;
> > +             cpu_addr += size;
> > +             idx += 2;
> > +
> > +             if (idx > MAX_NR_INBOUND_MAPS) {
> > +                     dev_err(pcie->dev, "Failed to map inbound regions!\n");
> > +                     return -EINVAL;
> > +             }
> > +     }
> > +     *index = idx;
> > +
> > +     return 0;
> > +}
> > +
> > +static int rcar_pcie_parse_map_dma_ranges(struct rcar_pcie *pcie,
> > +                                       struct device_node *np)
> > +{
> > +     struct of_pci_range_parser parser;
> > +     struct of_pci_range range;
> > +     int index = 0;
> > +     int err;
> > +
> > +     if (of_pci_dma_range_parser_init(&parser, np))
> > +             return -EINVAL;
> > +
> > +     /* Get the dma-ranges from DT */
> > +     for_each_of_pci_range(&parser, &range) {
> > +             u64 end = range.cpu_addr + range.size - 1;
> > +
> > +             dev_dbg(pcie->dev, "0x%08x 0x%016llx..0x%016llx -> 0x%016llx\n",
> > +                     range.flags, range.cpu_addr, end, range.pci_addr);
> > +
> > +             err = rcar_pcie_inbound_ranges(pcie, &range, &index);
> > +             if (err)
> > +                     return err;
> > +     }
> > +
> > +     return 0;
> > +}
> > +
> > +static const struct of_device_id rcar_pcie_of_match[] = {
> > +     { .compatible = "renesas,pcie-r8a7779",
> > +       .data = rcar_pcie_phy_init_h1 },
> > +     { .compatible = "renesas,pcie-r8a7790",
> > +       .data = rcar_pcie_phy_init_gen2 },
> > +     { .compatible = "renesas,pcie-r8a7791",
> > +       .data = rcar_pcie_phy_init_gen2 },
> > +     { .compatible = "renesas,pcie-rcar-gen2",
> > +       .data = rcar_pcie_phy_init_gen2 },
> > +     { .compatible = "renesas,pcie-r8a7795",
> > +       .data = rcar_pcie_phy_init_gen3 },
> > +     { .compatible = "renesas,pcie-rcar-gen3",
> > +       .data = rcar_pcie_phy_init_gen3 },
> > +     {},
> > +};
> > +
> > +static int rcar_pcie_probe(struct platform_device *pdev)
> > +{
> > +     int (*phy_init_fn)(struct rcar_pcie *pcie);
> > +     struct device *dev = &pdev->dev;
> > +     struct pci_host_bridge *bridge;
> > +     struct rcar_pcie *pcie;
> > +     u32 data;
> > +     int err;
> > +
> > +     bridge = pci_alloc_host_bridge(sizeof(*pcie));
> > +     if (!bridge)
> > +             return -ENOMEM;
> > +
> > +     pcie = pci_host_bridge_priv(bridge);
> > +
> > +     pcie->dev = dev;
> > +     platform_set_drvdata(pdev, pcie);
> > +
> > +     err = pci_parse_request_of_pci_ranges(dev, &pcie->resources, NULL);
> > +     if (err)
> > +             goto err_free_bridge;
> > +
> > +     pm_runtime_enable(pcie->dev);
> > +     err = pm_runtime_get_sync(pcie->dev);
> > +     if (err < 0) {
> > +             dev_err(pcie->dev, "pm_runtime_get_sync failed\n");
> > +             goto err_pm_disable;
> > +     }
> > +
> > +     err = rcar_pcie_get_resources(pcie);
> > +     if (err < 0) {
> > +             dev_err(dev, "failed to request resources: %d\n", err);
> > +             goto err_pm_put;
> > +     }
> > +
> > +     err = clk_prepare_enable(pcie->bus_clk);
> > +     if (err) {
> > +             dev_err(dev, "failed to enable bus clock: %d\n", err);
> > +             goto err_unmap_msi_irqs;
> > +     }
> > +
> > +     err = rcar_pcie_parse_map_dma_ranges(pcie, dev->of_node);
> > +     if (err)
> > +             goto err_clk_disable;
> > +
> > +     phy_init_fn = of_device_get_match_data(dev);
> > +     err = phy_init_fn(pcie);
> > +     if (err) {
> > +             dev_err(dev, "failed to init PCIe PHY\n");
> > +             goto err_clk_disable;
> > +     }
> > +
> > +     /* Failure to get a link might just be that no cards are inserted */
> > +     if (rcar_pcie_hw_init(pcie)) {
> > +             dev_info(dev, "PCIe link down\n");
> > +             err = -ENODEV;
> > +             goto err_phy_shutdown;
> > +     }
> > +
> > +     data = rcar_pci_read_reg(pcie->base, MACSR);
> > +     dev_info(dev, "PCIe x%d: link up\n", (data >> 20) & 0x3f);
> > +
> > +     if (IS_ENABLED(CONFIG_PCI_MSI)) {
> > +             err = rcar_pcie_enable_msi(pcie);
> > +             if (err < 0) {
> > +                     dev_err(dev,
> > +                             "failed to enable MSI support: %d\n",
> > +                             err);
> > +                     goto err_phy_shutdown;
> > +             }
> > +     }
> > +
> > +     err = rcar_pcie_enable(pcie);
> > +     if (err)
> > +             goto err_msi_teardown;
> > +
> > +     return 0;
> > +
> > +err_msi_teardown:
> > +     if (IS_ENABLED(CONFIG_PCI_MSI))
> > +             rcar_pcie_teardown_msi(pcie);
> > +
> > +err_phy_shutdown:
> > +     if (pcie->phy) {
> > +             phy_power_off(pcie->phy);
> > +             phy_exit(pcie->phy);
> > +     }
> > +
> > +err_clk_disable:
> > +     clk_disable_unprepare(pcie->bus_clk);
> > +
> > +err_unmap_msi_irqs:
> > +     irq_dispose_mapping(pcie->msi.irq2);
> > +     irq_dispose_mapping(pcie->msi.irq1);
> > +
> > +err_pm_put:
> > +     pm_runtime_put(dev);
> > +
> > +err_pm_disable:
> > +     pm_runtime_disable(dev);
> > +     pci_free_resource_list(&pcie->resources);
> > +
> > +err_free_bridge:
> > +     pci_free_host_bridge(bridge);
> > +
> > +     return err;
> > +}
> > +
> > +static int rcar_pcie_resume_noirq(struct device *dev)
> > +{
> > +     struct rcar_pcie *pcie = dev_get_drvdata(dev);
> > +
> > +     if (rcar_pci_read_reg(pcie->base, PMSR) &&
> > +         !(rcar_pci_read_reg(pcie->base, PCIETCTLR) & DL_DOWN))
> > +             return 0;
> > +
> > +     /* Re-establish the PCIe link */
> > +     rcar_rmw32(pcie->base, MACCTLR, MACCTLR_RESERVED, 0);
> > +     rcar_pci_write_reg(pcie->base, CFINIT, PCIETCTLR);
> > +     return rcar_pcie_wait_for_dl(pcie->base);
> > +}
> > +
> > +static const struct dev_pm_ops rcar_pcie_pm_ops = {
> > +     .resume_noirq = rcar_pcie_resume_noirq,
> > +};
> > +
> > +static struct platform_driver rcar_pcie_driver = {
> > +     .driver = {
> > +             .name = "rcar-pcie",
> > +             .of_match_table = rcar_pcie_of_match,
> > +             .pm = &rcar_pcie_pm_ops,
> > +             .suppress_bind_attrs = true,
> > +     },
> > +     .probe = rcar_pcie_probe,
> > +};
> > +builtin_platform_driver(rcar_pcie_driver);
> > diff --git a/drivers/pci/controller/pcie-rcar.c b/drivers/pci/controller/pcie-rcar.c
> > index e45bb2a7bfa5..1008ae6a896c 100644
> > --- a/drivers/pci/controller/pcie-rcar.c
> > +++ b/drivers/pci/controller/pcie-rcar.c
> > @@ -1,7 +1,7 @@
> >  // SPDX-License-Identifier: GPL-2.0
> >  /*
> >   * PCIe driver for Renesas R-Car SoCs
> > - *  Copyright (C) 2014 Renesas Electronics Europe Ltd
> > + *  Copyright (C) 2014-2019 Renesas Electronics Europe Ltd
> >   *
> >   * Based on:
> >   *  arch/sh/drivers/pci/pcie-sh7786.c
> > @@ -11,534 +11,80 @@
> >   * Author: Phil Edworthy <phil.edworthy@renesas.com>
> >   */
> >
> > -#include <linux/bitops.h>
> > -#include <linux/clk.h>
> >  #include <linux/delay.h>
> > -#include <linux/interrupt.h>
> > -#include <linux/irq.h>
> > -#include <linux/irqdomain.h>
> > -#include <linux/kernel.h>
> > -#include <linux/init.h>
> > -#include <linux/msi.h>
> > -#include <linux/of_address.h>
> > -#include <linux/of_irq.h>
> > -#include <linux/of_pci.h>
> > -#include <linux/of_platform.h>
> >  #include <linux/pci.h>
> > -#include <linux/phy/phy.h>
> > -#include <linux/platform_device.h>
> > -#include <linux/pm_runtime.h>
> > -#include <linux/slab.h>
> >
> > -#define PCIECAR                      0x000010
> > -#define PCIECCTLR            0x000018
> > -#define  CONFIG_SEND_ENABLE  BIT(31)
> > -#define  TYPE0                       (0 << 8)
> > -#define  TYPE1                       BIT(8)
> > -#define PCIECDR                      0x000020
> > -#define PCIEMSR                      0x000028
> > -#define PCIEINTXR            0x000400
> > -#define PCIEPHYSR            0x0007f0
> > -#define  PHYRDY                      BIT(0)
> > -#define PCIEMSITXR           0x000840
> > +#include "pcie-rcar.h"
> >
> > -/* Transfer control */
> > -#define PCIETCTLR            0x02000
> > -#define  DL_DOWN             BIT(3)
> > -#define  CFINIT                      BIT(0)
> > -#define PCIETSTR             0x02004
> > -#define  DATA_LINK_ACTIVE    BIT(0)
> > -#define PCIEERRFR            0x02020
> > -#define  UNSUPPORTED_REQUEST BIT(4)
> > -#define PCIEMSIFR            0x02044
> > -#define PCIEMSIALR           0x02048
> > -#define  MSIFE                       BIT(0)
> > -#define PCIEMSIAUR           0x0204c
> > -#define PCIEMSIIER           0x02050
> > -
> > -/* root port address */
> > -#define PCIEPRAR(x)          (0x02080 + ((x) * 0x4))
> > -
> > -/* local address reg & mask */
> > -#define PCIELAR(x)           (0x02200 + ((x) * 0x20))
> > -#define PCIELAMR(x)          (0x02208 + ((x) * 0x20))
> > -#define  LAM_PREFETCH                BIT(3)
> > -#define  LAM_64BIT           BIT(2)
> > -#define  LAR_ENABLE          BIT(1)
> > -
> > -/* PCIe address reg & mask */
> > -#define PCIEPALR(x)          (0x03400 + ((x) * 0x20))
> > -#define PCIEPAUR(x)          (0x03404 + ((x) * 0x20))
> > -#define PCIEPAMR(x)          (0x03408 + ((x) * 0x20))
> > -#define PCIEPTCTLR(x)                (0x0340c + ((x) * 0x20))
> > -#define  PAR_ENABLE          BIT(31)
> > -#define  IO_SPACE            BIT(8)
> > -
> > -/* Configuration */
> > -#define PCICONF(x)           (0x010000 + ((x) * 0x4))
> > -#define PMCAP(x)             (0x010040 + ((x) * 0x4))
> > -#define EXPCAP(x)            (0x010070 + ((x) * 0x4))
> > -#define VCCAP(x)             (0x010100 + ((x) * 0x4))
> > -
> > -/* link layer */
> > -#define IDSETR1                      0x011004
> > -#define TLCTLR                       0x011048
> > -#define MACSR                        0x011054
> > -#define  SPCHGFIN            BIT(4)
> > -#define  SPCHGFAIL           BIT(6)
> > -#define  SPCHGSUC            BIT(7)
> > -#define  LINK_SPEED          (0xf << 16)
> > -#define  LINK_SPEED_2_5GTS   (1 << 16)
> > -#define  LINK_SPEED_5_0GTS   (2 << 16)
> > -#define MACCTLR                      0x011058
> > -#define  MACCTLR_RESERVED    BIT(0)
> > -#define  SPEED_CHANGE                BIT(24)
> > -#define  SCRAMBLE_DISABLE    BIT(27)
> > -#define PMSR                 0x01105c
> > -#define MACS2R                       0x011078
> > -#define MACCGSPSETR          0x011084
> > -#define  SPCNGRSN            BIT(31)
> > -
> > -/* R-Car H1 PHY */
> > -#define H1_PCIEPHYADRR               0x04000c
> > -#define  WRITE_CMD           BIT(16)
> > -#define  PHY_ACK             BIT(24)
> > -#define  RATE_POS            12
> > -#define  LANE_POS            8
> > -#define  ADR_POS             0
> > -#define H1_PCIEPHYDOUTR              0x040014
> > -
> > -/* R-Car Gen2 PHY */
> > -#define GEN2_PCIEPHYADDR     0x780
> > -#define GEN2_PCIEPHYDATA     0x784
> > -#define GEN2_PCIEPHYCTRL     0x78c
> > -
> > -#define INT_PCI_MSI_NR               32
> > -
> > -#define RCONF(x)             (PCICONF(0) + (x))
> > -#define RPMCAP(x)            (PMCAP(0) + (x))
> > -#define REXPCAP(x)           (EXPCAP(0) + (x))
> > -#define RVCCAP(x)            (VCCAP(0) + (x))
> > -
> > -#define PCIE_CONF_BUS(b)     (((b) & 0xff) << 24)
> > -#define PCIE_CONF_DEV(d)     (((d) & 0x1f) << 19)
> > -#define PCIE_CONF_FUNC(f)    (((f) & 0x7) << 16)
> > -
> > -#define RCAR_PCI_MAX_RESOURCES       4
> > -#define MAX_NR_INBOUND_MAPS  6
> > -
> > -struct rcar_msi {
> > -     DECLARE_BITMAP(used, INT_PCI_MSI_NR);
> > -     struct irq_domain *domain;
> > -     struct msi_controller chip;
> > -     unsigned long pages;
> > -     struct mutex lock;
> > -     int irq1;
> > -     int irq2;
> > -};
> > -
> > -static inline struct rcar_msi *to_rcar_msi(struct msi_controller *chip)
> > -{
> > -     return container_of(chip, struct rcar_msi, chip);
> > -}
> > -
> > -/* Structure representing the PCIe interface */
> > -struct rcar_pcie {
> > -     struct device           *dev;
> > -     struct phy              *phy;
> > -     void __iomem            *base;
> > -     struct list_head        resources;
> > -     int                     root_bus_nr;
> > -     struct clk              *bus_clk;
> > -     struct                  rcar_msi msi;
> > -};
> > -
> > -static void rcar_pci_write_reg(struct rcar_pcie *pcie, u32 val,
> > -                            unsigned int reg)
> > +void rcar_pci_write_reg(void __iomem *base, u32 val, unsigned int reg)
> >  {
> > -     writel(val, pcie->base + reg);
> > +     writel(val, base + reg);
> >  }
> >
> > -static u32 rcar_pci_read_reg(struct rcar_pcie *pcie, unsigned int reg)
> > +u32 rcar_pci_read_reg(void __iomem *base, unsigned int reg)
> >  {
> > -     return readl(pcie->base + reg);
> > +     return readl(base + reg);
> >  }
> >
> > -enum {
> > -     RCAR_PCI_ACCESS_READ,
> > -     RCAR_PCI_ACCESS_WRITE,
> > -};
> > -
> > -static void rcar_rmw32(struct rcar_pcie *pcie, int where, u32 mask, u32 data)
> > +void rcar_rmw32(void __iomem *base, int where, u32 mask, u32 data)
> >  {
> >       unsigned int shift = BITS_PER_BYTE * (where & 3);
> > -     u32 val = rcar_pci_read_reg(pcie, where & ~3);
> > +     u32 val = rcar_pci_read_reg(base, where & ~3);
> >
> >       val &= ~(mask << shift);
> >       val |= data << shift;
> > -     rcar_pci_write_reg(pcie, val, where & ~3);
> > -}
> > -
> > -static u32 rcar_read_conf(struct rcar_pcie *pcie, int where)
> > -{
> > -     unsigned int shift = BITS_PER_BYTE * (where & 3);
> > -     u32 val = rcar_pci_read_reg(pcie, where & ~3);
> > -
> > -     return val >> shift;
> > +     rcar_pci_write_reg(base, val, where & ~3);
> >  }
> >
> > -/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
> > -static int rcar_pcie_config_access(struct rcar_pcie *pcie,
> > -             unsigned char access_type, struct pci_bus *bus,
> > -             unsigned int devfn, int where, u32 *data)
> > -{
> > -     unsigned int dev, func, reg, index;
> > -
> > -     dev = PCI_SLOT(devfn);
> > -     func = PCI_FUNC(devfn);
> > -     reg = where & ~3;
> > -     index = reg / 4;
> > -
> > -     /*
> > -      * While each channel has its own memory-mapped extended config
> > -      * space, it's generally only accessible when in endpoint mode.
> > -      * When in root complex mode, the controller is unable to target
> > -      * itself with either type 0 or type 1 accesses, and indeed, any
> > -      * controller initiated target transfer to its own config space
> > -      * result in a completer abort.
> > -      *
> > -      * Each channel effectively only supports a single device, but as
> > -      * the same channel <-> device access works for any PCI_SLOT()
> > -      * value, we cheat a bit here and bind the controller's config
> > -      * space to devfn 0 in order to enable self-enumeration. In this
> > -      * case the regular ECAR/ECDR path is sidelined and the mangled
> > -      * config access itself is initiated as an internal bus transaction.
> > -      */
> > -     if (pci_is_root_bus(bus)) {
> > -             if (dev != 0)
> > -                     return PCIBIOS_DEVICE_NOT_FOUND;
> > -
> > -             if (access_type == RCAR_PCI_ACCESS_READ) {
> > -                     *data = rcar_pci_read_reg(pcie, PCICONF(index));
> > -             } else {
> > -                     /* Keep an eye out for changes to the root bus number */
> > -                     if (pci_is_root_bus(bus) && (reg == PCI_PRIMARY_BUS))
> > -                             pcie->root_bus_nr = *data & 0xff;
> > -
> > -                     rcar_pci_write_reg(pcie, *data, PCICONF(index));
> > -             }
> > -
> > -             return PCIBIOS_SUCCESSFUL;
> > -     }
> > -
> > -     if (pcie->root_bus_nr < 0)
> > -             return PCIBIOS_DEVICE_NOT_FOUND;
> > -
> > -     /* Clear errors */
> > -     rcar_pci_write_reg(pcie, rcar_pci_read_reg(pcie, PCIEERRFR), PCIEERRFR);
> > -
> > -     /* Set the PIO address */
> > -     rcar_pci_write_reg(pcie, PCIE_CONF_BUS(bus->number) |
> > -             PCIE_CONF_DEV(dev) | PCIE_CONF_FUNC(func) | reg, PCIECAR);
> > -
> > -     /* Enable the configuration access */
> > -     if (bus->parent->number == pcie->root_bus_nr)
> > -             rcar_pci_write_reg(pcie, CONFIG_SEND_ENABLE | TYPE0, PCIECCTLR);
> > -     else
> > -             rcar_pci_write_reg(pcie, CONFIG_SEND_ENABLE | TYPE1, PCIECCTLR);
> > -
> > -     /* Check for errors */
> > -     if (rcar_pci_read_reg(pcie, PCIEERRFR) & UNSUPPORTED_REQUEST)
> > -             return PCIBIOS_DEVICE_NOT_FOUND;
> > -
> > -     /* Check for master and target aborts */
> > -     if (rcar_read_conf(pcie, RCONF(PCI_STATUS)) &
> > -             (PCI_STATUS_REC_MASTER_ABORT | PCI_STATUS_REC_TARGET_ABORT))
> > -             return PCIBIOS_DEVICE_NOT_FOUND;
> > -
> > -     if (access_type == RCAR_PCI_ACCESS_READ)
> > -             *data = rcar_pci_read_reg(pcie, PCIECDR);
> > -     else
> > -             rcar_pci_write_reg(pcie, *data, PCIECDR);
> > -
> > -     /* Disable the configuration access */
> > -     rcar_pci_write_reg(pcie, 0, PCIECCTLR);
> > -
> > -     return PCIBIOS_SUCCESSFUL;
> > -}
> > -
> > -static int rcar_pcie_read_conf(struct pci_bus *bus, unsigned int devfn,
> > -                            int where, int size, u32 *val)
> > -{
> > -     struct rcar_pcie *pcie = bus->sysdata;
> > -     int ret;
> > -
> > -     ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
> > -                                   bus, devfn, where, val);
> > -     if (ret != PCIBIOS_SUCCESSFUL) {
> > -             *val = 0xffffffff;
> > -             return ret;
> > -     }
> > -
> > -     if (size == 1)
> > -             *val = (*val >> (BITS_PER_BYTE * (where & 3))) & 0xff;
> > -     else if (size == 2)
> > -             *val = (*val >> (BITS_PER_BYTE * (where & 2))) & 0xffff;
> > -
> > -     dev_dbg(&bus->dev, "pcie-config-read: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
> > -             bus->number, devfn, where, size, *val);
> > -
> > -     return ret;
> > -}
> > -
> > -/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
> > -static int rcar_pcie_write_conf(struct pci_bus *bus, unsigned int devfn,
> > -                             int where, int size, u32 val)
> > -{
> > -     struct rcar_pcie *pcie = bus->sysdata;
> > -     unsigned int shift;
> > -     u32 data;
> > -     int ret;
> > -
> > -     ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
> > -                                   bus, devfn, where, &data);
> > -     if (ret != PCIBIOS_SUCCESSFUL)
> > -             return ret;
> > -
> > -     dev_dbg(&bus->dev, "pcie-config-write: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
> > -             bus->number, devfn, where, size, val);
> > -
> > -     if (size == 1) {
> > -             shift = BITS_PER_BYTE * (where & 3);
> > -             data &= ~(0xff << shift);
> > -             data |= ((val & 0xff) << shift);
> > -     } else if (size == 2) {
> > -             shift = BITS_PER_BYTE * (where & 2);
> > -             data &= ~(0xffff << shift);
> > -             data |= ((val & 0xffff) << shift);
> > -     } else
> > -             data = val;
> > -
> > -     ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_WRITE,
> > -                                   bus, devfn, where, &data);
> > -
> > -     return ret;
> > -}
> > -
> > -static struct pci_ops rcar_pcie_ops = {
> > -     .read   = rcar_pcie_read_conf,
> > -     .write  = rcar_pcie_write_conf,
> > -};
> > -
> > -static void rcar_pcie_setup_window(int win, struct rcar_pcie *pcie,
> > -                                struct resource *res)
> > +void rcar_pcie_set_outbound(int win, void __iomem *base,
> > +                         struct resource *res, bool host)
> >  {
> >       /* Setup PCIe address space mappings for each resource */
> > -     resource_size_t size;
> > -     resource_size_t res_start;
> > -     u32 mask;
> > -
> > -     rcar_pci_write_reg(pcie, 0x00000000, PCIEPTCTLR(win));
> > +     resource_size_t size = 0;
> > +     resource_size_t res_start = 0;
> > +     u32 mask = 0x0;
> >
> > +     rcar_pci_write_reg(base, mask, PCIEPTCTLR(win));
> >       /*
> >        * The PAMR mask is calculated in units of 128Bytes, which
> >        * keeps things pretty simple.
> >        */
> >       size = resource_size(res);
> > -     mask = (roundup_pow_of_two(size) / SZ_128) - 1;
> > -     rcar_pci_write_reg(pcie, mask << 7, PCIEPAMR(win));
> > +     if (size > 128)
> > +             mask = (roundup_pow_of_two(size) / SZ_128) - 1;
> > +     rcar_pci_write_reg(base, mask << 7, PCIEPAMR(win));
> >
> > -     if (res->flags & IORESOURCE_IO)
> > -             res_start = pci_pio_to_address(res->start);
> > -     else
> > +     if (!host) {
> >               res_start = res->start;
> > +     } else {
> > +             if (res->flags & IORESOURCE_IO)
> > +                     res_start = pci_pio_to_address(res->start);
> > +             else
> > +                     res_start = res->start;
> > +     }
> >
> > -     rcar_pci_write_reg(pcie, upper_32_bits(res_start), PCIEPAUR(win));
> > -     rcar_pci_write_reg(pcie, lower_32_bits(res_start) & ~0x7F,
> > +     rcar_pci_write_reg(base, upper_32_bits(res_start), PCIEPAUR(win));
> > +     rcar_pci_write_reg(base, lower_32_bits(res_start) & ~0x7F,
> >                          PCIEPALR(win));
> >
> > -     /* First resource is for IO */
> > -     mask = PAR_ENABLE;
> > +     mask = 0x0;
> > +     if (res->start)
> > +             mask = PAR_ENABLE;
> > +
> >       if (res->flags & IORESOURCE_IO)
> >               mask |= IO_SPACE;
> >
> > -     rcar_pci_write_reg(pcie, mask, PCIEPTCTLR(win));
> > -}
> > -
> > -static int rcar_pcie_setup(struct list_head *resource, struct rcar_pcie *pci)
> > -{
> > -     struct resource_entry *win;
> > -     int i = 0;
> > -
> > -     /* Setup PCI resources */
> > -     resource_list_for_each_entry(win, &pci->resources) {
> > -             struct resource *res = win->res;
> > -
> > -             if (!res->flags)
> > -                     continue;
> > -
> > -             switch (resource_type(res)) {
> > -             case IORESOURCE_IO:
> > -             case IORESOURCE_MEM:
> > -                     rcar_pcie_setup_window(i, pci, res);
> > -                     i++;
> > -                     break;
> > -             case IORESOURCE_BUS:
> > -                     pci->root_bus_nr = res->start;
> > -                     break;
> > -             default:
> > -                     continue;
> > -             }
> > -
> > -             pci_add_resource(resource, res);
> > -     }
> > -
> > -     return 1;
> > -}
> > -
> > -static void rcar_pcie_force_speedup(struct rcar_pcie *pcie)
> > -{
> > -     struct device *dev = pcie->dev;
> > -     unsigned int timeout = 1000;
> > -     u32 macsr;
> > -
> > -     if ((rcar_pci_read_reg(pcie, MACS2R) & LINK_SPEED) != LINK_SPEED_5_0GTS)
> > -             return;
> > -
> > -     if (rcar_pci_read_reg(pcie, MACCTLR) & SPEED_CHANGE) {
> > -             dev_err(dev, "Speed change already in progress\n");
> > -             return;
> > -     }
> > -
> > -     macsr = rcar_pci_read_reg(pcie, MACSR);
> > -     if ((macsr & LINK_SPEED) == LINK_SPEED_5_0GTS)
> > -             goto done;
> > -
> > -     /* Set target link speed to 5.0 GT/s */
> > -     rcar_rmw32(pcie, EXPCAP(12), PCI_EXP_LNKSTA_CLS,
> > -                PCI_EXP_LNKSTA_CLS_5_0GB);
> > -
> > -     /* Set speed change reason as intentional factor */
> > -     rcar_rmw32(pcie, MACCGSPSETR, SPCNGRSN, 0);
> > -
> > -     /* Clear SPCHGFIN, SPCHGSUC, and SPCHGFAIL */
> > -     if (macsr & (SPCHGFIN | SPCHGSUC | SPCHGFAIL))
> > -             rcar_pci_write_reg(pcie, macsr, MACSR);
> > -
> > -     /* Start link speed change */
> > -     rcar_rmw32(pcie, MACCTLR, SPEED_CHANGE, SPEED_CHANGE);
> > -
> > -     while (timeout--) {
> > -             macsr = rcar_pci_read_reg(pcie, MACSR);
> > -             if (macsr & SPCHGFIN) {
> > -                     /* Clear the interrupt bits */
> > -                     rcar_pci_write_reg(pcie, macsr, MACSR);
> > -
> > -                     if (macsr & SPCHGFAIL)
> > -                             dev_err(dev, "Speed change failed\n");
> > -
> > -                     goto done;
> > -             }
> > -
> > -             msleep(1);
> > -     }
> > -
> > -     dev_err(dev, "Speed change timed out\n");
> > -
> > -done:
> > -     dev_info(dev, "Current link speed is %s GT/s\n",
> > -              (macsr & LINK_SPEED) == LINK_SPEED_5_0GTS ? "5" : "2.5");
> > -}
> > -
> > -static int rcar_pcie_enable(struct rcar_pcie *pcie)
> > -{
> > -     struct device *dev = pcie->dev;
> > -     struct pci_host_bridge *bridge = pci_host_bridge_from_priv(pcie);
> > -     struct pci_bus *bus, *child;
> > -     int ret;
> > -
> > -     /* Try setting 5 GT/s link speed */
> > -     rcar_pcie_force_speedup(pcie);
> > -
> > -     rcar_pcie_setup(&bridge->windows, pcie);
> > -
> > -     pci_add_flags(PCI_REASSIGN_ALL_BUS);
> > -
> > -     bridge->dev.parent = dev;
> > -     bridge->sysdata = pcie;
> > -     bridge->busnr = pcie->root_bus_nr;
> > -     bridge->ops = &rcar_pcie_ops;
> > -     bridge->map_irq = of_irq_parse_and_map_pci;
> > -     bridge->swizzle_irq = pci_common_swizzle;
> > -     if (IS_ENABLED(CONFIG_PCI_MSI))
> > -             bridge->msi = &pcie->msi.chip;
> > -
> > -     ret = pci_scan_root_bus_bridge(bridge);
> > -     if (ret < 0)
> > -             return ret;
> > -
> > -     bus = bridge->bus;
> > -
> > -     pci_bus_size_bridges(bus);
> > -     pci_bus_assign_resources(bus);
> > -
> > -     list_for_each_entry(child, &bus->children, node)
> > -             pcie_bus_configure_settings(child);
> > -
> > -     pci_bus_add_devices(bus);
> > -
> > -     return 0;
> > -}
> > -
> > -static int phy_wait_for_ack(struct rcar_pcie *pcie)
> > -{
> > -     struct device *dev = pcie->dev;
> > -     unsigned int timeout = 100;
> > -
> > -     while (timeout--) {
> > -             if (rcar_pci_read_reg(pcie, H1_PCIEPHYADRR) & PHY_ACK)
> > -                     return 0;
> > -
> > -             udelay(100);
> > -     }
> > -
> > -     dev_err(dev, "Access to PCIe phy timed out\n");
> > -
> > -     return -ETIMEDOUT;
> > -}
> > -
> > -static void phy_write_reg(struct rcar_pcie *pcie,
> > -                       unsigned int rate, u32 addr,
> > -                       unsigned int lane, u32 data)
> > -{
> > -     u32 phyaddr;
> > -
> > -     phyaddr = WRITE_CMD |
> > -             ((rate & 1) << RATE_POS) |
> > -             ((lane & 0xf) << LANE_POS) |
> > -             ((addr & 0xff) << ADR_POS);
> > -
> > -     /* Set write data */
> > -     rcar_pci_write_reg(pcie, data, H1_PCIEPHYDOUTR);
> > -     rcar_pci_write_reg(pcie, phyaddr, H1_PCIEPHYADRR);
> > -
> > -     /* Ignore errors as they will be dealt with if the data link is down */
> > -     phy_wait_for_ack(pcie);
> > -
> > -     /* Clear command */
> > -     rcar_pci_write_reg(pcie, 0, H1_PCIEPHYDOUTR);
> > -     rcar_pci_write_reg(pcie, 0, H1_PCIEPHYADRR);
> > -
> > -     /* Ignore errors as they will be dealt with if the data link is down */
> > -     phy_wait_for_ack(pcie);
> > +     rcar_pci_write_reg(base, mask, PCIEPTCTLR(win));
> > +     /* flush modifications */
> > +     wmb();
> >  }
> >
> > -static int rcar_pcie_wait_for_phyrdy(struct rcar_pcie *pcie)
> > +int rcar_pcie_wait_for_phyrdy(void __iomem *base)
> >  {
> >       unsigned int timeout = 10;
> >
> >       while (timeout--) {
> > -             if (rcar_pci_read_reg(pcie, PCIEPHYSR) & PHYRDY)
> > +             if (rcar_pci_read_reg(base, PCIEPHYSR) & PHYRDY)
> >                       return 0;
> >
> >               msleep(5);
> > @@ -547,12 +93,12 @@ static int rcar_pcie_wait_for_phyrdy(struct rcar_pcie *pcie)
> >       return -ETIMEDOUT;
> >  }
> >
> > -static int rcar_pcie_wait_for_dl(struct rcar_pcie *pcie)
> > +int rcar_pcie_wait_for_dl(void __iomem *base)
> >  {
> >       unsigned int timeout = 10000;
> >
> >       while (timeout--) {
> > -             if ((rcar_pci_read_reg(pcie, PCIETSTR) & DATA_LINK_ACTIVE))
> > +             if ((rcar_pci_read_reg(base, PCIETSTR) & DATA_LINK_ACTIVE))
> >                       return 0;
> >
> >               udelay(5);
> > @@ -562,698 +108,27 @@ static int rcar_pcie_wait_for_dl(struct rcar_pcie *pcie)
> >       return -ETIMEDOUT;
> >  }
> >
> > -static int rcar_pcie_hw_init(struct rcar_pcie *pcie)
> > -{
> > -     int err;
> > -
> > -     /* Begin initialization */
> > -     rcar_pci_write_reg(pcie, 0, PCIETCTLR);
> > -
> > -     /* Set mode */
> > -     rcar_pci_write_reg(pcie, 1, PCIEMSR);
> > -
> > -     err = rcar_pcie_wait_for_phyrdy(pcie);
> > -     if (err)
> > -             return err;
> > -
> > -     /*
> > -      * Initial header for port config space is type 1, set the device
> > -      * class to match. Hardware takes care of propagating the IDSETR
> > -      * settings, so there is no need to bother with a quirk.
> > -      */
> > -     rcar_pci_write_reg(pcie, PCI_CLASS_BRIDGE_PCI << 16, IDSETR1);
> > -
> > -     /*
> > -      * Setup Secondary Bus Number & Subordinate Bus Number, even though
> > -      * they aren't used, to avoid bridge being detected as broken.
> > -      */
> > -     rcar_rmw32(pcie, RCONF(PCI_SECONDARY_BUS), 0xff, 1);
> > -     rcar_rmw32(pcie, RCONF(PCI_SUBORDINATE_BUS), 0xff, 1);
> > -
> > -     /* Initialize default capabilities. */
> > -     rcar_rmw32(pcie, REXPCAP(0), 0xff, PCI_CAP_ID_EXP);
> > -     rcar_rmw32(pcie, REXPCAP(PCI_EXP_FLAGS),
> > -             PCI_EXP_FLAGS_TYPE, PCI_EXP_TYPE_ROOT_PORT << 4);
> > -     rcar_rmw32(pcie, RCONF(PCI_HEADER_TYPE), 0x7f,
> > -             PCI_HEADER_TYPE_BRIDGE);
> > -
> > -     /* Enable data link layer active state reporting */
> > -     rcar_rmw32(pcie, REXPCAP(PCI_EXP_LNKCAP), PCI_EXP_LNKCAP_DLLLARC,
> > -             PCI_EXP_LNKCAP_DLLLARC);
> > -
> > -     /* Write out the physical slot number = 0 */
> > -     rcar_rmw32(pcie, REXPCAP(PCI_EXP_SLTCAP), PCI_EXP_SLTCAP_PSN, 0);
> > -
> > -     /* Set the completion timer timeout to the maximum 50ms. */
> > -     rcar_rmw32(pcie, TLCTLR + 1, 0x3f, 50);
> > -
> > -     /* Terminate list of capabilities (Next Capability Offset=0) */
> > -     rcar_rmw32(pcie, RVCCAP(0), 0xfff00000, 0);
> > -
> > -     /* Enable MSI */
> > -     if (IS_ENABLED(CONFIG_PCI_MSI))
> > -             rcar_pci_write_reg(pcie, 0x801f0000, PCIEMSITXR);
> > -
> > -     rcar_rmw32(pcie, MACCTLR, MACCTLR_RESERVED, 0);
> > -
> > -     /* Finish initialization - establish a PCI Express link */
> > -     rcar_pci_write_reg(pcie, CFINIT, PCIETCTLR);
> > -
> > -     /* This will timeout if we don't have a link. */
> > -     err = rcar_pcie_wait_for_dl(pcie);
> > -     if (err)
> > -             return err;
> > -
> > -     /* Enable INTx interrupts */
> > -     rcar_rmw32(pcie, PCIEINTXR, 0, 0xF << 8);
> > -
> > -     wmb();
> > -
> > -     return 0;
> > -}
> > -
> > -static int rcar_pcie_phy_init_h1(struct rcar_pcie *pcie)
> > -{
> > -     /* Initialize the phy */
> > -     phy_write_reg(pcie, 0, 0x42, 0x1, 0x0EC34191);
> > -     phy_write_reg(pcie, 1, 0x42, 0x1, 0x0EC34180);
> > -     phy_write_reg(pcie, 0, 0x43, 0x1, 0x00210188);
> > -     phy_write_reg(pcie, 1, 0x43, 0x1, 0x00210188);
> > -     phy_write_reg(pcie, 0, 0x44, 0x1, 0x015C0014);
> > -     phy_write_reg(pcie, 1, 0x44, 0x1, 0x015C0014);
> > -     phy_write_reg(pcie, 1, 0x4C, 0x1, 0x786174A0);
> > -     phy_write_reg(pcie, 1, 0x4D, 0x1, 0x048000BB);
> > -     phy_write_reg(pcie, 0, 0x51, 0x1, 0x079EC062);
> > -     phy_write_reg(pcie, 0, 0x52, 0x1, 0x20000000);
> > -     phy_write_reg(pcie, 1, 0x52, 0x1, 0x20000000);
> > -     phy_write_reg(pcie, 1, 0x56, 0x1, 0x00003806);
> > -
> > -     phy_write_reg(pcie, 0, 0x60, 0x1, 0x004B03A5);
> > -     phy_write_reg(pcie, 0, 0x64, 0x1, 0x3F0F1F0F);
> > -     phy_write_reg(pcie, 0, 0x66, 0x1, 0x00008000);
> > -
> > -     return 0;
> > -}
> > -
> > -static int rcar_pcie_phy_init_gen2(struct rcar_pcie *pcie)
> > -{
> > -     /*
> > -      * These settings come from the R-Car Series, 2nd Generation User's
> > -      * Manual, section 50.3.1 (2) Initialization of the physical layer.
> > -      */
> > -     rcar_pci_write_reg(pcie, 0x000f0030, GEN2_PCIEPHYADDR);
> > -     rcar_pci_write_reg(pcie, 0x00381203, GEN2_PCIEPHYDATA);
> > -     rcar_pci_write_reg(pcie, 0x00000001, GEN2_PCIEPHYCTRL);
> > -     rcar_pci_write_reg(pcie, 0x00000006, GEN2_PCIEPHYCTRL);
> > -
> > -     rcar_pci_write_reg(pcie, 0x000f0054, GEN2_PCIEPHYADDR);
> > -     /* The following value is for DC connection, no termination resistor */
> > -     rcar_pci_write_reg(pcie, 0x13802007, GEN2_PCIEPHYDATA);
> > -     rcar_pci_write_reg(pcie, 0x00000001, GEN2_PCIEPHYCTRL);
> > -     rcar_pci_write_reg(pcie, 0x00000006, GEN2_PCIEPHYCTRL);
> > -
> > -     return 0;
> > -}
> > -
> > -static int rcar_pcie_phy_init_gen3(struct rcar_pcie *pcie)
> > -{
> > -     int err;
> > -
> > -     err = phy_init(pcie->phy);
> > -     if (err)
> > -             return err;
> > -
> > -     err = phy_power_on(pcie->phy);
> > -     if (err)
> > -             phy_exit(pcie->phy);
> > -
> > -     return err;
> > -}
> > -
> > -static int rcar_msi_alloc(struct rcar_msi *chip)
> > -{
> > -     int msi;
> > -
> > -     mutex_lock(&chip->lock);
> > -
> > -     msi = find_first_zero_bit(chip->used, INT_PCI_MSI_NR);
> > -     if (msi < INT_PCI_MSI_NR)
> > -             set_bit(msi, chip->used);
> > -     else
> > -             msi = -ENOSPC;
> > -
> > -     mutex_unlock(&chip->lock);
> > -
> > -     return msi;
> > -}
> > -
> > -static int rcar_msi_alloc_region(struct rcar_msi *chip, int no_irqs)
> > -{
> > -     int msi;
> > -
> > -     mutex_lock(&chip->lock);
> > -     msi = bitmap_find_free_region(chip->used, INT_PCI_MSI_NR,
> > -                                   order_base_2(no_irqs));
> > -     mutex_unlock(&chip->lock);
> > -
> > -     return msi;
> > -}
> > -
> > -static void rcar_msi_free(struct rcar_msi *chip, unsigned long irq)
> > -{
> > -     mutex_lock(&chip->lock);
> > -     clear_bit(irq, chip->used);
> > -     mutex_unlock(&chip->lock);
> > -}
> > -
> > -static irqreturn_t rcar_pcie_msi_irq(int irq, void *data)
> > -{
> > -     struct rcar_pcie *pcie = data;
> > -     struct rcar_msi *msi = &pcie->msi;
> > -     struct device *dev = pcie->dev;
> > -     unsigned long reg;
> > -
> > -     reg = rcar_pci_read_reg(pcie, PCIEMSIFR);
> > -
> > -     /* MSI & INTx share an interrupt - we only handle MSI here */
> > -     if (!reg)
> > -             return IRQ_NONE;
> > -
> > -     while (reg) {
> > -             unsigned int index = find_first_bit(&reg, 32);
> > -             unsigned int msi_irq;
> > -
> > -             /* clear the interrupt */
> > -             rcar_pci_write_reg(pcie, 1 << index, PCIEMSIFR);
> > -
> > -             msi_irq = irq_find_mapping(msi->domain, index);
> > -             if (msi_irq) {
> > -                     if (test_bit(index, msi->used))
> > -                             generic_handle_irq(msi_irq);
> > -                     else
> > -                             dev_info(dev, "unhandled MSI\n");
> > -             } else {
> > -                     /* Unknown MSI, just clear it */
> > -                     dev_dbg(dev, "unexpected MSI\n");
> > -             }
> > -
> > -             /* see if there's any more pending in this vector */
> > -             reg = rcar_pci_read_reg(pcie, PCIEMSIFR);
> > -     }
> > -
> > -     return IRQ_HANDLED;
> > -}
> > -
> > -static int rcar_msi_setup_irq(struct msi_controller *chip, struct pci_dev *pdev,
> > -                           struct msi_desc *desc)
> > -{
> > -     struct rcar_msi *msi = to_rcar_msi(chip);
> > -     struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
> > -     struct msi_msg msg;
> > -     unsigned int irq;
> > -     int hwirq;
> > -
> > -     hwirq = rcar_msi_alloc(msi);
> > -     if (hwirq < 0)
> > -             return hwirq;
> > -
> > -     irq = irq_find_mapping(msi->domain, hwirq);
> > -     if (!irq) {
> > -             rcar_msi_free(msi, hwirq);
> > -             return -EINVAL;
> > -     }
> > -
> > -     irq_set_msi_desc(irq, desc);
> > -
> > -     msg.address_lo = rcar_pci_read_reg(pcie, PCIEMSIALR) & ~MSIFE;
> > -     msg.address_hi = rcar_pci_read_reg(pcie, PCIEMSIAUR);
> > -     msg.data = hwirq;
> > -
> > -     pci_write_msi_msg(irq, &msg);
> > -
> > -     return 0;
> > -}
> > -
> > -static int rcar_msi_setup_irqs(struct msi_controller *chip,
> > -                            struct pci_dev *pdev, int nvec, int type)
> > -{
> > -     struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
> > -     struct rcar_msi *msi = to_rcar_msi(chip);
> > -     struct msi_desc *desc;
> > -     struct msi_msg msg;
> > -     unsigned int irq;
> > -     int hwirq;
> > -     int i;
> > -
> > -     /* MSI-X interrupts are not supported */
> > -     if (type == PCI_CAP_ID_MSIX)
> > -             return -EINVAL;
> > -
> > -     WARN_ON(!list_is_singular(&pdev->dev.msi_list));
> > -     desc = list_entry(pdev->dev.msi_list.next, struct msi_desc, list);
> > -
> > -     hwirq = rcar_msi_alloc_region(msi, nvec);
> > -     if (hwirq < 0)
> > -             return -ENOSPC;
> > -
> > -     irq = irq_find_mapping(msi->domain, hwirq);
> > -     if (!irq)
> > -             return -ENOSPC;
> > -
> > -     for (i = 0; i < nvec; i++) {
> > -             /*
> > -              * irq_create_mapping() called from rcar_pcie_probe() pre-
> > -              * allocates descs,  so there is no need to allocate descs here.
> > -              * We can therefore assume that if irq_find_mapping() above
> > -              * returns non-zero, then the descs are also successfully
> > -              * allocated.
> > -              */
> > -             if (irq_set_msi_desc_off(irq, i, desc)) {
> > -                     /* TODO: clear */
> > -                     return -EINVAL;
> > -             }
> > -     }
> > -
> > -     desc->nvec_used = nvec;
> > -     desc->msi_attrib.multiple = order_base_2(nvec);
> > -
> > -     msg.address_lo = rcar_pci_read_reg(pcie, PCIEMSIALR) & ~MSIFE;
> > -     msg.address_hi = rcar_pci_read_reg(pcie, PCIEMSIAUR);
> > -     msg.data = hwirq;
> > -
> > -     pci_write_msi_msg(irq, &msg);
> > -
> > -     return 0;
> > -}
> > -
> > -static void rcar_msi_teardown_irq(struct msi_controller *chip, unsigned int irq)
> > -{
> > -     struct rcar_msi *msi = to_rcar_msi(chip);
> > -     struct irq_data *d = irq_get_irq_data(irq);
> > -
> > -     rcar_msi_free(msi, d->hwirq);
> > -}
> > -
> > -static struct irq_chip rcar_msi_irq_chip = {
> > -     .name = "R-Car PCIe MSI",
> > -     .irq_enable = pci_msi_unmask_irq,
> > -     .irq_disable = pci_msi_mask_irq,
> > -     .irq_mask = pci_msi_mask_irq,
> > -     .irq_unmask = pci_msi_unmask_irq,
> > -};
> > -
> > -static int rcar_msi_map(struct irq_domain *domain, unsigned int irq,
> > -                     irq_hw_number_t hwirq)
> > -{
> > -     irq_set_chip_and_handler(irq, &rcar_msi_irq_chip, handle_simple_irq);
> > -     irq_set_chip_data(irq, domain->host_data);
> > -
> > -     return 0;
> > -}
> > -
> > -static const struct irq_domain_ops msi_domain_ops = {
> > -     .map = rcar_msi_map,
> > -};
> > -
> > -static void rcar_pcie_unmap_msi(struct rcar_pcie *pcie)
> > -{
> > -     struct rcar_msi *msi = &pcie->msi;
> > -     int i, irq;
> > -
> > -     for (i = 0; i < INT_PCI_MSI_NR; i++) {
> > -             irq = irq_find_mapping(msi->domain, i);
> > -             if (irq > 0)
> > -                     irq_dispose_mapping(irq);
> > -     }
> > -
> > -     irq_domain_remove(msi->domain);
> > -}
> > -
> > -static int rcar_pcie_enable_msi(struct rcar_pcie *pcie)
> > -{
> > -     struct device *dev = pcie->dev;
> > -     struct rcar_msi *msi = &pcie->msi;
> > -     phys_addr_t base;
> > -     int err, i;
> > -
> > -     mutex_init(&msi->lock);
> > -
> > -     msi->chip.dev = dev;
> > -     msi->chip.setup_irq = rcar_msi_setup_irq;
> > -     msi->chip.setup_irqs = rcar_msi_setup_irqs;
> > -     msi->chip.teardown_irq = rcar_msi_teardown_irq;
> > -
> > -     msi->domain = irq_domain_add_linear(dev->of_node, INT_PCI_MSI_NR,
> > -                                         &msi_domain_ops, &msi->chip);
> > -     if (!msi->domain) {
> > -             dev_err(dev, "failed to create IRQ domain\n");
> > -             return -ENOMEM;
> > -     }
> > -
> > -     for (i = 0; i < INT_PCI_MSI_NR; i++)
> > -             irq_create_mapping(msi->domain, i);
> > -
> > -     /* Two irqs are for MSI, but they are also used for non-MSI irqs */
> > -     err = devm_request_irq(dev, msi->irq1, rcar_pcie_msi_irq,
> > -                            IRQF_SHARED | IRQF_NO_THREAD,
> > -                            rcar_msi_irq_chip.name, pcie);
> > -     if (err < 0) {
> > -             dev_err(dev, "failed to request IRQ: %d\n", err);
> > -             goto err;
> > -     }
> > -
> > -     err = devm_request_irq(dev, msi->irq2, rcar_pcie_msi_irq,
> > -                            IRQF_SHARED | IRQF_NO_THREAD,
> > -                            rcar_msi_irq_chip.name, pcie);
> > -     if (err < 0) {
> > -             dev_err(dev, "failed to request IRQ: %d\n", err);
> > -             goto err;
> > -     }
> > -
> > -     /* setup MSI data target */
> > -     msi->pages = __get_free_pages(GFP_KERNEL, 0);
> > -     if (!msi->pages) {
> > -             err = -ENOMEM;
> > -             goto err;
> > -     }
> > -     base = virt_to_phys((void *)msi->pages);
> > -
> > -     rcar_pci_write_reg(pcie, lower_32_bits(base) | MSIFE, PCIEMSIALR);
> > -     rcar_pci_write_reg(pcie, upper_32_bits(base), PCIEMSIAUR);
> > -
> > -     /* enable all MSI interrupts */
> > -     rcar_pci_write_reg(pcie, 0xffffffff, PCIEMSIIER);
> > -
> > -     return 0;
> > -
> > -err:
> > -     rcar_pcie_unmap_msi(pcie);
> > -     return err;
> > -}
> > -
> > -static void rcar_pcie_teardown_msi(struct rcar_pcie *pcie)
> > -{
> > -     struct rcar_msi *msi = &pcie->msi;
> > -
> > -     /* Disable all MSI interrupts */
> > -     rcar_pci_write_reg(pcie, 0, PCIEMSIIER);
> > -
> > -     /* Disable address decoding of the MSI interrupt, MSIFE */
> > -     rcar_pci_write_reg(pcie, 0, PCIEMSIALR);
> > -
> > -     free_pages(msi->pages, 0);
> > -
> > -     rcar_pcie_unmap_msi(pcie);
> > -}
> > -
> > -static int rcar_pcie_get_resources(struct rcar_pcie *pcie)
> > -{
> > -     struct device *dev = pcie->dev;
> > -     struct resource res;
> > -     int err, i;
> > -
> > -     pcie->phy = devm_phy_optional_get(dev, "pcie");
> > -     if (IS_ERR(pcie->phy))
> > -             return PTR_ERR(pcie->phy);
> > -
> > -     err = of_address_to_resource(dev->of_node, 0, &res);
> > -     if (err)
> > -             return err;
> > -
> > -     pcie->base = devm_ioremap_resource(dev, &res);
> > -     if (IS_ERR(pcie->base))
> > -             return PTR_ERR(pcie->base);
> > -
> > -     pcie->bus_clk = devm_clk_get(dev, "pcie_bus");
> > -     if (IS_ERR(pcie->bus_clk)) {
> > -             dev_err(dev, "cannot get pcie bus clock\n");
> > -             return PTR_ERR(pcie->bus_clk);
> > -     }
> > -
> > -     i = irq_of_parse_and_map(dev->of_node, 0);
> > -     if (!i) {
> > -             dev_err(dev, "cannot get platform resources for msi interrupt\n");
> > -             err = -ENOENT;
> > -             goto err_irq1;
> > -     }
> > -     pcie->msi.irq1 = i;
> > -
> > -     i = irq_of_parse_and_map(dev->of_node, 1);
> > -     if (!i) {
> > -             dev_err(dev, "cannot get platform resources for msi interrupt\n");
> > -             err = -ENOENT;
> > -             goto err_irq2;
> > -     }
> > -     pcie->msi.irq2 = i;
> > -
> > -     return 0;
> > -
> > -err_irq2:
> > -     irq_dispose_mapping(pcie->msi.irq1);
> > -err_irq1:
> > -     return err;
> > -}
> > -
> > -static int rcar_pcie_inbound_ranges(struct rcar_pcie *pcie,
> > -                                 struct of_pci_range *range,
> > -                                 int *index)
> > +void rcar_pcie_set_inbound(void __iomem *base,
> > +                        u64 cpu_addr, u64 pci_addr,
> > +                        u64 mask, int idx, bool host)
> >  {
> > -     u64 restype = range->flags;
> > -     u64 cpu_addr = range->cpu_addr;
> > -     u64 cpu_end = range->cpu_addr + range->size;
> > -     u64 pci_addr = range->pci_addr;
> > -     u32 flags = LAM_64BIT | LAR_ENABLE;
> > -     u64 mask;
> > -     u64 size;
> > -     int idx = *index;
> > -
> > -     if (restype & IORESOURCE_PREFETCH)
> > -             flags |= LAM_PREFETCH;
> > -
> >       /*
> > -      * If the size of the range is larger than the alignment of the start
> > -      * address, we have to use multiple entries to perform the mapping.
> > -      */
> > -     if (cpu_addr > 0) {
> > -             unsigned long nr_zeros = __ffs64(cpu_addr);
> > -             u64 alignment = 1ULL << nr_zeros;
> > -
> > -             size = min(range->size, alignment);
> > -     } else {
> > -             size = range->size;
> > -     }
> > -     /* Hardware supports max 4GiB inbound region */
> > -     size = min(size, 1ULL << 32);
> > -
> > -     mask = roundup_pow_of_two(size) - 1;
> > -     mask &= ~0xf;
> > -
> > -     while (cpu_addr < cpu_end) {
> > -             /*
> > -              * Set up 64-bit inbound regions as the range parser doesn't
> > -              * distinguish between 32 and 64-bit types.
> > -              */
> > -             rcar_pci_write_reg(pcie, lower_32_bits(pci_addr),
> > +     * Set up 64-bit inbound regions as the range parser doesn't
> > +     * distinguish between 32 and 64-bit types.
> > +     */
> > +     if (host) {
> > +             rcar_pci_write_reg(base, lower_32_bits(pci_addr),
> >                                  PCIEPRAR(idx));
> > -             rcar_pci_write_reg(pcie, lower_32_bits(cpu_addr), PCIELAR(idx));
> > -             rcar_pci_write_reg(pcie, lower_32_bits(mask) | flags,
> > -                                PCIELAMR(idx));
> > -
> > -             rcar_pci_write_reg(pcie, upper_32_bits(pci_addr),
> > +             rcar_pci_write_reg(base, upper_32_bits(pci_addr),
> >                                  PCIEPRAR(idx + 1));
> > -             rcar_pci_write_reg(pcie, upper_32_bits(cpu_addr),
> > -                                PCIELAR(idx + 1));
> > -             rcar_pci_write_reg(pcie, 0, PCIELAMR(idx + 1));
> > -
> > -             pci_addr += size;
> > -             cpu_addr += size;
> > -             idx += 2;
> > -
> > -             if (idx > MAX_NR_INBOUND_MAPS) {
> > -                     dev_err(pcie->dev, "Failed to map inbound regions!\n");
> > -                     return -EINVAL;
> > -             }
> >       }
> > -     *index = idx;
> > -
> > -     return 0;
> > -}
> >
> > -static int rcar_pcie_parse_map_dma_ranges(struct rcar_pcie *pcie,
> > -                                       struct device_node *np)
> > -{
> > -     struct of_pci_range range;
> > -     struct of_pci_range_parser parser;
> > -     int index = 0;
> > -     int err;
> > -
> > -     if (of_pci_dma_range_parser_init(&parser, np))
> > -             return -EINVAL;
> > -
> > -     /* Get the dma-ranges from DT */
> > -     for_each_of_pci_range(&parser, &range) {
> > -             u64 end = range.cpu_addr + range.size - 1;
> > +     rcar_pci_write_reg(base, lower_32_bits(cpu_addr), PCIELAR(idx));
> > +     rcar_pci_write_reg(base, lower_32_bits(mask), PCIELAMR(idx));
> >
> > -             dev_dbg(pcie->dev, "0x%08x 0x%016llx..0x%016llx -> 0x%016llx\n",
> > -                     range.flags, range.cpu_addr, end, range.pci_addr);
> > -
> > -             err = rcar_pcie_inbound_ranges(pcie, &range, &index);
> > -             if (err)
> > -                     return err;
> > -     }
> > +     rcar_pci_write_reg(base, upper_32_bits(cpu_addr), PCIELAR(idx + 1));
> > +     rcar_pci_write_reg(base, 0, PCIELAMR(idx + 1));
> >
> > -     return 0;
> > -}
> > -
> > -static const struct of_device_id rcar_pcie_of_match[] = {
> > -     { .compatible = "renesas,pcie-r8a7779",
> > -       .data = rcar_pcie_phy_init_h1 },
> > -     { .compatible = "renesas,pcie-r8a7790",
> > -       .data = rcar_pcie_phy_init_gen2 },
> > -     { .compatible = "renesas,pcie-r8a7791",
> > -       .data = rcar_pcie_phy_init_gen2 },
> > -     { .compatible = "renesas,pcie-rcar-gen2",
> > -       .data = rcar_pcie_phy_init_gen2 },
> > -     { .compatible = "renesas,pcie-r8a7795",
> > -       .data = rcar_pcie_phy_init_gen3 },
> > -     { .compatible = "renesas,pcie-rcar-gen3",
> > -       .data = rcar_pcie_phy_init_gen3 },
> > -     {},
> > -};
> > -
> > -static int rcar_pcie_probe(struct platform_device *pdev)
> > -{
> > -     struct device *dev = &pdev->dev;
> > -     struct rcar_pcie *pcie;
> > -     u32 data;
> > -     int err;
> > -     int (*phy_init_fn)(struct rcar_pcie *);
> > -     struct pci_host_bridge *bridge;
> > -
> > -     bridge = pci_alloc_host_bridge(sizeof(*pcie));
> > -     if (!bridge)
> > -             return -ENOMEM;
> > -
> > -     pcie = pci_host_bridge_priv(bridge);
> > -
> > -     pcie->dev = dev;
> > -     platform_set_drvdata(pdev, pcie);
> > -
> > -     err = pci_parse_request_of_pci_ranges(dev, &pcie->resources, NULL);
> > -     if (err)
> > -             goto err_free_bridge;
> > -
> > -     pm_runtime_enable(pcie->dev);
> > -     err = pm_runtime_get_sync(pcie->dev);
> > -     if (err < 0) {
> > -             dev_err(pcie->dev, "pm_runtime_get_sync failed\n");
> > -             goto err_pm_disable;
> > -     }
> > -
> > -     err = rcar_pcie_get_resources(pcie);
> > -     if (err < 0) {
> > -             dev_err(dev, "failed to request resources: %d\n", err);
> > -             goto err_pm_put;
> > -     }
> > -
> > -     err = clk_prepare_enable(pcie->bus_clk);
> > -     if (err) {
> > -             dev_err(dev, "failed to enable bus clock: %d\n", err);
> > -             goto err_unmap_msi_irqs;
> > -     }
> > -
> > -     err = rcar_pcie_parse_map_dma_ranges(pcie, dev->of_node);
> > -     if (err)
> > -             goto err_clk_disable;
> > -
> > -     phy_init_fn = of_device_get_match_data(dev);
> > -     err = phy_init_fn(pcie);
> > -     if (err) {
> > -             dev_err(dev, "failed to init PCIe PHY\n");
> > -             goto err_clk_disable;
> > -     }
> > -
> > -     /* Failure to get a link might just be that no cards are inserted */
> > -     if (rcar_pcie_hw_init(pcie)) {
> > -             dev_info(dev, "PCIe link down\n");
> > -             err = -ENODEV;
> > -             goto err_phy_shutdown;
> > -     }
> > -
> > -     data = rcar_pci_read_reg(pcie, MACSR);
> > -     dev_info(dev, "PCIe x%d: link up\n", (data >> 20) & 0x3f);
> > -
> > -     if (IS_ENABLED(CONFIG_PCI_MSI)) {
> > -             err = rcar_pcie_enable_msi(pcie);
> > -             if (err < 0) {
> > -                     dev_err(dev,
> > -                             "failed to enable MSI support: %d\n",
> > -                             err);
> > -                     goto err_phy_shutdown;
> > -             }
> > -     }
> > -
> > -     err = rcar_pcie_enable(pcie);
> > -     if (err)
> > -             goto err_msi_teardown;
> > -
> > -     return 0;
> > -
> > -err_msi_teardown:
> > -     if (IS_ENABLED(CONFIG_PCI_MSI))
> > -             rcar_pcie_teardown_msi(pcie);
> > -
> > -err_phy_shutdown:
> > -     if (pcie->phy) {
> > -             phy_power_off(pcie->phy);
> > -             phy_exit(pcie->phy);
> > -     }
> > -
> > -err_clk_disable:
> > -     clk_disable_unprepare(pcie->bus_clk);
> > -
> > -err_unmap_msi_irqs:
> > -     irq_dispose_mapping(pcie->msi.irq2);
> > -     irq_dispose_mapping(pcie->msi.irq1);
> > -
> > -err_pm_put:
> > -     pm_runtime_put(dev);
> > -
> > -err_pm_disable:
> > -     pm_runtime_disable(dev);
> > -     pci_free_resource_list(&pcie->resources);
> > -
> > -err_free_bridge:
> > -     pci_free_host_bridge(bridge);
> > -
> > -     return err;
> > -}
> > -
> > -static int rcar_pcie_resume_noirq(struct device *dev)
> > -{
> > -     struct rcar_pcie *pcie = dev_get_drvdata(dev);
> > -
> > -     if (rcar_pci_read_reg(pcie, PMSR) &&
> > -         !(rcar_pci_read_reg(pcie, PCIETCTLR) & DL_DOWN))
> > -             return 0;
> > -
> > -     /* Re-establish the PCIe link */
> > -     rcar_rmw32(pcie, MACCTLR, MACCTLR_RESERVED, 0);
> > -     rcar_pci_write_reg(pcie, CFINIT, PCIETCTLR);
> > -     return rcar_pcie_wait_for_dl(pcie);
> > +     /* flush modifications */
> > +     wmb();
> >  }
> > -
> > -static const struct dev_pm_ops rcar_pcie_pm_ops = {
> > -     .resume_noirq = rcar_pcie_resume_noirq,
> > -};
> > -
> > -static struct platform_driver rcar_pcie_driver = {
> > -     .driver = {
> > -             .name = "rcar-pcie",
> > -             .of_match_table = rcar_pcie_of_match,
> > -             .pm = &rcar_pcie_pm_ops,
> > -             .suppress_bind_attrs = true,
> > -     },
> > -     .probe = rcar_pcie_probe,
> > -};
> > -builtin_platform_driver(rcar_pcie_driver);
> > diff --git a/drivers/pci/controller/pcie-rcar.h b/drivers/pci/controller/pcie-rcar.h
> > new file mode 100644
> > index 000000000000..502621d4e6ac
> > --- /dev/null
> > +++ b/drivers/pci/controller/pcie-rcar.h
> > @@ -0,0 +1,129 @@
> > +/* SPDX-License-Identifier: GPL-2.0 */
> > +/*
> > + * PCIe driver for Renesas R-Car SoCs
> > + *  Copyright (C) 2014-2019 Renesas Electronics Europe Ltd
> > + *
> > + * Author: Phil Edworthy <phil.edworthy@renesas.com>
> > + */
> > +#ifndef _PCIE_RCAR_H
> > +#define _PCIE_RCAR_H
> > +
> > +#define PCIECAR                      0x000010
> > +#define PCIECCTLR            0x000018
> > +#define  CONFIG_SEND_ENABLE  BIT(31)
> > +#define  TYPE0                       (0 << 8)
> > +#define  TYPE1                       BIT(8)
> > +#define PCIECDR                      0x000020
> > +#define PCIEMSR                      0x000028
> > +#define PCIEINTXR            0x000400
> > +#define PCIEPHYSR            0x0007f0
> > +#define  PHYRDY                      BIT(0)
> > +#define PCIEMSITXR           0x000840
> > +
> > +/* Transfer control */
> > +#define PCIETCTLR            0x02000
> > +#define  DL_DOWN             BIT(3)
> > +#define  CFINIT                      BIT(0)
> > +#define PCIETSTR             0x02004
> > +#define  DATA_LINK_ACTIVE    BIT(0)
> > +#define PCIEERRFR            0x02020
> > +#define  UNSUPPORTED_REQUEST BIT(4)
> > +#define PCIEMSIFR            0x02044
> > +#define PCIEMSIALR           0x02048
> > +#define  MSIFE                       BIT(0)
> > +#define PCIEMSIAUR           0x0204c
> > +#define PCIEMSIIER           0x02050
> > +
> > +/* root port address */
> > +#define PCIEPRAR(x)          (0x02080 + ((x) * 0x4))
> > +
> > +/* local address reg & mask */
> > +#define PCIELAR(x)           (0x02200 + ((x) * 0x20))
> > +#define PCIELAMR(x)          (0x02208 + ((x) * 0x20))
> > +#define  LAM_PREFETCH                BIT(3)
> > +#define  LAM_64BIT           BIT(2)
> > +#define  LAR_ENABLE          BIT(1)
> > +
> > +/* PCIe address reg & mask */
> > +#define PCIEPALR(x)          (0x03400 + ((x) * 0x20))
> > +#define PCIEPAUR(x)          (0x03404 + ((x) * 0x20))
> > +#define PCIEPAMR(x)          (0x03408 + ((x) * 0x20))
> > +#define PCIEPTCTLR(x)                (0x0340c + ((x) * 0x20))
> > +#define  PAR_ENABLE          BIT(31)
> > +#define  IO_SPACE            BIT(8)
> > +
> > +/* Configuration */
> > +#define PCICONF(x)           (0x010000 + ((x) * 0x4))
> > +#define PMCAP(x)             (0x010040 + ((x) * 0x4))
> > +#define EXPCAP(x)            (0x010070 + ((x) * 0x4))
> > +#define VCCAP(x)             (0x010100 + ((x) * 0x4))
> > +
> > +/* link layer */
> > +#define IDSETR0                      0x011000
> > +#define  DEVICE_ID_SHFIT     16
> > +#define IDSETR1                      0x011004
> > +#define SUBIDSETR            0x011024
> > +#define TLCTLR                       0x011048
> > +#define MACSR                        0x011054
> > +#define  SPCHGFIN            BIT(4)
> > +#define  SPCHGFAIL           BIT(6)
> > +#define  SPCHGSUC            BIT(7)
> > +#define  LINK_SPEED          (0xf << 16)
> > +#define  LINK_SPEED_2_5GTS   BIT(16)
> > +#define  LINK_SPEED_5_0GTS   BIT(17)
> > +#define MACCTLR                      0x011058
> > +#define  MACCTLR_RESERVED    BIT(0)
> > +#define  SPEED_CHANGE                BIT(24)
> > +#define  SCRAMBLE_DISABLE    BIT(27)
> > +#define PMSR                 0x01105c
> > +#define MACS2R                       0x011078
> > +#define MACCGSPSETR          0x011084
> > +#define  SPCNGRSN            BIT(31)
> > +
> > +/* R-Car H1 PHY */
> > +#define H1_PCIEPHYADRR               0x04000c
> > +#define  WRITE_CMD           BIT(16)
> > +#define  PHY_ACK             BIT(24)
> > +#define  RATE_POS            12
> > +#define  LANE_POS            8
> > +#define  ADR_POS             0
> > +#define H1_PCIEPHYDOUTR              0x040014
> > +
> > +/* R-Car Gen2 PHY */
> > +#define GEN2_PCIEPHYADDR     0x780
> > +#define GEN2_PCIEPHYDATA     0x784
> > +#define GEN2_PCIEPHYCTRL     0x78c
> > +
> > +#define INT_PCI_MSI_NR               32
> > +
> > +#define RCONF(x)             (PCICONF(0) + (x))
> > +#define RPMCAP(x)            (PMCAP(0) + (x))
> > +#define REXPCAP(x)           (EXPCAP(0) + (x))
> > +#define RVCCAP(x)            (VCCAP(0) + (x))
> > +
> > +#define INTDIS_SHIFT         BIT(10)
> > +#define ASTINTX_SHIFT                BIT(16)
> > +
> > +#define PCIE_CONF_BUS(b)     (((b) & 0xff) << 24)
> > +#define PCIE_CONF_DEV(d)     (((d) & 0x1f) << 19)
> > +#define PCIE_CONF_FUNC(f)    (((f) & 0x7) << 16)
> > +
> > +#define RCAR_PCI_MAX_RESOURCES       4
> > +#define MAX_NR_INBOUND_MAPS  6
> > +
> > +enum {
> > +     RCAR_PCI_ACCESS_READ,
> > +     RCAR_PCI_ACCESS_WRITE,
> > +};
> > +
> > +void rcar_pci_write_reg(void __iomem *base, u32 val, unsigned int reg);
> > +u32 rcar_pci_read_reg(void __iomem *base, unsigned int reg);
> > +void rcar_rmw32(void __iomem *base, int where, u32 mask, u32 data);
> > +int rcar_pcie_wait_for_phyrdy(void __iomem *base);
> > +int rcar_pcie_wait_for_dl(void __iomem *base);
> > +void rcar_pcie_set_outbound(int win, void __iomem *base,
> > +                         struct resource *res, bool host);
> > +void rcar_pcie_set_inbound(void __iomem *base, u64 cpu_addr, u64 pci_addr,
> > +                        u64 mask, int idx, bool host);
> > +
> > +#endif
> >

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 1/5] pci: pcie-rcar: preparation for adding endpoint support
@ 2019-11-27 20:51       ` Lad, Prabhakar
  0 siblings, 0 replies; 60+ messages in thread
From: Lad, Prabhakar @ 2019-11-27 20:51 UTC (permalink / raw)
  To: Kishon Vijay Abraham I
  Cc: Mark Rutland,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Chris Paterson, Lorenzo Pieralisi, Arnd Bergmann,
	Geert Uytterhoeven, Greg Kroah-Hartman, linux-pci,
	Yoshihiro Shimoda, Magnus Damm, LKML, Lad, Prabhakar,
	Linux-Renesas, Rob Herring, Catalin Marinas, Bjorn Helgaas,
	Andrew Murray, Will Deacon, LAK, Marek Vasut

Hi Kishon,

On Wed, Nov 27, 2019 at 4:56 AM Kishon Vijay Abraham I <kishon@ti.com> wrote:
>
> Hi Prabhakar,
>
> On 07/11/19 1:06 AM, Lad Prabhakar wrote:
> > From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >
> > this patch prepares for adding endpoint support to rcar controller,
> > there are no functional changes with this patch, a common file is
> > created so that it can be shared with endpoint driver.
> >
> > Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > ---
> >  arch/arm64/configs/defconfig            |    2 +-
> >  arch/arm64/configs/renesas_defconfig    |    2 +-
> >  drivers/pci/controller/Kconfig          |    4 +-
> >  drivers/pci/controller/Makefile         |    2 +-
> >  drivers/pci/controller/pcie-rcar-host.c | 1058 +++++++++++++++++++
> >  drivers/pci/controller/pcie-rcar.c      | 1231 +----------------------
>
> Can you create this patch using "git format-patch -M", so that only whatever
> changed appears here.
>
sure will do that in the next version.

Cheers,
--Prabhakar

> Thanks
> Kishon
>
> >  drivers/pci/controller/pcie-rcar.h      |  129 +++
> >  7 files changed, 1245 insertions(+), 1183 deletions(-)
> >  create mode 100644 drivers/pci/controller/pcie-rcar-host.c
> >  create mode 100644 drivers/pci/controller/pcie-rcar.h
> >
> > diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
> > index 7d16f10a9251..f67ce0ccea1d 100644
> > --- a/arch/arm64/configs/defconfig
> > +++ b/arch/arm64/configs/defconfig
> > @@ -186,7 +186,7 @@ CONFIG_HOTPLUG_PCI=y
> >  CONFIG_HOTPLUG_PCI_ACPI=y
> >  CONFIG_PCI_AARDVARK=y
> >  CONFIG_PCI_TEGRA=y
> > -CONFIG_PCIE_RCAR=y
> > +CONFIG_PCIE_RCAR_HOST=y
> >  CONFIG_PCI_HOST_GENERIC=y
> >  CONFIG_PCI_XGENE=y
> >  CONFIG_PCIE_ALTERA=y
> > diff --git a/arch/arm64/configs/renesas_defconfig b/arch/arm64/configs/renesas_defconfig
> > index f7411e8807d6..75fc418377c5 100644
> > --- a/arch/arm64/configs/renesas_defconfig
> > +++ b/arch/arm64/configs/renesas_defconfig
> > @@ -91,7 +91,7 @@ CONFIG_PCIEPORTBUS=y
> >  CONFIG_HOTPLUG_PCI_PCIE=y
> >  CONFIG_PCI_IOV=y
> >  CONFIG_HOTPLUG_PCI=y
> > -CONFIG_PCIE_RCAR=y
> > +CONFIG_PCIE_RCAR_HOST=y
> >  CONFIG_PCI_HOST_GENERIC=y
> >  CONFIG_DEVTMPFS=y
> >  CONFIG_DEVTMPFS_MOUNT=y
> > diff --git a/drivers/pci/controller/Kconfig b/drivers/pci/controller/Kconfig
> > index 70e078238899..01c238e5ea1e 100644
> > --- a/drivers/pci/controller/Kconfig
> > +++ b/drivers/pci/controller/Kconfig
> > @@ -82,12 +82,12 @@ config PCI_RCAR_GEN2
> >         There are 3 internal PCI controllers available with a single
> >         built-in EHCI/OHCI host controller present on each one.
> >
> > -config PCIE_RCAR
> > +config PCIE_RCAR_HOST
> >       bool "Renesas R-Car PCIe controller"
> >       depends on ARCH_RENESAS || COMPILE_TEST
> >       depends on PCI_MSI_IRQ_DOMAIN
> >       help
> > -       Say Y here if you want PCIe controller support on R-Car SoCs.
> > +       Say Y here if you want PCIe controller support on R-Car SoCs in host mode.
> >
> >  config PCI_HOST_COMMON
> >       bool
> > diff --git a/drivers/pci/controller/Makefile b/drivers/pci/controller/Makefile
> > index a2a22c9d91af..3577902a0d40 100644
> > --- a/drivers/pci/controller/Makefile
> > +++ b/drivers/pci/controller/Makefile
> > @@ -9,7 +9,7 @@ obj-$(CONFIG_PCI_MVEBU) += pci-mvebu.o
> >  obj-$(CONFIG_PCI_AARDVARK) += pci-aardvark.o
> >  obj-$(CONFIG_PCI_TEGRA) += pci-tegra.o
> >  obj-$(CONFIG_PCI_RCAR_GEN2) += pci-rcar-gen2.o
> > -obj-$(CONFIG_PCIE_RCAR) += pcie-rcar.o
> > +obj-$(CONFIG_PCIE_RCAR_HOST) += pcie-rcar.o pcie-rcar-host.o
> >  obj-$(CONFIG_PCI_HOST_COMMON) += pci-host-common.o
> >  obj-$(CONFIG_PCI_HOST_GENERIC) += pci-host-generic.o
> >  obj-$(CONFIG_PCIE_XILINX) += pcie-xilinx.o
> > diff --git a/drivers/pci/controller/pcie-rcar-host.c b/drivers/pci/controller/pcie-rcar-host.c
> > new file mode 100644
> > index 000000000000..645ef3e41e27
> > --- /dev/null
> > +++ b/drivers/pci/controller/pcie-rcar-host.c
> > @@ -0,0 +1,1058 @@
> > +// SPDX-License-Identifier: GPL-2.0
> > +/*
> > + * PCIe driver for Renesas R-Car SoCs
> > + *  Copyright (C) 2014-2019 Renesas Electronics Europe Ltd
> > + *
> > + * Based on:
> > + *  arch/sh/drivers/pci/pcie-sh7786.c
> > + *  arch/sh/drivers/pci/ops-sh7786.c
> > + *  Copyright (C) 2009 - 2011  Paul Mundt
> > + *
> > + * Author: Phil Edworthy <phil.edworthy@renesas.com>
> > + */
> > +
> > +#include <linux/clk.h>
> > +#include <linux/delay.h>
> > +#include <linux/msi.h>
> > +#include <linux/of_address.h>
> > +#include <linux/of_irq.h>
> > +#include <linux/of_pci.h>
> > +#include <linux/of_platform.h>
> > +#include <linux/pci.h>
> > +#include <linux/phy/phy.h>
> > +
> > +#include "pcie-rcar.h"
> > +
> > +struct rcar_msi {
> > +     DECLARE_BITMAP(used, INT_PCI_MSI_NR);
> > +     struct irq_domain *domain;
> > +     struct msi_controller chip;
> > +     unsigned long pages;
> > +     struct mutex lock; /* serializes msi */
> > +     int irq1;
> > +     int irq2;
> > +};
> > +
> > +static inline struct rcar_msi *to_rcar_msi(struct msi_controller *chip)
> > +{
> > +     return container_of(chip, struct rcar_msi, chip);
> > +}
> > +
> > +/* Structure representing the PCIe interface */
> > +struct rcar_pcie {
> > +     struct device           *dev;
> > +     struct phy              *phy;
> > +     void __iomem            *base;
> > +     struct list_head        resources;
> > +     int                     root_bus_nr;
> > +     struct clk              *bus_clk;
> > +     struct                  rcar_msi msi;
> > +};
> > +
> > +static u32 rcar_read_conf(struct rcar_pcie *pcie, int where)
> > +{
> > +     unsigned int shift = BITS_PER_BYTE * (where & 3);
> > +     u32 val = rcar_pci_read_reg(pcie->base, where & ~3);
> > +
> > +     return val >> shift;
> > +}
> > +
> > +/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
> > +static int
> > +rcar_pcie_config_access(struct rcar_pcie *pcie,
> > +                     unsigned char access_type, struct pci_bus *bus,
> > +                     unsigned int devfn, int where, u32 *data)
> > +{
> > +     unsigned int dev, func, reg, index;
> > +
> > +     dev = PCI_SLOT(devfn);
> > +     func = PCI_FUNC(devfn);
> > +     reg = where & ~3;
> > +     index = reg / 4;
> > +
> > +     /*
> > +      * While each channel has its own memory-mapped extended config
> > +      * space, it's generally only accessible when in endpoint mode.
> > +      * When in root complex mode, the controller is unable to target
> > +      * itself with either type 0 or type 1 accesses, and indeed, any
> > +      * controller initiated target transfer to its own config space
> > +      * result in a completer abort.
> > +      *
> > +      * Each channel effectively only supports a single device, but as
> > +      * the same channel <-> device access works for any PCI_SLOT()
> > +      * value, we cheat a bit here and bind the controller's config
> > +      * space to devfn 0 in order to enable self-enumeration. In this
> > +      * case the regular ECAR/ECDR path is sidelined and the mangled
> > +      * config access itself is initiated as an internal bus transaction.
> > +      */
> > +     if (pci_is_root_bus(bus)) {
> > +             if (dev != 0)
> > +                     return PCIBIOS_DEVICE_NOT_FOUND;
> > +
> > +             if (access_type == RCAR_PCI_ACCESS_READ) {
> > +                     *data = rcar_pci_read_reg(pcie->base, PCICONF(index));
> > +             } else {
> > +                     /* Keep an eye out for changes to the root bus number */
> > +                     if (pci_is_root_bus(bus) && reg == PCI_PRIMARY_BUS)
> > +                             pcie->root_bus_nr = *data & 0xff;
> > +
> > +                     rcar_pci_write_reg(pcie->base, *data, PCICONF(index));
> > +             }
> > +
> > +             return PCIBIOS_SUCCESSFUL;
> > +     }
> > +
> > +     if (pcie->root_bus_nr < 0)
> > +             return PCIBIOS_DEVICE_NOT_FOUND;
> > +
> > +     /* Clear errors */
> > +     rcar_pci_write_reg(pcie->base, rcar_pci_read_reg(pcie->base, PCIEERRFR),
> > +                        PCIEERRFR);
> > +
> > +     /* Set the PIO address */
> > +     rcar_pci_write_reg(pcie->base, PCIE_CONF_BUS(bus->number) |
> > +             PCIE_CONF_DEV(dev) | PCIE_CONF_FUNC(func) | reg, PCIECAR);
> > +
> > +     /* Enable the configuration access */
> > +     if (bus->parent->number == pcie->root_bus_nr)
> > +             rcar_pci_write_reg(pcie->base,
> > +                                CONFIG_SEND_ENABLE | TYPE0, PCIECCTLR);
> > +     else
> > +             rcar_pci_write_reg(pcie->base,
> > +                                CONFIG_SEND_ENABLE | TYPE1, PCIECCTLR);
> > +
> > +     /* Check for errors */
> > +     if (rcar_pci_read_reg(pcie->base, PCIEERRFR) & UNSUPPORTED_REQUEST)
> > +             return PCIBIOS_DEVICE_NOT_FOUND;
> > +
> > +     /* Check for master and target aborts */
> > +     if (rcar_read_conf(pcie, RCONF(PCI_STATUS)) &
> > +             (PCI_STATUS_REC_MASTER_ABORT | PCI_STATUS_REC_TARGET_ABORT))
> > +             return PCIBIOS_DEVICE_NOT_FOUND;
> > +
> > +     if (access_type == RCAR_PCI_ACCESS_READ)
> > +             *data = rcar_pci_read_reg(pcie->base, PCIECDR);
> > +     else
> > +             rcar_pci_write_reg(pcie->base, *data, PCIECDR);
> > +
> > +     /* Disable the configuration access */
> > +     rcar_pci_write_reg(pcie->base, 0, PCIECCTLR);
> > +
> > +     return PCIBIOS_SUCCESSFUL;
> > +}
> > +
> > +static int rcar_pcie_read_conf(struct pci_bus *bus, unsigned int devfn,
> > +                            int where, int size, u32 *val)
> > +{
> > +     struct rcar_pcie *pcie = bus->sysdata;
> > +     int ret;
> > +
> > +     ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
> > +                                   bus, devfn, where, val);
> > +     if (ret != PCIBIOS_SUCCESSFUL) {
> > +             *val = 0xffffffff;
> > +             return ret;
> > +     }
> > +
> > +     if (size == 1)
> > +             *val = (*val >> (BITS_PER_BYTE * (where & 3))) & 0xff;
> > +     else if (size == 2)
> > +             *val = (*val >> (BITS_PER_BYTE * (where & 2))) & 0xffff;
> > +
> > +     dev_dbg(&bus->dev, "pcie-config-read: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
> > +             bus->number, devfn, where, size, *val);
> > +
> > +     return ret;
> > +}
> > +
> > +/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
> > +static int rcar_pcie_write_conf(struct pci_bus *bus, unsigned int devfn,
> > +                             int where, int size, u32 val)
> > +{
> > +     struct rcar_pcie *pcie = bus->sysdata;
> > +     unsigned int shift;
> > +     u32 data;
> > +     int ret;
> > +
> > +     ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
> > +                                   bus, devfn, where, &data);
> > +     if (ret != PCIBIOS_SUCCESSFUL)
> > +             return ret;
> > +
> > +     dev_dbg(&bus->dev, "pcie-config-write: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
> > +             bus->number, devfn, where, size, val);
> > +
> > +     if (size == 1) {
> > +             shift = BITS_PER_BYTE * (where & 3);
> > +             data &= ~(0xff << shift);
> > +             data |= ((val & 0xff) << shift);
> > +     } else if (size == 2) {
> > +             shift = BITS_PER_BYTE * (where & 2);
> > +             data &= ~(0xffff << shift);
> > +             data |= ((val & 0xffff) << shift);
> > +     } else {
> > +             data = val;
> > +     }
> > +
> > +     ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_WRITE,
> > +                                   bus, devfn, where, &data);
> > +
> > +     return ret;
> > +}
> > +
> > +static struct pci_ops rcar_pcie_ops = {
> > +     .read   = rcar_pcie_read_conf,
> > +     .write  = rcar_pcie_write_conf,
> > +};
> > +
> > +static int rcar_pcie_setup(struct list_head *resource, struct rcar_pcie *pci)
> > +{
> > +     struct resource_entry *win;
> > +     int i = 0;
> > +
> > +     /* Setup PCI resources */
> > +     resource_list_for_each_entry(win, &pci->resources) {
> > +             struct resource *res = win->res;
> > +
> > +             if (!res->flags)
> > +                     continue;
> > +
> > +             switch (resource_type(res)) {
> > +             case IORESOURCE_IO:
> > +             case IORESOURCE_MEM:
> > +                     rcar_pcie_set_outbound(i, pci->base, res, true);
> > +                     i++;
> > +                     break;
> > +             case IORESOURCE_BUS:
> > +                     pci->root_bus_nr = res->start;
> > +                     break;
> > +             default:
> > +                     continue;
> > +             }
> > +
> > +             pci_add_resource(resource, res);
> > +     }
> > +
> > +     return 1;
> > +}
> > +
> > +static void rcar_pcie_force_speedup(struct rcar_pcie *pcie)
> > +{
> > +     struct device *dev = pcie->dev;
> > +     unsigned int timeout = 1000;
> > +     u32 macsr;
> > +
> > +     if ((rcar_pci_read_reg(pcie->base, MACS2R) & LINK_SPEED) !=
> > +             LINK_SPEED_5_0GTS)
> > +             return;
> > +
> > +     if (rcar_pci_read_reg(pcie->base, MACCTLR) & SPEED_CHANGE) {
> > +             dev_err(dev, "Speed change already in progress\n");
> > +             return;
> > +     }
> > +
> > +     macsr = rcar_pci_read_reg(pcie->base, MACSR);
> > +     if ((macsr & LINK_SPEED) == LINK_SPEED_5_0GTS)
> > +             goto done;
> > +
> > +     /* Set target link speed to 5.0 GT/s */
> > +     rcar_rmw32(pcie->base, EXPCAP(12), PCI_EXP_LNKSTA_CLS,
> > +                PCI_EXP_LNKSTA_CLS_5_0GB);
> > +
> > +     /* Set speed change reason as intentional factor */
> > +     rcar_rmw32(pcie->base, MACCGSPSETR, SPCNGRSN, 0);
> > +
> > +     /* Clear SPCHGFIN, SPCHGSUC, and SPCHGFAIL */
> > +     if (macsr & (SPCHGFIN | SPCHGSUC | SPCHGFAIL))
> > +             rcar_pci_write_reg(pcie->base, macsr, MACSR);
> > +
> > +     /* Start link speed change */
> > +     rcar_rmw32(pcie->base, MACCTLR, SPEED_CHANGE, SPEED_CHANGE);
> > +
> > +     while (timeout--) {
> > +             macsr = rcar_pci_read_reg(pcie->base, MACSR);
> > +             if (macsr & SPCHGFIN) {
> > +                     /* Clear the interrupt bits */
> > +                     rcar_pci_write_reg(pcie->base, macsr, MACSR);
> > +
> > +                     if (macsr & SPCHGFAIL)
> > +                             dev_err(dev, "Speed change failed\n");
> > +
> > +                     goto done;
> > +             }
> > +
> > +             usleep_range(1000, 2000);
> > +     }
> > +
> > +     dev_err(dev, "Speed change timed out\n");
> > +
> > +done:
> > +     dev_info(dev, "Current link speed is %s GT/s\n",
> > +              (macsr & LINK_SPEED) == LINK_SPEED_5_0GTS ? "5" : "2.5");
> > +}
> > +
> > +static int rcar_pcie_enable(struct rcar_pcie *pcie)
> > +{
> > +     struct pci_host_bridge *bridge = pci_host_bridge_from_priv(pcie);
> > +     struct device *dev = pcie->dev;
> > +     struct pci_bus *bus, *child;
> > +     int ret;
> > +
> > +     /* Try setting 5 GT/s link speed */
> > +     rcar_pcie_force_speedup(pcie);
> > +
> > +     rcar_pcie_setup(&bridge->windows, pcie);
> > +
> > +     pci_add_flags(PCI_REASSIGN_ALL_BUS);
> > +
> > +     bridge->dev.parent = dev;
> > +     bridge->sysdata = pcie;
> > +     bridge->busnr = pcie->root_bus_nr;
> > +     bridge->ops = &rcar_pcie_ops;
> > +     bridge->map_irq = of_irq_parse_and_map_pci;
> > +     bridge->swizzle_irq = pci_common_swizzle;
> > +     if (IS_ENABLED(CONFIG_PCI_MSI))
> > +             bridge->msi = &pcie->msi.chip;
> > +
> > +     ret = pci_scan_root_bus_bridge(bridge);
> > +     if (ret < 0)
> > +             return ret;
> > +
> > +     bus = bridge->bus;
> > +
> > +     pci_bus_size_bridges(bus);
> > +     pci_bus_assign_resources(bus);
> > +
> > +     list_for_each_entry(child, &bus->children, node)
> > +             pcie_bus_configure_settings(child);
> > +
> > +     pci_bus_add_devices(bus);
> > +
> > +     return 0;
> > +}
> > +
> > +static int phy_wait_for_ack(struct rcar_pcie *pcie)
> > +{
> > +     struct device *dev = pcie->dev;
> > +     unsigned int timeout = 100;
> > +
> > +     while (timeout--) {
> > +             if (rcar_pci_read_reg(pcie->base, H1_PCIEPHYADRR) & PHY_ACK)
> > +                     return 0;
> > +
> > +             usleep_range(90, 110);
> > +     }
> > +
> > +     dev_err(dev, "Access to PCIe phy timed out\n");
> > +
> > +     return -ETIMEDOUT;
> > +}
> > +
> > +static void phy_write_reg(struct rcar_pcie *pcie,
> > +                       unsigned int rate, u32 addr,
> > +                       unsigned int lane, u32 data)
> > +{
> > +     u32 phyaddr;
> > +
> > +     phyaddr = WRITE_CMD |
> > +             ((rate & 1) << RATE_POS) |
> > +             ((lane & 0xf) << LANE_POS) |
> > +             ((addr & 0xff) << ADR_POS);
> > +
> > +     /* Set write data */
> > +     rcar_pci_write_reg(pcie->base, data, H1_PCIEPHYDOUTR);
> > +     rcar_pci_write_reg(pcie->base, phyaddr, H1_PCIEPHYADRR);
> > +
> > +     /* Ignore errors as they will be dealt with if the data link is down */
> > +     phy_wait_for_ack(pcie);
> > +
> > +     /* Clear command */
> > +     rcar_pci_write_reg(pcie->base, 0, H1_PCIEPHYDOUTR);
> > +     rcar_pci_write_reg(pcie->base, 0, H1_PCIEPHYADRR);
> > +
> > +     /* Ignore errors as they will be dealt with if the data link is down */
> > +     phy_wait_for_ack(pcie);
> > +}
> > +
> > +static int rcar_pcie_hw_init(struct rcar_pcie *pcie)
> > +{
> > +     int err;
> > +
> > +     /* Begin initialization */
> > +     rcar_pci_write_reg(pcie->base, 0, PCIETCTLR);
> > +
> > +     /* Set mode */
> > +     rcar_pci_write_reg(pcie->base, 1, PCIEMSR);
> > +
> > +     err = rcar_pcie_wait_for_phyrdy(pcie->base);
> > +     if (err)
> > +             return err;
> > +
> > +     /*
> > +      * Initial header for port config space is type 1, set the device
> > +      * class to match. Hardware takes care of propagating the IDSETR
> > +      * settings, so there is no need to bother with a quirk.
> > +      */
> > +     rcar_pci_write_reg(pcie->base, PCI_CLASS_BRIDGE_PCI << 16, IDSETR1);
> > +
> > +     /*
> > +      * Setup Secondary Bus Number & Subordinate Bus Number, even though
> > +      * they aren't used, to avoid bridge being detected as broken.
> > +      */
> > +     rcar_rmw32(pcie->base, RCONF(PCI_SECONDARY_BUS), 0xff, 1);
> > +     rcar_rmw32(pcie->base, RCONF(PCI_SUBORDINATE_BUS), 0xff, 1);
> > +
> > +     /* Initialize default capabilities. */
> > +     rcar_rmw32(pcie->base, REXPCAP(0), 0xff, PCI_CAP_ID_EXP);
> > +     rcar_rmw32(pcie->base, REXPCAP(PCI_EXP_FLAGS),
> > +                PCI_EXP_FLAGS_TYPE, PCI_EXP_TYPE_ROOT_PORT << 4);
> > +     rcar_rmw32(pcie->base, RCONF(PCI_HEADER_TYPE), 0x7f,
> > +                PCI_HEADER_TYPE_BRIDGE);
> > +
> > +     /* Enable data link layer active state reporting */
> > +     rcar_rmw32(pcie->base, REXPCAP(PCI_EXP_LNKCAP), PCI_EXP_LNKCAP_DLLLARC,
> > +                PCI_EXP_LNKCAP_DLLLARC);
> > +
> > +     /* Write out the physical slot number = 0 */
> > +     rcar_rmw32(pcie->base, REXPCAP(PCI_EXP_SLTCAP), PCI_EXP_SLTCAP_PSN, 0);
> > +
> > +     /* Set the completion timer timeout to the maximum 50ms. */
> > +     rcar_rmw32(pcie->base, TLCTLR + 1, 0x3f, 50);
> > +
> > +     /* Terminate list of capabilities (Next Capability Offset=0) */
> > +     rcar_rmw32(pcie->base, RVCCAP(0), 0xfff00000, 0);
> > +
> > +     /* Enable MSI */
> > +     if (IS_ENABLED(CONFIG_PCI_MSI))
> > +             rcar_pci_write_reg(pcie->base, 0x801f0000, PCIEMSITXR);
> > +
> > +     rcar_rmw32(pcie->base, MACCTLR, MACCTLR_RESERVED, 0);
> > +
> > +     /* Finish initialization - establish a PCI Express link */
> > +     rcar_pci_write_reg(pcie->base, CFINIT, PCIETCTLR);
> > +
> > +     /* This will timeout if we don't have a link. */
> > +     err = rcar_pcie_wait_for_dl(pcie->base);
> > +     if (err)
> > +             return err;
> > +
> > +     /* Enable INTx interrupts */
> > +     rcar_rmw32(pcie->base, PCIEINTXR, 0, 0xF << 8);
> > +
> > +     /* flush modifications */
> > +     wmb();
> > +
> > +     return 0;
> > +}
> > +
> > +static int rcar_pcie_phy_init_h1(struct rcar_pcie *pcie)
> > +{
> > +     /* Initialize the phy */
> > +     phy_write_reg(pcie, 0, 0x42, 0x1, 0x0EC34191);
> > +     phy_write_reg(pcie, 1, 0x42, 0x1, 0x0EC34180);
> > +     phy_write_reg(pcie, 0, 0x43, 0x1, 0x00210188);
> > +     phy_write_reg(pcie, 1, 0x43, 0x1, 0x00210188);
> > +     phy_write_reg(pcie, 0, 0x44, 0x1, 0x015C0014);
> > +     phy_write_reg(pcie, 1, 0x44, 0x1, 0x015C0014);
> > +     phy_write_reg(pcie, 1, 0x4C, 0x1, 0x786174A0);
> > +     phy_write_reg(pcie, 1, 0x4D, 0x1, 0x048000BB);
> > +     phy_write_reg(pcie, 0, 0x51, 0x1, 0x079EC062);
> > +     phy_write_reg(pcie, 0, 0x52, 0x1, 0x20000000);
> > +     phy_write_reg(pcie, 1, 0x52, 0x1, 0x20000000);
> > +     phy_write_reg(pcie, 1, 0x56, 0x1, 0x00003806);
> > +
> > +     phy_write_reg(pcie, 0, 0x60, 0x1, 0x004B03A5);
> > +     phy_write_reg(pcie, 0, 0x64, 0x1, 0x3F0F1F0F);
> > +     phy_write_reg(pcie, 0, 0x66, 0x1, 0x00008000);
> > +
> > +     return 0;
> > +}
> > +
> > +static int rcar_pcie_phy_init_gen2(struct rcar_pcie *pcie)
> > +{
> > +     /*
> > +      * These settings come from the R-Car Series, 2nd Generation User's
> > +      * Manual, section 50.3.1 (2) Initialization of the physical layer.
> > +      */
> > +     rcar_pci_write_reg(pcie->base, 0x000f0030, GEN2_PCIEPHYADDR);
> > +     rcar_pci_write_reg(pcie->base, 0x00381203, GEN2_PCIEPHYDATA);
> > +     rcar_pci_write_reg(pcie->base, 0x00000001, GEN2_PCIEPHYCTRL);
> > +     rcar_pci_write_reg(pcie->base, 0x00000006, GEN2_PCIEPHYCTRL);
> > +
> > +     rcar_pci_write_reg(pcie->base, 0x000f0054, GEN2_PCIEPHYADDR);
> > +     /* The following value is for DC connection, no termination resistor */
> > +     rcar_pci_write_reg(pcie->base, 0x13802007, GEN2_PCIEPHYDATA);
> > +     rcar_pci_write_reg(pcie->base, 0x00000001, GEN2_PCIEPHYCTRL);
> > +     rcar_pci_write_reg(pcie->base, 0x00000006, GEN2_PCIEPHYCTRL);
> > +
> > +     return 0;
> > +}
> > +
> > +static int rcar_pcie_phy_init_gen3(struct rcar_pcie *pcie)
> > +{
> > +     int err;
> > +
> > +     err = phy_init(pcie->phy);
> > +     if (err)
> > +             return err;
> > +
> > +     err = phy_power_on(pcie->phy);
> > +     if (err)
> > +             phy_exit(pcie->phy);
> > +
> > +     return err;
> > +}
> > +
> > +static int rcar_msi_alloc(struct rcar_msi *chip)
> > +{
> > +     int msi;
> > +
> > +     mutex_lock(&chip->lock);
> > +
> > +     msi = find_first_zero_bit(chip->used, INT_PCI_MSI_NR);
> > +     if (msi < INT_PCI_MSI_NR)
> > +             set_bit(msi, chip->used);
> > +     else
> > +             msi = -ENOSPC;
> > +
> > +     mutex_unlock(&chip->lock);
> > +
> > +     return msi;
> > +}
> > +
> > +static int rcar_msi_alloc_region(struct rcar_msi *chip, int no_irqs)
> > +{
> > +     int msi;
> > +
> > +     mutex_lock(&chip->lock);
> > +     msi = bitmap_find_free_region(chip->used, INT_PCI_MSI_NR,
> > +                                   order_base_2(no_irqs));
> > +     mutex_unlock(&chip->lock);
> > +
> > +     return msi;
> > +}
> > +
> > +static void rcar_msi_free(struct rcar_msi *chip, unsigned long irq)
> > +{
> > +     mutex_lock(&chip->lock);
> > +     clear_bit(irq, chip->used);
> > +     mutex_unlock(&chip->lock);
> > +}
> > +
> > +static irqreturn_t rcar_pcie_msi_irq(int irq, void *data)
> > +{
> > +     struct rcar_pcie *pcie = data;
> > +     struct rcar_msi *msi = &pcie->msi;
> > +     struct device *dev = pcie->dev;
> > +     unsigned long reg;
> > +
> > +     reg = rcar_pci_read_reg(pcie->base, PCIEMSIFR);
> > +
> > +     /* MSI & INTx share an interrupt - we only handle MSI here */
> > +     if (!reg)
> > +             return IRQ_NONE;
> > +
> > +     while (reg) {
> > +             unsigned int index = find_first_bit(&reg, 32);
> > +             unsigned int msi_irq;
> > +
> > +             /* clear the interrupt */
> > +             rcar_pci_write_reg(pcie->base, 1 << index, PCIEMSIFR);
> > +
> > +             msi_irq = irq_find_mapping(msi->domain, index);
> > +             if (msi_irq) {
> > +                     if (test_bit(index, msi->used))
> > +                             generic_handle_irq(msi_irq);
> > +                     else
> > +                             dev_info(dev, "unhandled MSI\n");
> > +             } else {
> > +                     /* Unknown MSI, just clear it */
> > +                     dev_dbg(dev, "unexpected MSI\n");
> > +             }
> > +
> > +             /* see if there's any more pending in this vector */
> > +             reg = rcar_pci_read_reg(pcie->base, PCIEMSIFR);
> > +     }
> > +
> > +     return IRQ_HANDLED;
> > +}
> > +
> > +static int rcar_msi_setup_irq(struct msi_controller *chip, struct pci_dev *pdev,
> > +                           struct msi_desc *desc)
> > +{
> > +     struct rcar_msi *msi = to_rcar_msi(chip);
> > +     struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
> > +     struct msi_msg msg;
> > +     unsigned int irq;
> > +     int hwirq;
> > +
> > +     hwirq = rcar_msi_alloc(msi);
> > +     if (hwirq < 0)
> > +             return hwirq;
> > +
> > +     irq = irq_find_mapping(msi->domain, hwirq);
> > +     if (!irq) {
> > +             rcar_msi_free(msi, hwirq);
> > +             return -EINVAL;
> > +     }
> > +
> > +     irq_set_msi_desc(irq, desc);
> > +
> > +     msg.address_lo = rcar_pci_read_reg(pcie->base, PCIEMSIALR) & ~MSIFE;
> > +     msg.address_hi = rcar_pci_read_reg(pcie->base, PCIEMSIAUR);
> > +     msg.data = hwirq;
> > +
> > +     pci_write_msi_msg(irq, &msg);
> > +
> > +     return 0;
> > +}
> > +
> > +static int rcar_msi_setup_irqs(struct msi_controller *chip,
> > +                            struct pci_dev *pdev, int nvec, int type)
> > +{
> > +     struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
> > +     struct rcar_msi *msi = to_rcar_msi(chip);
> > +     struct msi_desc *desc;
> > +     struct msi_msg msg;
> > +     unsigned int irq;
> > +     int hwirq;
> > +     int i;
> > +
> > +     /* MSI-X interrupts are not supported */
> > +     if (type == PCI_CAP_ID_MSIX)
> > +             return -EINVAL;
> > +
> > +     WARN_ON(!list_is_singular(&pdev->dev.msi_list));
> > +     desc = list_entry(pdev->dev.msi_list.next, struct msi_desc, list);
> > +
> > +     hwirq = rcar_msi_alloc_region(msi, nvec);
> > +     if (hwirq < 0)
> > +             return -ENOSPC;
> > +
> > +     irq = irq_find_mapping(msi->domain, hwirq);
> > +     if (!irq)
> > +             return -ENOSPC;
> > +
> > +     for (i = 0; i < nvec; i++) {
> > +             /*
> > +              * irq_create_mapping() called from rcar_pcie_probe() pre-
> > +              * allocates descs,  so there is no need to allocate descs here.
> > +              * We can therefore assume that if irq_find_mapping() above
> > +              * returns non-zero, then the descs are also successfully
> > +              * allocated.
> > +              */
> > +             if (irq_set_msi_desc_off(irq, i, desc)) {
> > +                     /* TODO: clear */
> > +                     return -EINVAL;
> > +             }
> > +     }
> > +
> > +     desc->nvec_used = nvec;
> > +     desc->msi_attrib.multiple = order_base_2(nvec);
> > +
> > +     msg.address_lo = rcar_pci_read_reg(pcie->base, PCIEMSIALR) & ~MSIFE;
> > +     msg.address_hi = rcar_pci_read_reg(pcie->base, PCIEMSIAUR);
> > +     msg.data = hwirq;
> > +
> > +     pci_write_msi_msg(irq, &msg);
> > +
> > +     return 0;
> > +}
> > +
> > +static void rcar_msi_teardown_irq(struct msi_controller *chip, unsigned int irq)
> > +{
> > +     struct irq_data *d = irq_get_irq_data(irq);
> > +     struct rcar_msi *msi = to_rcar_msi(chip);
> > +
> > +     rcar_msi_free(msi, d->hwirq);
> > +}
> > +
> > +static struct irq_chip rcar_msi_irq_chip = {
> > +     .name = "R-Car PCIe MSI",
> > +     .irq_enable = pci_msi_unmask_irq,
> > +     .irq_disable = pci_msi_mask_irq,
> > +     .irq_mask = pci_msi_mask_irq,
> > +     .irq_unmask = pci_msi_unmask_irq,
> > +};
> > +
> > +static int rcar_msi_map(struct irq_domain *domain, unsigned int irq,
> > +                     irq_hw_number_t hwirq)
> > +{
> > +     irq_set_chip_and_handler(irq, &rcar_msi_irq_chip, handle_simple_irq);
> > +     irq_set_chip_data(irq, domain->host_data);
> > +
> > +     return 0;
> > +}
> > +
> > +static const struct irq_domain_ops msi_domain_ops = {
> > +     .map = rcar_msi_map,
> > +};
> > +
> > +static void rcar_pcie_unmap_msi(struct rcar_pcie *pcie)
> > +{
> > +     struct rcar_msi *msi = &pcie->msi;
> > +     int i, irq;
> > +
> > +     for (i = 0; i < INT_PCI_MSI_NR; i++) {
> > +             irq = irq_find_mapping(msi->domain, i);
> > +             if (irq > 0)
> > +                     irq_dispose_mapping(irq);
> > +     }
> > +
> > +     irq_domain_remove(msi->domain);
> > +}
> > +
> > +static int rcar_pcie_enable_msi(struct rcar_pcie *pcie)
> > +{
> > +     struct rcar_msi *msi = &pcie->msi;
> > +     struct device *dev = pcie->dev;
> > +     phys_addr_t base;
> > +     int err, i;
> > +
> > +     mutex_init(&msi->lock);
> > +
> > +     msi->chip.dev = dev;
> > +     msi->chip.setup_irq = rcar_msi_setup_irq;
> > +     msi->chip.setup_irqs = rcar_msi_setup_irqs;
> > +     msi->chip.teardown_irq = rcar_msi_teardown_irq;
> > +
> > +     msi->domain = irq_domain_add_linear(dev->of_node, INT_PCI_MSI_NR,
> > +                                         &msi_domain_ops, &msi->chip);
> > +     if (!msi->domain) {
> > +             dev_err(dev, "failed to create IRQ domain\n");
> > +             return -ENOMEM;
> > +     }
> > +
> > +     for (i = 0; i < INT_PCI_MSI_NR; i++)
> > +             irq_create_mapping(msi->domain, i);
> > +
> > +     /* Two irqs are for MSI, but they are also used for non-MSI irqs */
> > +     err = devm_request_irq(dev, msi->irq1, rcar_pcie_msi_irq,
> > +                            IRQF_SHARED | IRQF_NO_THREAD,
> > +                            rcar_msi_irq_chip.name, pcie);
> > +     if (err < 0) {
> > +             dev_err(dev, "failed to request IRQ: %d\n", err);
> > +             goto err;
> > +     }
> > +
> > +     err = devm_request_irq(dev, msi->irq2, rcar_pcie_msi_irq,
> > +                            IRQF_SHARED | IRQF_NO_THREAD,
> > +                            rcar_msi_irq_chip.name, pcie);
> > +     if (err < 0) {
> > +             dev_err(dev, "failed to request IRQ: %d\n", err);
> > +             goto err;
> > +     }
> > +
> > +     /* setup MSI data target */
> > +     msi->pages = __get_free_pages(GFP_KERNEL, 0);
> > +     if (!msi->pages) {
> > +             err = -ENOMEM;
> > +             goto err;
> > +     }
> > +     base = virt_to_phys((void *)msi->pages);
> > +
> > +     rcar_pci_write_reg(pcie->base, lower_32_bits(base) | MSIFE, PCIEMSIALR);
> > +     rcar_pci_write_reg(pcie->base, upper_32_bits(base), PCIEMSIAUR);
> > +
> > +     /* enable all MSI interrupts */
> > +     rcar_pci_write_reg(pcie->base, 0xffffffff, PCIEMSIIER);
> > +
> > +     return 0;
> > +
> > +err:
> > +     rcar_pcie_unmap_msi(pcie);
> > +     return err;
> > +}
> > +
> > +static void rcar_pcie_teardown_msi(struct rcar_pcie *pcie)
> > +{
> > +     struct rcar_msi *msi = &pcie->msi;
> > +
> > +     /* Disable all MSI interrupts */
> > +     rcar_pci_write_reg(pcie->base, 0, PCIEMSIIER);
> > +
> > +     /* Disable address decoding of the MSI interrupt, MSIFE */
> > +     rcar_pci_write_reg(pcie->base, 0, PCIEMSIALR);
> > +
> > +     free_pages(msi->pages, 0);
> > +
> > +     rcar_pcie_unmap_msi(pcie);
> > +}
> > +
> > +static int rcar_pcie_get_resources(struct rcar_pcie *pcie)
> > +{
> > +     struct device *dev = pcie->dev;
> > +     struct resource res;
> > +     int err, i;
> > +
> > +     pcie->phy = devm_phy_optional_get(dev, "pcie");
> > +     if (IS_ERR(pcie->phy))
> > +             return PTR_ERR(pcie->phy);
> > +
> > +     err = of_address_to_resource(dev->of_node, 0, &res);
> > +     if (err)
> > +             return err;
> > +
> > +     pcie->base = devm_ioremap_resource(dev, &res);
> > +     if (IS_ERR(pcie->base))
> > +             return PTR_ERR(pcie->base);
> > +
> > +     pcie->bus_clk = devm_clk_get(dev, "pcie_bus");
> > +     if (IS_ERR(pcie->bus_clk)) {
> > +             dev_err(dev, "cannot get pcie bus clock\n");
> > +             return PTR_ERR(pcie->bus_clk);
> > +     }
> > +
> > +     i = irq_of_parse_and_map(dev->of_node, 0);
> > +     if (!i) {
> > +             dev_err(dev, "cannot get platform resources for msi interrupt\n");
> > +             err = -ENOENT;
> > +             goto err_irq1;
> > +     }
> > +     pcie->msi.irq1 = i;
> > +
> > +     i = irq_of_parse_and_map(dev->of_node, 1);
> > +     if (!i) {
> > +             dev_err(dev, "cannot get platform resources for msi interrupt\n");
> > +             err = -ENOENT;
> > +             goto err_irq2;
> > +     }
> > +     pcie->msi.irq2 = i;
> > +
> > +     return 0;
> > +
> > +err_irq2:
> > +     irq_dispose_mapping(pcie->msi.irq1);
> > +err_irq1:
> > +     return err;
> > +}
> > +
> > +static int rcar_pcie_inbound_ranges(struct rcar_pcie *pcie,
> > +                                 struct of_pci_range *range,
> > +                                 int *index)
> > +{
> > +     u64 restype = range->flags;
> > +     u64 cpu_addr = range->cpu_addr;
> > +     u64 cpu_end = range->cpu_addr + range->size;
> > +     u64 pci_addr = range->pci_addr;
> > +     u32 flags = LAM_64BIT | LAR_ENABLE;
> > +     u64 mask;
> > +     u64 size;
> > +     int idx = *index;
> > +
> > +     if (restype & IORESOURCE_PREFETCH)
> > +             flags |= LAM_PREFETCH;
> > +
> > +     /*
> > +      * If the size of the range is larger than the alignment of the start
> > +      * address, we have to use multiple entries to perform the mapping.
> > +      */
> > +     if (cpu_addr > 0) {
> > +             unsigned long nr_zeros = __ffs64(cpu_addr);
> > +             u64 alignment = 1ULL << nr_zeros;
> > +
> > +             size = min(range->size, alignment);
> > +     } else {
> > +             size = range->size;
> > +     }
> > +     /* Hardware supports max 4GiB inbound region */
> > +     size = min(size, 1ULL << 32);
> > +
> > +     mask = roundup_pow_of_two(size) - 1;
> > +     mask &= ~0xf;
> > +
> > +     while (cpu_addr < cpu_end) {
> > +             rcar_pcie_set_inbound(pcie->base, cpu_addr, pci_addr,
> > +                                   mask | flags, idx, true);
> > +             pci_addr += size;
> > +             cpu_addr += size;
> > +             idx += 2;
> > +
> > +             if (idx > MAX_NR_INBOUND_MAPS) {
> > +                     dev_err(pcie->dev, "Failed to map inbound regions!\n");
> > +                     return -EINVAL;
> > +             }
> > +     }
> > +     *index = idx;
> > +
> > +     return 0;
> > +}
> > +
> > +static int rcar_pcie_parse_map_dma_ranges(struct rcar_pcie *pcie,
> > +                                       struct device_node *np)
> > +{
> > +     struct of_pci_range_parser parser;
> > +     struct of_pci_range range;
> > +     int index = 0;
> > +     int err;
> > +
> > +     if (of_pci_dma_range_parser_init(&parser, np))
> > +             return -EINVAL;
> > +
> > +     /* Get the dma-ranges from DT */
> > +     for_each_of_pci_range(&parser, &range) {
> > +             u64 end = range.cpu_addr + range.size - 1;
> > +
> > +             dev_dbg(pcie->dev, "0x%08x 0x%016llx..0x%016llx -> 0x%016llx\n",
> > +                     range.flags, range.cpu_addr, end, range.pci_addr);
> > +
> > +             err = rcar_pcie_inbound_ranges(pcie, &range, &index);
> > +             if (err)
> > +                     return err;
> > +     }
> > +
> > +     return 0;
> > +}
> > +
> > +static const struct of_device_id rcar_pcie_of_match[] = {
> > +     { .compatible = "renesas,pcie-r8a7779",
> > +       .data = rcar_pcie_phy_init_h1 },
> > +     { .compatible = "renesas,pcie-r8a7790",
> > +       .data = rcar_pcie_phy_init_gen2 },
> > +     { .compatible = "renesas,pcie-r8a7791",
> > +       .data = rcar_pcie_phy_init_gen2 },
> > +     { .compatible = "renesas,pcie-rcar-gen2",
> > +       .data = rcar_pcie_phy_init_gen2 },
> > +     { .compatible = "renesas,pcie-r8a7795",
> > +       .data = rcar_pcie_phy_init_gen3 },
> > +     { .compatible = "renesas,pcie-rcar-gen3",
> > +       .data = rcar_pcie_phy_init_gen3 },
> > +     {},
> > +};
> > +
> > +static int rcar_pcie_probe(struct platform_device *pdev)
> > +{
> > +     int (*phy_init_fn)(struct rcar_pcie *pcie);
> > +     struct device *dev = &pdev->dev;
> > +     struct pci_host_bridge *bridge;
> > +     struct rcar_pcie *pcie;
> > +     u32 data;
> > +     int err;
> > +
> > +     bridge = pci_alloc_host_bridge(sizeof(*pcie));
> > +     if (!bridge)
> > +             return -ENOMEM;
> > +
> > +     pcie = pci_host_bridge_priv(bridge);
> > +
> > +     pcie->dev = dev;
> > +     platform_set_drvdata(pdev, pcie);
> > +
> > +     err = pci_parse_request_of_pci_ranges(dev, &pcie->resources, NULL);
> > +     if (err)
> > +             goto err_free_bridge;
> > +
> > +     pm_runtime_enable(pcie->dev);
> > +     err = pm_runtime_get_sync(pcie->dev);
> > +     if (err < 0) {
> > +             dev_err(pcie->dev, "pm_runtime_get_sync failed\n");
> > +             goto err_pm_disable;
> > +     }
> > +
> > +     err = rcar_pcie_get_resources(pcie);
> > +     if (err < 0) {
> > +             dev_err(dev, "failed to request resources: %d\n", err);
> > +             goto err_pm_put;
> > +     }
> > +
> > +     err = clk_prepare_enable(pcie->bus_clk);
> > +     if (err) {
> > +             dev_err(dev, "failed to enable bus clock: %d\n", err);
> > +             goto err_unmap_msi_irqs;
> > +     }
> > +
> > +     err = rcar_pcie_parse_map_dma_ranges(pcie, dev->of_node);
> > +     if (err)
> > +             goto err_clk_disable;
> > +
> > +     phy_init_fn = of_device_get_match_data(dev);
> > +     err = phy_init_fn(pcie);
> > +     if (err) {
> > +             dev_err(dev, "failed to init PCIe PHY\n");
> > +             goto err_clk_disable;
> > +     }
> > +
> > +     /* Failure to get a link might just be that no cards are inserted */
> > +     if (rcar_pcie_hw_init(pcie)) {
> > +             dev_info(dev, "PCIe link down\n");
> > +             err = -ENODEV;
> > +             goto err_phy_shutdown;
> > +     }
> > +
> > +     data = rcar_pci_read_reg(pcie->base, MACSR);
> > +     dev_info(dev, "PCIe x%d: link up\n", (data >> 20) & 0x3f);
> > +
> > +     if (IS_ENABLED(CONFIG_PCI_MSI)) {
> > +             err = rcar_pcie_enable_msi(pcie);
> > +             if (err < 0) {
> > +                     dev_err(dev,
> > +                             "failed to enable MSI support: %d\n",
> > +                             err);
> > +                     goto err_phy_shutdown;
> > +             }
> > +     }
> > +
> > +     err = rcar_pcie_enable(pcie);
> > +     if (err)
> > +             goto err_msi_teardown;
> > +
> > +     return 0;
> > +
> > +err_msi_teardown:
> > +     if (IS_ENABLED(CONFIG_PCI_MSI))
> > +             rcar_pcie_teardown_msi(pcie);
> > +
> > +err_phy_shutdown:
> > +     if (pcie->phy) {
> > +             phy_power_off(pcie->phy);
> > +             phy_exit(pcie->phy);
> > +     }
> > +
> > +err_clk_disable:
> > +     clk_disable_unprepare(pcie->bus_clk);
> > +
> > +err_unmap_msi_irqs:
> > +     irq_dispose_mapping(pcie->msi.irq2);
> > +     irq_dispose_mapping(pcie->msi.irq1);
> > +
> > +err_pm_put:
> > +     pm_runtime_put(dev);
> > +
> > +err_pm_disable:
> > +     pm_runtime_disable(dev);
> > +     pci_free_resource_list(&pcie->resources);
> > +
> > +err_free_bridge:
> > +     pci_free_host_bridge(bridge);
> > +
> > +     return err;
> > +}
> > +
> > +static int rcar_pcie_resume_noirq(struct device *dev)
> > +{
> > +     struct rcar_pcie *pcie = dev_get_drvdata(dev);
> > +
> > +     if (rcar_pci_read_reg(pcie->base, PMSR) &&
> > +         !(rcar_pci_read_reg(pcie->base, PCIETCTLR) & DL_DOWN))
> > +             return 0;
> > +
> > +     /* Re-establish the PCIe link */
> > +     rcar_rmw32(pcie->base, MACCTLR, MACCTLR_RESERVED, 0);
> > +     rcar_pci_write_reg(pcie->base, CFINIT, PCIETCTLR);
> > +     return rcar_pcie_wait_for_dl(pcie->base);
> > +}
> > +
> > +static const struct dev_pm_ops rcar_pcie_pm_ops = {
> > +     .resume_noirq = rcar_pcie_resume_noirq,
> > +};
> > +
> > +static struct platform_driver rcar_pcie_driver = {
> > +     .driver = {
> > +             .name = "rcar-pcie",
> > +             .of_match_table = rcar_pcie_of_match,
> > +             .pm = &rcar_pcie_pm_ops,
> > +             .suppress_bind_attrs = true,
> > +     },
> > +     .probe = rcar_pcie_probe,
> > +};
> > +builtin_platform_driver(rcar_pcie_driver);
> > diff --git a/drivers/pci/controller/pcie-rcar.c b/drivers/pci/controller/pcie-rcar.c
> > index e45bb2a7bfa5..1008ae6a896c 100644
> > --- a/drivers/pci/controller/pcie-rcar.c
> > +++ b/drivers/pci/controller/pcie-rcar.c
> > @@ -1,7 +1,7 @@
> >  // SPDX-License-Identifier: GPL-2.0
> >  /*
> >   * PCIe driver for Renesas R-Car SoCs
> > - *  Copyright (C) 2014 Renesas Electronics Europe Ltd
> > + *  Copyright (C) 2014-2019 Renesas Electronics Europe Ltd
> >   *
> >   * Based on:
> >   *  arch/sh/drivers/pci/pcie-sh7786.c
> > @@ -11,534 +11,80 @@
> >   * Author: Phil Edworthy <phil.edworthy@renesas.com>
> >   */
> >
> > -#include <linux/bitops.h>
> > -#include <linux/clk.h>
> >  #include <linux/delay.h>
> > -#include <linux/interrupt.h>
> > -#include <linux/irq.h>
> > -#include <linux/irqdomain.h>
> > -#include <linux/kernel.h>
> > -#include <linux/init.h>
> > -#include <linux/msi.h>
> > -#include <linux/of_address.h>
> > -#include <linux/of_irq.h>
> > -#include <linux/of_pci.h>
> > -#include <linux/of_platform.h>
> >  #include <linux/pci.h>
> > -#include <linux/phy/phy.h>
> > -#include <linux/platform_device.h>
> > -#include <linux/pm_runtime.h>
> > -#include <linux/slab.h>
> >
> > -#define PCIECAR                      0x000010
> > -#define PCIECCTLR            0x000018
> > -#define  CONFIG_SEND_ENABLE  BIT(31)
> > -#define  TYPE0                       (0 << 8)
> > -#define  TYPE1                       BIT(8)
> > -#define PCIECDR                      0x000020
> > -#define PCIEMSR                      0x000028
> > -#define PCIEINTXR            0x000400
> > -#define PCIEPHYSR            0x0007f0
> > -#define  PHYRDY                      BIT(0)
> > -#define PCIEMSITXR           0x000840
> > +#include "pcie-rcar.h"
> >
> > -/* Transfer control */
> > -#define PCIETCTLR            0x02000
> > -#define  DL_DOWN             BIT(3)
> > -#define  CFINIT                      BIT(0)
> > -#define PCIETSTR             0x02004
> > -#define  DATA_LINK_ACTIVE    BIT(0)
> > -#define PCIEERRFR            0x02020
> > -#define  UNSUPPORTED_REQUEST BIT(4)
> > -#define PCIEMSIFR            0x02044
> > -#define PCIEMSIALR           0x02048
> > -#define  MSIFE                       BIT(0)
> > -#define PCIEMSIAUR           0x0204c
> > -#define PCIEMSIIER           0x02050
> > -
> > -/* root port address */
> > -#define PCIEPRAR(x)          (0x02080 + ((x) * 0x4))
> > -
> > -/* local address reg & mask */
> > -#define PCIELAR(x)           (0x02200 + ((x) * 0x20))
> > -#define PCIELAMR(x)          (0x02208 + ((x) * 0x20))
> > -#define  LAM_PREFETCH                BIT(3)
> > -#define  LAM_64BIT           BIT(2)
> > -#define  LAR_ENABLE          BIT(1)
> > -
> > -/* PCIe address reg & mask */
> > -#define PCIEPALR(x)          (0x03400 + ((x) * 0x20))
> > -#define PCIEPAUR(x)          (0x03404 + ((x) * 0x20))
> > -#define PCIEPAMR(x)          (0x03408 + ((x) * 0x20))
> > -#define PCIEPTCTLR(x)                (0x0340c + ((x) * 0x20))
> > -#define  PAR_ENABLE          BIT(31)
> > -#define  IO_SPACE            BIT(8)
> > -
> > -/* Configuration */
> > -#define PCICONF(x)           (0x010000 + ((x) * 0x4))
> > -#define PMCAP(x)             (0x010040 + ((x) * 0x4))
> > -#define EXPCAP(x)            (0x010070 + ((x) * 0x4))
> > -#define VCCAP(x)             (0x010100 + ((x) * 0x4))
> > -
> > -/* link layer */
> > -#define IDSETR1                      0x011004
> > -#define TLCTLR                       0x011048
> > -#define MACSR                        0x011054
> > -#define  SPCHGFIN            BIT(4)
> > -#define  SPCHGFAIL           BIT(6)
> > -#define  SPCHGSUC            BIT(7)
> > -#define  LINK_SPEED          (0xf << 16)
> > -#define  LINK_SPEED_2_5GTS   (1 << 16)
> > -#define  LINK_SPEED_5_0GTS   (2 << 16)
> > -#define MACCTLR                      0x011058
> > -#define  MACCTLR_RESERVED    BIT(0)
> > -#define  SPEED_CHANGE                BIT(24)
> > -#define  SCRAMBLE_DISABLE    BIT(27)
> > -#define PMSR                 0x01105c
> > -#define MACS2R                       0x011078
> > -#define MACCGSPSETR          0x011084
> > -#define  SPCNGRSN            BIT(31)
> > -
> > -/* R-Car H1 PHY */
> > -#define H1_PCIEPHYADRR               0x04000c
> > -#define  WRITE_CMD           BIT(16)
> > -#define  PHY_ACK             BIT(24)
> > -#define  RATE_POS            12
> > -#define  LANE_POS            8
> > -#define  ADR_POS             0
> > -#define H1_PCIEPHYDOUTR              0x040014
> > -
> > -/* R-Car Gen2 PHY */
> > -#define GEN2_PCIEPHYADDR     0x780
> > -#define GEN2_PCIEPHYDATA     0x784
> > -#define GEN2_PCIEPHYCTRL     0x78c
> > -
> > -#define INT_PCI_MSI_NR               32
> > -
> > -#define RCONF(x)             (PCICONF(0) + (x))
> > -#define RPMCAP(x)            (PMCAP(0) + (x))
> > -#define REXPCAP(x)           (EXPCAP(0) + (x))
> > -#define RVCCAP(x)            (VCCAP(0) + (x))
> > -
> > -#define PCIE_CONF_BUS(b)     (((b) & 0xff) << 24)
> > -#define PCIE_CONF_DEV(d)     (((d) & 0x1f) << 19)
> > -#define PCIE_CONF_FUNC(f)    (((f) & 0x7) << 16)
> > -
> > -#define RCAR_PCI_MAX_RESOURCES       4
> > -#define MAX_NR_INBOUND_MAPS  6
> > -
> > -struct rcar_msi {
> > -     DECLARE_BITMAP(used, INT_PCI_MSI_NR);
> > -     struct irq_domain *domain;
> > -     struct msi_controller chip;
> > -     unsigned long pages;
> > -     struct mutex lock;
> > -     int irq1;
> > -     int irq2;
> > -};
> > -
> > -static inline struct rcar_msi *to_rcar_msi(struct msi_controller *chip)
> > -{
> > -     return container_of(chip, struct rcar_msi, chip);
> > -}
> > -
> > -/* Structure representing the PCIe interface */
> > -struct rcar_pcie {
> > -     struct device           *dev;
> > -     struct phy              *phy;
> > -     void __iomem            *base;
> > -     struct list_head        resources;
> > -     int                     root_bus_nr;
> > -     struct clk              *bus_clk;
> > -     struct                  rcar_msi msi;
> > -};
> > -
> > -static void rcar_pci_write_reg(struct rcar_pcie *pcie, u32 val,
> > -                            unsigned int reg)
> > +void rcar_pci_write_reg(void __iomem *base, u32 val, unsigned int reg)
> >  {
> > -     writel(val, pcie->base + reg);
> > +     writel(val, base + reg);
> >  }
> >
> > -static u32 rcar_pci_read_reg(struct rcar_pcie *pcie, unsigned int reg)
> > +u32 rcar_pci_read_reg(void __iomem *base, unsigned int reg)
> >  {
> > -     return readl(pcie->base + reg);
> > +     return readl(base + reg);
> >  }
> >
> > -enum {
> > -     RCAR_PCI_ACCESS_READ,
> > -     RCAR_PCI_ACCESS_WRITE,
> > -};
> > -
> > -static void rcar_rmw32(struct rcar_pcie *pcie, int where, u32 mask, u32 data)
> > +void rcar_rmw32(void __iomem *base, int where, u32 mask, u32 data)
> >  {
> >       unsigned int shift = BITS_PER_BYTE * (where & 3);
> > -     u32 val = rcar_pci_read_reg(pcie, where & ~3);
> > +     u32 val = rcar_pci_read_reg(base, where & ~3);
> >
> >       val &= ~(mask << shift);
> >       val |= data << shift;
> > -     rcar_pci_write_reg(pcie, val, where & ~3);
> > -}
> > -
> > -static u32 rcar_read_conf(struct rcar_pcie *pcie, int where)
> > -{
> > -     unsigned int shift = BITS_PER_BYTE * (where & 3);
> > -     u32 val = rcar_pci_read_reg(pcie, where & ~3);
> > -
> > -     return val >> shift;
> > +     rcar_pci_write_reg(base, val, where & ~3);
> >  }
> >
> > -/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
> > -static int rcar_pcie_config_access(struct rcar_pcie *pcie,
> > -             unsigned char access_type, struct pci_bus *bus,
> > -             unsigned int devfn, int where, u32 *data)
> > -{
> > -     unsigned int dev, func, reg, index;
> > -
> > -     dev = PCI_SLOT(devfn);
> > -     func = PCI_FUNC(devfn);
> > -     reg = where & ~3;
> > -     index = reg / 4;
> > -
> > -     /*
> > -      * While each channel has its own memory-mapped extended config
> > -      * space, it's generally only accessible when in endpoint mode.
> > -      * When in root complex mode, the controller is unable to target
> > -      * itself with either type 0 or type 1 accesses, and indeed, any
> > -      * controller initiated target transfer to its own config space
> > -      * result in a completer abort.
> > -      *
> > -      * Each channel effectively only supports a single device, but as
> > -      * the same channel <-> device access works for any PCI_SLOT()
> > -      * value, we cheat a bit here and bind the controller's config
> > -      * space to devfn 0 in order to enable self-enumeration. In this
> > -      * case the regular ECAR/ECDR path is sidelined and the mangled
> > -      * config access itself is initiated as an internal bus transaction.
> > -      */
> > -     if (pci_is_root_bus(bus)) {
> > -             if (dev != 0)
> > -                     return PCIBIOS_DEVICE_NOT_FOUND;
> > -
> > -             if (access_type == RCAR_PCI_ACCESS_READ) {
> > -                     *data = rcar_pci_read_reg(pcie, PCICONF(index));
> > -             } else {
> > -                     /* Keep an eye out for changes to the root bus number */
> > -                     if (pci_is_root_bus(bus) && (reg == PCI_PRIMARY_BUS))
> > -                             pcie->root_bus_nr = *data & 0xff;
> > -
> > -                     rcar_pci_write_reg(pcie, *data, PCICONF(index));
> > -             }
> > -
> > -             return PCIBIOS_SUCCESSFUL;
> > -     }
> > -
> > -     if (pcie->root_bus_nr < 0)
> > -             return PCIBIOS_DEVICE_NOT_FOUND;
> > -
> > -     /* Clear errors */
> > -     rcar_pci_write_reg(pcie, rcar_pci_read_reg(pcie, PCIEERRFR), PCIEERRFR);
> > -
> > -     /* Set the PIO address */
> > -     rcar_pci_write_reg(pcie, PCIE_CONF_BUS(bus->number) |
> > -             PCIE_CONF_DEV(dev) | PCIE_CONF_FUNC(func) | reg, PCIECAR);
> > -
> > -     /* Enable the configuration access */
> > -     if (bus->parent->number == pcie->root_bus_nr)
> > -             rcar_pci_write_reg(pcie, CONFIG_SEND_ENABLE | TYPE0, PCIECCTLR);
> > -     else
> > -             rcar_pci_write_reg(pcie, CONFIG_SEND_ENABLE | TYPE1, PCIECCTLR);
> > -
> > -     /* Check for errors */
> > -     if (rcar_pci_read_reg(pcie, PCIEERRFR) & UNSUPPORTED_REQUEST)
> > -             return PCIBIOS_DEVICE_NOT_FOUND;
> > -
> > -     /* Check for master and target aborts */
> > -     if (rcar_read_conf(pcie, RCONF(PCI_STATUS)) &
> > -             (PCI_STATUS_REC_MASTER_ABORT | PCI_STATUS_REC_TARGET_ABORT))
> > -             return PCIBIOS_DEVICE_NOT_FOUND;
> > -
> > -     if (access_type == RCAR_PCI_ACCESS_READ)
> > -             *data = rcar_pci_read_reg(pcie, PCIECDR);
> > -     else
> > -             rcar_pci_write_reg(pcie, *data, PCIECDR);
> > -
> > -     /* Disable the configuration access */
> > -     rcar_pci_write_reg(pcie, 0, PCIECCTLR);
> > -
> > -     return PCIBIOS_SUCCESSFUL;
> > -}
> > -
> > -static int rcar_pcie_read_conf(struct pci_bus *bus, unsigned int devfn,
> > -                            int where, int size, u32 *val)
> > -{
> > -     struct rcar_pcie *pcie = bus->sysdata;
> > -     int ret;
> > -
> > -     ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
> > -                                   bus, devfn, where, val);
> > -     if (ret != PCIBIOS_SUCCESSFUL) {
> > -             *val = 0xffffffff;
> > -             return ret;
> > -     }
> > -
> > -     if (size == 1)
> > -             *val = (*val >> (BITS_PER_BYTE * (where & 3))) & 0xff;
> > -     else if (size == 2)
> > -             *val = (*val >> (BITS_PER_BYTE * (where & 2))) & 0xffff;
> > -
> > -     dev_dbg(&bus->dev, "pcie-config-read: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
> > -             bus->number, devfn, where, size, *val);
> > -
> > -     return ret;
> > -}
> > -
> > -/* Serialization is provided by 'pci_lock' in drivers/pci/access.c */
> > -static int rcar_pcie_write_conf(struct pci_bus *bus, unsigned int devfn,
> > -                             int where, int size, u32 val)
> > -{
> > -     struct rcar_pcie *pcie = bus->sysdata;
> > -     unsigned int shift;
> > -     u32 data;
> > -     int ret;
> > -
> > -     ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_READ,
> > -                                   bus, devfn, where, &data);
> > -     if (ret != PCIBIOS_SUCCESSFUL)
> > -             return ret;
> > -
> > -     dev_dbg(&bus->dev, "pcie-config-write: bus=%3d devfn=0x%04x where=0x%04x size=%d val=0x%08x\n",
> > -             bus->number, devfn, where, size, val);
> > -
> > -     if (size == 1) {
> > -             shift = BITS_PER_BYTE * (where & 3);
> > -             data &= ~(0xff << shift);
> > -             data |= ((val & 0xff) << shift);
> > -     } else if (size == 2) {
> > -             shift = BITS_PER_BYTE * (where & 2);
> > -             data &= ~(0xffff << shift);
> > -             data |= ((val & 0xffff) << shift);
> > -     } else
> > -             data = val;
> > -
> > -     ret = rcar_pcie_config_access(pcie, RCAR_PCI_ACCESS_WRITE,
> > -                                   bus, devfn, where, &data);
> > -
> > -     return ret;
> > -}
> > -
> > -static struct pci_ops rcar_pcie_ops = {
> > -     .read   = rcar_pcie_read_conf,
> > -     .write  = rcar_pcie_write_conf,
> > -};
> > -
> > -static void rcar_pcie_setup_window(int win, struct rcar_pcie *pcie,
> > -                                struct resource *res)
> > +void rcar_pcie_set_outbound(int win, void __iomem *base,
> > +                         struct resource *res, bool host)
> >  {
> >       /* Setup PCIe address space mappings for each resource */
> > -     resource_size_t size;
> > -     resource_size_t res_start;
> > -     u32 mask;
> > -
> > -     rcar_pci_write_reg(pcie, 0x00000000, PCIEPTCTLR(win));
> > +     resource_size_t size = 0;
> > +     resource_size_t res_start = 0;
> > +     u32 mask = 0x0;
> >
> > +     rcar_pci_write_reg(base, mask, PCIEPTCTLR(win));
> >       /*
> >        * The PAMR mask is calculated in units of 128Bytes, which
> >        * keeps things pretty simple.
> >        */
> >       size = resource_size(res);
> > -     mask = (roundup_pow_of_two(size) / SZ_128) - 1;
> > -     rcar_pci_write_reg(pcie, mask << 7, PCIEPAMR(win));
> > +     if (size > 128)
> > +             mask = (roundup_pow_of_two(size) / SZ_128) - 1;
> > +     rcar_pci_write_reg(base, mask << 7, PCIEPAMR(win));
> >
> > -     if (res->flags & IORESOURCE_IO)
> > -             res_start = pci_pio_to_address(res->start);
> > -     else
> > +     if (!host) {
> >               res_start = res->start;
> > +     } else {
> > +             if (res->flags & IORESOURCE_IO)
> > +                     res_start = pci_pio_to_address(res->start);
> > +             else
> > +                     res_start = res->start;
> > +     }
> >
> > -     rcar_pci_write_reg(pcie, upper_32_bits(res_start), PCIEPAUR(win));
> > -     rcar_pci_write_reg(pcie, lower_32_bits(res_start) & ~0x7F,
> > +     rcar_pci_write_reg(base, upper_32_bits(res_start), PCIEPAUR(win));
> > +     rcar_pci_write_reg(base, lower_32_bits(res_start) & ~0x7F,
> >                          PCIEPALR(win));
> >
> > -     /* First resource is for IO */
> > -     mask = PAR_ENABLE;
> > +     mask = 0x0;
> > +     if (res->start)
> > +             mask = PAR_ENABLE;
> > +
> >       if (res->flags & IORESOURCE_IO)
> >               mask |= IO_SPACE;
> >
> > -     rcar_pci_write_reg(pcie, mask, PCIEPTCTLR(win));
> > -}
> > -
> > -static int rcar_pcie_setup(struct list_head *resource, struct rcar_pcie *pci)
> > -{
> > -     struct resource_entry *win;
> > -     int i = 0;
> > -
> > -     /* Setup PCI resources */
> > -     resource_list_for_each_entry(win, &pci->resources) {
> > -             struct resource *res = win->res;
> > -
> > -             if (!res->flags)
> > -                     continue;
> > -
> > -             switch (resource_type(res)) {
> > -             case IORESOURCE_IO:
> > -             case IORESOURCE_MEM:
> > -                     rcar_pcie_setup_window(i, pci, res);
> > -                     i++;
> > -                     break;
> > -             case IORESOURCE_BUS:
> > -                     pci->root_bus_nr = res->start;
> > -                     break;
> > -             default:
> > -                     continue;
> > -             }
> > -
> > -             pci_add_resource(resource, res);
> > -     }
> > -
> > -     return 1;
> > -}
> > -
> > -static void rcar_pcie_force_speedup(struct rcar_pcie *pcie)
> > -{
> > -     struct device *dev = pcie->dev;
> > -     unsigned int timeout = 1000;
> > -     u32 macsr;
> > -
> > -     if ((rcar_pci_read_reg(pcie, MACS2R) & LINK_SPEED) != LINK_SPEED_5_0GTS)
> > -             return;
> > -
> > -     if (rcar_pci_read_reg(pcie, MACCTLR) & SPEED_CHANGE) {
> > -             dev_err(dev, "Speed change already in progress\n");
> > -             return;
> > -     }
> > -
> > -     macsr = rcar_pci_read_reg(pcie, MACSR);
> > -     if ((macsr & LINK_SPEED) == LINK_SPEED_5_0GTS)
> > -             goto done;
> > -
> > -     /* Set target link speed to 5.0 GT/s */
> > -     rcar_rmw32(pcie, EXPCAP(12), PCI_EXP_LNKSTA_CLS,
> > -                PCI_EXP_LNKSTA_CLS_5_0GB);
> > -
> > -     /* Set speed change reason as intentional factor */
> > -     rcar_rmw32(pcie, MACCGSPSETR, SPCNGRSN, 0);
> > -
> > -     /* Clear SPCHGFIN, SPCHGSUC, and SPCHGFAIL */
> > -     if (macsr & (SPCHGFIN | SPCHGSUC | SPCHGFAIL))
> > -             rcar_pci_write_reg(pcie, macsr, MACSR);
> > -
> > -     /* Start link speed change */
> > -     rcar_rmw32(pcie, MACCTLR, SPEED_CHANGE, SPEED_CHANGE);
> > -
> > -     while (timeout--) {
> > -             macsr = rcar_pci_read_reg(pcie, MACSR);
> > -             if (macsr & SPCHGFIN) {
> > -                     /* Clear the interrupt bits */
> > -                     rcar_pci_write_reg(pcie, macsr, MACSR);
> > -
> > -                     if (macsr & SPCHGFAIL)
> > -                             dev_err(dev, "Speed change failed\n");
> > -
> > -                     goto done;
> > -             }
> > -
> > -             msleep(1);
> > -     }
> > -
> > -     dev_err(dev, "Speed change timed out\n");
> > -
> > -done:
> > -     dev_info(dev, "Current link speed is %s GT/s\n",
> > -              (macsr & LINK_SPEED) == LINK_SPEED_5_0GTS ? "5" : "2.5");
> > -}
> > -
> > -static int rcar_pcie_enable(struct rcar_pcie *pcie)
> > -{
> > -     struct device *dev = pcie->dev;
> > -     struct pci_host_bridge *bridge = pci_host_bridge_from_priv(pcie);
> > -     struct pci_bus *bus, *child;
> > -     int ret;
> > -
> > -     /* Try setting 5 GT/s link speed */
> > -     rcar_pcie_force_speedup(pcie);
> > -
> > -     rcar_pcie_setup(&bridge->windows, pcie);
> > -
> > -     pci_add_flags(PCI_REASSIGN_ALL_BUS);
> > -
> > -     bridge->dev.parent = dev;
> > -     bridge->sysdata = pcie;
> > -     bridge->busnr = pcie->root_bus_nr;
> > -     bridge->ops = &rcar_pcie_ops;
> > -     bridge->map_irq = of_irq_parse_and_map_pci;
> > -     bridge->swizzle_irq = pci_common_swizzle;
> > -     if (IS_ENABLED(CONFIG_PCI_MSI))
> > -             bridge->msi = &pcie->msi.chip;
> > -
> > -     ret = pci_scan_root_bus_bridge(bridge);
> > -     if (ret < 0)
> > -             return ret;
> > -
> > -     bus = bridge->bus;
> > -
> > -     pci_bus_size_bridges(bus);
> > -     pci_bus_assign_resources(bus);
> > -
> > -     list_for_each_entry(child, &bus->children, node)
> > -             pcie_bus_configure_settings(child);
> > -
> > -     pci_bus_add_devices(bus);
> > -
> > -     return 0;
> > -}
> > -
> > -static int phy_wait_for_ack(struct rcar_pcie *pcie)
> > -{
> > -     struct device *dev = pcie->dev;
> > -     unsigned int timeout = 100;
> > -
> > -     while (timeout--) {
> > -             if (rcar_pci_read_reg(pcie, H1_PCIEPHYADRR) & PHY_ACK)
> > -                     return 0;
> > -
> > -             udelay(100);
> > -     }
> > -
> > -     dev_err(dev, "Access to PCIe phy timed out\n");
> > -
> > -     return -ETIMEDOUT;
> > -}
> > -
> > -static void phy_write_reg(struct rcar_pcie *pcie,
> > -                       unsigned int rate, u32 addr,
> > -                       unsigned int lane, u32 data)
> > -{
> > -     u32 phyaddr;
> > -
> > -     phyaddr = WRITE_CMD |
> > -             ((rate & 1) << RATE_POS) |
> > -             ((lane & 0xf) << LANE_POS) |
> > -             ((addr & 0xff) << ADR_POS);
> > -
> > -     /* Set write data */
> > -     rcar_pci_write_reg(pcie, data, H1_PCIEPHYDOUTR);
> > -     rcar_pci_write_reg(pcie, phyaddr, H1_PCIEPHYADRR);
> > -
> > -     /* Ignore errors as they will be dealt with if the data link is down */
> > -     phy_wait_for_ack(pcie);
> > -
> > -     /* Clear command */
> > -     rcar_pci_write_reg(pcie, 0, H1_PCIEPHYDOUTR);
> > -     rcar_pci_write_reg(pcie, 0, H1_PCIEPHYADRR);
> > -
> > -     /* Ignore errors as they will be dealt with if the data link is down */
> > -     phy_wait_for_ack(pcie);
> > +     rcar_pci_write_reg(base, mask, PCIEPTCTLR(win));
> > +     /* flush modifications */
> > +     wmb();
> >  }
> >
> > -static int rcar_pcie_wait_for_phyrdy(struct rcar_pcie *pcie)
> > +int rcar_pcie_wait_for_phyrdy(void __iomem *base)
> >  {
> >       unsigned int timeout = 10;
> >
> >       while (timeout--) {
> > -             if (rcar_pci_read_reg(pcie, PCIEPHYSR) & PHYRDY)
> > +             if (rcar_pci_read_reg(base, PCIEPHYSR) & PHYRDY)
> >                       return 0;
> >
> >               msleep(5);
> > @@ -547,12 +93,12 @@ static int rcar_pcie_wait_for_phyrdy(struct rcar_pcie *pcie)
> >       return -ETIMEDOUT;
> >  }
> >
> > -static int rcar_pcie_wait_for_dl(struct rcar_pcie *pcie)
> > +int rcar_pcie_wait_for_dl(void __iomem *base)
> >  {
> >       unsigned int timeout = 10000;
> >
> >       while (timeout--) {
> > -             if ((rcar_pci_read_reg(pcie, PCIETSTR) & DATA_LINK_ACTIVE))
> > +             if ((rcar_pci_read_reg(base, PCIETSTR) & DATA_LINK_ACTIVE))
> >                       return 0;
> >
> >               udelay(5);
> > @@ -562,698 +108,27 @@ static int rcar_pcie_wait_for_dl(struct rcar_pcie *pcie)
> >       return -ETIMEDOUT;
> >  }
> >
> > -static int rcar_pcie_hw_init(struct rcar_pcie *pcie)
> > -{
> > -     int err;
> > -
> > -     /* Begin initialization */
> > -     rcar_pci_write_reg(pcie, 0, PCIETCTLR);
> > -
> > -     /* Set mode */
> > -     rcar_pci_write_reg(pcie, 1, PCIEMSR);
> > -
> > -     err = rcar_pcie_wait_for_phyrdy(pcie);
> > -     if (err)
> > -             return err;
> > -
> > -     /*
> > -      * Initial header for port config space is type 1, set the device
> > -      * class to match. Hardware takes care of propagating the IDSETR
> > -      * settings, so there is no need to bother with a quirk.
> > -      */
> > -     rcar_pci_write_reg(pcie, PCI_CLASS_BRIDGE_PCI << 16, IDSETR1);
> > -
> > -     /*
> > -      * Setup Secondary Bus Number & Subordinate Bus Number, even though
> > -      * they aren't used, to avoid bridge being detected as broken.
> > -      */
> > -     rcar_rmw32(pcie, RCONF(PCI_SECONDARY_BUS), 0xff, 1);
> > -     rcar_rmw32(pcie, RCONF(PCI_SUBORDINATE_BUS), 0xff, 1);
> > -
> > -     /* Initialize default capabilities. */
> > -     rcar_rmw32(pcie, REXPCAP(0), 0xff, PCI_CAP_ID_EXP);
> > -     rcar_rmw32(pcie, REXPCAP(PCI_EXP_FLAGS),
> > -             PCI_EXP_FLAGS_TYPE, PCI_EXP_TYPE_ROOT_PORT << 4);
> > -     rcar_rmw32(pcie, RCONF(PCI_HEADER_TYPE), 0x7f,
> > -             PCI_HEADER_TYPE_BRIDGE);
> > -
> > -     /* Enable data link layer active state reporting */
> > -     rcar_rmw32(pcie, REXPCAP(PCI_EXP_LNKCAP), PCI_EXP_LNKCAP_DLLLARC,
> > -             PCI_EXP_LNKCAP_DLLLARC);
> > -
> > -     /* Write out the physical slot number = 0 */
> > -     rcar_rmw32(pcie, REXPCAP(PCI_EXP_SLTCAP), PCI_EXP_SLTCAP_PSN, 0);
> > -
> > -     /* Set the completion timer timeout to the maximum 50ms. */
> > -     rcar_rmw32(pcie, TLCTLR + 1, 0x3f, 50);
> > -
> > -     /* Terminate list of capabilities (Next Capability Offset=0) */
> > -     rcar_rmw32(pcie, RVCCAP(0), 0xfff00000, 0);
> > -
> > -     /* Enable MSI */
> > -     if (IS_ENABLED(CONFIG_PCI_MSI))
> > -             rcar_pci_write_reg(pcie, 0x801f0000, PCIEMSITXR);
> > -
> > -     rcar_rmw32(pcie, MACCTLR, MACCTLR_RESERVED, 0);
> > -
> > -     /* Finish initialization - establish a PCI Express link */
> > -     rcar_pci_write_reg(pcie, CFINIT, PCIETCTLR);
> > -
> > -     /* This will timeout if we don't have a link. */
> > -     err = rcar_pcie_wait_for_dl(pcie);
> > -     if (err)
> > -             return err;
> > -
> > -     /* Enable INTx interrupts */
> > -     rcar_rmw32(pcie, PCIEINTXR, 0, 0xF << 8);
> > -
> > -     wmb();
> > -
> > -     return 0;
> > -}
> > -
> > -static int rcar_pcie_phy_init_h1(struct rcar_pcie *pcie)
> > -{
> > -     /* Initialize the phy */
> > -     phy_write_reg(pcie, 0, 0x42, 0x1, 0x0EC34191);
> > -     phy_write_reg(pcie, 1, 0x42, 0x1, 0x0EC34180);
> > -     phy_write_reg(pcie, 0, 0x43, 0x1, 0x00210188);
> > -     phy_write_reg(pcie, 1, 0x43, 0x1, 0x00210188);
> > -     phy_write_reg(pcie, 0, 0x44, 0x1, 0x015C0014);
> > -     phy_write_reg(pcie, 1, 0x44, 0x1, 0x015C0014);
> > -     phy_write_reg(pcie, 1, 0x4C, 0x1, 0x786174A0);
> > -     phy_write_reg(pcie, 1, 0x4D, 0x1, 0x048000BB);
> > -     phy_write_reg(pcie, 0, 0x51, 0x1, 0x079EC062);
> > -     phy_write_reg(pcie, 0, 0x52, 0x1, 0x20000000);
> > -     phy_write_reg(pcie, 1, 0x52, 0x1, 0x20000000);
> > -     phy_write_reg(pcie, 1, 0x56, 0x1, 0x00003806);
> > -
> > -     phy_write_reg(pcie, 0, 0x60, 0x1, 0x004B03A5);
> > -     phy_write_reg(pcie, 0, 0x64, 0x1, 0x3F0F1F0F);
> > -     phy_write_reg(pcie, 0, 0x66, 0x1, 0x00008000);
> > -
> > -     return 0;
> > -}
> > -
> > -static int rcar_pcie_phy_init_gen2(struct rcar_pcie *pcie)
> > -{
> > -     /*
> > -      * These settings come from the R-Car Series, 2nd Generation User's
> > -      * Manual, section 50.3.1 (2) Initialization of the physical layer.
> > -      */
> > -     rcar_pci_write_reg(pcie, 0x000f0030, GEN2_PCIEPHYADDR);
> > -     rcar_pci_write_reg(pcie, 0x00381203, GEN2_PCIEPHYDATA);
> > -     rcar_pci_write_reg(pcie, 0x00000001, GEN2_PCIEPHYCTRL);
> > -     rcar_pci_write_reg(pcie, 0x00000006, GEN2_PCIEPHYCTRL);
> > -
> > -     rcar_pci_write_reg(pcie, 0x000f0054, GEN2_PCIEPHYADDR);
> > -     /* The following value is for DC connection, no termination resistor */
> > -     rcar_pci_write_reg(pcie, 0x13802007, GEN2_PCIEPHYDATA);
> > -     rcar_pci_write_reg(pcie, 0x00000001, GEN2_PCIEPHYCTRL);
> > -     rcar_pci_write_reg(pcie, 0x00000006, GEN2_PCIEPHYCTRL);
> > -
> > -     return 0;
> > -}
> > -
> > -static int rcar_pcie_phy_init_gen3(struct rcar_pcie *pcie)
> > -{
> > -     int err;
> > -
> > -     err = phy_init(pcie->phy);
> > -     if (err)
> > -             return err;
> > -
> > -     err = phy_power_on(pcie->phy);
> > -     if (err)
> > -             phy_exit(pcie->phy);
> > -
> > -     return err;
> > -}
> > -
> > -static int rcar_msi_alloc(struct rcar_msi *chip)
> > -{
> > -     int msi;
> > -
> > -     mutex_lock(&chip->lock);
> > -
> > -     msi = find_first_zero_bit(chip->used, INT_PCI_MSI_NR);
> > -     if (msi < INT_PCI_MSI_NR)
> > -             set_bit(msi, chip->used);
> > -     else
> > -             msi = -ENOSPC;
> > -
> > -     mutex_unlock(&chip->lock);
> > -
> > -     return msi;
> > -}
> > -
> > -static int rcar_msi_alloc_region(struct rcar_msi *chip, int no_irqs)
> > -{
> > -     int msi;
> > -
> > -     mutex_lock(&chip->lock);
> > -     msi = bitmap_find_free_region(chip->used, INT_PCI_MSI_NR,
> > -                                   order_base_2(no_irqs));
> > -     mutex_unlock(&chip->lock);
> > -
> > -     return msi;
> > -}
> > -
> > -static void rcar_msi_free(struct rcar_msi *chip, unsigned long irq)
> > -{
> > -     mutex_lock(&chip->lock);
> > -     clear_bit(irq, chip->used);
> > -     mutex_unlock(&chip->lock);
> > -}
> > -
> > -static irqreturn_t rcar_pcie_msi_irq(int irq, void *data)
> > -{
> > -     struct rcar_pcie *pcie = data;
> > -     struct rcar_msi *msi = &pcie->msi;
> > -     struct device *dev = pcie->dev;
> > -     unsigned long reg;
> > -
> > -     reg = rcar_pci_read_reg(pcie, PCIEMSIFR);
> > -
> > -     /* MSI & INTx share an interrupt - we only handle MSI here */
> > -     if (!reg)
> > -             return IRQ_NONE;
> > -
> > -     while (reg) {
> > -             unsigned int index = find_first_bit(&reg, 32);
> > -             unsigned int msi_irq;
> > -
> > -             /* clear the interrupt */
> > -             rcar_pci_write_reg(pcie, 1 << index, PCIEMSIFR);
> > -
> > -             msi_irq = irq_find_mapping(msi->domain, index);
> > -             if (msi_irq) {
> > -                     if (test_bit(index, msi->used))
> > -                             generic_handle_irq(msi_irq);
> > -                     else
> > -                             dev_info(dev, "unhandled MSI\n");
> > -             } else {
> > -                     /* Unknown MSI, just clear it */
> > -                     dev_dbg(dev, "unexpected MSI\n");
> > -             }
> > -
> > -             /* see if there's any more pending in this vector */
> > -             reg = rcar_pci_read_reg(pcie, PCIEMSIFR);
> > -     }
> > -
> > -     return IRQ_HANDLED;
> > -}
> > -
> > -static int rcar_msi_setup_irq(struct msi_controller *chip, struct pci_dev *pdev,
> > -                           struct msi_desc *desc)
> > -{
> > -     struct rcar_msi *msi = to_rcar_msi(chip);
> > -     struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
> > -     struct msi_msg msg;
> > -     unsigned int irq;
> > -     int hwirq;
> > -
> > -     hwirq = rcar_msi_alloc(msi);
> > -     if (hwirq < 0)
> > -             return hwirq;
> > -
> > -     irq = irq_find_mapping(msi->domain, hwirq);
> > -     if (!irq) {
> > -             rcar_msi_free(msi, hwirq);
> > -             return -EINVAL;
> > -     }
> > -
> > -     irq_set_msi_desc(irq, desc);
> > -
> > -     msg.address_lo = rcar_pci_read_reg(pcie, PCIEMSIALR) & ~MSIFE;
> > -     msg.address_hi = rcar_pci_read_reg(pcie, PCIEMSIAUR);
> > -     msg.data = hwirq;
> > -
> > -     pci_write_msi_msg(irq, &msg);
> > -
> > -     return 0;
> > -}
> > -
> > -static int rcar_msi_setup_irqs(struct msi_controller *chip,
> > -                            struct pci_dev *pdev, int nvec, int type)
> > -{
> > -     struct rcar_pcie *pcie = container_of(chip, struct rcar_pcie, msi.chip);
> > -     struct rcar_msi *msi = to_rcar_msi(chip);
> > -     struct msi_desc *desc;
> > -     struct msi_msg msg;
> > -     unsigned int irq;
> > -     int hwirq;
> > -     int i;
> > -
> > -     /* MSI-X interrupts are not supported */
> > -     if (type == PCI_CAP_ID_MSIX)
> > -             return -EINVAL;
> > -
> > -     WARN_ON(!list_is_singular(&pdev->dev.msi_list));
> > -     desc = list_entry(pdev->dev.msi_list.next, struct msi_desc, list);
> > -
> > -     hwirq = rcar_msi_alloc_region(msi, nvec);
> > -     if (hwirq < 0)
> > -             return -ENOSPC;
> > -
> > -     irq = irq_find_mapping(msi->domain, hwirq);
> > -     if (!irq)
> > -             return -ENOSPC;
> > -
> > -     for (i = 0; i < nvec; i++) {
> > -             /*
> > -              * irq_create_mapping() called from rcar_pcie_probe() pre-
> > -              * allocates descs,  so there is no need to allocate descs here.
> > -              * We can therefore assume that if irq_find_mapping() above
> > -              * returns non-zero, then the descs are also successfully
> > -              * allocated.
> > -              */
> > -             if (irq_set_msi_desc_off(irq, i, desc)) {
> > -                     /* TODO: clear */
> > -                     return -EINVAL;
> > -             }
> > -     }
> > -
> > -     desc->nvec_used = nvec;
> > -     desc->msi_attrib.multiple = order_base_2(nvec);
> > -
> > -     msg.address_lo = rcar_pci_read_reg(pcie, PCIEMSIALR) & ~MSIFE;
> > -     msg.address_hi = rcar_pci_read_reg(pcie, PCIEMSIAUR);
> > -     msg.data = hwirq;
> > -
> > -     pci_write_msi_msg(irq, &msg);
> > -
> > -     return 0;
> > -}
> > -
> > -static void rcar_msi_teardown_irq(struct msi_controller *chip, unsigned int irq)
> > -{
> > -     struct rcar_msi *msi = to_rcar_msi(chip);
> > -     struct irq_data *d = irq_get_irq_data(irq);
> > -
> > -     rcar_msi_free(msi, d->hwirq);
> > -}
> > -
> > -static struct irq_chip rcar_msi_irq_chip = {
> > -     .name = "R-Car PCIe MSI",
> > -     .irq_enable = pci_msi_unmask_irq,
> > -     .irq_disable = pci_msi_mask_irq,
> > -     .irq_mask = pci_msi_mask_irq,
> > -     .irq_unmask = pci_msi_unmask_irq,
> > -};
> > -
> > -static int rcar_msi_map(struct irq_domain *domain, unsigned int irq,
> > -                     irq_hw_number_t hwirq)
> > -{
> > -     irq_set_chip_and_handler(irq, &rcar_msi_irq_chip, handle_simple_irq);
> > -     irq_set_chip_data(irq, domain->host_data);
> > -
> > -     return 0;
> > -}
> > -
> > -static const struct irq_domain_ops msi_domain_ops = {
> > -     .map = rcar_msi_map,
> > -};
> > -
> > -static void rcar_pcie_unmap_msi(struct rcar_pcie *pcie)
> > -{
> > -     struct rcar_msi *msi = &pcie->msi;
> > -     int i, irq;
> > -
> > -     for (i = 0; i < INT_PCI_MSI_NR; i++) {
> > -             irq = irq_find_mapping(msi->domain, i);
> > -             if (irq > 0)
> > -                     irq_dispose_mapping(irq);
> > -     }
> > -
> > -     irq_domain_remove(msi->domain);
> > -}
> > -
> > -static int rcar_pcie_enable_msi(struct rcar_pcie *pcie)
> > -{
> > -     struct device *dev = pcie->dev;
> > -     struct rcar_msi *msi = &pcie->msi;
> > -     phys_addr_t base;
> > -     int err, i;
> > -
> > -     mutex_init(&msi->lock);
> > -
> > -     msi->chip.dev = dev;
> > -     msi->chip.setup_irq = rcar_msi_setup_irq;
> > -     msi->chip.setup_irqs = rcar_msi_setup_irqs;
> > -     msi->chip.teardown_irq = rcar_msi_teardown_irq;
> > -
> > -     msi->domain = irq_domain_add_linear(dev->of_node, INT_PCI_MSI_NR,
> > -                                         &msi_domain_ops, &msi->chip);
> > -     if (!msi->domain) {
> > -             dev_err(dev, "failed to create IRQ domain\n");
> > -             return -ENOMEM;
> > -     }
> > -
> > -     for (i = 0; i < INT_PCI_MSI_NR; i++)
> > -             irq_create_mapping(msi->domain, i);
> > -
> > -     /* Two irqs are for MSI, but they are also used for non-MSI irqs */
> > -     err = devm_request_irq(dev, msi->irq1, rcar_pcie_msi_irq,
> > -                            IRQF_SHARED | IRQF_NO_THREAD,
> > -                            rcar_msi_irq_chip.name, pcie);
> > -     if (err < 0) {
> > -             dev_err(dev, "failed to request IRQ: %d\n", err);
> > -             goto err;
> > -     }
> > -
> > -     err = devm_request_irq(dev, msi->irq2, rcar_pcie_msi_irq,
> > -                            IRQF_SHARED | IRQF_NO_THREAD,
> > -                            rcar_msi_irq_chip.name, pcie);
> > -     if (err < 0) {
> > -             dev_err(dev, "failed to request IRQ: %d\n", err);
> > -             goto err;
> > -     }
> > -
> > -     /* setup MSI data target */
> > -     msi->pages = __get_free_pages(GFP_KERNEL, 0);
> > -     if (!msi->pages) {
> > -             err = -ENOMEM;
> > -             goto err;
> > -     }
> > -     base = virt_to_phys((void *)msi->pages);
> > -
> > -     rcar_pci_write_reg(pcie, lower_32_bits(base) | MSIFE, PCIEMSIALR);
> > -     rcar_pci_write_reg(pcie, upper_32_bits(base), PCIEMSIAUR);
> > -
> > -     /* enable all MSI interrupts */
> > -     rcar_pci_write_reg(pcie, 0xffffffff, PCIEMSIIER);
> > -
> > -     return 0;
> > -
> > -err:
> > -     rcar_pcie_unmap_msi(pcie);
> > -     return err;
> > -}
> > -
> > -static void rcar_pcie_teardown_msi(struct rcar_pcie *pcie)
> > -{
> > -     struct rcar_msi *msi = &pcie->msi;
> > -
> > -     /* Disable all MSI interrupts */
> > -     rcar_pci_write_reg(pcie, 0, PCIEMSIIER);
> > -
> > -     /* Disable address decoding of the MSI interrupt, MSIFE */
> > -     rcar_pci_write_reg(pcie, 0, PCIEMSIALR);
> > -
> > -     free_pages(msi->pages, 0);
> > -
> > -     rcar_pcie_unmap_msi(pcie);
> > -}
> > -
> > -static int rcar_pcie_get_resources(struct rcar_pcie *pcie)
> > -{
> > -     struct device *dev = pcie->dev;
> > -     struct resource res;
> > -     int err, i;
> > -
> > -     pcie->phy = devm_phy_optional_get(dev, "pcie");
> > -     if (IS_ERR(pcie->phy))
> > -             return PTR_ERR(pcie->phy);
> > -
> > -     err = of_address_to_resource(dev->of_node, 0, &res);
> > -     if (err)
> > -             return err;
> > -
> > -     pcie->base = devm_ioremap_resource(dev, &res);
> > -     if (IS_ERR(pcie->base))
> > -             return PTR_ERR(pcie->base);
> > -
> > -     pcie->bus_clk = devm_clk_get(dev, "pcie_bus");
> > -     if (IS_ERR(pcie->bus_clk)) {
> > -             dev_err(dev, "cannot get pcie bus clock\n");
> > -             return PTR_ERR(pcie->bus_clk);
> > -     }
> > -
> > -     i = irq_of_parse_and_map(dev->of_node, 0);
> > -     if (!i) {
> > -             dev_err(dev, "cannot get platform resources for msi interrupt\n");
> > -             err = -ENOENT;
> > -             goto err_irq1;
> > -     }
> > -     pcie->msi.irq1 = i;
> > -
> > -     i = irq_of_parse_and_map(dev->of_node, 1);
> > -     if (!i) {
> > -             dev_err(dev, "cannot get platform resources for msi interrupt\n");
> > -             err = -ENOENT;
> > -             goto err_irq2;
> > -     }
> > -     pcie->msi.irq2 = i;
> > -
> > -     return 0;
> > -
> > -err_irq2:
> > -     irq_dispose_mapping(pcie->msi.irq1);
> > -err_irq1:
> > -     return err;
> > -}
> > -
> > -static int rcar_pcie_inbound_ranges(struct rcar_pcie *pcie,
> > -                                 struct of_pci_range *range,
> > -                                 int *index)
> > +void rcar_pcie_set_inbound(void __iomem *base,
> > +                        u64 cpu_addr, u64 pci_addr,
> > +                        u64 mask, int idx, bool host)
> >  {
> > -     u64 restype = range->flags;
> > -     u64 cpu_addr = range->cpu_addr;
> > -     u64 cpu_end = range->cpu_addr + range->size;
> > -     u64 pci_addr = range->pci_addr;
> > -     u32 flags = LAM_64BIT | LAR_ENABLE;
> > -     u64 mask;
> > -     u64 size;
> > -     int idx = *index;
> > -
> > -     if (restype & IORESOURCE_PREFETCH)
> > -             flags |= LAM_PREFETCH;
> > -
> >       /*
> > -      * If the size of the range is larger than the alignment of the start
> > -      * address, we have to use multiple entries to perform the mapping.
> > -      */
> > -     if (cpu_addr > 0) {
> > -             unsigned long nr_zeros = __ffs64(cpu_addr);
> > -             u64 alignment = 1ULL << nr_zeros;
> > -
> > -             size = min(range->size, alignment);
> > -     } else {
> > -             size = range->size;
> > -     }
> > -     /* Hardware supports max 4GiB inbound region */
> > -     size = min(size, 1ULL << 32);
> > -
> > -     mask = roundup_pow_of_two(size) - 1;
> > -     mask &= ~0xf;
> > -
> > -     while (cpu_addr < cpu_end) {
> > -             /*
> > -              * Set up 64-bit inbound regions as the range parser doesn't
> > -              * distinguish between 32 and 64-bit types.
> > -              */
> > -             rcar_pci_write_reg(pcie, lower_32_bits(pci_addr),
> > +     * Set up 64-bit inbound regions as the range parser doesn't
> > +     * distinguish between 32 and 64-bit types.
> > +     */
> > +     if (host) {
> > +             rcar_pci_write_reg(base, lower_32_bits(pci_addr),
> >                                  PCIEPRAR(idx));
> > -             rcar_pci_write_reg(pcie, lower_32_bits(cpu_addr), PCIELAR(idx));
> > -             rcar_pci_write_reg(pcie, lower_32_bits(mask) | flags,
> > -                                PCIELAMR(idx));
> > -
> > -             rcar_pci_write_reg(pcie, upper_32_bits(pci_addr),
> > +             rcar_pci_write_reg(base, upper_32_bits(pci_addr),
> >                                  PCIEPRAR(idx + 1));
> > -             rcar_pci_write_reg(pcie, upper_32_bits(cpu_addr),
> > -                                PCIELAR(idx + 1));
> > -             rcar_pci_write_reg(pcie, 0, PCIELAMR(idx + 1));
> > -
> > -             pci_addr += size;
> > -             cpu_addr += size;
> > -             idx += 2;
> > -
> > -             if (idx > MAX_NR_INBOUND_MAPS) {
> > -                     dev_err(pcie->dev, "Failed to map inbound regions!\n");
> > -                     return -EINVAL;
> > -             }
> >       }
> > -     *index = idx;
> > -
> > -     return 0;
> > -}
> >
> > -static int rcar_pcie_parse_map_dma_ranges(struct rcar_pcie *pcie,
> > -                                       struct device_node *np)
> > -{
> > -     struct of_pci_range range;
> > -     struct of_pci_range_parser parser;
> > -     int index = 0;
> > -     int err;
> > -
> > -     if (of_pci_dma_range_parser_init(&parser, np))
> > -             return -EINVAL;
> > -
> > -     /* Get the dma-ranges from DT */
> > -     for_each_of_pci_range(&parser, &range) {
> > -             u64 end = range.cpu_addr + range.size - 1;
> > +     rcar_pci_write_reg(base, lower_32_bits(cpu_addr), PCIELAR(idx));
> > +     rcar_pci_write_reg(base, lower_32_bits(mask), PCIELAMR(idx));
> >
> > -             dev_dbg(pcie->dev, "0x%08x 0x%016llx..0x%016llx -> 0x%016llx\n",
> > -                     range.flags, range.cpu_addr, end, range.pci_addr);
> > -
> > -             err = rcar_pcie_inbound_ranges(pcie, &range, &index);
> > -             if (err)
> > -                     return err;
> > -     }
> > +     rcar_pci_write_reg(base, upper_32_bits(cpu_addr), PCIELAR(idx + 1));
> > +     rcar_pci_write_reg(base, 0, PCIELAMR(idx + 1));
> >
> > -     return 0;
> > -}
> > -
> > -static const struct of_device_id rcar_pcie_of_match[] = {
> > -     { .compatible = "renesas,pcie-r8a7779",
> > -       .data = rcar_pcie_phy_init_h1 },
> > -     { .compatible = "renesas,pcie-r8a7790",
> > -       .data = rcar_pcie_phy_init_gen2 },
> > -     { .compatible = "renesas,pcie-r8a7791",
> > -       .data = rcar_pcie_phy_init_gen2 },
> > -     { .compatible = "renesas,pcie-rcar-gen2",
> > -       .data = rcar_pcie_phy_init_gen2 },
> > -     { .compatible = "renesas,pcie-r8a7795",
> > -       .data = rcar_pcie_phy_init_gen3 },
> > -     { .compatible = "renesas,pcie-rcar-gen3",
> > -       .data = rcar_pcie_phy_init_gen3 },
> > -     {},
> > -};
> > -
> > -static int rcar_pcie_probe(struct platform_device *pdev)
> > -{
> > -     struct device *dev = &pdev->dev;
> > -     struct rcar_pcie *pcie;
> > -     u32 data;
> > -     int err;
> > -     int (*phy_init_fn)(struct rcar_pcie *);
> > -     struct pci_host_bridge *bridge;
> > -
> > -     bridge = pci_alloc_host_bridge(sizeof(*pcie));
> > -     if (!bridge)
> > -             return -ENOMEM;
> > -
> > -     pcie = pci_host_bridge_priv(bridge);
> > -
> > -     pcie->dev = dev;
> > -     platform_set_drvdata(pdev, pcie);
> > -
> > -     err = pci_parse_request_of_pci_ranges(dev, &pcie->resources, NULL);
> > -     if (err)
> > -             goto err_free_bridge;
> > -
> > -     pm_runtime_enable(pcie->dev);
> > -     err = pm_runtime_get_sync(pcie->dev);
> > -     if (err < 0) {
> > -             dev_err(pcie->dev, "pm_runtime_get_sync failed\n");
> > -             goto err_pm_disable;
> > -     }
> > -
> > -     err = rcar_pcie_get_resources(pcie);
> > -     if (err < 0) {
> > -             dev_err(dev, "failed to request resources: %d\n", err);
> > -             goto err_pm_put;
> > -     }
> > -
> > -     err = clk_prepare_enable(pcie->bus_clk);
> > -     if (err) {
> > -             dev_err(dev, "failed to enable bus clock: %d\n", err);
> > -             goto err_unmap_msi_irqs;
> > -     }
> > -
> > -     err = rcar_pcie_parse_map_dma_ranges(pcie, dev->of_node);
> > -     if (err)
> > -             goto err_clk_disable;
> > -
> > -     phy_init_fn = of_device_get_match_data(dev);
> > -     err = phy_init_fn(pcie);
> > -     if (err) {
> > -             dev_err(dev, "failed to init PCIe PHY\n");
> > -             goto err_clk_disable;
> > -     }
> > -
> > -     /* Failure to get a link might just be that no cards are inserted */
> > -     if (rcar_pcie_hw_init(pcie)) {
> > -             dev_info(dev, "PCIe link down\n");
> > -             err = -ENODEV;
> > -             goto err_phy_shutdown;
> > -     }
> > -
> > -     data = rcar_pci_read_reg(pcie, MACSR);
> > -     dev_info(dev, "PCIe x%d: link up\n", (data >> 20) & 0x3f);
> > -
> > -     if (IS_ENABLED(CONFIG_PCI_MSI)) {
> > -             err = rcar_pcie_enable_msi(pcie);
> > -             if (err < 0) {
> > -                     dev_err(dev,
> > -                             "failed to enable MSI support: %d\n",
> > -                             err);
> > -                     goto err_phy_shutdown;
> > -             }
> > -     }
> > -
> > -     err = rcar_pcie_enable(pcie);
> > -     if (err)
> > -             goto err_msi_teardown;
> > -
> > -     return 0;
> > -
> > -err_msi_teardown:
> > -     if (IS_ENABLED(CONFIG_PCI_MSI))
> > -             rcar_pcie_teardown_msi(pcie);
> > -
> > -err_phy_shutdown:
> > -     if (pcie->phy) {
> > -             phy_power_off(pcie->phy);
> > -             phy_exit(pcie->phy);
> > -     }
> > -
> > -err_clk_disable:
> > -     clk_disable_unprepare(pcie->bus_clk);
> > -
> > -err_unmap_msi_irqs:
> > -     irq_dispose_mapping(pcie->msi.irq2);
> > -     irq_dispose_mapping(pcie->msi.irq1);
> > -
> > -err_pm_put:
> > -     pm_runtime_put(dev);
> > -
> > -err_pm_disable:
> > -     pm_runtime_disable(dev);
> > -     pci_free_resource_list(&pcie->resources);
> > -
> > -err_free_bridge:
> > -     pci_free_host_bridge(bridge);
> > -
> > -     return err;
> > -}
> > -
> > -static int rcar_pcie_resume_noirq(struct device *dev)
> > -{
> > -     struct rcar_pcie *pcie = dev_get_drvdata(dev);
> > -
> > -     if (rcar_pci_read_reg(pcie, PMSR) &&
> > -         !(rcar_pci_read_reg(pcie, PCIETCTLR) & DL_DOWN))
> > -             return 0;
> > -
> > -     /* Re-establish the PCIe link */
> > -     rcar_rmw32(pcie, MACCTLR, MACCTLR_RESERVED, 0);
> > -     rcar_pci_write_reg(pcie, CFINIT, PCIETCTLR);
> > -     return rcar_pcie_wait_for_dl(pcie);
> > +     /* flush modifications */
> > +     wmb();
> >  }
> > -
> > -static const struct dev_pm_ops rcar_pcie_pm_ops = {
> > -     .resume_noirq = rcar_pcie_resume_noirq,
> > -};
> > -
> > -static struct platform_driver rcar_pcie_driver = {
> > -     .driver = {
> > -             .name = "rcar-pcie",
> > -             .of_match_table = rcar_pcie_of_match,
> > -             .pm = &rcar_pcie_pm_ops,
> > -             .suppress_bind_attrs = true,
> > -     },
> > -     .probe = rcar_pcie_probe,
> > -};
> > -builtin_platform_driver(rcar_pcie_driver);
> > diff --git a/drivers/pci/controller/pcie-rcar.h b/drivers/pci/controller/pcie-rcar.h
> > new file mode 100644
> > index 000000000000..502621d4e6ac
> > --- /dev/null
> > +++ b/drivers/pci/controller/pcie-rcar.h
> > @@ -0,0 +1,129 @@
> > +/* SPDX-License-Identifier: GPL-2.0 */
> > +/*
> > + * PCIe driver for Renesas R-Car SoCs
> > + *  Copyright (C) 2014-2019 Renesas Electronics Europe Ltd
> > + *
> > + * Author: Phil Edworthy <phil.edworthy@renesas.com>
> > + */
> > +#ifndef _PCIE_RCAR_H
> > +#define _PCIE_RCAR_H
> > +
> > +#define PCIECAR                      0x000010
> > +#define PCIECCTLR            0x000018
> > +#define  CONFIG_SEND_ENABLE  BIT(31)
> > +#define  TYPE0                       (0 << 8)
> > +#define  TYPE1                       BIT(8)
> > +#define PCIECDR                      0x000020
> > +#define PCIEMSR                      0x000028
> > +#define PCIEINTXR            0x000400
> > +#define PCIEPHYSR            0x0007f0
> > +#define  PHYRDY                      BIT(0)
> > +#define PCIEMSITXR           0x000840
> > +
> > +/* Transfer control */
> > +#define PCIETCTLR            0x02000
> > +#define  DL_DOWN             BIT(3)
> > +#define  CFINIT                      BIT(0)
> > +#define PCIETSTR             0x02004
> > +#define  DATA_LINK_ACTIVE    BIT(0)
> > +#define PCIEERRFR            0x02020
> > +#define  UNSUPPORTED_REQUEST BIT(4)
> > +#define PCIEMSIFR            0x02044
> > +#define PCIEMSIALR           0x02048
> > +#define  MSIFE                       BIT(0)
> > +#define PCIEMSIAUR           0x0204c
> > +#define PCIEMSIIER           0x02050
> > +
> > +/* root port address */
> > +#define PCIEPRAR(x)          (0x02080 + ((x) * 0x4))
> > +
> > +/* local address reg & mask */
> > +#define PCIELAR(x)           (0x02200 + ((x) * 0x20))
> > +#define PCIELAMR(x)          (0x02208 + ((x) * 0x20))
> > +#define  LAM_PREFETCH                BIT(3)
> > +#define  LAM_64BIT           BIT(2)
> > +#define  LAR_ENABLE          BIT(1)
> > +
> > +/* PCIe address reg & mask */
> > +#define PCIEPALR(x)          (0x03400 + ((x) * 0x20))
> > +#define PCIEPAUR(x)          (0x03404 + ((x) * 0x20))
> > +#define PCIEPAMR(x)          (0x03408 + ((x) * 0x20))
> > +#define PCIEPTCTLR(x)                (0x0340c + ((x) * 0x20))
> > +#define  PAR_ENABLE          BIT(31)
> > +#define  IO_SPACE            BIT(8)
> > +
> > +/* Configuration */
> > +#define PCICONF(x)           (0x010000 + ((x) * 0x4))
> > +#define PMCAP(x)             (0x010040 + ((x) * 0x4))
> > +#define EXPCAP(x)            (0x010070 + ((x) * 0x4))
> > +#define VCCAP(x)             (0x010100 + ((x) * 0x4))
> > +
> > +/* link layer */
> > +#define IDSETR0                      0x011000
> > +#define  DEVICE_ID_SHFIT     16
> > +#define IDSETR1                      0x011004
> > +#define SUBIDSETR            0x011024
> > +#define TLCTLR                       0x011048
> > +#define MACSR                        0x011054
> > +#define  SPCHGFIN            BIT(4)
> > +#define  SPCHGFAIL           BIT(6)
> > +#define  SPCHGSUC            BIT(7)
> > +#define  LINK_SPEED          (0xf << 16)
> > +#define  LINK_SPEED_2_5GTS   BIT(16)
> > +#define  LINK_SPEED_5_0GTS   BIT(17)
> > +#define MACCTLR                      0x011058
> > +#define  MACCTLR_RESERVED    BIT(0)
> > +#define  SPEED_CHANGE                BIT(24)
> > +#define  SCRAMBLE_DISABLE    BIT(27)
> > +#define PMSR                 0x01105c
> > +#define MACS2R                       0x011078
> > +#define MACCGSPSETR          0x011084
> > +#define  SPCNGRSN            BIT(31)
> > +
> > +/* R-Car H1 PHY */
> > +#define H1_PCIEPHYADRR               0x04000c
> > +#define  WRITE_CMD           BIT(16)
> > +#define  PHY_ACK             BIT(24)
> > +#define  RATE_POS            12
> > +#define  LANE_POS            8
> > +#define  ADR_POS             0
> > +#define H1_PCIEPHYDOUTR              0x040014
> > +
> > +/* R-Car Gen2 PHY */
> > +#define GEN2_PCIEPHYADDR     0x780
> > +#define GEN2_PCIEPHYDATA     0x784
> > +#define GEN2_PCIEPHYCTRL     0x78c
> > +
> > +#define INT_PCI_MSI_NR               32
> > +
> > +#define RCONF(x)             (PCICONF(0) + (x))
> > +#define RPMCAP(x)            (PMCAP(0) + (x))
> > +#define REXPCAP(x)           (EXPCAP(0) + (x))
> > +#define RVCCAP(x)            (VCCAP(0) + (x))
> > +
> > +#define INTDIS_SHIFT         BIT(10)
> > +#define ASTINTX_SHIFT                BIT(16)
> > +
> > +#define PCIE_CONF_BUS(b)     (((b) & 0xff) << 24)
> > +#define PCIE_CONF_DEV(d)     (((d) & 0x1f) << 19)
> > +#define PCIE_CONF_FUNC(f)    (((f) & 0x7) << 16)
> > +
> > +#define RCAR_PCI_MAX_RESOURCES       4
> > +#define MAX_NR_INBOUND_MAPS  6
> > +
> > +enum {
> > +     RCAR_PCI_ACCESS_READ,
> > +     RCAR_PCI_ACCESS_WRITE,
> > +};
> > +
> > +void rcar_pci_write_reg(void __iomem *base, u32 val, unsigned int reg);
> > +u32 rcar_pci_read_reg(void __iomem *base, unsigned int reg);
> > +void rcar_rmw32(void __iomem *base, int where, u32 mask, u32 data);
> > +int rcar_pcie_wait_for_phyrdy(void __iomem *base);
> > +int rcar_pcie_wait_for_dl(void __iomem *base);
> > +void rcar_pcie_set_outbound(int win, void __iomem *base,
> > +                         struct resource *res, bool host);
> > +void rcar_pcie_set_inbound(void __iomem *base, u64 cpu_addr, u64 pci_addr,
> > +                        u64 mask, int idx, bool host);
> > +
> > +#endif
> >

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
  2019-11-27  5:44             ` Kishon Vijay Abraham I
@ 2019-11-27 21:00               ` Lad, Prabhakar
  -1 siblings, 0 replies; 60+ messages in thread
From: Lad, Prabhakar @ 2019-11-27 21:00 UTC (permalink / raw)
  To: Kishon Vijay Abraham I
  Cc: Rob Herring, Geert Uytterhoeven, Bjorn Helgaas, Mark Rutland,
	Magnus Damm, Marek Vasut, Yoshihiro Shimoda, linux-pci,
	Catalin Marinas, Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Linux Kernel Mailing List, Linux ARM, Linux-Renesas,
	Chris Paterson, Lad, Prabhakar

Hi Kishon,

On Wed, Nov 27, 2019 at 5:45 AM Kishon Vijay Abraham I <kishon@ti.com> wrote:
>
> Hi,
>
> On 13/11/19 9:38 AM, Rob Herring wrote:
> > On Thu, Nov 07, 2019 at 09:08:35PM +0100, Geert Uytterhoeven wrote:
> >> Hi Prabhakar,
> >>
> >> On Thu, Nov 7, 2019 at 10:26 AM Lad, Prabhakar
> >> <prabhakar.csengg@gmail.com> wrote:
> >>> On Thu, Nov 7, 2019 at 8:44 AM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> >>>> On Wed, Nov 6, 2019 at 8:36 PM Lad Prabhakar <prabhakar.csengg@gmail.com> wrote:
> >>>>> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >>>>>
> >>>>> This patch adds the bindings for the R-Car PCIe endpoint driver.
> >>>>>
> >>>>> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >>>>
> >>>> Thanks for your patch!
> >>>>
> >>>>> --- /dev/null
> >>>>> +++ b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> >>>>> @@ -0,0 +1,43 @@
> >>>>> +* Renesas R-Car PCIe Endpoint Controller DT description
> >>>>> +
> >>>>> +Required properties:
> >>>>> +           "renesas,pcie-ep-r8a774c0" for the R8A774C0 SoC;
> >>>>> +           "renesas,pcie-ep-rcar-gen3" for a generic R-Car Gen3 or
> >>>>> +                                    RZ/G2 compatible device.
> >>>>
> >>>> Unless I'm missing something, this is for the exact same hardware block as
> >>>> Documentation/devicetree/bindings/pci/rcar-pci.txt?
> >>>> So shouldn't you amend those bindings, instead of adding new compatible
> >>>> values?
> >>>> Please remember that DT describes hardware, not software policy.
> >>>> So IMHO choosing between host and endpoint is purely a configuration
> >>>> issue, and could be indicated by the presence or lack of some DT properties.
> >>>> E.g. host mode requires both "bus-range" and "device_type" properties,
> >>>> so their absence could indicate endpoint mode.
> >>>>
> >>> yes its the same hardware block as described in the rcar-pci.txt, I
> >>> did think about amending it
> >>> but  it might turn out to be bit messy,
> >>>
> >>> required properties host ======required properties Endpoint
> >>> ====================||==================
> >>> 1: reg                                || reg
> >>> 2:bus-range                      || reg names
> >>> 3: device_type                  || resets
> >>> 4: ranges                          || clocks
> >>> 5: dma-ranges                  || clock-names
> >>> 6: interrupts                      ||
> >>> 7: interrupt-cells               ||
> >>> 8: interrupt-map-mask     ||
> >>> 9: clocks                          ||
> >>> 10: clock-names             ||
> >>
> >> We have a similar situation with SPI, where a controller can operate in
> >> master or slave mode, based on the absence or presence of the
> >> "spi-slave" DT property.
> >>
> >>> and if I go ahead with the same compatible string that would mean to
> >>> add support for endpoint
> >>> mode in the host driver itself. I did follow the examples of
> >>
> >> You can still have two separate drivers, binding against the same
> >> compatible value.  Just let the .probe() function return -ENODEV if it
> >> discovers (by looking at DT properties) if the node is configured for
> >> the other mode.
> >> Which brings us to my next questions: is there any code that could be
> >> shared between the drivers for the two modes?
> >>
> >>> rockchip/cadence/designware where
> >>> its the same hardware block but has two different binding files one
> >>> for host mode and other for
> >>> endpoint mode.
> >>
> >> Having two separate DT binding documents sounds fine to me, if unifying
> >> them makes things too complex.
> >> However, I think they should use the same compatible value, because the
> >> hardware block is the same, but just used in a different mode.
> >>
> >> Rob/Mark: Any input from the DT maintainers?
> >
> > Separate files makes sense because different modes will want to
> > include different common schemas. We've generally been doing different
> > compatibles too which makes validating the node has the right set of
> > properties easier.
> >
> >>>>> +- reg: Five register ranges as listed in the reg-names property
> >>>>> +- reg-names: Must include the following names
> >>>>> +       - "apb-base"
> >>>>> +       - "memory0"
> >>>>> +       - "memory1"
> >>>>> +       - "memory2"
> >>>>> +       - "memory3"
> >>>>
> >>>> What is the purpose of the last 4 regions?
> >>>> Can they be chosen by the driver, at runtime?
> >>>>
> >>> no the driver cannot choose them at runtime, as these are the only
> >>> PCIE memory(0/1/2/3) ranges
> >>> in the AXI address space where host memory can be mapped.
> >>
> >> Are they fixed by the PCIe hardware, i.e. could they be looked up by the
> >> driver based on the compatible value?
> >
> > That would be strange for a memory range.
> >
> > Sounds like like 'ranges' though I'm not sure if 'ranges' for an EP
> > makes sense or what that should look like.
>
> These are similar to "memory node" with multiple address, size pairs. I'm
> thinking if these should be added as a subnode within PCIe EP controller device
> tree node?
>
+1

something similar like below ?

       pcie_ep: pcie_ep@fe000000 {
               compatible = "renesas,pcie-r8a7791", "renesas,pcie-rcar-gen2";
               reg = <0 0xfe000000 0 0x80000>;
               reg-names = "apb-base";
               clocks = <&cpg CPG_MOD 319>;
               clock-names = "pcie";
               power-domains = <&sysc R8A774C0_PD_ALWAYS_ON>;
               resets = <&cpg 319>;
               mem-region {
                       base =  <0x0 0xfe100000 0 0x100000>,
                                    <0x0 0xfe200000 0 0x200000>,
                                    <0x0 0x30000000 0 0x8000000>,
                                    <0x0 0x38000000 0 0x8000000>;
                };
       };

Cheers,
--Prabhakar

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings
@ 2019-11-27 21:00               ` Lad, Prabhakar
  0 siblings, 0 replies; 60+ messages in thread
From: Lad, Prabhakar @ 2019-11-27 21:00 UTC (permalink / raw)
  To: Kishon Vijay Abraham I
  Cc: Mark Rutland, Rob Herring, Chris Paterson, Lorenzo Pieralisi,
	Arnd Bergmann, Greg Kroah-Hartman, linux-pci, Yoshihiro Shimoda,
	Magnus Damm, Linux Kernel Mailing List, Lad, Prabhakar,
	Linux-Renesas,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Geert Uytterhoeven, Catalin Marinas, Bjorn Helgaas,
	Andrew Murray, Will Deacon, Linux ARM, Marek Vasut

Hi Kishon,

On Wed, Nov 27, 2019 at 5:45 AM Kishon Vijay Abraham I <kishon@ti.com> wrote:
>
> Hi,
>
> On 13/11/19 9:38 AM, Rob Herring wrote:
> > On Thu, Nov 07, 2019 at 09:08:35PM +0100, Geert Uytterhoeven wrote:
> >> Hi Prabhakar,
> >>
> >> On Thu, Nov 7, 2019 at 10:26 AM Lad, Prabhakar
> >> <prabhakar.csengg@gmail.com> wrote:
> >>> On Thu, Nov 7, 2019 at 8:44 AM Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> >>>> On Wed, Nov 6, 2019 at 8:36 PM Lad Prabhakar <prabhakar.csengg@gmail.com> wrote:
> >>>>> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >>>>>
> >>>>> This patch adds the bindings for the R-Car PCIe endpoint driver.
> >>>>>
> >>>>> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >>>>
> >>>> Thanks for your patch!
> >>>>
> >>>>> --- /dev/null
> >>>>> +++ b/Documentation/devicetree/bindings/pci/rcar-pci-ep.txt
> >>>>> @@ -0,0 +1,43 @@
> >>>>> +* Renesas R-Car PCIe Endpoint Controller DT description
> >>>>> +
> >>>>> +Required properties:
> >>>>> +           "renesas,pcie-ep-r8a774c0" for the R8A774C0 SoC;
> >>>>> +           "renesas,pcie-ep-rcar-gen3" for a generic R-Car Gen3 or
> >>>>> +                                    RZ/G2 compatible device.
> >>>>
> >>>> Unless I'm missing something, this is for the exact same hardware block as
> >>>> Documentation/devicetree/bindings/pci/rcar-pci.txt?
> >>>> So shouldn't you amend those bindings, instead of adding new compatible
> >>>> values?
> >>>> Please remember that DT describes hardware, not software policy.
> >>>> So IMHO choosing between host and endpoint is purely a configuration
> >>>> issue, and could be indicated by the presence or lack of some DT properties.
> >>>> E.g. host mode requires both "bus-range" and "device_type" properties,
> >>>> so their absence could indicate endpoint mode.
> >>>>
> >>> yes its the same hardware block as described in the rcar-pci.txt, I
> >>> did think about amending it
> >>> but  it might turn out to be bit messy,
> >>>
> >>> required properties host ======required properties Endpoint
> >>> ====================||==================
> >>> 1: reg                                || reg
> >>> 2:bus-range                      || reg names
> >>> 3: device_type                  || resets
> >>> 4: ranges                          || clocks
> >>> 5: dma-ranges                  || clock-names
> >>> 6: interrupts                      ||
> >>> 7: interrupt-cells               ||
> >>> 8: interrupt-map-mask     ||
> >>> 9: clocks                          ||
> >>> 10: clock-names             ||
> >>
> >> We have a similar situation with SPI, where a controller can operate in
> >> master or slave mode, based on the absence or presence of the
> >> "spi-slave" DT property.
> >>
> >>> and if I go ahead with the same compatible string that would mean to
> >>> add support for endpoint
> >>> mode in the host driver itself. I did follow the examples of
> >>
> >> You can still have two separate drivers, binding against the same
> >> compatible value.  Just let the .probe() function return -ENODEV if it
> >> discovers (by looking at DT properties) if the node is configured for
> >> the other mode.
> >> Which brings us to my next questions: is there any code that could be
> >> shared between the drivers for the two modes?
> >>
> >>> rockchip/cadence/designware where
> >>> its the same hardware block but has two different binding files one
> >>> for host mode and other for
> >>> endpoint mode.
> >>
> >> Having two separate DT binding documents sounds fine to me, if unifying
> >> them makes things too complex.
> >> However, I think they should use the same compatible value, because the
> >> hardware block is the same, but just used in a different mode.
> >>
> >> Rob/Mark: Any input from the DT maintainers?
> >
> > Separate files makes sense because different modes will want to
> > include different common schemas. We've generally been doing different
> > compatibles too which makes validating the node has the right set of
> > properties easier.
> >
> >>>>> +- reg: Five register ranges as listed in the reg-names property
> >>>>> +- reg-names: Must include the following names
> >>>>> +       - "apb-base"
> >>>>> +       - "memory0"
> >>>>> +       - "memory1"
> >>>>> +       - "memory2"
> >>>>> +       - "memory3"
> >>>>
> >>>> What is the purpose of the last 4 regions?
> >>>> Can they be chosen by the driver, at runtime?
> >>>>
> >>> no the driver cannot choose them at runtime, as these are the only
> >>> PCIE memory(0/1/2/3) ranges
> >>> in the AXI address space where host memory can be mapped.
> >>
> >> Are they fixed by the PCIe hardware, i.e. could they be looked up by the
> >> driver based on the compatible value?
> >
> > That would be strange for a memory range.
> >
> > Sounds like like 'ranges' though I'm not sure if 'ranges' for an EP
> > makes sense or what that should look like.
>
> These are similar to "memory node" with multiple address, size pairs. I'm
> thinking if these should be added as a subnode within PCIe EP controller device
> tree node?
>
+1

something similar like below ?

       pcie_ep: pcie_ep@fe000000 {
               compatible = "renesas,pcie-r8a7791", "renesas,pcie-rcar-gen2";
               reg = <0 0xfe000000 0 0x80000>;
               reg-names = "apb-base";
               clocks = <&cpg CPG_MOD 319>;
               clock-names = "pcie";
               power-domains = <&sysc R8A774C0_PD_ALWAYS_ON>;
               resets = <&cpg 319>;
               mem-region {
                       base =  <0x0 0xfe100000 0 0x100000>,
                                    <0x0 0xfe200000 0 0x200000>,
                                    <0x0 0x30000000 0 0x8000000>,
                                    <0x0 0x38000000 0 0x8000000>;
                };
       };

Cheers,
--Prabhakar

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 2/5] pci: endpoint: add support to handle multiple base for mapping outbound memory
  2019-11-27  5:14     ` Kishon Vijay Abraham I
  (?)
@ 2019-11-27 21:21       ` Lad, Prabhakar
  -1 siblings, 0 replies; 60+ messages in thread
From: Lad, Prabhakar @ 2019-11-27 21:21 UTC (permalink / raw)
  To: Kishon Vijay Abraham I
  Cc: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Marek Vasut, Yoshihiro Shimoda, linux-pci,
	Catalin Marinas, Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS, LKML,
	LAK, Linux-Renesas, Chris Paterson, Lad, Prabhakar,
	linux-rockchip, Shawn Lin, Heiko Stuebner, Tom Joseph,
	Jingoo Han, Gustavo Pimentel

Hi Kishon,

Thank you for the review.

On Wed, Nov 27, 2019 at 5:15 AM Kishon Vijay Abraham I <kishon@ti.com> wrote:
>
> Hi Prabhakar,
>
> On 07/11/19 1:06 AM, Lad Prabhakar wrote:
> > From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >
> > rcar pcie controller has support to map multiple memory regions
> > for mapping the outbound memory in local system, this feature
> > inspires to add support for handling multiple memory bases in
> > endpoint framework. In case of multiple memory regions only chunk
> > or complete region can be mapped and this window needs to be
> > passed to the controller driver.
> >
> > Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > Cc: <linux-rockchip@lists.infradead.org>
> > Cc: Shawn Lin <shawn.lin@rock-chips.com>
> > Cc: Heiko Stuebner <heiko@sntech.de>
> > Cc: Tom Joseph <tjoseph@cadence.com>
> > Cc: Jingoo Han <jingoohan1@gmail.com>
> > Cc: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
> > ---
> >  .../pci/controller/dwc/pcie-designware-ep.c   |  30 ++-
> >  drivers/pci/controller/pcie-cadence-ep.c      |  11 +-
> >  drivers/pci/controller/pcie-rockchip-ep.c     |  13 +-
> >  drivers/pci/endpoint/functions/pci-epf-test.c |  29 +--
> >  drivers/pci/endpoint/pci-epc-core.c           |   7 +-
> >  drivers/pci/endpoint/pci-epc-mem.c            | 189 ++++++++++++++----
> >  include/linux/pci-epc.h                       |  43 ++--
> >  7 files changed, 234 insertions(+), 88 deletions(-)
> >
> > diff --git a/drivers/pci/controller/dwc/pcie-designware-ep.c b/drivers/pci/controller/dwc/pcie-designware-ep.c
> > index 3dd2e2697294..8d23c20b9afd 100644
> > --- a/drivers/pci/controller/dwc/pcie-designware-ep.c
> > +++ b/drivers/pci/controller/dwc/pcie-designware-ep.c
> > @@ -195,7 +195,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, u8 func_no,
> >  }
> >
> >  static int dw_pcie_ep_map_addr(struct pci_epc *epc, u8 func_no,
> > -                            phys_addr_t addr,
> > +                            phys_addr_t addr, int window,
> >                              u64 pci_addr, size_t size)
> >  {
> >       int ret;
> > @@ -367,6 +367,7 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
> >       unsigned int aligned_offset;
> >       u16 msg_ctrl, msg_data;
> >       u32 msg_addr_lower, msg_addr_upper, reg;
> > +     int window = PCI_EPC_DEFAULT_WINDOW;
> >       u64 msg_addr;
> >       bool has_upper;
> >       int ret;
> > @@ -390,11 +391,11 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
> >               reg = ep->msi_cap + PCI_MSI_DATA_32;
> >               msg_data = dw_pcie_readw_dbi(pci, reg);
> >       }
> > -     aligned_offset = msg_addr_lower & (epc->mem->page_size - 1);
> > +     aligned_offset = msg_addr_lower & (epc->mem[window]->page_size - 1);
> >       msg_addr = ((u64)msg_addr_upper) << 32 |
> >                       (msg_addr_lower & ~aligned_offset);
> > -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> > -                               epc->mem->page_size);
> > +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> > +                               msg_addr, epc->mem[window]->page_size);
> >       if (ret)
> >               return ret;
> >
> > @@ -416,6 +417,7 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >       u32 reg, msg_data, vec_ctrl;
> >       u64 tbl_addr, msg_addr, reg_u64;
> >       void __iomem *msix_tbl;
> > +     int window = PCI_EPC_DEFAULT_WINDOW;
> >       int ret;
> >
> >       reg = ep->msix_cap + PCI_MSIX_TABLE;
> > @@ -452,8 +454,8 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >               return -EPERM;
> >       }
> >
> > -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> > -                               epc->mem->page_size);
> > +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> > +                               msg_addr, epc->mem[window]->page_size);
> >       if (ret)
> >               return ret;
> >
> > @@ -466,10 +468,11 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >
> >  void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
> >  {
> > +     int window = PCI_EPC_DEFAULT_WINDOW;
> >       struct pci_epc *epc = ep->epc;
> >
> >       pci_epc_mem_free_addr(epc, ep->msi_mem_phys, ep->msi_mem,
> > -                           epc->mem->page_size);
> > +                           epc->mem[window]->page_size);
> >
> >       pci_epc_mem_exit(epc);
> >  }
> > @@ -499,9 +502,12 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
> >       u32 reg;
> >       void *addr;
> >       u8 hdr_type;
> > +     int window;
> >       unsigned int nbars;
> >       unsigned int offset;
> >       struct pci_epc *epc;
> > +     size_t msi_page_size;
> > +     struct pci_epc_mem_window mem_window;
> >       struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
> >       struct device *dev = pci->dev;
> >       struct device_node *np = dev->of_node;
> > @@ -574,15 +580,17 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
> >       if (ret < 0)
> >               epc->max_functions = 1;
> >
> > -     ret = __pci_epc_mem_init(epc, ep->phys_base, ep->addr_size,
> > -                              ep->page_size);
> > +     mem_window.phys_base = ep->phys_base;
> > +     mem_window.size = ep->addr_size;
> > +     ret = __pci_epc_mem_init(epc, &mem_window, 1, ep->page_size);
> >       if (ret < 0) {
> >               dev_err(dev, "Failed to initialize address space\n");
> >               return ret;
> >       }
> >
> > -     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys,
> > -                                          epc->mem->page_size);
> > +     msi_page_size = epc->mem[PCI_EPC_DEFAULT_WINDOW]->page_size;
> > +     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys, &window,
> > +                                          msi_page_size);
> >       if (!ep->msi_mem) {
> >               dev_err(dev, "Failed to reserve memory for MSI/MSI-X\n");
> >               return -ENOMEM;
> > diff --git a/drivers/pci/controller/pcie-cadence-ep.c b/drivers/pci/controller/pcie-cadence-ep.c
> > index def7820cb824..7991b38a5350 100644
> > --- a/drivers/pci/controller/pcie-cadence-ep.c
> > +++ b/drivers/pci/controller/pcie-cadence-ep.c
> > @@ -172,7 +172,7 @@ static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
> >  }
> >
> >  static int cdns_pcie_ep_map_addr(struct pci_epc *epc, u8 fn, phys_addr_t addr,
> > -                              u64 pci_addr, size_t size)
> > +                              int window, u64 pci_addr, size_t size)
> >  {
> >       struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
> >       struct cdns_pcie *pcie = &ep->pcie;
> > @@ -434,12 +434,14 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
> >  {
> >       struct device *dev = &pdev->dev;
> >       struct device_node *np = dev->of_node;
> > +     struct pci_epc_mem_window mem_window;
> >       struct cdns_pcie_ep *ep;
> >       struct cdns_pcie *pcie;
> >       struct pci_epc *epc;
> >       struct resource *res;
> >       int ret;
> >       int phy_count;
> > +     int window;
> >
> >       ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> >       if (!ep)
> > @@ -502,15 +504,16 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
> >       if (of_property_read_u8(np, "max-functions", &epc->max_functions) < 0)
> >               epc->max_functions = 1;
> >
> > -     ret = pci_epc_mem_init(epc, pcie->mem_res->start,
> > -                            resource_size(pcie->mem_res));
> > +     mem_window.phys_base = pcie->mem_res->start;
> > +     mem_window.size = resource_size(pcie->mem_res);
> > +     ret = pci_epc_mem_init(epc, &mem_window, 1);
> >       if (ret < 0) {
> >               dev_err(dev, "failed to initialize the memory space\n");
> >               goto err_init;
> >       }
> >
> >       ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> > -                                               SZ_128K);
> > +                                               &window, SZ_128K);
> >       if (!ep->irq_cpu_addr) {
> >               dev_err(dev, "failed to reserve memory space for MSI\n");
> >               ret = -ENOMEM;
> > diff --git a/drivers/pci/controller/pcie-rockchip-ep.c b/drivers/pci/controller/pcie-rockchip-ep.c
> > index d743b0a48988..d59e85c8d319 100644
> > --- a/drivers/pci/controller/pcie-rockchip-ep.c
> > +++ b/drivers/pci/controller/pcie-rockchip-ep.c
> > @@ -256,8 +256,8 @@ static void rockchip_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
> >  }
> >
> >  static int rockchip_pcie_ep_map_addr(struct pci_epc *epc, u8 fn,
> > -                                  phys_addr_t addr, u64 pci_addr,
> > -                                  size_t size)
> > +                                  phys_addr_t addr, int window,
> > +                                  u64 pci_addr, size_t size)
> >  {
> >       struct rockchip_pcie_ep *ep = epc_get_drvdata(epc);
> >       struct rockchip_pcie *pcie = &ep->rockchip;
> > @@ -562,11 +562,13 @@ static const struct of_device_id rockchip_pcie_ep_of_match[] = {
> >
> >  static int rockchip_pcie_ep_probe(struct platform_device *pdev)
> >  {
> > +     struct pci_epc_mem_window mem_window;
> >       struct device *dev = &pdev->dev;
> >       struct rockchip_pcie_ep *ep;
> >       struct rockchip_pcie *rockchip;
> >       struct pci_epc *epc;
> >       size_t max_regions;
> > +     int window;
> >       int err;
> >
> >       ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> > @@ -614,15 +616,16 @@ static int rockchip_pcie_ep_probe(struct platform_device *pdev)
> >       /* Only enable function 0 by default */
> >       rockchip_pcie_write(rockchip, BIT(0), PCIE_CORE_PHY_FUNC_CFG);
> >
> > -     err = pci_epc_mem_init(epc, rockchip->mem_res->start,
> > -                            resource_size(rockchip->mem_res));
> > +     mem_window.phys_base = rockchip->mem_res->start;
> > +     mem_window.size = resource_size(rockchip->mem_res);
> > +     err = pci_epc_mem_init(epc, &mem_window, 1);
> >       if (err < 0) {
> >               dev_err(dev, "failed to initialize the memory space\n");
> >               goto err_uninit_port;
> >       }
> >
> >       ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> > -                                               SZ_128K);
> > +                                               &window, SZ_128K);
> >       if (!ep->irq_cpu_addr) {
> >               dev_err(dev, "failed to reserve memory space for MSI\n");
> >               err = -ENOMEM;
> > diff --git a/drivers/pci/endpoint/functions/pci-epf-test.c b/drivers/pci/endpoint/functions/pci-epf-test.c
> > index 5d74f81ddfe4..475228011703 100644
> > --- a/drivers/pci/endpoint/functions/pci-epf-test.c
> > +++ b/drivers/pci/endpoint/functions/pci-epf-test.c
> > @@ -84,8 +84,10 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >       struct pci_epc *epc = epf->epc;
> >       enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >       struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> > +     int window;
> >
> > -     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr, reg->size);
> > +     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr,
> > +                                       &window, reg->size);
> >       if (!src_addr) {
> >               dev_err(dev, "Failed to allocate source address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> > @@ -93,15 +95,16 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >               goto err;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, reg->src_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, window,
> > +                            reg->src_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map source address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> >               goto err_src_addr;
> >       }
> >
> > -     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr, reg->size);
> > +     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr,
> > +                                       &window, reg->size);
> >       if (!dst_addr) {
> >               dev_err(dev, "Failed to allocate destination address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > @@ -109,8 +112,8 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >               goto err_src_map_addr;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, reg->dst_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, window,
> > +                            reg->dst_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map destination address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > @@ -146,8 +149,9 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
> >       struct pci_epc *epc = epf->epc;
> >       enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >       struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> > +     int window;
> >
> > -     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> > +     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
> >       if (!src_addr) {
> >               dev_err(dev, "Failed to allocate address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> > @@ -155,8 +159,8 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
> >               goto err;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->src_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> > +                            reg->src_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> > @@ -193,13 +197,14 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
> >       void __iomem *dst_addr;
> >       void *buf;
> >       phys_addr_t phys_addr;
> > +     int window;
> >       struct pci_epf *epf = epf_test->epf;
> >       struct device *dev = &epf->dev;
> >       struct pci_epc *epc = epf->epc;
> >       enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >       struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> >
> > -     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> > +     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
> >       if (!dst_addr) {
> >               dev_err(dev, "Failed to allocate address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > @@ -207,8 +212,8 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
> >               goto err;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->dst_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> > +                            reg->dst_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > diff --git a/drivers/pci/endpoint/pci-epc-core.c b/drivers/pci/endpoint/pci-epc-core.c
> > index 2091508c1620..289c266c2d90 100644
> > --- a/drivers/pci/endpoint/pci-epc-core.c
> > +++ b/drivers/pci/endpoint/pci-epc-core.c
> > @@ -358,13 +358,15 @@ EXPORT_SYMBOL_GPL(pci_epc_unmap_addr);
> >   * @epc: the EPC device on which address is allocated
> >   * @func_no: the endpoint function number in the EPC device
> >   * @phys_addr: physical address of the local system
> > + * @window: index to the window region where PCI address will be mapped
> >   * @pci_addr: PCI address to which the physical address should be mapped
> >   * @size: the size of the allocation
> >   *
> >   * Invoke to map CPU address with PCI address.
> >   */
> >  int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> > -                  phys_addr_t phys_addr, u64 pci_addr, size_t size)
> > +                  phys_addr_t phys_addr, int window,
> > +                  u64 pci_addr, size_t size)
> >  {
> >       int ret;
> >       unsigned long flags;
> > @@ -376,7 +378,8 @@ int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> >               return 0;
> >
> >       spin_lock_irqsave(&epc->lock, flags);
> > -     ret = epc->ops->map_addr(epc, func_no, phys_addr, pci_addr, size);
> > +     ret = epc->ops->map_addr(epc, func_no, phys_addr,
> > +                              window, pci_addr, size);
> >       spin_unlock_irqrestore(&epc->lock, flags);
> >
> >       return ret;
> > diff --git a/drivers/pci/endpoint/pci-epc-mem.c b/drivers/pci/endpoint/pci-epc-mem.c
> > index d2b174ce15de..c955f2c97944 100644
> > --- a/drivers/pci/endpoint/pci-epc-mem.c
> > +++ b/drivers/pci/endpoint/pci-epc-mem.c
> > @@ -39,56 +39,77 @@ static int pci_epc_mem_get_order(struct pci_epc_mem *mem, size_t size)
> >   * __pci_epc_mem_init() - initialize the pci_epc_mem structure
> >   * @epc: the EPC device that invoked pci_epc_mem_init
> >   * @phys_base: the physical address of the base
> > - * @size: the size of the address space
> > + * @num_windows: number of windows device supports
>
> struct pci_epc_mem_window is missing here.
oops my bad will fix that.

> >   * @page_size: size of each page
> >   *
> >   * Invoke to initialize the pci_epc_mem structure used by the
> >   * endpoint functions to allocate mapped PCI address.
> >   */
> > -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_base, size_t size,
> > -                    size_t page_size)
> > +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *windows,
> > +                    int num_windows, size_t page_size)
> >  {
> > -     int ret;
> > -     struct pci_epc_mem *mem;
> > -     unsigned long *bitmap;
> > +     struct pci_epc_mem *mem = NULL;
> > +     unsigned long *bitmap = NULL;
> >       unsigned int page_shift;
> > -     int pages;
> >       int bitmap_size;
> > +     int pages;
> > +     int ret;
> > +     int i;
> > +
> > +     epc->mem_windows = 0;
> > +
> > +     if (!windows)
> > +             return -EINVAL;
> > +
> > +     if (num_windows <= 0)
> > +             return -EINVAL;
> >
> >       if (page_size < PAGE_SIZE)
> >               page_size = PAGE_SIZE;
> >
> >       page_shift = ilog2(page_size);
> > -     pages = size >> page_shift;
> > -     bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
> >
> > -     mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> > -     if (!mem) {
> > -             ret = -ENOMEM;
> > -             goto err;
> > -     }
> > +     epc->mem = kcalloc(num_windows, sizeof(*mem), GFP_KERNEL);
> > +     if (!epc->mem)
> > +             return -EINVAL;
> >
> > -     bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> > -     if (!bitmap) {
> > -             ret = -ENOMEM;
> > -             goto err_mem;
> > -     }
> > +     for (i = 0; i < num_windows; i++) {
> > +             pages = windows[i].phys_base >> page_shift;
> > +             bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
> >
> > -     mem->bitmap = bitmap;
> > -     mem->phys_base = phys_base;
> > -     mem->page_size = page_size;
> > -     mem->pages = pages;
> > -     mem->size = size;
> > +             mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> > +             if (!mem) {
> > +                     ret = -ENOMEM;
> > +                     goto err_mem;
> > +             }
> >
> > -     epc->mem = mem;
> > +             bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> > +             if (!bitmap) {
> > +                     ret = -ENOMEM;
> > +                     goto err_mem;
> > +             }
> > +
> > +             mem->bitmap = bitmap;
> > +             mem->window.phys_base = windows[i].phys_base;
> > +             mem->page_size = page_size;
> > +             mem->pages = pages;
> > +             mem->window.size = windows[i].size;
> > +             mem->window.map_size = 0;
> > +
> > +             epc->mem[i] = mem;
> > +     }
> > +     epc->mem_windows = num_windows;
> >
> >       return 0;
> >
> >  err_mem:
> > -     kfree(mem);
> > +     for (; i >= 0; i--) {
> > +             kfree(mem->bitmap);
> > +             kfree(epc->mem[i]);
> > +     }
> > +     kfree(epc->mem);
> >
> > -err:
> > -return ret;
> > +     return ret;
> >  }
> >  EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
> >
> > @@ -101,48 +122,126 @@ EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
> >   */
> >  void pci_epc_mem_exit(struct pci_epc *epc)
> >  {
> > -     struct pci_epc_mem *mem = epc->mem;
> > +     struct pci_epc_mem *mem;
> > +     int i;
> > +
> > +     if (!epc->mem_windows)
> > +             return;
> > +
> > +     for (i = 0; i <= epc->mem_windows; i--) {
> > +             mem = epc->mem[i];
> > +             kfree(mem->bitmap);
> > +             kfree(epc->mem[i]);
> > +     }
> > +     kfree(epc->mem);
> >
> >       epc->mem = NULL;
> > -     kfree(mem->bitmap);
> > -     kfree(mem);
> > +     epc->mem_windows = 0;
> >  }
> >  EXPORT_SYMBOL_GPL(pci_epc_mem_exit);
> >
> > +static int pci_epc_find_best_fit_window(struct pci_epc *epc, size_t size)
> > +{
> > +     size_t window_least_size = 0;
> > +     int best_fit_window = -1;
> > +     struct pci_epc_mem *mem;
> > +     size_t actual_size;
> > +     int i;
> > +
> > +     for (i = 0; i < epc->mem_windows; i++) {
> > +             mem = epc->mem[i];
> > +
> > +             /* if chunk from this region is already used skip it */
> > +             if (mem->window.map_size)
> > +                     continue;
> > +
> > +             actual_size = ALIGN(size, mem->page_size);
> > +
> > +             if (best_fit_window == -1) {
> > +                     best_fit_window = i;
> > +                     window_least_size = mem->window.size;
> > +             } else {
> > +                     if (actual_size <= mem->window.size &&
> > +                         mem->window.size < window_least_size) {
> > +                             best_fit_window = i;
> > +                             window_least_size = mem->window.size;
> > +                     }
> > +             }
> > +     }
> > +
> > +     return best_fit_window;
> > +}
> > +
> >  /**
> >   * pci_epc_mem_alloc_addr() - allocate memory address from EPC addr space
> >   * @epc: the EPC device on which memory has to be allocated
> >   * @phys_addr: populate the allocated physical address here
> > + * @window: populate the window here which will be used to map PCI address
> >   * @size: the size of the address space that has to be allocated
> >   *
> >   * Invoke to allocate memory address from the EPC address space. This
> >   * is usually done to map the remote RC address into the local system.
> >   */
> >  void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
> > -                                  phys_addr_t *phys_addr, size_t size)
> > +                                  phys_addr_t *phys_addr,
> > +                                  int *window, size_t size)
> >  {
> > +     int best_fit = PCI_EPC_DEFAULT_WINDOW;
> > +     void __iomem *virt_addr = NULL;
> > +     struct pci_epc_mem *mem;
> > +     unsigned int page_shift;
> >       int pageno;
> > -     void __iomem *virt_addr;
> > -     struct pci_epc_mem *mem = epc->mem;
> > -     unsigned int page_shift = ilog2(mem->page_size);
> >       int order;
> >
> > +     if (epc->mem_windows <= 0)
> > +             return NULL;
> > +
> > +     if (epc->mem_windows > 1) {
> > +             best_fit = pci_epc_find_best_fit_window(epc, size);
> > +             if (best_fit < 0)
> > +                     return NULL;
> > +     }
> > +
> > +     mem = epc->mem[best_fit];
> >       size = ALIGN(size, mem->page_size);
> > +     if (size > (mem->window.size - mem->window.map_size))
> > +             return NULL;
>
> Assume I have two mem regions, first region is of size 128MB and the second
> region is of size 4GB. If there is a allocation request for 4GB, will the
> allocation succeed?
>
yes it would succeed, the pci_epc_find_best_fit_window() would get the
corresponding window.

Cheers,
--Prabhakar Lad

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 2/5] pci: endpoint: add support to handle multiple base for mapping outbound memory
@ 2019-11-27 21:21       ` Lad, Prabhakar
  0 siblings, 0 replies; 60+ messages in thread
From: Lad, Prabhakar @ 2019-11-27 21:21 UTC (permalink / raw)
  To: Kishon Vijay Abraham I
  Cc: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Marek Vasut, Yoshihiro Shimoda, linux-pci,
	Catalin Marinas, Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS, LKML,
	LAK

Hi Kishon,

Thank you for the review.

On Wed, Nov 27, 2019 at 5:15 AM Kishon Vijay Abraham I <kishon@ti.com> wrote:
>
> Hi Prabhakar,
>
> On 07/11/19 1:06 AM, Lad Prabhakar wrote:
> > From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >
> > rcar pcie controller has support to map multiple memory regions
> > for mapping the outbound memory in local system, this feature
> > inspires to add support for handling multiple memory bases in
> > endpoint framework. In case of multiple memory regions only chunk
> > or complete region can be mapped and this window needs to be
> > passed to the controller driver.
> >
> > Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > Cc: <linux-rockchip@lists.infradead.org>
> > Cc: Shawn Lin <shawn.lin@rock-chips.com>
> > Cc: Heiko Stuebner <heiko@sntech.de>
> > Cc: Tom Joseph <tjoseph@cadence.com>
> > Cc: Jingoo Han <jingoohan1@gmail.com>
> > Cc: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
> > ---
> >  .../pci/controller/dwc/pcie-designware-ep.c   |  30 ++-
> >  drivers/pci/controller/pcie-cadence-ep.c      |  11 +-
> >  drivers/pci/controller/pcie-rockchip-ep.c     |  13 +-
> >  drivers/pci/endpoint/functions/pci-epf-test.c |  29 +--
> >  drivers/pci/endpoint/pci-epc-core.c           |   7 +-
> >  drivers/pci/endpoint/pci-epc-mem.c            | 189 ++++++++++++++----
> >  include/linux/pci-epc.h                       |  43 ++--
> >  7 files changed, 234 insertions(+), 88 deletions(-)
> >
> > diff --git a/drivers/pci/controller/dwc/pcie-designware-ep.c b/drivers/pci/controller/dwc/pcie-designware-ep.c
> > index 3dd2e2697294..8d23c20b9afd 100644
> > --- a/drivers/pci/controller/dwc/pcie-designware-ep.c
> > +++ b/drivers/pci/controller/dwc/pcie-designware-ep.c
> > @@ -195,7 +195,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, u8 func_no,
> >  }
> >
> >  static int dw_pcie_ep_map_addr(struct pci_epc *epc, u8 func_no,
> > -                            phys_addr_t addr,
> > +                            phys_addr_t addr, int window,
> >                              u64 pci_addr, size_t size)
> >  {
> >       int ret;
> > @@ -367,6 +367,7 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
> >       unsigned int aligned_offset;
> >       u16 msg_ctrl, msg_data;
> >       u32 msg_addr_lower, msg_addr_upper, reg;
> > +     int window = PCI_EPC_DEFAULT_WINDOW;
> >       u64 msg_addr;
> >       bool has_upper;
> >       int ret;
> > @@ -390,11 +391,11 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
> >               reg = ep->msi_cap + PCI_MSI_DATA_32;
> >               msg_data = dw_pcie_readw_dbi(pci, reg);
> >       }
> > -     aligned_offset = msg_addr_lower & (epc->mem->page_size - 1);
> > +     aligned_offset = msg_addr_lower & (epc->mem[window]->page_size - 1);
> >       msg_addr = ((u64)msg_addr_upper) << 32 |
> >                       (msg_addr_lower & ~aligned_offset);
> > -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> > -                               epc->mem->page_size);
> > +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> > +                               msg_addr, epc->mem[window]->page_size);
> >       if (ret)
> >               return ret;
> >
> > @@ -416,6 +417,7 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >       u32 reg, msg_data, vec_ctrl;
> >       u64 tbl_addr, msg_addr, reg_u64;
> >       void __iomem *msix_tbl;
> > +     int window = PCI_EPC_DEFAULT_WINDOW;
> >       int ret;
> >
> >       reg = ep->msix_cap + PCI_MSIX_TABLE;
> > @@ -452,8 +454,8 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >               return -EPERM;
> >       }
> >
> > -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> > -                               epc->mem->page_size);
> > +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> > +                               msg_addr, epc->mem[window]->page_size);
> >       if (ret)
> >               return ret;
> >
> > @@ -466,10 +468,11 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >
> >  void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
> >  {
> > +     int window = PCI_EPC_DEFAULT_WINDOW;
> >       struct pci_epc *epc = ep->epc;
> >
> >       pci_epc_mem_free_addr(epc, ep->msi_mem_phys, ep->msi_mem,
> > -                           epc->mem->page_size);
> > +                           epc->mem[window]->page_size);
> >
> >       pci_epc_mem_exit(epc);
> >  }
> > @@ -499,9 +502,12 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
> >       u32 reg;
> >       void *addr;
> >       u8 hdr_type;
> > +     int window;
> >       unsigned int nbars;
> >       unsigned int offset;
> >       struct pci_epc *epc;
> > +     size_t msi_page_size;
> > +     struct pci_epc_mem_window mem_window;
> >       struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
> >       struct device *dev = pci->dev;
> >       struct device_node *np = dev->of_node;
> > @@ -574,15 +580,17 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
> >       if (ret < 0)
> >               epc->max_functions = 1;
> >
> > -     ret = __pci_epc_mem_init(epc, ep->phys_base, ep->addr_size,
> > -                              ep->page_size);
> > +     mem_window.phys_base = ep->phys_base;
> > +     mem_window.size = ep->addr_size;
> > +     ret = __pci_epc_mem_init(epc, &mem_window, 1, ep->page_size);
> >       if (ret < 0) {
> >               dev_err(dev, "Failed to initialize address space\n");
> >               return ret;
> >       }
> >
> > -     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys,
> > -                                          epc->mem->page_size);
> > +     msi_page_size = epc->mem[PCI_EPC_DEFAULT_WINDOW]->page_size;
> > +     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys, &window,
> > +                                          msi_page_size);
> >       if (!ep->msi_mem) {
> >               dev_err(dev, "Failed to reserve memory for MSI/MSI-X\n");
> >               return -ENOMEM;
> > diff --git a/drivers/pci/controller/pcie-cadence-ep.c b/drivers/pci/controller/pcie-cadence-ep.c
> > index def7820cb824..7991b38a5350 100644
> > --- a/drivers/pci/controller/pcie-cadence-ep.c
> > +++ b/drivers/pci/controller/pcie-cadence-ep.c
> > @@ -172,7 +172,7 @@ static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
> >  }
> >
> >  static int cdns_pcie_ep_map_addr(struct pci_epc *epc, u8 fn, phys_addr_t addr,
> > -                              u64 pci_addr, size_t size)
> > +                              int window, u64 pci_addr, size_t size)
> >  {
> >       struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
> >       struct cdns_pcie *pcie = &ep->pcie;
> > @@ -434,12 +434,14 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
> >  {
> >       struct device *dev = &pdev->dev;
> >       struct device_node *np = dev->of_node;
> > +     struct pci_epc_mem_window mem_window;
> >       struct cdns_pcie_ep *ep;
> >       struct cdns_pcie *pcie;
> >       struct pci_epc *epc;
> >       struct resource *res;
> >       int ret;
> >       int phy_count;
> > +     int window;
> >
> >       ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> >       if (!ep)
> > @@ -502,15 +504,16 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
> >       if (of_property_read_u8(np, "max-functions", &epc->max_functions) < 0)
> >               epc->max_functions = 1;
> >
> > -     ret = pci_epc_mem_init(epc, pcie->mem_res->start,
> > -                            resource_size(pcie->mem_res));
> > +     mem_window.phys_base = pcie->mem_res->start;
> > +     mem_window.size = resource_size(pcie->mem_res);
> > +     ret = pci_epc_mem_init(epc, &mem_window, 1);
> >       if (ret < 0) {
> >               dev_err(dev, "failed to initialize the memory space\n");
> >               goto err_init;
> >       }
> >
> >       ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> > -                                               SZ_128K);
> > +                                               &window, SZ_128K);
> >       if (!ep->irq_cpu_addr) {
> >               dev_err(dev, "failed to reserve memory space for MSI\n");
> >               ret = -ENOMEM;
> > diff --git a/drivers/pci/controller/pcie-rockchip-ep.c b/drivers/pci/controller/pcie-rockchip-ep.c
> > index d743b0a48988..d59e85c8d319 100644
> > --- a/drivers/pci/controller/pcie-rockchip-ep.c
> > +++ b/drivers/pci/controller/pcie-rockchip-ep.c
> > @@ -256,8 +256,8 @@ static void rockchip_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
> >  }
> >
> >  static int rockchip_pcie_ep_map_addr(struct pci_epc *epc, u8 fn,
> > -                                  phys_addr_t addr, u64 pci_addr,
> > -                                  size_t size)
> > +                                  phys_addr_t addr, int window,
> > +                                  u64 pci_addr, size_t size)
> >  {
> >       struct rockchip_pcie_ep *ep = epc_get_drvdata(epc);
> >       struct rockchip_pcie *pcie = &ep->rockchip;
> > @@ -562,11 +562,13 @@ static const struct of_device_id rockchip_pcie_ep_of_match[] = {
> >
> >  static int rockchip_pcie_ep_probe(struct platform_device *pdev)
> >  {
> > +     struct pci_epc_mem_window mem_window;
> >       struct device *dev = &pdev->dev;
> >       struct rockchip_pcie_ep *ep;
> >       struct rockchip_pcie *rockchip;
> >       struct pci_epc *epc;
> >       size_t max_regions;
> > +     int window;
> >       int err;
> >
> >       ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> > @@ -614,15 +616,16 @@ static int rockchip_pcie_ep_probe(struct platform_device *pdev)
> >       /* Only enable function 0 by default */
> >       rockchip_pcie_write(rockchip, BIT(0), PCIE_CORE_PHY_FUNC_CFG);
> >
> > -     err = pci_epc_mem_init(epc, rockchip->mem_res->start,
> > -                            resource_size(rockchip->mem_res));
> > +     mem_window.phys_base = rockchip->mem_res->start;
> > +     mem_window.size = resource_size(rockchip->mem_res);
> > +     err = pci_epc_mem_init(epc, &mem_window, 1);
> >       if (err < 0) {
> >               dev_err(dev, "failed to initialize the memory space\n");
> >               goto err_uninit_port;
> >       }
> >
> >       ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> > -                                               SZ_128K);
> > +                                               &window, SZ_128K);
> >       if (!ep->irq_cpu_addr) {
> >               dev_err(dev, "failed to reserve memory space for MSI\n");
> >               err = -ENOMEM;
> > diff --git a/drivers/pci/endpoint/functions/pci-epf-test.c b/drivers/pci/endpoint/functions/pci-epf-test.c
> > index 5d74f81ddfe4..475228011703 100644
> > --- a/drivers/pci/endpoint/functions/pci-epf-test.c
> > +++ b/drivers/pci/endpoint/functions/pci-epf-test.c
> > @@ -84,8 +84,10 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >       struct pci_epc *epc = epf->epc;
> >       enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >       struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> > +     int window;
> >
> > -     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr, reg->size);
> > +     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr,
> > +                                       &window, reg->size);
> >       if (!src_addr) {
> >               dev_err(dev, "Failed to allocate source address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> > @@ -93,15 +95,16 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >               goto err;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, reg->src_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, window,
> > +                            reg->src_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map source address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> >               goto err_src_addr;
> >       }
> >
> > -     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr, reg->size);
> > +     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr,
> > +                                       &window, reg->size);
> >       if (!dst_addr) {
> >               dev_err(dev, "Failed to allocate destination address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > @@ -109,8 +112,8 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >               goto err_src_map_addr;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, reg->dst_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, window,
> > +                            reg->dst_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map destination address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > @@ -146,8 +149,9 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
> >       struct pci_epc *epc = epf->epc;
> >       enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >       struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> > +     int window;
> >
> > -     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> > +     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
> >       if (!src_addr) {
> >               dev_err(dev, "Failed to allocate address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> > @@ -155,8 +159,8 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
> >               goto err;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->src_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> > +                            reg->src_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> > @@ -193,13 +197,14 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
> >       void __iomem *dst_addr;
> >       void *buf;
> >       phys_addr_t phys_addr;
> > +     int window;
> >       struct pci_epf *epf = epf_test->epf;
> >       struct device *dev = &epf->dev;
> >       struct pci_epc *epc = epf->epc;
> >       enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >       struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> >
> > -     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> > +     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
> >       if (!dst_addr) {
> >               dev_err(dev, "Failed to allocate address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > @@ -207,8 +212,8 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
> >               goto err;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->dst_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> > +                            reg->dst_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > diff --git a/drivers/pci/endpoint/pci-epc-core.c b/drivers/pci/endpoint/pci-epc-core.c
> > index 2091508c1620..289c266c2d90 100644
> > --- a/drivers/pci/endpoint/pci-epc-core.c
> > +++ b/drivers/pci/endpoint/pci-epc-core.c
> > @@ -358,13 +358,15 @@ EXPORT_SYMBOL_GPL(pci_epc_unmap_addr);
> >   * @epc: the EPC device on which address is allocated
> >   * @func_no: the endpoint function number in the EPC device
> >   * @phys_addr: physical address of the local system
> > + * @window: index to the window region where PCI address will be mapped
> >   * @pci_addr: PCI address to which the physical address should be mapped
> >   * @size: the size of the allocation
> >   *
> >   * Invoke to map CPU address with PCI address.
> >   */
> >  int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> > -                  phys_addr_t phys_addr, u64 pci_addr, size_t size)
> > +                  phys_addr_t phys_addr, int window,
> > +                  u64 pci_addr, size_t size)
> >  {
> >       int ret;
> >       unsigned long flags;
> > @@ -376,7 +378,8 @@ int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> >               return 0;
> >
> >       spin_lock_irqsave(&epc->lock, flags);
> > -     ret = epc->ops->map_addr(epc, func_no, phys_addr, pci_addr, size);
> > +     ret = epc->ops->map_addr(epc, func_no, phys_addr,
> > +                              window, pci_addr, size);
> >       spin_unlock_irqrestore(&epc->lock, flags);
> >
> >       return ret;
> > diff --git a/drivers/pci/endpoint/pci-epc-mem.c b/drivers/pci/endpoint/pci-epc-mem.c
> > index d2b174ce15de..c955f2c97944 100644
> > --- a/drivers/pci/endpoint/pci-epc-mem.c
> > +++ b/drivers/pci/endpoint/pci-epc-mem.c
> > @@ -39,56 +39,77 @@ static int pci_epc_mem_get_order(struct pci_epc_mem *mem, size_t size)
> >   * __pci_epc_mem_init() - initialize the pci_epc_mem structure
> >   * @epc: the EPC device that invoked pci_epc_mem_init
> >   * @phys_base: the physical address of the base
> > - * @size: the size of the address space
> > + * @num_windows: number of windows device supports
>
> struct pci_epc_mem_window is missing here.
oops my bad will fix that.

> >   * @page_size: size of each page
> >   *
> >   * Invoke to initialize the pci_epc_mem structure used by the
> >   * endpoint functions to allocate mapped PCI address.
> >   */
> > -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_base, size_t size,
> > -                    size_t page_size)
> > +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *windows,
> > +                    int num_windows, size_t page_size)
> >  {
> > -     int ret;
> > -     struct pci_epc_mem *mem;
> > -     unsigned long *bitmap;
> > +     struct pci_epc_mem *mem = NULL;
> > +     unsigned long *bitmap = NULL;
> >       unsigned int page_shift;
> > -     int pages;
> >       int bitmap_size;
> > +     int pages;
> > +     int ret;
> > +     int i;
> > +
> > +     epc->mem_windows = 0;
> > +
> > +     if (!windows)
> > +             return -EINVAL;
> > +
> > +     if (num_windows <= 0)
> > +             return -EINVAL;
> >
> >       if (page_size < PAGE_SIZE)
> >               page_size = PAGE_SIZE;
> >
> >       page_shift = ilog2(page_size);
> > -     pages = size >> page_shift;
> > -     bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
> >
> > -     mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> > -     if (!mem) {
> > -             ret = -ENOMEM;
> > -             goto err;
> > -     }
> > +     epc->mem = kcalloc(num_windows, sizeof(*mem), GFP_KERNEL);
> > +     if (!epc->mem)
> > +             return -EINVAL;
> >
> > -     bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> > -     if (!bitmap) {
> > -             ret = -ENOMEM;
> > -             goto err_mem;
> > -     }
> > +     for (i = 0; i < num_windows; i++) {
> > +             pages = windows[i].phys_base >> page_shift;
> > +             bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
> >
> > -     mem->bitmap = bitmap;
> > -     mem->phys_base = phys_base;
> > -     mem->page_size = page_size;
> > -     mem->pages = pages;
> > -     mem->size = size;
> > +             mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> > +             if (!mem) {
> > +                     ret = -ENOMEM;
> > +                     goto err_mem;
> > +             }
> >
> > -     epc->mem = mem;
> > +             bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> > +             if (!bitmap) {
> > +                     ret = -ENOMEM;
> > +                     goto err_mem;
> > +             }
> > +
> > +             mem->bitmap = bitmap;
> > +             mem->window.phys_base = windows[i].phys_base;
> > +             mem->page_size = page_size;
> > +             mem->pages = pages;
> > +             mem->window.size = windows[i].size;
> > +             mem->window.map_size = 0;
> > +
> > +             epc->mem[i] = mem;
> > +     }
> > +     epc->mem_windows = num_windows;
> >
> >       return 0;
> >
> >  err_mem:
> > -     kfree(mem);
> > +     for (; i >= 0; i--) {
> > +             kfree(mem->bitmap);
> > +             kfree(epc->mem[i]);
> > +     }
> > +     kfree(epc->mem);
> >
> > -err:
> > -return ret;
> > +     return ret;
> >  }
> >  EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
> >
> > @@ -101,48 +122,126 @@ EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
> >   */
> >  void pci_epc_mem_exit(struct pci_epc *epc)
> >  {
> > -     struct pci_epc_mem *mem = epc->mem;
> > +     struct pci_epc_mem *mem;
> > +     int i;
> > +
> > +     if (!epc->mem_windows)
> > +             return;
> > +
> > +     for (i = 0; i <= epc->mem_windows; i--) {
> > +             mem = epc->mem[i];
> > +             kfree(mem->bitmap);
> > +             kfree(epc->mem[i]);
> > +     }
> > +     kfree(epc->mem);
> >
> >       epc->mem = NULL;
> > -     kfree(mem->bitmap);
> > -     kfree(mem);
> > +     epc->mem_windows = 0;
> >  }
> >  EXPORT_SYMBOL_GPL(pci_epc_mem_exit);
> >
> > +static int pci_epc_find_best_fit_window(struct pci_epc *epc, size_t size)
> > +{
> > +     size_t window_least_size = 0;
> > +     int best_fit_window = -1;
> > +     struct pci_epc_mem *mem;
> > +     size_t actual_size;
> > +     int i;
> > +
> > +     for (i = 0; i < epc->mem_windows; i++) {
> > +             mem = epc->mem[i];
> > +
> > +             /* if chunk from this region is already used skip it */
> > +             if (mem->window.map_size)
> > +                     continue;
> > +
> > +             actual_size = ALIGN(size, mem->page_size);
> > +
> > +             if (best_fit_window == -1) {
> > +                     best_fit_window = i;
> > +                     window_least_size = mem->window.size;
> > +             } else {
> > +                     if (actual_size <= mem->window.size &&
> > +                         mem->window.size < window_least_size) {
> > +                             best_fit_window = i;
> > +                             window_least_size = mem->window.size;
> > +                     }
> > +             }
> > +     }
> > +
> > +     return best_fit_window;
> > +}
> > +
> >  /**
> >   * pci_epc_mem_alloc_addr() - allocate memory address from EPC addr space
> >   * @epc: the EPC device on which memory has to be allocated
> >   * @phys_addr: populate the allocated physical address here
> > + * @window: populate the window here which will be used to map PCI address
> >   * @size: the size of the address space that has to be allocated
> >   *
> >   * Invoke to allocate memory address from the EPC address space. This
> >   * is usually done to map the remote RC address into the local system.
> >   */
> >  void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
> > -                                  phys_addr_t *phys_addr, size_t size)
> > +                                  phys_addr_t *phys_addr,
> > +                                  int *window, size_t size)
> >  {
> > +     int best_fit = PCI_EPC_DEFAULT_WINDOW;
> > +     void __iomem *virt_addr = NULL;
> > +     struct pci_epc_mem *mem;
> > +     unsigned int page_shift;
> >       int pageno;
> > -     void __iomem *virt_addr;
> > -     struct pci_epc_mem *mem = epc->mem;
> > -     unsigned int page_shift = ilog2(mem->page_size);
> >       int order;
> >
> > +     if (epc->mem_windows <= 0)
> > +             return NULL;
> > +
> > +     if (epc->mem_windows > 1) {
> > +             best_fit = pci_epc_find_best_fit_window(epc, size);
> > +             if (best_fit < 0)
> > +                     return NULL;
> > +     }
> > +
> > +     mem = epc->mem[best_fit];
> >       size = ALIGN(size, mem->page_size);
> > +     if (size > (mem->window.size - mem->window.map_size))
> > +             return NULL;
>
> Assume I have two mem regions, first region is of size 128MB and the second
> region is of size 4GB. If there is a allocation request for 4GB, will the
> allocation succeed?
>
yes it would succeed, the pci_epc_find_best_fit_window() would get the
corresponding window.

Cheers,
--Prabhakar Lad

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 2/5] pci: endpoint: add support to handle multiple base for mapping outbound memory
@ 2019-11-27 21:21       ` Lad, Prabhakar
  0 siblings, 0 replies; 60+ messages in thread
From: Lad, Prabhakar @ 2019-11-27 21:21 UTC (permalink / raw)
  To: Kishon Vijay Abraham I
  Cc: Mark Rutland, Heiko Stuebner, Geert Uytterhoeven, linux-pci,
	Shawn Lin, Will Deacon, Marek Vasut, Lorenzo Pieralisi,
	Magnus Damm, linux-rockchip, Catalin Marinas,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Chris Paterson, Arnd Bergmann, Lad, Prabhakar, Rob Herring,
	Bjorn Helgaas, LAK, Greg Kroah-Hartman, Yoshihiro Shimoda, LKML,
	Linux-Renesas, Tom Joseph, Jingoo Han, Andrew Murray,
	Gustavo Pimentel

Hi Kishon,

Thank you for the review.

On Wed, Nov 27, 2019 at 5:15 AM Kishon Vijay Abraham I <kishon@ti.com> wrote:
>
> Hi Prabhakar,
>
> On 07/11/19 1:06 AM, Lad Prabhakar wrote:
> > From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >
> > rcar pcie controller has support to map multiple memory regions
> > for mapping the outbound memory in local system, this feature
> > inspires to add support for handling multiple memory bases in
> > endpoint framework. In case of multiple memory regions only chunk
> > or complete region can be mapped and this window needs to be
> > passed to the controller driver.
> >
> > Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > Cc: <linux-rockchip@lists.infradead.org>
> > Cc: Shawn Lin <shawn.lin@rock-chips.com>
> > Cc: Heiko Stuebner <heiko@sntech.de>
> > Cc: Tom Joseph <tjoseph@cadence.com>
> > Cc: Jingoo Han <jingoohan1@gmail.com>
> > Cc: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
> > ---
> >  .../pci/controller/dwc/pcie-designware-ep.c   |  30 ++-
> >  drivers/pci/controller/pcie-cadence-ep.c      |  11 +-
> >  drivers/pci/controller/pcie-rockchip-ep.c     |  13 +-
> >  drivers/pci/endpoint/functions/pci-epf-test.c |  29 +--
> >  drivers/pci/endpoint/pci-epc-core.c           |   7 +-
> >  drivers/pci/endpoint/pci-epc-mem.c            | 189 ++++++++++++++----
> >  include/linux/pci-epc.h                       |  43 ++--
> >  7 files changed, 234 insertions(+), 88 deletions(-)
> >
> > diff --git a/drivers/pci/controller/dwc/pcie-designware-ep.c b/drivers/pci/controller/dwc/pcie-designware-ep.c
> > index 3dd2e2697294..8d23c20b9afd 100644
> > --- a/drivers/pci/controller/dwc/pcie-designware-ep.c
> > +++ b/drivers/pci/controller/dwc/pcie-designware-ep.c
> > @@ -195,7 +195,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, u8 func_no,
> >  }
> >
> >  static int dw_pcie_ep_map_addr(struct pci_epc *epc, u8 func_no,
> > -                            phys_addr_t addr,
> > +                            phys_addr_t addr, int window,
> >                              u64 pci_addr, size_t size)
> >  {
> >       int ret;
> > @@ -367,6 +367,7 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
> >       unsigned int aligned_offset;
> >       u16 msg_ctrl, msg_data;
> >       u32 msg_addr_lower, msg_addr_upper, reg;
> > +     int window = PCI_EPC_DEFAULT_WINDOW;
> >       u64 msg_addr;
> >       bool has_upper;
> >       int ret;
> > @@ -390,11 +391,11 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
> >               reg = ep->msi_cap + PCI_MSI_DATA_32;
> >               msg_data = dw_pcie_readw_dbi(pci, reg);
> >       }
> > -     aligned_offset = msg_addr_lower & (epc->mem->page_size - 1);
> > +     aligned_offset = msg_addr_lower & (epc->mem[window]->page_size - 1);
> >       msg_addr = ((u64)msg_addr_upper) << 32 |
> >                       (msg_addr_lower & ~aligned_offset);
> > -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> > -                               epc->mem->page_size);
> > +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> > +                               msg_addr, epc->mem[window]->page_size);
> >       if (ret)
> >               return ret;
> >
> > @@ -416,6 +417,7 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >       u32 reg, msg_data, vec_ctrl;
> >       u64 tbl_addr, msg_addr, reg_u64;
> >       void __iomem *msix_tbl;
> > +     int window = PCI_EPC_DEFAULT_WINDOW;
> >       int ret;
> >
> >       reg = ep->msix_cap + PCI_MSIX_TABLE;
> > @@ -452,8 +454,8 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >               return -EPERM;
> >       }
> >
> > -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> > -                               epc->mem->page_size);
> > +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> > +                               msg_addr, epc->mem[window]->page_size);
> >       if (ret)
> >               return ret;
> >
> > @@ -466,10 +468,11 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >
> >  void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
> >  {
> > +     int window = PCI_EPC_DEFAULT_WINDOW;
> >       struct pci_epc *epc = ep->epc;
> >
> >       pci_epc_mem_free_addr(epc, ep->msi_mem_phys, ep->msi_mem,
> > -                           epc->mem->page_size);
> > +                           epc->mem[window]->page_size);
> >
> >       pci_epc_mem_exit(epc);
> >  }
> > @@ -499,9 +502,12 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
> >       u32 reg;
> >       void *addr;
> >       u8 hdr_type;
> > +     int window;
> >       unsigned int nbars;
> >       unsigned int offset;
> >       struct pci_epc *epc;
> > +     size_t msi_page_size;
> > +     struct pci_epc_mem_window mem_window;
> >       struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
> >       struct device *dev = pci->dev;
> >       struct device_node *np = dev->of_node;
> > @@ -574,15 +580,17 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
> >       if (ret < 0)
> >               epc->max_functions = 1;
> >
> > -     ret = __pci_epc_mem_init(epc, ep->phys_base, ep->addr_size,
> > -                              ep->page_size);
> > +     mem_window.phys_base = ep->phys_base;
> > +     mem_window.size = ep->addr_size;
> > +     ret = __pci_epc_mem_init(epc, &mem_window, 1, ep->page_size);
> >       if (ret < 0) {
> >               dev_err(dev, "Failed to initialize address space\n");
> >               return ret;
> >       }
> >
> > -     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys,
> > -                                          epc->mem->page_size);
> > +     msi_page_size = epc->mem[PCI_EPC_DEFAULT_WINDOW]->page_size;
> > +     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys, &window,
> > +                                          msi_page_size);
> >       if (!ep->msi_mem) {
> >               dev_err(dev, "Failed to reserve memory for MSI/MSI-X\n");
> >               return -ENOMEM;
> > diff --git a/drivers/pci/controller/pcie-cadence-ep.c b/drivers/pci/controller/pcie-cadence-ep.c
> > index def7820cb824..7991b38a5350 100644
> > --- a/drivers/pci/controller/pcie-cadence-ep.c
> > +++ b/drivers/pci/controller/pcie-cadence-ep.c
> > @@ -172,7 +172,7 @@ static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
> >  }
> >
> >  static int cdns_pcie_ep_map_addr(struct pci_epc *epc, u8 fn, phys_addr_t addr,
> > -                              u64 pci_addr, size_t size)
> > +                              int window, u64 pci_addr, size_t size)
> >  {
> >       struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
> >       struct cdns_pcie *pcie = &ep->pcie;
> > @@ -434,12 +434,14 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
> >  {
> >       struct device *dev = &pdev->dev;
> >       struct device_node *np = dev->of_node;
> > +     struct pci_epc_mem_window mem_window;
> >       struct cdns_pcie_ep *ep;
> >       struct cdns_pcie *pcie;
> >       struct pci_epc *epc;
> >       struct resource *res;
> >       int ret;
> >       int phy_count;
> > +     int window;
> >
> >       ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> >       if (!ep)
> > @@ -502,15 +504,16 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
> >       if (of_property_read_u8(np, "max-functions", &epc->max_functions) < 0)
> >               epc->max_functions = 1;
> >
> > -     ret = pci_epc_mem_init(epc, pcie->mem_res->start,
> > -                            resource_size(pcie->mem_res));
> > +     mem_window.phys_base = pcie->mem_res->start;
> > +     mem_window.size = resource_size(pcie->mem_res);
> > +     ret = pci_epc_mem_init(epc, &mem_window, 1);
> >       if (ret < 0) {
> >               dev_err(dev, "failed to initialize the memory space\n");
> >               goto err_init;
> >       }
> >
> >       ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> > -                                               SZ_128K);
> > +                                               &window, SZ_128K);
> >       if (!ep->irq_cpu_addr) {
> >               dev_err(dev, "failed to reserve memory space for MSI\n");
> >               ret = -ENOMEM;
> > diff --git a/drivers/pci/controller/pcie-rockchip-ep.c b/drivers/pci/controller/pcie-rockchip-ep.c
> > index d743b0a48988..d59e85c8d319 100644
> > --- a/drivers/pci/controller/pcie-rockchip-ep.c
> > +++ b/drivers/pci/controller/pcie-rockchip-ep.c
> > @@ -256,8 +256,8 @@ static void rockchip_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
> >  }
> >
> >  static int rockchip_pcie_ep_map_addr(struct pci_epc *epc, u8 fn,
> > -                                  phys_addr_t addr, u64 pci_addr,
> > -                                  size_t size)
> > +                                  phys_addr_t addr, int window,
> > +                                  u64 pci_addr, size_t size)
> >  {
> >       struct rockchip_pcie_ep *ep = epc_get_drvdata(epc);
> >       struct rockchip_pcie *pcie = &ep->rockchip;
> > @@ -562,11 +562,13 @@ static const struct of_device_id rockchip_pcie_ep_of_match[] = {
> >
> >  static int rockchip_pcie_ep_probe(struct platform_device *pdev)
> >  {
> > +     struct pci_epc_mem_window mem_window;
> >       struct device *dev = &pdev->dev;
> >       struct rockchip_pcie_ep *ep;
> >       struct rockchip_pcie *rockchip;
> >       struct pci_epc *epc;
> >       size_t max_regions;
> > +     int window;
> >       int err;
> >
> >       ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> > @@ -614,15 +616,16 @@ static int rockchip_pcie_ep_probe(struct platform_device *pdev)
> >       /* Only enable function 0 by default */
> >       rockchip_pcie_write(rockchip, BIT(0), PCIE_CORE_PHY_FUNC_CFG);
> >
> > -     err = pci_epc_mem_init(epc, rockchip->mem_res->start,
> > -                            resource_size(rockchip->mem_res));
> > +     mem_window.phys_base = rockchip->mem_res->start;
> > +     mem_window.size = resource_size(rockchip->mem_res);
> > +     err = pci_epc_mem_init(epc, &mem_window, 1);
> >       if (err < 0) {
> >               dev_err(dev, "failed to initialize the memory space\n");
> >               goto err_uninit_port;
> >       }
> >
> >       ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> > -                                               SZ_128K);
> > +                                               &window, SZ_128K);
> >       if (!ep->irq_cpu_addr) {
> >               dev_err(dev, "failed to reserve memory space for MSI\n");
> >               err = -ENOMEM;
> > diff --git a/drivers/pci/endpoint/functions/pci-epf-test.c b/drivers/pci/endpoint/functions/pci-epf-test.c
> > index 5d74f81ddfe4..475228011703 100644
> > --- a/drivers/pci/endpoint/functions/pci-epf-test.c
> > +++ b/drivers/pci/endpoint/functions/pci-epf-test.c
> > @@ -84,8 +84,10 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >       struct pci_epc *epc = epf->epc;
> >       enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >       struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> > +     int window;
> >
> > -     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr, reg->size);
> > +     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr,
> > +                                       &window, reg->size);
> >       if (!src_addr) {
> >               dev_err(dev, "Failed to allocate source address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> > @@ -93,15 +95,16 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >               goto err;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, reg->src_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, window,
> > +                            reg->src_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map source address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> >               goto err_src_addr;
> >       }
> >
> > -     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr, reg->size);
> > +     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr,
> > +                                       &window, reg->size);
> >       if (!dst_addr) {
> >               dev_err(dev, "Failed to allocate destination address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > @@ -109,8 +112,8 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >               goto err_src_map_addr;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, reg->dst_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, window,
> > +                            reg->dst_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map destination address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > @@ -146,8 +149,9 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
> >       struct pci_epc *epc = epf->epc;
> >       enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >       struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> > +     int window;
> >
> > -     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> > +     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
> >       if (!src_addr) {
> >               dev_err(dev, "Failed to allocate address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> > @@ -155,8 +159,8 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
> >               goto err;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->src_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> > +                            reg->src_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> > @@ -193,13 +197,14 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
> >       void __iomem *dst_addr;
> >       void *buf;
> >       phys_addr_t phys_addr;
> > +     int window;
> >       struct pci_epf *epf = epf_test->epf;
> >       struct device *dev = &epf->dev;
> >       struct pci_epc *epc = epf->epc;
> >       enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >       struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> >
> > -     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> > +     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
> >       if (!dst_addr) {
> >               dev_err(dev, "Failed to allocate address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > @@ -207,8 +212,8 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
> >               goto err;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->dst_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> > +                            reg->dst_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > diff --git a/drivers/pci/endpoint/pci-epc-core.c b/drivers/pci/endpoint/pci-epc-core.c
> > index 2091508c1620..289c266c2d90 100644
> > --- a/drivers/pci/endpoint/pci-epc-core.c
> > +++ b/drivers/pci/endpoint/pci-epc-core.c
> > @@ -358,13 +358,15 @@ EXPORT_SYMBOL_GPL(pci_epc_unmap_addr);
> >   * @epc: the EPC device on which address is allocated
> >   * @func_no: the endpoint function number in the EPC device
> >   * @phys_addr: physical address of the local system
> > + * @window: index to the window region where PCI address will be mapped
> >   * @pci_addr: PCI address to which the physical address should be mapped
> >   * @size: the size of the allocation
> >   *
> >   * Invoke to map CPU address with PCI address.
> >   */
> >  int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> > -                  phys_addr_t phys_addr, u64 pci_addr, size_t size)
> > +                  phys_addr_t phys_addr, int window,
> > +                  u64 pci_addr, size_t size)
> >  {
> >       int ret;
> >       unsigned long flags;
> > @@ -376,7 +378,8 @@ int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> >               return 0;
> >
> >       spin_lock_irqsave(&epc->lock, flags);
> > -     ret = epc->ops->map_addr(epc, func_no, phys_addr, pci_addr, size);
> > +     ret = epc->ops->map_addr(epc, func_no, phys_addr,
> > +                              window, pci_addr, size);
> >       spin_unlock_irqrestore(&epc->lock, flags);
> >
> >       return ret;
> > diff --git a/drivers/pci/endpoint/pci-epc-mem.c b/drivers/pci/endpoint/pci-epc-mem.c
> > index d2b174ce15de..c955f2c97944 100644
> > --- a/drivers/pci/endpoint/pci-epc-mem.c
> > +++ b/drivers/pci/endpoint/pci-epc-mem.c
> > @@ -39,56 +39,77 @@ static int pci_epc_mem_get_order(struct pci_epc_mem *mem, size_t size)
> >   * __pci_epc_mem_init() - initialize the pci_epc_mem structure
> >   * @epc: the EPC device that invoked pci_epc_mem_init
> >   * @phys_base: the physical address of the base
> > - * @size: the size of the address space
> > + * @num_windows: number of windows device supports
>
> struct pci_epc_mem_window is missing here.
oops my bad will fix that.

> >   * @page_size: size of each page
> >   *
> >   * Invoke to initialize the pci_epc_mem structure used by the
> >   * endpoint functions to allocate mapped PCI address.
> >   */
> > -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_base, size_t size,
> > -                    size_t page_size)
> > +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *windows,
> > +                    int num_windows, size_t page_size)
> >  {
> > -     int ret;
> > -     struct pci_epc_mem *mem;
> > -     unsigned long *bitmap;
> > +     struct pci_epc_mem *mem = NULL;
> > +     unsigned long *bitmap = NULL;
> >       unsigned int page_shift;
> > -     int pages;
> >       int bitmap_size;
> > +     int pages;
> > +     int ret;
> > +     int i;
> > +
> > +     epc->mem_windows = 0;
> > +
> > +     if (!windows)
> > +             return -EINVAL;
> > +
> > +     if (num_windows <= 0)
> > +             return -EINVAL;
> >
> >       if (page_size < PAGE_SIZE)
> >               page_size = PAGE_SIZE;
> >
> >       page_shift = ilog2(page_size);
> > -     pages = size >> page_shift;
> > -     bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
> >
> > -     mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> > -     if (!mem) {
> > -             ret = -ENOMEM;
> > -             goto err;
> > -     }
> > +     epc->mem = kcalloc(num_windows, sizeof(*mem), GFP_KERNEL);
> > +     if (!epc->mem)
> > +             return -EINVAL;
> >
> > -     bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> > -     if (!bitmap) {
> > -             ret = -ENOMEM;
> > -             goto err_mem;
> > -     }
> > +     for (i = 0; i < num_windows; i++) {
> > +             pages = windows[i].phys_base >> page_shift;
> > +             bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
> >
> > -     mem->bitmap = bitmap;
> > -     mem->phys_base = phys_base;
> > -     mem->page_size = page_size;
> > -     mem->pages = pages;
> > -     mem->size = size;
> > +             mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> > +             if (!mem) {
> > +                     ret = -ENOMEM;
> > +                     goto err_mem;
> > +             }
> >
> > -     epc->mem = mem;
> > +             bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> > +             if (!bitmap) {
> > +                     ret = -ENOMEM;
> > +                     goto err_mem;
> > +             }
> > +
> > +             mem->bitmap = bitmap;
> > +             mem->window.phys_base = windows[i].phys_base;
> > +             mem->page_size = page_size;
> > +             mem->pages = pages;
> > +             mem->window.size = windows[i].size;
> > +             mem->window.map_size = 0;
> > +
> > +             epc->mem[i] = mem;
> > +     }
> > +     epc->mem_windows = num_windows;
> >
> >       return 0;
> >
> >  err_mem:
> > -     kfree(mem);
> > +     for (; i >= 0; i--) {
> > +             kfree(mem->bitmap);
> > +             kfree(epc->mem[i]);
> > +     }
> > +     kfree(epc->mem);
> >
> > -err:
> > -return ret;
> > +     return ret;
> >  }
> >  EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
> >
> > @@ -101,48 +122,126 @@ EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
> >   */
> >  void pci_epc_mem_exit(struct pci_epc *epc)
> >  {
> > -     struct pci_epc_mem *mem = epc->mem;
> > +     struct pci_epc_mem *mem;
> > +     int i;
> > +
> > +     if (!epc->mem_windows)
> > +             return;
> > +
> > +     for (i = 0; i <= epc->mem_windows; i--) {
> > +             mem = epc->mem[i];
> > +             kfree(mem->bitmap);
> > +             kfree(epc->mem[i]);
> > +     }
> > +     kfree(epc->mem);
> >
> >       epc->mem = NULL;
> > -     kfree(mem->bitmap);
> > -     kfree(mem);
> > +     epc->mem_windows = 0;
> >  }
> >  EXPORT_SYMBOL_GPL(pci_epc_mem_exit);
> >
> > +static int pci_epc_find_best_fit_window(struct pci_epc *epc, size_t size)
> > +{
> > +     size_t window_least_size = 0;
> > +     int best_fit_window = -1;
> > +     struct pci_epc_mem *mem;
> > +     size_t actual_size;
> > +     int i;
> > +
> > +     for (i = 0; i < epc->mem_windows; i++) {
> > +             mem = epc->mem[i];
> > +
> > +             /* if chunk from this region is already used skip it */
> > +             if (mem->window.map_size)
> > +                     continue;
> > +
> > +             actual_size = ALIGN(size, mem->page_size);
> > +
> > +             if (best_fit_window == -1) {
> > +                     best_fit_window = i;
> > +                     window_least_size = mem->window.size;
> > +             } else {
> > +                     if (actual_size <= mem->window.size &&
> > +                         mem->window.size < window_least_size) {
> > +                             best_fit_window = i;
> > +                             window_least_size = mem->window.size;
> > +                     }
> > +             }
> > +     }
> > +
> > +     return best_fit_window;
> > +}
> > +
> >  /**
> >   * pci_epc_mem_alloc_addr() - allocate memory address from EPC addr space
> >   * @epc: the EPC device on which memory has to be allocated
> >   * @phys_addr: populate the allocated physical address here
> > + * @window: populate the window here which will be used to map PCI address
> >   * @size: the size of the address space that has to be allocated
> >   *
> >   * Invoke to allocate memory address from the EPC address space. This
> >   * is usually done to map the remote RC address into the local system.
> >   */
> >  void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
> > -                                  phys_addr_t *phys_addr, size_t size)
> > +                                  phys_addr_t *phys_addr,
> > +                                  int *window, size_t size)
> >  {
> > +     int best_fit = PCI_EPC_DEFAULT_WINDOW;
> > +     void __iomem *virt_addr = NULL;
> > +     struct pci_epc_mem *mem;
> > +     unsigned int page_shift;
> >       int pageno;
> > -     void __iomem *virt_addr;
> > -     struct pci_epc_mem *mem = epc->mem;
> > -     unsigned int page_shift = ilog2(mem->page_size);
> >       int order;
> >
> > +     if (epc->mem_windows <= 0)
> > +             return NULL;
> > +
> > +     if (epc->mem_windows > 1) {
> > +             best_fit = pci_epc_find_best_fit_window(epc, size);
> > +             if (best_fit < 0)
> > +                     return NULL;
> > +     }
> > +
> > +     mem = epc->mem[best_fit];
> >       size = ALIGN(size, mem->page_size);
> > +     if (size > (mem->window.size - mem->window.map_size))
> > +             return NULL;
>
> Assume I have two mem regions, first region is of size 128MB and the second
> region is of size 4GB. If there is a allocation request for 4GB, will the
> allocation succeed?
>
yes it would succeed, the pci_epc_find_best_fit_window() would get the
corresponding window.

Cheers,
--Prabhakar Lad

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 2/5] pci: endpoint: add support to handle multiple base for mapping outbound memory
  2019-11-27  6:05     ` Kishon Vijay Abraham I
  (?)
@ 2019-11-27 21:28       ` Lad, Prabhakar
  -1 siblings, 0 replies; 60+ messages in thread
From: Lad, Prabhakar @ 2019-11-27 21:28 UTC (permalink / raw)
  To: Kishon Vijay Abraham I
  Cc: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Marek Vasut, Yoshihiro Shimoda, linux-pci,
	Catalin Marinas, Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS, LKML,
	LAK, Linux-Renesas, Chris Paterson, Lad, Prabhakar,
	linux-rockchip, Shawn Lin, Heiko Stuebner, Tom Joseph,
	Jingoo Han, Gustavo Pimentel

Hi Kishon,

On Wed, Nov 27, 2019 at 6:06 AM Kishon Vijay Abraham I <kishon@ti.com> wrote:
>
> Hi Prabhakar,
>
> On 07/11/19 1:06 AM, Lad Prabhakar wrote:
> > From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >
> > rcar pcie controller has support to map multiple memory regions
> > for mapping the outbound memory in local system, this feature
> > inspires to add support for handling multiple memory bases in
> > endpoint framework. In case of multiple memory regions only chunk
> > or complete region can be mapped and this window needs to be
> > passed to the controller driver.
>
> Wouldn't this result in lot of unused address and exhausting the memory regions
> quickly? I think we could have a "type" associated with each of these regions
> and smaller allocations could use one "type" and larger allocation could use a
> different "type"? For instance we allocate memory regions for MSI interrupt and
> for data transfer. We could use smaller allocations for MSI and larger
> allocations for data transfers.
>
> We could also extend the "type" for certain special requirements. For instance
> I've seen platforms which have dedicated address space for high priority data
> transfers (uses higher virtual channel numbers) and dedicated address space for
> low priority data transfers (uses lower virtual channel numbers).
>
> Here too we could give a special flag or type for higher priority address
> regions and different type for lower priority address. While allocating, the
> function drivers should provide a flag specifying the region where memory
> should be allocated.
>
agreed but in this controller if chunk is allocated from the window
this cannot be re-used
for further allocations, I haven't come across a where a window could
be used with multiple
chunks being allocated. so as you suggest I could extend the type to
have additional flag
which would indicate if window can be used for multiple chunk allocations.

> Can you add flags in struct pci_epc_mem_window and see how that comes out?
>
sure will do that.

Cheers,
--Prabhakar

> Thanks
> Kishon
>
> >
> > Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > Cc: <linux-rockchip@lists.infradead.org>
> > Cc: Shawn Lin <shawn.lin@rock-chips.com>
> > Cc: Heiko Stuebner <heiko@sntech.de>
> > Cc: Tom Joseph <tjoseph@cadence.com>
> > Cc: Jingoo Han <jingoohan1@gmail.com>
> > Cc: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
> > ---
> >  .../pci/controller/dwc/pcie-designware-ep.c   |  30 ++-
> >  drivers/pci/controller/pcie-cadence-ep.c      |  11 +-
> >  drivers/pci/controller/pcie-rockchip-ep.c     |  13 +-
> >  drivers/pci/endpoint/functions/pci-epf-test.c |  29 +--
> >  drivers/pci/endpoint/pci-epc-core.c           |   7 +-
> >  drivers/pci/endpoint/pci-epc-mem.c            | 189 ++++++++++++++----
> >  include/linux/pci-epc.h                       |  43 ++--
> >  7 files changed, 234 insertions(+), 88 deletions(-)
> >
> > diff --git a/drivers/pci/controller/dwc/pcie-designware-ep.c b/drivers/pci/controller/dwc/pcie-designware-ep.c
> > index 3dd2e2697294..8d23c20b9afd 100644
> > --- a/drivers/pci/controller/dwc/pcie-designware-ep.c
> > +++ b/drivers/pci/controller/dwc/pcie-designware-ep.c
> > @@ -195,7 +195,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, u8 func_no,
> >  }
> >
> >  static int dw_pcie_ep_map_addr(struct pci_epc *epc, u8 func_no,
> > -                            phys_addr_t addr,
> > +                            phys_addr_t addr, int window,
> >                              u64 pci_addr, size_t size)
> >  {
> >       int ret;
> > @@ -367,6 +367,7 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
> >       unsigned int aligned_offset;
> >       u16 msg_ctrl, msg_data;
> >       u32 msg_addr_lower, msg_addr_upper, reg;
> > +     int window = PCI_EPC_DEFAULT_WINDOW;
> >       u64 msg_addr;
> >       bool has_upper;
> >       int ret;
> > @@ -390,11 +391,11 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
> >               reg = ep->msi_cap + PCI_MSI_DATA_32;
> >               msg_data = dw_pcie_readw_dbi(pci, reg);
> >       }
> > -     aligned_offset = msg_addr_lower & (epc->mem->page_size - 1);
> > +     aligned_offset = msg_addr_lower & (epc->mem[window]->page_size - 1);
> >       msg_addr = ((u64)msg_addr_upper) << 32 |
> >                       (msg_addr_lower & ~aligned_offset);
> > -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> > -                               epc->mem->page_size);
> > +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> > +                               msg_addr, epc->mem[window]->page_size);
> >       if (ret)
> >               return ret;
> >
> > @@ -416,6 +417,7 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >       u32 reg, msg_data, vec_ctrl;
> >       u64 tbl_addr, msg_addr, reg_u64;
> >       void __iomem *msix_tbl;
> > +     int window = PCI_EPC_DEFAULT_WINDOW;
> >       int ret;
> >
> >       reg = ep->msix_cap + PCI_MSIX_TABLE;
> > @@ -452,8 +454,8 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >               return -EPERM;
> >       }
> >
> > -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> > -                               epc->mem->page_size);
> > +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> > +                               msg_addr, epc->mem[window]->page_size);
> >       if (ret)
> >               return ret;
> >
> > @@ -466,10 +468,11 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >
> >  void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
> >  {
> > +     int window = PCI_EPC_DEFAULT_WINDOW;
> >       struct pci_epc *epc = ep->epc;
> >
> >       pci_epc_mem_free_addr(epc, ep->msi_mem_phys, ep->msi_mem,
> > -                           epc->mem->page_size);
> > +                           epc->mem[window]->page_size);
> >
> >       pci_epc_mem_exit(epc);
> >  }
> > @@ -499,9 +502,12 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
> >       u32 reg;
> >       void *addr;
> >       u8 hdr_type;
> > +     int window;
> >       unsigned int nbars;
> >       unsigned int offset;
> >       struct pci_epc *epc;
> > +     size_t msi_page_size;
> > +     struct pci_epc_mem_window mem_window;
> >       struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
> >       struct device *dev = pci->dev;
> >       struct device_node *np = dev->of_node;
> > @@ -574,15 +580,17 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
> >       if (ret < 0)
> >               epc->max_functions = 1;
> >
> > -     ret = __pci_epc_mem_init(epc, ep->phys_base, ep->addr_size,
> > -                              ep->page_size);
> > +     mem_window.phys_base = ep->phys_base;
> > +     mem_window.size = ep->addr_size;
> > +     ret = __pci_epc_mem_init(epc, &mem_window, 1, ep->page_size);
> >       if (ret < 0) {
> >               dev_err(dev, "Failed to initialize address space\n");
> >               return ret;
> >       }
> >
> > -     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys,
> > -                                          epc->mem->page_size);
> > +     msi_page_size = epc->mem[PCI_EPC_DEFAULT_WINDOW]->page_size;
> > +     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys, &window,
> > +                                          msi_page_size);
> >       if (!ep->msi_mem) {
> >               dev_err(dev, "Failed to reserve memory for MSI/MSI-X\n");
> >               return -ENOMEM;
> > diff --git a/drivers/pci/controller/pcie-cadence-ep.c b/drivers/pci/controller/pcie-cadence-ep.c
> > index def7820cb824..7991b38a5350 100644
> > --- a/drivers/pci/controller/pcie-cadence-ep.c
> > +++ b/drivers/pci/controller/pcie-cadence-ep.c
> > @@ -172,7 +172,7 @@ static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
> >  }
> >
> >  static int cdns_pcie_ep_map_addr(struct pci_epc *epc, u8 fn, phys_addr_t addr,
> > -                              u64 pci_addr, size_t size)
> > +                              int window, u64 pci_addr, size_t size)
> >  {
> >       struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
> >       struct cdns_pcie *pcie = &ep->pcie;
> > @@ -434,12 +434,14 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
> >  {
> >       struct device *dev = &pdev->dev;
> >       struct device_node *np = dev->of_node;
> > +     struct pci_epc_mem_window mem_window;
> >       struct cdns_pcie_ep *ep;
> >       struct cdns_pcie *pcie;
> >       struct pci_epc *epc;
> >       struct resource *res;
> >       int ret;
> >       int phy_count;
> > +     int window;
> >
> >       ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> >       if (!ep)
> > @@ -502,15 +504,16 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
> >       if (of_property_read_u8(np, "max-functions", &epc->max_functions) < 0)
> >               epc->max_functions = 1;
> >
> > -     ret = pci_epc_mem_init(epc, pcie->mem_res->start,
> > -                            resource_size(pcie->mem_res));
> > +     mem_window.phys_base = pcie->mem_res->start;
> > +     mem_window.size = resource_size(pcie->mem_res);
> > +     ret = pci_epc_mem_init(epc, &mem_window, 1);
> >       if (ret < 0) {
> >               dev_err(dev, "failed to initialize the memory space\n");
> >               goto err_init;
> >       }
> >
> >       ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> > -                                               SZ_128K);
> > +                                               &window, SZ_128K);
> >       if (!ep->irq_cpu_addr) {
> >               dev_err(dev, "failed to reserve memory space for MSI\n");
> >               ret = -ENOMEM;
> > diff --git a/drivers/pci/controller/pcie-rockchip-ep.c b/drivers/pci/controller/pcie-rockchip-ep.c
> > index d743b0a48988..d59e85c8d319 100644
> > --- a/drivers/pci/controller/pcie-rockchip-ep.c
> > +++ b/drivers/pci/controller/pcie-rockchip-ep.c
> > @@ -256,8 +256,8 @@ static void rockchip_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
> >  }
> >
> >  static int rockchip_pcie_ep_map_addr(struct pci_epc *epc, u8 fn,
> > -                                  phys_addr_t addr, u64 pci_addr,
> > -                                  size_t size)
> > +                                  phys_addr_t addr, int window,
> > +                                  u64 pci_addr, size_t size)
> >  {
> >       struct rockchip_pcie_ep *ep = epc_get_drvdata(epc);
> >       struct rockchip_pcie *pcie = &ep->rockchip;
> > @@ -562,11 +562,13 @@ static const struct of_device_id rockchip_pcie_ep_of_match[] = {
> >
> >  static int rockchip_pcie_ep_probe(struct platform_device *pdev)
> >  {
> > +     struct pci_epc_mem_window mem_window;
> >       struct device *dev = &pdev->dev;
> >       struct rockchip_pcie_ep *ep;
> >       struct rockchip_pcie *rockchip;
> >       struct pci_epc *epc;
> >       size_t max_regions;
> > +     int window;
> >       int err;
> >
> >       ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> > @@ -614,15 +616,16 @@ static int rockchip_pcie_ep_probe(struct platform_device *pdev)
> >       /* Only enable function 0 by default */
> >       rockchip_pcie_write(rockchip, BIT(0), PCIE_CORE_PHY_FUNC_CFG);
> >
> > -     err = pci_epc_mem_init(epc, rockchip->mem_res->start,
> > -                            resource_size(rockchip->mem_res));
> > +     mem_window.phys_base = rockchip->mem_res->start;
> > +     mem_window.size = resource_size(rockchip->mem_res);
> > +     err = pci_epc_mem_init(epc, &mem_window, 1);
> >       if (err < 0) {
> >               dev_err(dev, "failed to initialize the memory space\n");
> >               goto err_uninit_port;
> >       }
> >
> >       ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> > -                                               SZ_128K);
> > +                                               &window, SZ_128K);
> >       if (!ep->irq_cpu_addr) {
> >               dev_err(dev, "failed to reserve memory space for MSI\n");
> >               err = -ENOMEM;
> > diff --git a/drivers/pci/endpoint/functions/pci-epf-test.c b/drivers/pci/endpoint/functions/pci-epf-test.c
> > index 5d74f81ddfe4..475228011703 100644
> > --- a/drivers/pci/endpoint/functions/pci-epf-test.c
> > +++ b/drivers/pci/endpoint/functions/pci-epf-test.c
> > @@ -84,8 +84,10 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >       struct pci_epc *epc = epf->epc;
> >       enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >       struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> > +     int window;
> >
> > -     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr, reg->size);
> > +     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr,
> > +                                       &window, reg->size);
> >       if (!src_addr) {
> >               dev_err(dev, "Failed to allocate source address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> > @@ -93,15 +95,16 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >               goto err;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, reg->src_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, window,
> > +                            reg->src_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map source address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> >               goto err_src_addr;
> >       }
> >
> > -     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr, reg->size);
> > +     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr,
> > +                                       &window, reg->size);
> >       if (!dst_addr) {
> >               dev_err(dev, "Failed to allocate destination address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > @@ -109,8 +112,8 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >               goto err_src_map_addr;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, reg->dst_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, window,
> > +                            reg->dst_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map destination address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > @@ -146,8 +149,9 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
> >       struct pci_epc *epc = epf->epc;
> >       enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >       struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> > +     int window;
> >
> > -     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> > +     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
> >       if (!src_addr) {
> >               dev_err(dev, "Failed to allocate address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> > @@ -155,8 +159,8 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
> >               goto err;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->src_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> > +                            reg->src_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> > @@ -193,13 +197,14 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
> >       void __iomem *dst_addr;
> >       void *buf;
> >       phys_addr_t phys_addr;
> > +     int window;
> >       struct pci_epf *epf = epf_test->epf;
> >       struct device *dev = &epf->dev;
> >       struct pci_epc *epc = epf->epc;
> >       enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >       struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> >
> > -     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> > +     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
> >       if (!dst_addr) {
> >               dev_err(dev, "Failed to allocate address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > @@ -207,8 +212,8 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
> >               goto err;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->dst_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> > +                            reg->dst_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > diff --git a/drivers/pci/endpoint/pci-epc-core.c b/drivers/pci/endpoint/pci-epc-core.c
> > index 2091508c1620..289c266c2d90 100644
> > --- a/drivers/pci/endpoint/pci-epc-core.c
> > +++ b/drivers/pci/endpoint/pci-epc-core.c
> > @@ -358,13 +358,15 @@ EXPORT_SYMBOL_GPL(pci_epc_unmap_addr);
> >   * @epc: the EPC device on which address is allocated
> >   * @func_no: the endpoint function number in the EPC device
> >   * @phys_addr: physical address of the local system
> > + * @window: index to the window region where PCI address will be mapped
> >   * @pci_addr: PCI address to which the physical address should be mapped
> >   * @size: the size of the allocation
> >   *
> >   * Invoke to map CPU address with PCI address.
> >   */
> >  int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> > -                  phys_addr_t phys_addr, u64 pci_addr, size_t size)
> > +                  phys_addr_t phys_addr, int window,
> > +                  u64 pci_addr, size_t size)
> >  {
> >       int ret;
> >       unsigned long flags;
> > @@ -376,7 +378,8 @@ int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> >               return 0;
> >
> >       spin_lock_irqsave(&epc->lock, flags);
> > -     ret = epc->ops->map_addr(epc, func_no, phys_addr, pci_addr, size);
> > +     ret = epc->ops->map_addr(epc, func_no, phys_addr,
> > +                              window, pci_addr, size);
> >       spin_unlock_irqrestore(&epc->lock, flags);
> >
> >       return ret;
> > diff --git a/drivers/pci/endpoint/pci-epc-mem.c b/drivers/pci/endpoint/pci-epc-mem.c
> > index d2b174ce15de..c955f2c97944 100644
> > --- a/drivers/pci/endpoint/pci-epc-mem.c
> > +++ b/drivers/pci/endpoint/pci-epc-mem.c
> > @@ -39,56 +39,77 @@ static int pci_epc_mem_get_order(struct pci_epc_mem *mem, size_t size)
> >   * __pci_epc_mem_init() - initialize the pci_epc_mem structure
> >   * @epc: the EPC device that invoked pci_epc_mem_init
> >   * @phys_base: the physical address of the base
> > - * @size: the size of the address space
> > + * @num_windows: number of windows device supports
> >   * @page_size: size of each page
> >   *
> >   * Invoke to initialize the pci_epc_mem structure used by the
> >   * endpoint functions to allocate mapped PCI address.
> >   */
> > -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_base, size_t size,
> > -                    size_t page_size)
> > +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *windows,
> > +                    int num_windows, size_t page_size)
> >  {
> > -     int ret;
> > -     struct pci_epc_mem *mem;
> > -     unsigned long *bitmap;
> > +     struct pci_epc_mem *mem = NULL;
> > +     unsigned long *bitmap = NULL;
> >       unsigned int page_shift;
> > -     int pages;
> >       int bitmap_size;
> > +     int pages;
> > +     int ret;
> > +     int i;
> > +
> > +     epc->mem_windows = 0;
> > +
> > +     if (!windows)
> > +             return -EINVAL;
> > +
> > +     if (num_windows <= 0)
> > +             return -EINVAL;
> >
> >       if (page_size < PAGE_SIZE)
> >               page_size = PAGE_SIZE;
> >
> >       page_shift = ilog2(page_size);
> > -     pages = size >> page_shift;
> > -     bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
> >
> > -     mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> > -     if (!mem) {
> > -             ret = -ENOMEM;
> > -             goto err;
> > -     }
> > +     epc->mem = kcalloc(num_windows, sizeof(*mem), GFP_KERNEL);
> > +     if (!epc->mem)
> > +             return -EINVAL;
> >
> > -     bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> > -     if (!bitmap) {
> > -             ret = -ENOMEM;
> > -             goto err_mem;
> > -     }
> > +     for (i = 0; i < num_windows; i++) {
> > +             pages = windows[i].phys_base >> page_shift;
> > +             bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
> >
> > -     mem->bitmap = bitmap;
> > -     mem->phys_base = phys_base;
> > -     mem->page_size = page_size;
> > -     mem->pages = pages;
> > -     mem->size = size;
> > +             mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> > +             if (!mem) {
> > +                     ret = -ENOMEM;
> > +                     goto err_mem;
> > +             }
> >
> > -     epc->mem = mem;
> > +             bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> > +             if (!bitmap) {
> > +                     ret = -ENOMEM;
> > +                     goto err_mem;
> > +             }
> > +
> > +             mem->bitmap = bitmap;
> > +             mem->window.phys_base = windows[i].phys_base;
> > +             mem->page_size = page_size;
> > +             mem->pages = pages;
> > +             mem->window.size = windows[i].size;
> > +             mem->window.map_size = 0;
> > +
> > +             epc->mem[i] = mem;
> > +     }
> > +     epc->mem_windows = num_windows;
> >
> >       return 0;
> >
> >  err_mem:
> > -     kfree(mem);
> > +     for (; i >= 0; i--) {
> > +             kfree(mem->bitmap);
> > +             kfree(epc->mem[i]);
> > +     }
> > +     kfree(epc->mem);
> >
> > -err:
> > -return ret;
> > +     return ret;
> >  }
> >  EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
> >
> > @@ -101,48 +122,126 @@ EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
> >   */
> >  void pci_epc_mem_exit(struct pci_epc *epc)
> >  {
> > -     struct pci_epc_mem *mem = epc->mem;
> > +     struct pci_epc_mem *mem;
> > +     int i;
> > +
> > +     if (!epc->mem_windows)
> > +             return;
> > +
> > +     for (i = 0; i <= epc->mem_windows; i--) {
> > +             mem = epc->mem[i];
> > +             kfree(mem->bitmap);
> > +             kfree(epc->mem[i]);
> > +     }
> > +     kfree(epc->mem);
> >
> >       epc->mem = NULL;
> > -     kfree(mem->bitmap);
> > -     kfree(mem);
> > +     epc->mem_windows = 0;
> >  }
> >  EXPORT_SYMBOL_GPL(pci_epc_mem_exit);
> >
> > +static int pci_epc_find_best_fit_window(struct pci_epc *epc, size_t size)
> > +{
> > +     size_t window_least_size = 0;
> > +     int best_fit_window = -1;
> > +     struct pci_epc_mem *mem;
> > +     size_t actual_size;
> > +     int i;
> > +
> > +     for (i = 0; i < epc->mem_windows; i++) {
> > +             mem = epc->mem[i];
> > +
> > +             /* if chunk from this region is already used skip it */
> > +             if (mem->window.map_size)
> > +                     continue;
> > +
> > +             actual_size = ALIGN(size, mem->page_size);
> > +
> > +             if (best_fit_window == -1) {
> > +                     best_fit_window = i;
> > +                     window_least_size = mem->window.size;
> > +             } else {
> > +                     if (actual_size <= mem->window.size &&
> > +                         mem->window.size < window_least_size) {
> > +                             best_fit_window = i;
> > +                             window_least_size = mem->window.size;
> > +                     }
> > +             }
> > +     }
> > +
> > +     return best_fit_window;
> > +}
> > +
> >  /**
> >   * pci_epc_mem_alloc_addr() - allocate memory address from EPC addr space
> >   * @epc: the EPC device on which memory has to be allocated
> >   * @phys_addr: populate the allocated physical address here
> > + * @window: populate the window here which will be used to map PCI address
> >   * @size: the size of the address space that has to be allocated
> >   *
> >   * Invoke to allocate memory address from the EPC address space. This
> >   * is usually done to map the remote RC address into the local system.
> >   */
> >  void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
> > -                                  phys_addr_t *phys_addr, size_t size)
> > +                                  phys_addr_t *phys_addr,
> > +                                  int *window, size_t size)
> >  {
> > +     int best_fit = PCI_EPC_DEFAULT_WINDOW;
> > +     void __iomem *virt_addr = NULL;
> > +     struct pci_epc_mem *mem;
> > +     unsigned int page_shift;
> >       int pageno;
> > -     void __iomem *virt_addr;
> > -     struct pci_epc_mem *mem = epc->mem;
> > -     unsigned int page_shift = ilog2(mem->page_size);
> >       int order;
> >
> > +     if (epc->mem_windows <= 0)
> > +             return NULL;
> > +
> > +     if (epc->mem_windows > 1) {
> > +             best_fit = pci_epc_find_best_fit_window(epc, size);
> > +             if (best_fit < 0)
> > +                     return NULL;
> > +     }
> > +
> > +     mem = epc->mem[best_fit];
> >       size = ALIGN(size, mem->page_size);
> > +     if (size > (mem->window.size - mem->window.map_size))
> > +             return NULL;
> > +     page_shift = ilog2(mem->page_size);
> >       order = pci_epc_mem_get_order(mem, size);
> >
> >       pageno = bitmap_find_free_region(mem->bitmap, mem->pages, order);
> >       if (pageno < 0)
> >               return NULL;
> >
> > -     *phys_addr = mem->phys_base + ((phys_addr_t)pageno << page_shift);
> > +     *phys_addr = mem->window.phys_base + (pageno << page_shift);
> >       virt_addr = ioremap(*phys_addr, size);
> > -     if (!virt_addr)
> > +     if (!virt_addr) {
> >               bitmap_release_region(mem->bitmap, pageno, order);
> > +     } else {
> > +             mem->window.map_size += size;
> > +             *window = best_fit;
> > +     }
> >
> >       return virt_addr;
> >  }
> >  EXPORT_SYMBOL_GPL(pci_epc_mem_alloc_addr);
> >
> > +static int pci_epc_get_matching_window(struct pci_epc *epc,
> > +                                    phys_addr_t phys_addr)
> > +{
> > +     struct pci_epc_mem *mem;
> > +     int i;
> > +
> > +     for (i = 0; i < epc->mem_windows; i++) {
> > +             mem = epc->mem[i];
> > +
> > +             if (mem->window.phys_base == phys_addr)
> > +                     return i;
> > +     }
> > +
> > +     return -EINVAL;
> > +}
> > +
> >  /**
> >   * pci_epc_mem_free_addr() - free the allocated memory address
> >   * @epc: the EPC device on which memory was allocated
> > @@ -155,16 +254,26 @@ EXPORT_SYMBOL_GPL(pci_epc_mem_alloc_addr);
> >  void pci_epc_mem_free_addr(struct pci_epc *epc, phys_addr_t phys_addr,
> >                          void __iomem *virt_addr, size_t size)
> >  {
> > +     struct pci_epc_mem *mem;
> > +     unsigned int page_shift;
> > +     int window = 0;
> >       int pageno;
> > -     struct pci_epc_mem *mem = epc->mem;
> > -     unsigned int page_shift = ilog2(mem->page_size);
> >       int order;
> >
> > +     if (epc->mem_windows > 1) {
> > +             window = pci_epc_get_matching_window(epc, phys_addr);
> > +             if (window < 0)
> > +                     return;
> > +     }
> > +
> > +     mem = epc->mem[window];
> > +     page_shift = ilog2(mem->page_size);
> >       iounmap(virt_addr);
> > -     pageno = (phys_addr - mem->phys_base) >> page_shift;
> > +     pageno = (phys_addr - mem->window.phys_base) >> page_shift;
> >       size = ALIGN(size, mem->page_size);
> >       order = pci_epc_mem_get_order(mem, size);
> >       bitmap_release_region(mem->bitmap, pageno, order);
> > +     mem->window.map_size -= size;
> >  }
> >  EXPORT_SYMBOL_GPL(pci_epc_mem_free_addr);
> >
> > diff --git a/include/linux/pci-epc.h b/include/linux/pci-epc.h
> > index 56f1846b9d39..5e86d22dd5c5 100644
> > --- a/include/linux/pci-epc.h
> > +++ b/include/linux/pci-epc.h
> > @@ -48,7 +48,8 @@ struct pci_epc_ops {
> >       void    (*clear_bar)(struct pci_epc *epc, u8 func_no,
> >                            struct pci_epf_bar *epf_bar);
> >       int     (*map_addr)(struct pci_epc *epc, u8 func_no,
> > -                         phys_addr_t addr, u64 pci_addr, size_t size);
> > +                         phys_addr_t addr, int window,
> > +                         u64 pci_addr, size_t size);
> >       void    (*unmap_addr)(struct pci_epc *epc, u8 func_no,
> >                             phys_addr_t addr);
> >       int     (*set_msi)(struct pci_epc *epc, u8 func_no, u8 interrupts);
> > @@ -64,17 +65,28 @@ struct pci_epc_ops {
> >       struct module *owner;
> >  };
> >
> > +#define PCI_EPC_DEFAULT_WINDOW               0
> > +
> > +/**
> > + * struct pci_epc_mem_window - address window of the endpoint controller
> > + * @phys_base: physical base address of the PCI address window
> > + * @size: the size of the PCI address window
> > + */
> > +struct pci_epc_mem_window {
> > +     phys_addr_t     phys_base;
> > +     size_t          size;
> > +     size_t          map_size;
> > +};
> > +
> >  /**
> >   * struct pci_epc_mem - address space of the endpoint controller
> > - * @phys_base: physical base address of the PCI address space
> > - * @size: the size of the PCI address space
> > + * @window: address window of the endpoint controller
> >   * @bitmap: bitmap to manage the PCI address space
> > - * @pages: number of bits representing the address region
> >   * @page_size: size of each page
> > + * @pages: number of bits representing the address region
> >   */
> >  struct pci_epc_mem {
> > -     phys_addr_t     phys_base;
> > -     size_t          size;
> > +     struct pci_epc_mem_window window;
> >       unsigned long   *bitmap;
> >       size_t          page_size;
> >       int             pages;
> > @@ -85,7 +97,8 @@ struct pci_epc_mem {
> >   * @dev: PCI EPC device
> >   * @pci_epf: list of endpoint functions present in this EPC device
> >   * @ops: function pointers for performing endpoint operations
> > - * @mem: address space of the endpoint controller
> > + * @mem: array of address space of the endpoint controller
> > + * @mem_windows: number of windows supported by device
> >   * @max_functions: max number of functions that can be configured in this EPC
> >   * @group: configfs group representing the PCI EPC device
> >   * @lock: spinlock to protect pci_epc ops
> > @@ -94,7 +107,8 @@ struct pci_epc {
> >       struct device                   dev;
> >       struct list_head                pci_epf;
> >       const struct pci_epc_ops        *ops;
> > -     struct pci_epc_mem              *mem;
> > +     struct pci_epc_mem              **mem;
> > +     int                             mem_windows;
> >       u8                              max_functions;
> >       struct config_group             *group;
> >       /* spinlock to protect against concurrent access of EP controller */
> > @@ -128,8 +142,8 @@ struct pci_epc_features {
> >  #define devm_pci_epc_create(dev, ops)    \
> >               __devm_pci_epc_create((dev), (ops), THIS_MODULE)
> >
> > -#define pci_epc_mem_init(epc, phys_addr, size)       \
> > -             __pci_epc_mem_init((epc), (phys_addr), (size), PAGE_SIZE)
> > +#define pci_epc_mem_init(epc, windows, num_windows)  \
> > +             __pci_epc_mem_init((epc), windows, num_windows, PAGE_SIZE)
> >
> >  static inline void epc_set_drvdata(struct pci_epc *epc, void *data)
> >  {
> > @@ -159,7 +173,7 @@ int pci_epc_set_bar(struct pci_epc *epc, u8 func_no,
> >  void pci_epc_clear_bar(struct pci_epc *epc, u8 func_no,
> >                      struct pci_epf_bar *epf_bar);
> >  int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> > -                  phys_addr_t phys_addr,
> > +                  phys_addr_t phys_addr, int window,
> >                    u64 pci_addr, size_t size);
> >  void pci_epc_unmap_addr(struct pci_epc *epc, u8 func_no,
> >                       phys_addr_t phys_addr);
> > @@ -178,11 +192,12 @@ unsigned int pci_epc_get_first_free_bar(const struct pci_epc_features
> >  struct pci_epc *pci_epc_get(const char *epc_name);
> >  void pci_epc_put(struct pci_epc *epc);
> >
> > -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_addr, size_t size,
> > -                    size_t page_size);
> > +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *window,
> > +                    int num_windows, size_t page_size);
> >  void pci_epc_mem_exit(struct pci_epc *epc);
> >  void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
> > -                                  phys_addr_t *phys_addr, size_t size);
> > +                                  phys_addr_t *phys_addr,
> > +                                  int *window, size_t size);
> >  void pci_epc_mem_free_addr(struct pci_epc *epc, phys_addr_t phys_addr,
> >                          void __iomem *virt_addr, size_t size);
> >  #endif /* __LINUX_PCI_EPC_H */
> >

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 2/5] pci: endpoint: add support to handle multiple base for mapping outbound memory
@ 2019-11-27 21:28       ` Lad, Prabhakar
  0 siblings, 0 replies; 60+ messages in thread
From: Lad, Prabhakar @ 2019-11-27 21:28 UTC (permalink / raw)
  To: Kishon Vijay Abraham I
  Cc: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Marek Vasut, Yoshihiro Shimoda, linux-pci,
	Catalin Marinas, Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS, LKML,
	LAK

Hi Kishon,

On Wed, Nov 27, 2019 at 6:06 AM Kishon Vijay Abraham I <kishon@ti.com> wrote:
>
> Hi Prabhakar,
>
> On 07/11/19 1:06 AM, Lad Prabhakar wrote:
> > From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >
> > rcar pcie controller has support to map multiple memory regions
> > for mapping the outbound memory in local system, this feature
> > inspires to add support for handling multiple memory bases in
> > endpoint framework. In case of multiple memory regions only chunk
> > or complete region can be mapped and this window needs to be
> > passed to the controller driver.
>
> Wouldn't this result in lot of unused address and exhausting the memory regions
> quickly? I think we could have a "type" associated with each of these regions
> and smaller allocations could use one "type" and larger allocation could use a
> different "type"? For instance we allocate memory regions for MSI interrupt and
> for data transfer. We could use smaller allocations for MSI and larger
> allocations for data transfers.
>
> We could also extend the "type" for certain special requirements. For instance
> I've seen platforms which have dedicated address space for high priority data
> transfers (uses higher virtual channel numbers) and dedicated address space for
> low priority data transfers (uses lower virtual channel numbers).
>
> Here too we could give a special flag or type for higher priority address
> regions and different type for lower priority address. While allocating, the
> function drivers should provide a flag specifying the region where memory
> should be allocated.
>
agreed but in this controller if chunk is allocated from the window
this cannot be re-used
for further allocations, I haven't come across a where a window could
be used with multiple
chunks being allocated. so as you suggest I could extend the type to
have additional flag
which would indicate if window can be used for multiple chunk allocations.

> Can you add flags in struct pci_epc_mem_window and see how that comes out?
>
sure will do that.

Cheers,
--Prabhakar

> Thanks
> Kishon
>
> >
> > Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > Cc: <linux-rockchip@lists.infradead.org>
> > Cc: Shawn Lin <shawn.lin@rock-chips.com>
> > Cc: Heiko Stuebner <heiko@sntech.de>
> > Cc: Tom Joseph <tjoseph@cadence.com>
> > Cc: Jingoo Han <jingoohan1@gmail.com>
> > Cc: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
> > ---
> >  .../pci/controller/dwc/pcie-designware-ep.c   |  30 ++-
> >  drivers/pci/controller/pcie-cadence-ep.c      |  11 +-
> >  drivers/pci/controller/pcie-rockchip-ep.c     |  13 +-
> >  drivers/pci/endpoint/functions/pci-epf-test.c |  29 +--
> >  drivers/pci/endpoint/pci-epc-core.c           |   7 +-
> >  drivers/pci/endpoint/pci-epc-mem.c            | 189 ++++++++++++++----
> >  include/linux/pci-epc.h                       |  43 ++--
> >  7 files changed, 234 insertions(+), 88 deletions(-)
> >
> > diff --git a/drivers/pci/controller/dwc/pcie-designware-ep.c b/drivers/pci/controller/dwc/pcie-designware-ep.c
> > index 3dd2e2697294..8d23c20b9afd 100644
> > --- a/drivers/pci/controller/dwc/pcie-designware-ep.c
> > +++ b/drivers/pci/controller/dwc/pcie-designware-ep.c
> > @@ -195,7 +195,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, u8 func_no,
> >  }
> >
> >  static int dw_pcie_ep_map_addr(struct pci_epc *epc, u8 func_no,
> > -                            phys_addr_t addr,
> > +                            phys_addr_t addr, int window,
> >                              u64 pci_addr, size_t size)
> >  {
> >       int ret;
> > @@ -367,6 +367,7 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
> >       unsigned int aligned_offset;
> >       u16 msg_ctrl, msg_data;
> >       u32 msg_addr_lower, msg_addr_upper, reg;
> > +     int window = PCI_EPC_DEFAULT_WINDOW;
> >       u64 msg_addr;
> >       bool has_upper;
> >       int ret;
> > @@ -390,11 +391,11 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
> >               reg = ep->msi_cap + PCI_MSI_DATA_32;
> >               msg_data = dw_pcie_readw_dbi(pci, reg);
> >       }
> > -     aligned_offset = msg_addr_lower & (epc->mem->page_size - 1);
> > +     aligned_offset = msg_addr_lower & (epc->mem[window]->page_size - 1);
> >       msg_addr = ((u64)msg_addr_upper) << 32 |
> >                       (msg_addr_lower & ~aligned_offset);
> > -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> > -                               epc->mem->page_size);
> > +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> > +                               msg_addr, epc->mem[window]->page_size);
> >       if (ret)
> >               return ret;
> >
> > @@ -416,6 +417,7 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >       u32 reg, msg_data, vec_ctrl;
> >       u64 tbl_addr, msg_addr, reg_u64;
> >       void __iomem *msix_tbl;
> > +     int window = PCI_EPC_DEFAULT_WINDOW;
> >       int ret;
> >
> >       reg = ep->msix_cap + PCI_MSIX_TABLE;
> > @@ -452,8 +454,8 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >               return -EPERM;
> >       }
> >
> > -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> > -                               epc->mem->page_size);
> > +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> > +                               msg_addr, epc->mem[window]->page_size);
> >       if (ret)
> >               return ret;
> >
> > @@ -466,10 +468,11 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >
> >  void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
> >  {
> > +     int window = PCI_EPC_DEFAULT_WINDOW;
> >       struct pci_epc *epc = ep->epc;
> >
> >       pci_epc_mem_free_addr(epc, ep->msi_mem_phys, ep->msi_mem,
> > -                           epc->mem->page_size);
> > +                           epc->mem[window]->page_size);
> >
> >       pci_epc_mem_exit(epc);
> >  }
> > @@ -499,9 +502,12 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
> >       u32 reg;
> >       void *addr;
> >       u8 hdr_type;
> > +     int window;
> >       unsigned int nbars;
> >       unsigned int offset;
> >       struct pci_epc *epc;
> > +     size_t msi_page_size;
> > +     struct pci_epc_mem_window mem_window;
> >       struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
> >       struct device *dev = pci->dev;
> >       struct device_node *np = dev->of_node;
> > @@ -574,15 +580,17 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
> >       if (ret < 0)
> >               epc->max_functions = 1;
> >
> > -     ret = __pci_epc_mem_init(epc, ep->phys_base, ep->addr_size,
> > -                              ep->page_size);
> > +     mem_window.phys_base = ep->phys_base;
> > +     mem_window.size = ep->addr_size;
> > +     ret = __pci_epc_mem_init(epc, &mem_window, 1, ep->page_size);
> >       if (ret < 0) {
> >               dev_err(dev, "Failed to initialize address space\n");
> >               return ret;
> >       }
> >
> > -     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys,
> > -                                          epc->mem->page_size);
> > +     msi_page_size = epc->mem[PCI_EPC_DEFAULT_WINDOW]->page_size;
> > +     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys, &window,
> > +                                          msi_page_size);
> >       if (!ep->msi_mem) {
> >               dev_err(dev, "Failed to reserve memory for MSI/MSI-X\n");
> >               return -ENOMEM;
> > diff --git a/drivers/pci/controller/pcie-cadence-ep.c b/drivers/pci/controller/pcie-cadence-ep.c
> > index def7820cb824..7991b38a5350 100644
> > --- a/drivers/pci/controller/pcie-cadence-ep.c
> > +++ b/drivers/pci/controller/pcie-cadence-ep.c
> > @@ -172,7 +172,7 @@ static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
> >  }
> >
> >  static int cdns_pcie_ep_map_addr(struct pci_epc *epc, u8 fn, phys_addr_t addr,
> > -                              u64 pci_addr, size_t size)
> > +                              int window, u64 pci_addr, size_t size)
> >  {
> >       struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
> >       struct cdns_pcie *pcie = &ep->pcie;
> > @@ -434,12 +434,14 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
> >  {
> >       struct device *dev = &pdev->dev;
> >       struct device_node *np = dev->of_node;
> > +     struct pci_epc_mem_window mem_window;
> >       struct cdns_pcie_ep *ep;
> >       struct cdns_pcie *pcie;
> >       struct pci_epc *epc;
> >       struct resource *res;
> >       int ret;
> >       int phy_count;
> > +     int window;
> >
> >       ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> >       if (!ep)
> > @@ -502,15 +504,16 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
> >       if (of_property_read_u8(np, "max-functions", &epc->max_functions) < 0)
> >               epc->max_functions = 1;
> >
> > -     ret = pci_epc_mem_init(epc, pcie->mem_res->start,
> > -                            resource_size(pcie->mem_res));
> > +     mem_window.phys_base = pcie->mem_res->start;
> > +     mem_window.size = resource_size(pcie->mem_res);
> > +     ret = pci_epc_mem_init(epc, &mem_window, 1);
> >       if (ret < 0) {
> >               dev_err(dev, "failed to initialize the memory space\n");
> >               goto err_init;
> >       }
> >
> >       ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> > -                                               SZ_128K);
> > +                                               &window, SZ_128K);
> >       if (!ep->irq_cpu_addr) {
> >               dev_err(dev, "failed to reserve memory space for MSI\n");
> >               ret = -ENOMEM;
> > diff --git a/drivers/pci/controller/pcie-rockchip-ep.c b/drivers/pci/controller/pcie-rockchip-ep.c
> > index d743b0a48988..d59e85c8d319 100644
> > --- a/drivers/pci/controller/pcie-rockchip-ep.c
> > +++ b/drivers/pci/controller/pcie-rockchip-ep.c
> > @@ -256,8 +256,8 @@ static void rockchip_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
> >  }
> >
> >  static int rockchip_pcie_ep_map_addr(struct pci_epc *epc, u8 fn,
> > -                                  phys_addr_t addr, u64 pci_addr,
> > -                                  size_t size)
> > +                                  phys_addr_t addr, int window,
> > +                                  u64 pci_addr, size_t size)
> >  {
> >       struct rockchip_pcie_ep *ep = epc_get_drvdata(epc);
> >       struct rockchip_pcie *pcie = &ep->rockchip;
> > @@ -562,11 +562,13 @@ static const struct of_device_id rockchip_pcie_ep_of_match[] = {
> >
> >  static int rockchip_pcie_ep_probe(struct platform_device *pdev)
> >  {
> > +     struct pci_epc_mem_window mem_window;
> >       struct device *dev = &pdev->dev;
> >       struct rockchip_pcie_ep *ep;
> >       struct rockchip_pcie *rockchip;
> >       struct pci_epc *epc;
> >       size_t max_regions;
> > +     int window;
> >       int err;
> >
> >       ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> > @@ -614,15 +616,16 @@ static int rockchip_pcie_ep_probe(struct platform_device *pdev)
> >       /* Only enable function 0 by default */
> >       rockchip_pcie_write(rockchip, BIT(0), PCIE_CORE_PHY_FUNC_CFG);
> >
> > -     err = pci_epc_mem_init(epc, rockchip->mem_res->start,
> > -                            resource_size(rockchip->mem_res));
> > +     mem_window.phys_base = rockchip->mem_res->start;
> > +     mem_window.size = resource_size(rockchip->mem_res);
> > +     err = pci_epc_mem_init(epc, &mem_window, 1);
> >       if (err < 0) {
> >               dev_err(dev, "failed to initialize the memory space\n");
> >               goto err_uninit_port;
> >       }
> >
> >       ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> > -                                               SZ_128K);
> > +                                               &window, SZ_128K);
> >       if (!ep->irq_cpu_addr) {
> >               dev_err(dev, "failed to reserve memory space for MSI\n");
> >               err = -ENOMEM;
> > diff --git a/drivers/pci/endpoint/functions/pci-epf-test.c b/drivers/pci/endpoint/functions/pci-epf-test.c
> > index 5d74f81ddfe4..475228011703 100644
> > --- a/drivers/pci/endpoint/functions/pci-epf-test.c
> > +++ b/drivers/pci/endpoint/functions/pci-epf-test.c
> > @@ -84,8 +84,10 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >       struct pci_epc *epc = epf->epc;
> >       enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >       struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> > +     int window;
> >
> > -     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr, reg->size);
> > +     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr,
> > +                                       &window, reg->size);
> >       if (!src_addr) {
> >               dev_err(dev, "Failed to allocate source address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> > @@ -93,15 +95,16 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >               goto err;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, reg->src_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, window,
> > +                            reg->src_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map source address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> >               goto err_src_addr;
> >       }
> >
> > -     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr, reg->size);
> > +     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr,
> > +                                       &window, reg->size);
> >       if (!dst_addr) {
> >               dev_err(dev, "Failed to allocate destination address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > @@ -109,8 +112,8 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >               goto err_src_map_addr;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, reg->dst_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, window,
> > +                            reg->dst_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map destination address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > @@ -146,8 +149,9 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
> >       struct pci_epc *epc = epf->epc;
> >       enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >       struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> > +     int window;
> >
> > -     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> > +     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
> >       if (!src_addr) {
> >               dev_err(dev, "Failed to allocate address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> > @@ -155,8 +159,8 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
> >               goto err;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->src_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> > +                            reg->src_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> > @@ -193,13 +197,14 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
> >       void __iomem *dst_addr;
> >       void *buf;
> >       phys_addr_t phys_addr;
> > +     int window;
> >       struct pci_epf *epf = epf_test->epf;
> >       struct device *dev = &epf->dev;
> >       struct pci_epc *epc = epf->epc;
> >       enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >       struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> >
> > -     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> > +     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
> >       if (!dst_addr) {
> >               dev_err(dev, "Failed to allocate address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > @@ -207,8 +212,8 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
> >               goto err;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->dst_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> > +                            reg->dst_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > diff --git a/drivers/pci/endpoint/pci-epc-core.c b/drivers/pci/endpoint/pci-epc-core.c
> > index 2091508c1620..289c266c2d90 100644
> > --- a/drivers/pci/endpoint/pci-epc-core.c
> > +++ b/drivers/pci/endpoint/pci-epc-core.c
> > @@ -358,13 +358,15 @@ EXPORT_SYMBOL_GPL(pci_epc_unmap_addr);
> >   * @epc: the EPC device on which address is allocated
> >   * @func_no: the endpoint function number in the EPC device
> >   * @phys_addr: physical address of the local system
> > + * @window: index to the window region where PCI address will be mapped
> >   * @pci_addr: PCI address to which the physical address should be mapped
> >   * @size: the size of the allocation
> >   *
> >   * Invoke to map CPU address with PCI address.
> >   */
> >  int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> > -                  phys_addr_t phys_addr, u64 pci_addr, size_t size)
> > +                  phys_addr_t phys_addr, int window,
> > +                  u64 pci_addr, size_t size)
> >  {
> >       int ret;
> >       unsigned long flags;
> > @@ -376,7 +378,8 @@ int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> >               return 0;
> >
> >       spin_lock_irqsave(&epc->lock, flags);
> > -     ret = epc->ops->map_addr(epc, func_no, phys_addr, pci_addr, size);
> > +     ret = epc->ops->map_addr(epc, func_no, phys_addr,
> > +                              window, pci_addr, size);
> >       spin_unlock_irqrestore(&epc->lock, flags);
> >
> >       return ret;
> > diff --git a/drivers/pci/endpoint/pci-epc-mem.c b/drivers/pci/endpoint/pci-epc-mem.c
> > index d2b174ce15de..c955f2c97944 100644
> > --- a/drivers/pci/endpoint/pci-epc-mem.c
> > +++ b/drivers/pci/endpoint/pci-epc-mem.c
> > @@ -39,56 +39,77 @@ static int pci_epc_mem_get_order(struct pci_epc_mem *mem, size_t size)
> >   * __pci_epc_mem_init() - initialize the pci_epc_mem structure
> >   * @epc: the EPC device that invoked pci_epc_mem_init
> >   * @phys_base: the physical address of the base
> > - * @size: the size of the address space
> > + * @num_windows: number of windows device supports
> >   * @page_size: size of each page
> >   *
> >   * Invoke to initialize the pci_epc_mem structure used by the
> >   * endpoint functions to allocate mapped PCI address.
> >   */
> > -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_base, size_t size,
> > -                    size_t page_size)
> > +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *windows,
> > +                    int num_windows, size_t page_size)
> >  {
> > -     int ret;
> > -     struct pci_epc_mem *mem;
> > -     unsigned long *bitmap;
> > +     struct pci_epc_mem *mem = NULL;
> > +     unsigned long *bitmap = NULL;
> >       unsigned int page_shift;
> > -     int pages;
> >       int bitmap_size;
> > +     int pages;
> > +     int ret;
> > +     int i;
> > +
> > +     epc->mem_windows = 0;
> > +
> > +     if (!windows)
> > +             return -EINVAL;
> > +
> > +     if (num_windows <= 0)
> > +             return -EINVAL;
> >
> >       if (page_size < PAGE_SIZE)
> >               page_size = PAGE_SIZE;
> >
> >       page_shift = ilog2(page_size);
> > -     pages = size >> page_shift;
> > -     bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
> >
> > -     mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> > -     if (!mem) {
> > -             ret = -ENOMEM;
> > -             goto err;
> > -     }
> > +     epc->mem = kcalloc(num_windows, sizeof(*mem), GFP_KERNEL);
> > +     if (!epc->mem)
> > +             return -EINVAL;
> >
> > -     bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> > -     if (!bitmap) {
> > -             ret = -ENOMEM;
> > -             goto err_mem;
> > -     }
> > +     for (i = 0; i < num_windows; i++) {
> > +             pages = windows[i].phys_base >> page_shift;
> > +             bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
> >
> > -     mem->bitmap = bitmap;
> > -     mem->phys_base = phys_base;
> > -     mem->page_size = page_size;
> > -     mem->pages = pages;
> > -     mem->size = size;
> > +             mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> > +             if (!mem) {
> > +                     ret = -ENOMEM;
> > +                     goto err_mem;
> > +             }
> >
> > -     epc->mem = mem;
> > +             bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> > +             if (!bitmap) {
> > +                     ret = -ENOMEM;
> > +                     goto err_mem;
> > +             }
> > +
> > +             mem->bitmap = bitmap;
> > +             mem->window.phys_base = windows[i].phys_base;
> > +             mem->page_size = page_size;
> > +             mem->pages = pages;
> > +             mem->window.size = windows[i].size;
> > +             mem->window.map_size = 0;
> > +
> > +             epc->mem[i] = mem;
> > +     }
> > +     epc->mem_windows = num_windows;
> >
> >       return 0;
> >
> >  err_mem:
> > -     kfree(mem);
> > +     for (; i >= 0; i--) {
> > +             kfree(mem->bitmap);
> > +             kfree(epc->mem[i]);
> > +     }
> > +     kfree(epc->mem);
> >
> > -err:
> > -return ret;
> > +     return ret;
> >  }
> >  EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
> >
> > @@ -101,48 +122,126 @@ EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
> >   */
> >  void pci_epc_mem_exit(struct pci_epc *epc)
> >  {
> > -     struct pci_epc_mem *mem = epc->mem;
> > +     struct pci_epc_mem *mem;
> > +     int i;
> > +
> > +     if (!epc->mem_windows)
> > +             return;
> > +
> > +     for (i = 0; i <= epc->mem_windows; i--) {
> > +             mem = epc->mem[i];
> > +             kfree(mem->bitmap);
> > +             kfree(epc->mem[i]);
> > +     }
> > +     kfree(epc->mem);
> >
> >       epc->mem = NULL;
> > -     kfree(mem->bitmap);
> > -     kfree(mem);
> > +     epc->mem_windows = 0;
> >  }
> >  EXPORT_SYMBOL_GPL(pci_epc_mem_exit);
> >
> > +static int pci_epc_find_best_fit_window(struct pci_epc *epc, size_t size)
> > +{
> > +     size_t window_least_size = 0;
> > +     int best_fit_window = -1;
> > +     struct pci_epc_mem *mem;
> > +     size_t actual_size;
> > +     int i;
> > +
> > +     for (i = 0; i < epc->mem_windows; i++) {
> > +             mem = epc->mem[i];
> > +
> > +             /* if chunk from this region is already used skip it */
> > +             if (mem->window.map_size)
> > +                     continue;
> > +
> > +             actual_size = ALIGN(size, mem->page_size);
> > +
> > +             if (best_fit_window == -1) {
> > +                     best_fit_window = i;
> > +                     window_least_size = mem->window.size;
> > +             } else {
> > +                     if (actual_size <= mem->window.size &&
> > +                         mem->window.size < window_least_size) {
> > +                             best_fit_window = i;
> > +                             window_least_size = mem->window.size;
> > +                     }
> > +             }
> > +     }
> > +
> > +     return best_fit_window;
> > +}
> > +
> >  /**
> >   * pci_epc_mem_alloc_addr() - allocate memory address from EPC addr space
> >   * @epc: the EPC device on which memory has to be allocated
> >   * @phys_addr: populate the allocated physical address here
> > + * @window: populate the window here which will be used to map PCI address
> >   * @size: the size of the address space that has to be allocated
> >   *
> >   * Invoke to allocate memory address from the EPC address space. This
> >   * is usually done to map the remote RC address into the local system.
> >   */
> >  void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
> > -                                  phys_addr_t *phys_addr, size_t size)
> > +                                  phys_addr_t *phys_addr,
> > +                                  int *window, size_t size)
> >  {
> > +     int best_fit = PCI_EPC_DEFAULT_WINDOW;
> > +     void __iomem *virt_addr = NULL;
> > +     struct pci_epc_mem *mem;
> > +     unsigned int page_shift;
> >       int pageno;
> > -     void __iomem *virt_addr;
> > -     struct pci_epc_mem *mem = epc->mem;
> > -     unsigned int page_shift = ilog2(mem->page_size);
> >       int order;
> >
> > +     if (epc->mem_windows <= 0)
> > +             return NULL;
> > +
> > +     if (epc->mem_windows > 1) {
> > +             best_fit = pci_epc_find_best_fit_window(epc, size);
> > +             if (best_fit < 0)
> > +                     return NULL;
> > +     }
> > +
> > +     mem = epc->mem[best_fit];
> >       size = ALIGN(size, mem->page_size);
> > +     if (size > (mem->window.size - mem->window.map_size))
> > +             return NULL;
> > +     page_shift = ilog2(mem->page_size);
> >       order = pci_epc_mem_get_order(mem, size);
> >
> >       pageno = bitmap_find_free_region(mem->bitmap, mem->pages, order);
> >       if (pageno < 0)
> >               return NULL;
> >
> > -     *phys_addr = mem->phys_base + ((phys_addr_t)pageno << page_shift);
> > +     *phys_addr = mem->window.phys_base + (pageno << page_shift);
> >       virt_addr = ioremap(*phys_addr, size);
> > -     if (!virt_addr)
> > +     if (!virt_addr) {
> >               bitmap_release_region(mem->bitmap, pageno, order);
> > +     } else {
> > +             mem->window.map_size += size;
> > +             *window = best_fit;
> > +     }
> >
> >       return virt_addr;
> >  }
> >  EXPORT_SYMBOL_GPL(pci_epc_mem_alloc_addr);
> >
> > +static int pci_epc_get_matching_window(struct pci_epc *epc,
> > +                                    phys_addr_t phys_addr)
> > +{
> > +     struct pci_epc_mem *mem;
> > +     int i;
> > +
> > +     for (i = 0; i < epc->mem_windows; i++) {
> > +             mem = epc->mem[i];
> > +
> > +             if (mem->window.phys_base == phys_addr)
> > +                     return i;
> > +     }
> > +
> > +     return -EINVAL;
> > +}
> > +
> >  /**
> >   * pci_epc_mem_free_addr() - free the allocated memory address
> >   * @epc: the EPC device on which memory was allocated
> > @@ -155,16 +254,26 @@ EXPORT_SYMBOL_GPL(pci_epc_mem_alloc_addr);
> >  void pci_epc_mem_free_addr(struct pci_epc *epc, phys_addr_t phys_addr,
> >                          void __iomem *virt_addr, size_t size)
> >  {
> > +     struct pci_epc_mem *mem;
> > +     unsigned int page_shift;
> > +     int window = 0;
> >       int pageno;
> > -     struct pci_epc_mem *mem = epc->mem;
> > -     unsigned int page_shift = ilog2(mem->page_size);
> >       int order;
> >
> > +     if (epc->mem_windows > 1) {
> > +             window = pci_epc_get_matching_window(epc, phys_addr);
> > +             if (window < 0)
> > +                     return;
> > +     }
> > +
> > +     mem = epc->mem[window];
> > +     page_shift = ilog2(mem->page_size);
> >       iounmap(virt_addr);
> > -     pageno = (phys_addr - mem->phys_base) >> page_shift;
> > +     pageno = (phys_addr - mem->window.phys_base) >> page_shift;
> >       size = ALIGN(size, mem->page_size);
> >       order = pci_epc_mem_get_order(mem, size);
> >       bitmap_release_region(mem->bitmap, pageno, order);
> > +     mem->window.map_size -= size;
> >  }
> >  EXPORT_SYMBOL_GPL(pci_epc_mem_free_addr);
> >
> > diff --git a/include/linux/pci-epc.h b/include/linux/pci-epc.h
> > index 56f1846b9d39..5e86d22dd5c5 100644
> > --- a/include/linux/pci-epc.h
> > +++ b/include/linux/pci-epc.h
> > @@ -48,7 +48,8 @@ struct pci_epc_ops {
> >       void    (*clear_bar)(struct pci_epc *epc, u8 func_no,
> >                            struct pci_epf_bar *epf_bar);
> >       int     (*map_addr)(struct pci_epc *epc, u8 func_no,
> > -                         phys_addr_t addr, u64 pci_addr, size_t size);
> > +                         phys_addr_t addr, int window,
> > +                         u64 pci_addr, size_t size);
> >       void    (*unmap_addr)(struct pci_epc *epc, u8 func_no,
> >                             phys_addr_t addr);
> >       int     (*set_msi)(struct pci_epc *epc, u8 func_no, u8 interrupts);
> > @@ -64,17 +65,28 @@ struct pci_epc_ops {
> >       struct module *owner;
> >  };
> >
> > +#define PCI_EPC_DEFAULT_WINDOW               0
> > +
> > +/**
> > + * struct pci_epc_mem_window - address window of the endpoint controller
> > + * @phys_base: physical base address of the PCI address window
> > + * @size: the size of the PCI address window
> > + */
> > +struct pci_epc_mem_window {
> > +     phys_addr_t     phys_base;
> > +     size_t          size;
> > +     size_t          map_size;
> > +};
> > +
> >  /**
> >   * struct pci_epc_mem - address space of the endpoint controller
> > - * @phys_base: physical base address of the PCI address space
> > - * @size: the size of the PCI address space
> > + * @window: address window of the endpoint controller
> >   * @bitmap: bitmap to manage the PCI address space
> > - * @pages: number of bits representing the address region
> >   * @page_size: size of each page
> > + * @pages: number of bits representing the address region
> >   */
> >  struct pci_epc_mem {
> > -     phys_addr_t     phys_base;
> > -     size_t          size;
> > +     struct pci_epc_mem_window window;
> >       unsigned long   *bitmap;
> >       size_t          page_size;
> >       int             pages;
> > @@ -85,7 +97,8 @@ struct pci_epc_mem {
> >   * @dev: PCI EPC device
> >   * @pci_epf: list of endpoint functions present in this EPC device
> >   * @ops: function pointers for performing endpoint operations
> > - * @mem: address space of the endpoint controller
> > + * @mem: array of address space of the endpoint controller
> > + * @mem_windows: number of windows supported by device
> >   * @max_functions: max number of functions that can be configured in this EPC
> >   * @group: configfs group representing the PCI EPC device
> >   * @lock: spinlock to protect pci_epc ops
> > @@ -94,7 +107,8 @@ struct pci_epc {
> >       struct device                   dev;
> >       struct list_head                pci_epf;
> >       const struct pci_epc_ops        *ops;
> > -     struct pci_epc_mem              *mem;
> > +     struct pci_epc_mem              **mem;
> > +     int                             mem_windows;
> >       u8                              max_functions;
> >       struct config_group             *group;
> >       /* spinlock to protect against concurrent access of EP controller */
> > @@ -128,8 +142,8 @@ struct pci_epc_features {
> >  #define devm_pci_epc_create(dev, ops)    \
> >               __devm_pci_epc_create((dev), (ops), THIS_MODULE)
> >
> > -#define pci_epc_mem_init(epc, phys_addr, size)       \
> > -             __pci_epc_mem_init((epc), (phys_addr), (size), PAGE_SIZE)
> > +#define pci_epc_mem_init(epc, windows, num_windows)  \
> > +             __pci_epc_mem_init((epc), windows, num_windows, PAGE_SIZE)
> >
> >  static inline void epc_set_drvdata(struct pci_epc *epc, void *data)
> >  {
> > @@ -159,7 +173,7 @@ int pci_epc_set_bar(struct pci_epc *epc, u8 func_no,
> >  void pci_epc_clear_bar(struct pci_epc *epc, u8 func_no,
> >                      struct pci_epf_bar *epf_bar);
> >  int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> > -                  phys_addr_t phys_addr,
> > +                  phys_addr_t phys_addr, int window,
> >                    u64 pci_addr, size_t size);
> >  void pci_epc_unmap_addr(struct pci_epc *epc, u8 func_no,
> >                       phys_addr_t phys_addr);
> > @@ -178,11 +192,12 @@ unsigned int pci_epc_get_first_free_bar(const struct pci_epc_features
> >  struct pci_epc *pci_epc_get(const char *epc_name);
> >  void pci_epc_put(struct pci_epc *epc);
> >
> > -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_addr, size_t size,
> > -                    size_t page_size);
> > +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *window,
> > +                    int num_windows, size_t page_size);
> >  void pci_epc_mem_exit(struct pci_epc *epc);
> >  void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
> > -                                  phys_addr_t *phys_addr, size_t size);
> > +                                  phys_addr_t *phys_addr,
> > +                                  int *window, size_t size);
> >  void pci_epc_mem_free_addr(struct pci_epc *epc, phys_addr_t phys_addr,
> >                          void __iomem *virt_addr, size_t size);
> >  #endif /* __LINUX_PCI_EPC_H */
> >

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 2/5] pci: endpoint: add support to handle multiple base for mapping outbound memory
@ 2019-11-27 21:28       ` Lad, Prabhakar
  0 siblings, 0 replies; 60+ messages in thread
From: Lad, Prabhakar @ 2019-11-27 21:28 UTC (permalink / raw)
  To: Kishon Vijay Abraham I
  Cc: Mark Rutland, Heiko Stuebner, Geert Uytterhoeven, linux-pci,
	Shawn Lin, Will Deacon, Marek Vasut, Lorenzo Pieralisi,
	Magnus Damm, linux-rockchip, Catalin Marinas,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Chris Paterson, Arnd Bergmann, Lad, Prabhakar, Rob Herring,
	Bjorn Helgaas, LAK, Greg Kroah-Hartman, Yoshihiro Shimoda, LKML,
	Linux-Renesas, Tom Joseph, Jingoo Han, Andrew Murray,
	Gustavo Pimentel

Hi Kishon,

On Wed, Nov 27, 2019 at 6:06 AM Kishon Vijay Abraham I <kishon@ti.com> wrote:
>
> Hi Prabhakar,
>
> On 07/11/19 1:06 AM, Lad Prabhakar wrote:
> > From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >
> > rcar pcie controller has support to map multiple memory regions
> > for mapping the outbound memory in local system, this feature
> > inspires to add support for handling multiple memory bases in
> > endpoint framework. In case of multiple memory regions only chunk
> > or complete region can be mapped and this window needs to be
> > passed to the controller driver.
>
> Wouldn't this result in lot of unused address and exhausting the memory regions
> quickly? I think we could have a "type" associated with each of these regions
> and smaller allocations could use one "type" and larger allocation could use a
> different "type"? For instance we allocate memory regions for MSI interrupt and
> for data transfer. We could use smaller allocations for MSI and larger
> allocations for data transfers.
>
> We could also extend the "type" for certain special requirements. For instance
> I've seen platforms which have dedicated address space for high priority data
> transfers (uses higher virtual channel numbers) and dedicated address space for
> low priority data transfers (uses lower virtual channel numbers).
>
> Here too we could give a special flag or type for higher priority address
> regions and different type for lower priority address. While allocating, the
> function drivers should provide a flag specifying the region where memory
> should be allocated.
>
agreed but in this controller if chunk is allocated from the window
this cannot be re-used
for further allocations, I haven't come across a where a window could
be used with multiple
chunks being allocated. so as you suggest I could extend the type to
have additional flag
which would indicate if window can be used for multiple chunk allocations.

> Can you add flags in struct pci_epc_mem_window and see how that comes out?
>
sure will do that.

Cheers,
--Prabhakar

> Thanks
> Kishon
>
> >
> > Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> > Cc: <linux-rockchip@lists.infradead.org>
> > Cc: Shawn Lin <shawn.lin@rock-chips.com>
> > Cc: Heiko Stuebner <heiko@sntech.de>
> > Cc: Tom Joseph <tjoseph@cadence.com>
> > Cc: Jingoo Han <jingoohan1@gmail.com>
> > Cc: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
> > ---
> >  .../pci/controller/dwc/pcie-designware-ep.c   |  30 ++-
> >  drivers/pci/controller/pcie-cadence-ep.c      |  11 +-
> >  drivers/pci/controller/pcie-rockchip-ep.c     |  13 +-
> >  drivers/pci/endpoint/functions/pci-epf-test.c |  29 +--
> >  drivers/pci/endpoint/pci-epc-core.c           |   7 +-
> >  drivers/pci/endpoint/pci-epc-mem.c            | 189 ++++++++++++++----
> >  include/linux/pci-epc.h                       |  43 ++--
> >  7 files changed, 234 insertions(+), 88 deletions(-)
> >
> > diff --git a/drivers/pci/controller/dwc/pcie-designware-ep.c b/drivers/pci/controller/dwc/pcie-designware-ep.c
> > index 3dd2e2697294..8d23c20b9afd 100644
> > --- a/drivers/pci/controller/dwc/pcie-designware-ep.c
> > +++ b/drivers/pci/controller/dwc/pcie-designware-ep.c
> > @@ -195,7 +195,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, u8 func_no,
> >  }
> >
> >  static int dw_pcie_ep_map_addr(struct pci_epc *epc, u8 func_no,
> > -                            phys_addr_t addr,
> > +                            phys_addr_t addr, int window,
> >                              u64 pci_addr, size_t size)
> >  {
> >       int ret;
> > @@ -367,6 +367,7 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
> >       unsigned int aligned_offset;
> >       u16 msg_ctrl, msg_data;
> >       u32 msg_addr_lower, msg_addr_upper, reg;
> > +     int window = PCI_EPC_DEFAULT_WINDOW;
> >       u64 msg_addr;
> >       bool has_upper;
> >       int ret;
> > @@ -390,11 +391,11 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
> >               reg = ep->msi_cap + PCI_MSI_DATA_32;
> >               msg_data = dw_pcie_readw_dbi(pci, reg);
> >       }
> > -     aligned_offset = msg_addr_lower & (epc->mem->page_size - 1);
> > +     aligned_offset = msg_addr_lower & (epc->mem[window]->page_size - 1);
> >       msg_addr = ((u64)msg_addr_upper) << 32 |
> >                       (msg_addr_lower & ~aligned_offset);
> > -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> > -                               epc->mem->page_size);
> > +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> > +                               msg_addr, epc->mem[window]->page_size);
> >       if (ret)
> >               return ret;
> >
> > @@ -416,6 +417,7 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >       u32 reg, msg_data, vec_ctrl;
> >       u64 tbl_addr, msg_addr, reg_u64;
> >       void __iomem *msix_tbl;
> > +     int window = PCI_EPC_DEFAULT_WINDOW;
> >       int ret;
> >
> >       reg = ep->msix_cap + PCI_MSIX_TABLE;
> > @@ -452,8 +454,8 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >               return -EPERM;
> >       }
> >
> > -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> > -                               epc->mem->page_size);
> > +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> > +                               msg_addr, epc->mem[window]->page_size);
> >       if (ret)
> >               return ret;
> >
> > @@ -466,10 +468,11 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >
> >  void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
> >  {
> > +     int window = PCI_EPC_DEFAULT_WINDOW;
> >       struct pci_epc *epc = ep->epc;
> >
> >       pci_epc_mem_free_addr(epc, ep->msi_mem_phys, ep->msi_mem,
> > -                           epc->mem->page_size);
> > +                           epc->mem[window]->page_size);
> >
> >       pci_epc_mem_exit(epc);
> >  }
> > @@ -499,9 +502,12 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
> >       u32 reg;
> >       void *addr;
> >       u8 hdr_type;
> > +     int window;
> >       unsigned int nbars;
> >       unsigned int offset;
> >       struct pci_epc *epc;
> > +     size_t msi_page_size;
> > +     struct pci_epc_mem_window mem_window;
> >       struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
> >       struct device *dev = pci->dev;
> >       struct device_node *np = dev->of_node;
> > @@ -574,15 +580,17 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
> >       if (ret < 0)
> >               epc->max_functions = 1;
> >
> > -     ret = __pci_epc_mem_init(epc, ep->phys_base, ep->addr_size,
> > -                              ep->page_size);
> > +     mem_window.phys_base = ep->phys_base;
> > +     mem_window.size = ep->addr_size;
> > +     ret = __pci_epc_mem_init(epc, &mem_window, 1, ep->page_size);
> >       if (ret < 0) {
> >               dev_err(dev, "Failed to initialize address space\n");
> >               return ret;
> >       }
> >
> > -     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys,
> > -                                          epc->mem->page_size);
> > +     msi_page_size = epc->mem[PCI_EPC_DEFAULT_WINDOW]->page_size;
> > +     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys, &window,
> > +                                          msi_page_size);
> >       if (!ep->msi_mem) {
> >               dev_err(dev, "Failed to reserve memory for MSI/MSI-X\n");
> >               return -ENOMEM;
> > diff --git a/drivers/pci/controller/pcie-cadence-ep.c b/drivers/pci/controller/pcie-cadence-ep.c
> > index def7820cb824..7991b38a5350 100644
> > --- a/drivers/pci/controller/pcie-cadence-ep.c
> > +++ b/drivers/pci/controller/pcie-cadence-ep.c
> > @@ -172,7 +172,7 @@ static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
> >  }
> >
> >  static int cdns_pcie_ep_map_addr(struct pci_epc *epc, u8 fn, phys_addr_t addr,
> > -                              u64 pci_addr, size_t size)
> > +                              int window, u64 pci_addr, size_t size)
> >  {
> >       struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
> >       struct cdns_pcie *pcie = &ep->pcie;
> > @@ -434,12 +434,14 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
> >  {
> >       struct device *dev = &pdev->dev;
> >       struct device_node *np = dev->of_node;
> > +     struct pci_epc_mem_window mem_window;
> >       struct cdns_pcie_ep *ep;
> >       struct cdns_pcie *pcie;
> >       struct pci_epc *epc;
> >       struct resource *res;
> >       int ret;
> >       int phy_count;
> > +     int window;
> >
> >       ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> >       if (!ep)
> > @@ -502,15 +504,16 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
> >       if (of_property_read_u8(np, "max-functions", &epc->max_functions) < 0)
> >               epc->max_functions = 1;
> >
> > -     ret = pci_epc_mem_init(epc, pcie->mem_res->start,
> > -                            resource_size(pcie->mem_res));
> > +     mem_window.phys_base = pcie->mem_res->start;
> > +     mem_window.size = resource_size(pcie->mem_res);
> > +     ret = pci_epc_mem_init(epc, &mem_window, 1);
> >       if (ret < 0) {
> >               dev_err(dev, "failed to initialize the memory space\n");
> >               goto err_init;
> >       }
> >
> >       ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> > -                                               SZ_128K);
> > +                                               &window, SZ_128K);
> >       if (!ep->irq_cpu_addr) {
> >               dev_err(dev, "failed to reserve memory space for MSI\n");
> >               ret = -ENOMEM;
> > diff --git a/drivers/pci/controller/pcie-rockchip-ep.c b/drivers/pci/controller/pcie-rockchip-ep.c
> > index d743b0a48988..d59e85c8d319 100644
> > --- a/drivers/pci/controller/pcie-rockchip-ep.c
> > +++ b/drivers/pci/controller/pcie-rockchip-ep.c
> > @@ -256,8 +256,8 @@ static void rockchip_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
> >  }
> >
> >  static int rockchip_pcie_ep_map_addr(struct pci_epc *epc, u8 fn,
> > -                                  phys_addr_t addr, u64 pci_addr,
> > -                                  size_t size)
> > +                                  phys_addr_t addr, int window,
> > +                                  u64 pci_addr, size_t size)
> >  {
> >       struct rockchip_pcie_ep *ep = epc_get_drvdata(epc);
> >       struct rockchip_pcie *pcie = &ep->rockchip;
> > @@ -562,11 +562,13 @@ static const struct of_device_id rockchip_pcie_ep_of_match[] = {
> >
> >  static int rockchip_pcie_ep_probe(struct platform_device *pdev)
> >  {
> > +     struct pci_epc_mem_window mem_window;
> >       struct device *dev = &pdev->dev;
> >       struct rockchip_pcie_ep *ep;
> >       struct rockchip_pcie *rockchip;
> >       struct pci_epc *epc;
> >       size_t max_regions;
> > +     int window;
> >       int err;
> >
> >       ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> > @@ -614,15 +616,16 @@ static int rockchip_pcie_ep_probe(struct platform_device *pdev)
> >       /* Only enable function 0 by default */
> >       rockchip_pcie_write(rockchip, BIT(0), PCIE_CORE_PHY_FUNC_CFG);
> >
> > -     err = pci_epc_mem_init(epc, rockchip->mem_res->start,
> > -                            resource_size(rockchip->mem_res));
> > +     mem_window.phys_base = rockchip->mem_res->start;
> > +     mem_window.size = resource_size(rockchip->mem_res);
> > +     err = pci_epc_mem_init(epc, &mem_window, 1);
> >       if (err < 0) {
> >               dev_err(dev, "failed to initialize the memory space\n");
> >               goto err_uninit_port;
> >       }
> >
> >       ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> > -                                               SZ_128K);
> > +                                               &window, SZ_128K);
> >       if (!ep->irq_cpu_addr) {
> >               dev_err(dev, "failed to reserve memory space for MSI\n");
> >               err = -ENOMEM;
> > diff --git a/drivers/pci/endpoint/functions/pci-epf-test.c b/drivers/pci/endpoint/functions/pci-epf-test.c
> > index 5d74f81ddfe4..475228011703 100644
> > --- a/drivers/pci/endpoint/functions/pci-epf-test.c
> > +++ b/drivers/pci/endpoint/functions/pci-epf-test.c
> > @@ -84,8 +84,10 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >       struct pci_epc *epc = epf->epc;
> >       enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >       struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> > +     int window;
> >
> > -     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr, reg->size);
> > +     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr,
> > +                                       &window, reg->size);
> >       if (!src_addr) {
> >               dev_err(dev, "Failed to allocate source address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> > @@ -93,15 +95,16 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >               goto err;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, reg->src_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, window,
> > +                            reg->src_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map source address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> >               goto err_src_addr;
> >       }
> >
> > -     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr, reg->size);
> > +     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr,
> > +                                       &window, reg->size);
> >       if (!dst_addr) {
> >               dev_err(dev, "Failed to allocate destination address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > @@ -109,8 +112,8 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >               goto err_src_map_addr;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, reg->dst_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, window,
> > +                            reg->dst_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map destination address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > @@ -146,8 +149,9 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
> >       struct pci_epc *epc = epf->epc;
> >       enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >       struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> > +     int window;
> >
> > -     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> > +     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
> >       if (!src_addr) {
> >               dev_err(dev, "Failed to allocate address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> > @@ -155,8 +159,8 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
> >               goto err;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->src_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> > +                            reg->src_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map address\n");
> >               reg->status = STATUS_SRC_ADDR_INVALID;
> > @@ -193,13 +197,14 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
> >       void __iomem *dst_addr;
> >       void *buf;
> >       phys_addr_t phys_addr;
> > +     int window;
> >       struct pci_epf *epf = epf_test->epf;
> >       struct device *dev = &epf->dev;
> >       struct pci_epc *epc = epf->epc;
> >       enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >       struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> >
> > -     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> > +     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
> >       if (!dst_addr) {
> >               dev_err(dev, "Failed to allocate address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > @@ -207,8 +212,8 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
> >               goto err;
> >       }
> >
> > -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->dst_addr,
> > -                            reg->size);
> > +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> > +                            reg->dst_addr, reg->size);
> >       if (ret) {
> >               dev_err(dev, "Failed to map address\n");
> >               reg->status = STATUS_DST_ADDR_INVALID;
> > diff --git a/drivers/pci/endpoint/pci-epc-core.c b/drivers/pci/endpoint/pci-epc-core.c
> > index 2091508c1620..289c266c2d90 100644
> > --- a/drivers/pci/endpoint/pci-epc-core.c
> > +++ b/drivers/pci/endpoint/pci-epc-core.c
> > @@ -358,13 +358,15 @@ EXPORT_SYMBOL_GPL(pci_epc_unmap_addr);
> >   * @epc: the EPC device on which address is allocated
> >   * @func_no: the endpoint function number in the EPC device
> >   * @phys_addr: physical address of the local system
> > + * @window: index to the window region where PCI address will be mapped
> >   * @pci_addr: PCI address to which the physical address should be mapped
> >   * @size: the size of the allocation
> >   *
> >   * Invoke to map CPU address with PCI address.
> >   */
> >  int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> > -                  phys_addr_t phys_addr, u64 pci_addr, size_t size)
> > +                  phys_addr_t phys_addr, int window,
> > +                  u64 pci_addr, size_t size)
> >  {
> >       int ret;
> >       unsigned long flags;
> > @@ -376,7 +378,8 @@ int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> >               return 0;
> >
> >       spin_lock_irqsave(&epc->lock, flags);
> > -     ret = epc->ops->map_addr(epc, func_no, phys_addr, pci_addr, size);
> > +     ret = epc->ops->map_addr(epc, func_no, phys_addr,
> > +                              window, pci_addr, size);
> >       spin_unlock_irqrestore(&epc->lock, flags);
> >
> >       return ret;
> > diff --git a/drivers/pci/endpoint/pci-epc-mem.c b/drivers/pci/endpoint/pci-epc-mem.c
> > index d2b174ce15de..c955f2c97944 100644
> > --- a/drivers/pci/endpoint/pci-epc-mem.c
> > +++ b/drivers/pci/endpoint/pci-epc-mem.c
> > @@ -39,56 +39,77 @@ static int pci_epc_mem_get_order(struct pci_epc_mem *mem, size_t size)
> >   * __pci_epc_mem_init() - initialize the pci_epc_mem structure
> >   * @epc: the EPC device that invoked pci_epc_mem_init
> >   * @phys_base: the physical address of the base
> > - * @size: the size of the address space
> > + * @num_windows: number of windows device supports
> >   * @page_size: size of each page
> >   *
> >   * Invoke to initialize the pci_epc_mem structure used by the
> >   * endpoint functions to allocate mapped PCI address.
> >   */
> > -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_base, size_t size,
> > -                    size_t page_size)
> > +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *windows,
> > +                    int num_windows, size_t page_size)
> >  {
> > -     int ret;
> > -     struct pci_epc_mem *mem;
> > -     unsigned long *bitmap;
> > +     struct pci_epc_mem *mem = NULL;
> > +     unsigned long *bitmap = NULL;
> >       unsigned int page_shift;
> > -     int pages;
> >       int bitmap_size;
> > +     int pages;
> > +     int ret;
> > +     int i;
> > +
> > +     epc->mem_windows = 0;
> > +
> > +     if (!windows)
> > +             return -EINVAL;
> > +
> > +     if (num_windows <= 0)
> > +             return -EINVAL;
> >
> >       if (page_size < PAGE_SIZE)
> >               page_size = PAGE_SIZE;
> >
> >       page_shift = ilog2(page_size);
> > -     pages = size >> page_shift;
> > -     bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
> >
> > -     mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> > -     if (!mem) {
> > -             ret = -ENOMEM;
> > -             goto err;
> > -     }
> > +     epc->mem = kcalloc(num_windows, sizeof(*mem), GFP_KERNEL);
> > +     if (!epc->mem)
> > +             return -EINVAL;
> >
> > -     bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> > -     if (!bitmap) {
> > -             ret = -ENOMEM;
> > -             goto err_mem;
> > -     }
> > +     for (i = 0; i < num_windows; i++) {
> > +             pages = windows[i].phys_base >> page_shift;
> > +             bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
> >
> > -     mem->bitmap = bitmap;
> > -     mem->phys_base = phys_base;
> > -     mem->page_size = page_size;
> > -     mem->pages = pages;
> > -     mem->size = size;
> > +             mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> > +             if (!mem) {
> > +                     ret = -ENOMEM;
> > +                     goto err_mem;
> > +             }
> >
> > -     epc->mem = mem;
> > +             bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> > +             if (!bitmap) {
> > +                     ret = -ENOMEM;
> > +                     goto err_mem;
> > +             }
> > +
> > +             mem->bitmap = bitmap;
> > +             mem->window.phys_base = windows[i].phys_base;
> > +             mem->page_size = page_size;
> > +             mem->pages = pages;
> > +             mem->window.size = windows[i].size;
> > +             mem->window.map_size = 0;
> > +
> > +             epc->mem[i] = mem;
> > +     }
> > +     epc->mem_windows = num_windows;
> >
> >       return 0;
> >
> >  err_mem:
> > -     kfree(mem);
> > +     for (; i >= 0; i--) {
> > +             kfree(mem->bitmap);
> > +             kfree(epc->mem[i]);
> > +     }
> > +     kfree(epc->mem);
> >
> > -err:
> > -return ret;
> > +     return ret;
> >  }
> >  EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
> >
> > @@ -101,48 +122,126 @@ EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
> >   */
> >  void pci_epc_mem_exit(struct pci_epc *epc)
> >  {
> > -     struct pci_epc_mem *mem = epc->mem;
> > +     struct pci_epc_mem *mem;
> > +     int i;
> > +
> > +     if (!epc->mem_windows)
> > +             return;
> > +
> > +     for (i = 0; i <= epc->mem_windows; i--) {
> > +             mem = epc->mem[i];
> > +             kfree(mem->bitmap);
> > +             kfree(epc->mem[i]);
> > +     }
> > +     kfree(epc->mem);
> >
> >       epc->mem = NULL;
> > -     kfree(mem->bitmap);
> > -     kfree(mem);
> > +     epc->mem_windows = 0;
> >  }
> >  EXPORT_SYMBOL_GPL(pci_epc_mem_exit);
> >
> > +static int pci_epc_find_best_fit_window(struct pci_epc *epc, size_t size)
> > +{
> > +     size_t window_least_size = 0;
> > +     int best_fit_window = -1;
> > +     struct pci_epc_mem *mem;
> > +     size_t actual_size;
> > +     int i;
> > +
> > +     for (i = 0; i < epc->mem_windows; i++) {
> > +             mem = epc->mem[i];
> > +
> > +             /* if chunk from this region is already used skip it */
> > +             if (mem->window.map_size)
> > +                     continue;
> > +
> > +             actual_size = ALIGN(size, mem->page_size);
> > +
> > +             if (best_fit_window == -1) {
> > +                     best_fit_window = i;
> > +                     window_least_size = mem->window.size;
> > +             } else {
> > +                     if (actual_size <= mem->window.size &&
> > +                         mem->window.size < window_least_size) {
> > +                             best_fit_window = i;
> > +                             window_least_size = mem->window.size;
> > +                     }
> > +             }
> > +     }
> > +
> > +     return best_fit_window;
> > +}
> > +
> >  /**
> >   * pci_epc_mem_alloc_addr() - allocate memory address from EPC addr space
> >   * @epc: the EPC device on which memory has to be allocated
> >   * @phys_addr: populate the allocated physical address here
> > + * @window: populate the window here which will be used to map PCI address
> >   * @size: the size of the address space that has to be allocated
> >   *
> >   * Invoke to allocate memory address from the EPC address space. This
> >   * is usually done to map the remote RC address into the local system.
> >   */
> >  void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
> > -                                  phys_addr_t *phys_addr, size_t size)
> > +                                  phys_addr_t *phys_addr,
> > +                                  int *window, size_t size)
> >  {
> > +     int best_fit = PCI_EPC_DEFAULT_WINDOW;
> > +     void __iomem *virt_addr = NULL;
> > +     struct pci_epc_mem *mem;
> > +     unsigned int page_shift;
> >       int pageno;
> > -     void __iomem *virt_addr;
> > -     struct pci_epc_mem *mem = epc->mem;
> > -     unsigned int page_shift = ilog2(mem->page_size);
> >       int order;
> >
> > +     if (epc->mem_windows <= 0)
> > +             return NULL;
> > +
> > +     if (epc->mem_windows > 1) {
> > +             best_fit = pci_epc_find_best_fit_window(epc, size);
> > +             if (best_fit < 0)
> > +                     return NULL;
> > +     }
> > +
> > +     mem = epc->mem[best_fit];
> >       size = ALIGN(size, mem->page_size);
> > +     if (size > (mem->window.size - mem->window.map_size))
> > +             return NULL;
> > +     page_shift = ilog2(mem->page_size);
> >       order = pci_epc_mem_get_order(mem, size);
> >
> >       pageno = bitmap_find_free_region(mem->bitmap, mem->pages, order);
> >       if (pageno < 0)
> >               return NULL;
> >
> > -     *phys_addr = mem->phys_base + ((phys_addr_t)pageno << page_shift);
> > +     *phys_addr = mem->window.phys_base + (pageno << page_shift);
> >       virt_addr = ioremap(*phys_addr, size);
> > -     if (!virt_addr)
> > +     if (!virt_addr) {
> >               bitmap_release_region(mem->bitmap, pageno, order);
> > +     } else {
> > +             mem->window.map_size += size;
> > +             *window = best_fit;
> > +     }
> >
> >       return virt_addr;
> >  }
> >  EXPORT_SYMBOL_GPL(pci_epc_mem_alloc_addr);
> >
> > +static int pci_epc_get_matching_window(struct pci_epc *epc,
> > +                                    phys_addr_t phys_addr)
> > +{
> > +     struct pci_epc_mem *mem;
> > +     int i;
> > +
> > +     for (i = 0; i < epc->mem_windows; i++) {
> > +             mem = epc->mem[i];
> > +
> > +             if (mem->window.phys_base == phys_addr)
> > +                     return i;
> > +     }
> > +
> > +     return -EINVAL;
> > +}
> > +
> >  /**
> >   * pci_epc_mem_free_addr() - free the allocated memory address
> >   * @epc: the EPC device on which memory was allocated
> > @@ -155,16 +254,26 @@ EXPORT_SYMBOL_GPL(pci_epc_mem_alloc_addr);
> >  void pci_epc_mem_free_addr(struct pci_epc *epc, phys_addr_t phys_addr,
> >                          void __iomem *virt_addr, size_t size)
> >  {
> > +     struct pci_epc_mem *mem;
> > +     unsigned int page_shift;
> > +     int window = 0;
> >       int pageno;
> > -     struct pci_epc_mem *mem = epc->mem;
> > -     unsigned int page_shift = ilog2(mem->page_size);
> >       int order;
> >
> > +     if (epc->mem_windows > 1) {
> > +             window = pci_epc_get_matching_window(epc, phys_addr);
> > +             if (window < 0)
> > +                     return;
> > +     }
> > +
> > +     mem = epc->mem[window];
> > +     page_shift = ilog2(mem->page_size);
> >       iounmap(virt_addr);
> > -     pageno = (phys_addr - mem->phys_base) >> page_shift;
> > +     pageno = (phys_addr - mem->window.phys_base) >> page_shift;
> >       size = ALIGN(size, mem->page_size);
> >       order = pci_epc_mem_get_order(mem, size);
> >       bitmap_release_region(mem->bitmap, pageno, order);
> > +     mem->window.map_size -= size;
> >  }
> >  EXPORT_SYMBOL_GPL(pci_epc_mem_free_addr);
> >
> > diff --git a/include/linux/pci-epc.h b/include/linux/pci-epc.h
> > index 56f1846b9d39..5e86d22dd5c5 100644
> > --- a/include/linux/pci-epc.h
> > +++ b/include/linux/pci-epc.h
> > @@ -48,7 +48,8 @@ struct pci_epc_ops {
> >       void    (*clear_bar)(struct pci_epc *epc, u8 func_no,
> >                            struct pci_epf_bar *epf_bar);
> >       int     (*map_addr)(struct pci_epc *epc, u8 func_no,
> > -                         phys_addr_t addr, u64 pci_addr, size_t size);
> > +                         phys_addr_t addr, int window,
> > +                         u64 pci_addr, size_t size);
> >       void    (*unmap_addr)(struct pci_epc *epc, u8 func_no,
> >                             phys_addr_t addr);
> >       int     (*set_msi)(struct pci_epc *epc, u8 func_no, u8 interrupts);
> > @@ -64,17 +65,28 @@ struct pci_epc_ops {
> >       struct module *owner;
> >  };
> >
> > +#define PCI_EPC_DEFAULT_WINDOW               0
> > +
> > +/**
> > + * struct pci_epc_mem_window - address window of the endpoint controller
> > + * @phys_base: physical base address of the PCI address window
> > + * @size: the size of the PCI address window
> > + */
> > +struct pci_epc_mem_window {
> > +     phys_addr_t     phys_base;
> > +     size_t          size;
> > +     size_t          map_size;
> > +};
> > +
> >  /**
> >   * struct pci_epc_mem - address space of the endpoint controller
> > - * @phys_base: physical base address of the PCI address space
> > - * @size: the size of the PCI address space
> > + * @window: address window of the endpoint controller
> >   * @bitmap: bitmap to manage the PCI address space
> > - * @pages: number of bits representing the address region
> >   * @page_size: size of each page
> > + * @pages: number of bits representing the address region
> >   */
> >  struct pci_epc_mem {
> > -     phys_addr_t     phys_base;
> > -     size_t          size;
> > +     struct pci_epc_mem_window window;
> >       unsigned long   *bitmap;
> >       size_t          page_size;
> >       int             pages;
> > @@ -85,7 +97,8 @@ struct pci_epc_mem {
> >   * @dev: PCI EPC device
> >   * @pci_epf: list of endpoint functions present in this EPC device
> >   * @ops: function pointers for performing endpoint operations
> > - * @mem: address space of the endpoint controller
> > + * @mem: array of address space of the endpoint controller
> > + * @mem_windows: number of windows supported by device
> >   * @max_functions: max number of functions that can be configured in this EPC
> >   * @group: configfs group representing the PCI EPC device
> >   * @lock: spinlock to protect pci_epc ops
> > @@ -94,7 +107,8 @@ struct pci_epc {
> >       struct device                   dev;
> >       struct list_head                pci_epf;
> >       const struct pci_epc_ops        *ops;
> > -     struct pci_epc_mem              *mem;
> > +     struct pci_epc_mem              **mem;
> > +     int                             mem_windows;
> >       u8                              max_functions;
> >       struct config_group             *group;
> >       /* spinlock to protect against concurrent access of EP controller */
> > @@ -128,8 +142,8 @@ struct pci_epc_features {
> >  #define devm_pci_epc_create(dev, ops)    \
> >               __devm_pci_epc_create((dev), (ops), THIS_MODULE)
> >
> > -#define pci_epc_mem_init(epc, phys_addr, size)       \
> > -             __pci_epc_mem_init((epc), (phys_addr), (size), PAGE_SIZE)
> > +#define pci_epc_mem_init(epc, windows, num_windows)  \
> > +             __pci_epc_mem_init((epc), windows, num_windows, PAGE_SIZE)
> >
> >  static inline void epc_set_drvdata(struct pci_epc *epc, void *data)
> >  {
> > @@ -159,7 +173,7 @@ int pci_epc_set_bar(struct pci_epc *epc, u8 func_no,
> >  void pci_epc_clear_bar(struct pci_epc *epc, u8 func_no,
> >                      struct pci_epf_bar *epf_bar);
> >  int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> > -                  phys_addr_t phys_addr,
> > +                  phys_addr_t phys_addr, int window,
> >                    u64 pci_addr, size_t size);
> >  void pci_epc_unmap_addr(struct pci_epc *epc, u8 func_no,
> >                       phys_addr_t phys_addr);
> > @@ -178,11 +192,12 @@ unsigned int pci_epc_get_first_free_bar(const struct pci_epc_features
> >  struct pci_epc *pci_epc_get(const char *epc_name);
> >  void pci_epc_put(struct pci_epc *epc);
> >
> > -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_addr, size_t size,
> > -                    size_t page_size);
> > +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *window,
> > +                    int num_windows, size_t page_size);
> >  void pci_epc_mem_exit(struct pci_epc *epc);
> >  void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
> > -                                  phys_addr_t *phys_addr, size_t size);
> > +                                  phys_addr_t *phys_addr,
> > +                                  int *window, size_t size);
> >  void pci_epc_mem_free_addr(struct pci_epc *epc, phys_addr_t phys_addr,
> >                          void __iomem *virt_addr, size_t size);
> >  #endif /* __LINUX_PCI_EPC_H */
> >

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 2/5] pci: endpoint: add support to handle multiple base for mapping outbound memory
  2019-11-27 21:21       ` Lad, Prabhakar
  (?)
@ 2019-12-05 10:22         ` Kishon Vijay Abraham I
  -1 siblings, 0 replies; 60+ messages in thread
From: Kishon Vijay Abraham I @ 2019-12-05 10:22 UTC (permalink / raw)
  To: Lad, Prabhakar
  Cc: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Marek Vasut, Yoshihiro Shimoda, linux-pci,
	Catalin Marinas, Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS, LKML,
	LAK, Linux-Renesas, Chris Paterson, Lad, Prabhakar,
	linux-rockchip, Shawn Lin, Heiko Stuebner, Tom Joseph,
	Jingoo Han, Gustavo Pimentel

Hi,

On 28/11/19 2:51 am, Lad, Prabhakar wrote:
> Hi Kishon,
> 
> Thank you for the review.
> 
> On Wed, Nov 27, 2019 at 5:15 AM Kishon Vijay Abraham I <kishon@ti.com> wrote:
>>
>> Hi Prabhakar,
>>
>> On 07/11/19 1:06 AM, Lad Prabhakar wrote:
>>> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
>>>
>>> rcar pcie controller has support to map multiple memory regions
>>> for mapping the outbound memory in local system, this feature
>>> inspires to add support for handling multiple memory bases in
>>> endpoint framework. In case of multiple memory regions only chunk
>>> or complete region can be mapped and this window needs to be
>>> passed to the controller driver.
>>>
>>> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
>>> Cc: <linux-rockchip@lists.infradead.org>
>>> Cc: Shawn Lin <shawn.lin@rock-chips.com>
>>> Cc: Heiko Stuebner <heiko@sntech.de>
>>> Cc: Tom Joseph <tjoseph@cadence.com>
>>> Cc: Jingoo Han <jingoohan1@gmail.com>
>>> Cc: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
>>> ---
>>>   .../pci/controller/dwc/pcie-designware-ep.c   |  30 ++-
>>>   drivers/pci/controller/pcie-cadence-ep.c      |  11 +-
>>>   drivers/pci/controller/pcie-rockchip-ep.c     |  13 +-
>>>   drivers/pci/endpoint/functions/pci-epf-test.c |  29 +--
>>>   drivers/pci/endpoint/pci-epc-core.c           |   7 +-
>>>   drivers/pci/endpoint/pci-epc-mem.c            | 189 ++++++++++++++----
>>>   include/linux/pci-epc.h                       |  43 ++--
>>>   7 files changed, 234 insertions(+), 88 deletions(-)
>>>
>>> diff --git a/drivers/pci/controller/dwc/pcie-designware-ep.c b/drivers/pci/controller/dwc/pcie-designware-ep.c
>>> index 3dd2e2697294..8d23c20b9afd 100644
>>> --- a/drivers/pci/controller/dwc/pcie-designware-ep.c
>>> +++ b/drivers/pci/controller/dwc/pcie-designware-ep.c
>>> @@ -195,7 +195,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, u8 func_no,
>>>   }
>>>
>>>   static int dw_pcie_ep_map_addr(struct pci_epc *epc, u8 func_no,
>>> -                            phys_addr_t addr,
>>> +                            phys_addr_t addr, int window,
>>>                               u64 pci_addr, size_t size)
>>>   {
>>>        int ret;
>>> @@ -367,6 +367,7 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
>>>        unsigned int aligned_offset;
>>>        u16 msg_ctrl, msg_data;
>>>        u32 msg_addr_lower, msg_addr_upper, reg;
>>> +     int window = PCI_EPC_DEFAULT_WINDOW;
>>>        u64 msg_addr;
>>>        bool has_upper;
>>>        int ret;
>>> @@ -390,11 +391,11 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
>>>                reg = ep->msi_cap + PCI_MSI_DATA_32;
>>>                msg_data = dw_pcie_readw_dbi(pci, reg);
>>>        }
>>> -     aligned_offset = msg_addr_lower & (epc->mem->page_size - 1);
>>> +     aligned_offset = msg_addr_lower & (epc->mem[window]->page_size - 1);
>>>        msg_addr = ((u64)msg_addr_upper) << 32 |
>>>                        (msg_addr_lower & ~aligned_offset);
>>> -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
>>> -                               epc->mem->page_size);
>>> +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
>>> +                               msg_addr, epc->mem[window]->page_size);
>>>        if (ret)
>>>                return ret;
>>>
>>> @@ -416,6 +417,7 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>>>        u32 reg, msg_data, vec_ctrl;
>>>        u64 tbl_addr, msg_addr, reg_u64;
>>>        void __iomem *msix_tbl;
>>> +     int window = PCI_EPC_DEFAULT_WINDOW;
>>>        int ret;
>>>
>>>        reg = ep->msix_cap + PCI_MSIX_TABLE;
>>> @@ -452,8 +454,8 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>>>                return -EPERM;
>>>        }
>>>
>>> -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
>>> -                               epc->mem->page_size);
>>> +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
>>> +                               msg_addr, epc->mem[window]->page_size);
>>>        if (ret)
>>>                return ret;
>>>
>>> @@ -466,10 +468,11 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>>>
>>>   void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
>>>   {
>>> +     int window = PCI_EPC_DEFAULT_WINDOW;
>>>        struct pci_epc *epc = ep->epc;
>>>
>>>        pci_epc_mem_free_addr(epc, ep->msi_mem_phys, ep->msi_mem,
>>> -                           epc->mem->page_size);
>>> +                           epc->mem[window]->page_size);
>>>
>>>        pci_epc_mem_exit(epc);
>>>   }
>>> @@ -499,9 +502,12 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
>>>        u32 reg;
>>>        void *addr;
>>>        u8 hdr_type;
>>> +     int window;
>>>        unsigned int nbars;
>>>        unsigned int offset;
>>>        struct pci_epc *epc;
>>> +     size_t msi_page_size;
>>> +     struct pci_epc_mem_window mem_window;
>>>        struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
>>>        struct device *dev = pci->dev;
>>>        struct device_node *np = dev->of_node;
>>> @@ -574,15 +580,17 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
>>>        if (ret < 0)
>>>                epc->max_functions = 1;
>>>
>>> -     ret = __pci_epc_mem_init(epc, ep->phys_base, ep->addr_size,
>>> -                              ep->page_size);
>>> +     mem_window.phys_base = ep->phys_base;
>>> +     mem_window.size = ep->addr_size;
>>> +     ret = __pci_epc_mem_init(epc, &mem_window, 1, ep->page_size);
>>>        if (ret < 0) {
>>>                dev_err(dev, "Failed to initialize address space\n");
>>>                return ret;
>>>        }
>>>
>>> -     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys,
>>> -                                          epc->mem->page_size);
>>> +     msi_page_size = epc->mem[PCI_EPC_DEFAULT_WINDOW]->page_size;
>>> +     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys, &window,
>>> +                                          msi_page_size);
>>>        if (!ep->msi_mem) {
>>>                dev_err(dev, "Failed to reserve memory for MSI/MSI-X\n");
>>>                return -ENOMEM;
>>> diff --git a/drivers/pci/controller/pcie-cadence-ep.c b/drivers/pci/controller/pcie-cadence-ep.c
>>> index def7820cb824..7991b38a5350 100644
>>> --- a/drivers/pci/controller/pcie-cadence-ep.c
>>> +++ b/drivers/pci/controller/pcie-cadence-ep.c
>>> @@ -172,7 +172,7 @@ static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
>>>   }
>>>
>>>   static int cdns_pcie_ep_map_addr(struct pci_epc *epc, u8 fn, phys_addr_t addr,
>>> -                              u64 pci_addr, size_t size)
>>> +                              int window, u64 pci_addr, size_t size)
>>>   {
>>>        struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>>>        struct cdns_pcie *pcie = &ep->pcie;
>>> @@ -434,12 +434,14 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
>>>   {
>>>        struct device *dev = &pdev->dev;
>>>        struct device_node *np = dev->of_node;
>>> +     struct pci_epc_mem_window mem_window;
>>>        struct cdns_pcie_ep *ep;
>>>        struct cdns_pcie *pcie;
>>>        struct pci_epc *epc;
>>>        struct resource *res;
>>>        int ret;
>>>        int phy_count;
>>> +     int window;
>>>
>>>        ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
>>>        if (!ep)
>>> @@ -502,15 +504,16 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
>>>        if (of_property_read_u8(np, "max-functions", &epc->max_functions) < 0)
>>>                epc->max_functions = 1;
>>>
>>> -     ret = pci_epc_mem_init(epc, pcie->mem_res->start,
>>> -                            resource_size(pcie->mem_res));
>>> +     mem_window.phys_base = pcie->mem_res->start;
>>> +     mem_window.size = resource_size(pcie->mem_res);
>>> +     ret = pci_epc_mem_init(epc, &mem_window, 1);
>>>        if (ret < 0) {
>>>                dev_err(dev, "failed to initialize the memory space\n");
>>>                goto err_init;
>>>        }
>>>
>>>        ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
>>> -                                               SZ_128K);
>>> +                                               &window, SZ_128K);
>>>        if (!ep->irq_cpu_addr) {
>>>                dev_err(dev, "failed to reserve memory space for MSI\n");
>>>                ret = -ENOMEM;
>>> diff --git a/drivers/pci/controller/pcie-rockchip-ep.c b/drivers/pci/controller/pcie-rockchip-ep.c
>>> index d743b0a48988..d59e85c8d319 100644
>>> --- a/drivers/pci/controller/pcie-rockchip-ep.c
>>> +++ b/drivers/pci/controller/pcie-rockchip-ep.c
>>> @@ -256,8 +256,8 @@ static void rockchip_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
>>>   }
>>>
>>>   static int rockchip_pcie_ep_map_addr(struct pci_epc *epc, u8 fn,
>>> -                                  phys_addr_t addr, u64 pci_addr,
>>> -                                  size_t size)
>>> +                                  phys_addr_t addr, int window,
>>> +                                  u64 pci_addr, size_t size)
>>>   {
>>>        struct rockchip_pcie_ep *ep = epc_get_drvdata(epc);
>>>        struct rockchip_pcie *pcie = &ep->rockchip;
>>> @@ -562,11 +562,13 @@ static const struct of_device_id rockchip_pcie_ep_of_match[] = {
>>>
>>>   static int rockchip_pcie_ep_probe(struct platform_device *pdev)
>>>   {
>>> +     struct pci_epc_mem_window mem_window;
>>>        struct device *dev = &pdev->dev;
>>>        struct rockchip_pcie_ep *ep;
>>>        struct rockchip_pcie *rockchip;
>>>        struct pci_epc *epc;
>>>        size_t max_regions;
>>> +     int window;
>>>        int err;
>>>
>>>        ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
>>> @@ -614,15 +616,16 @@ static int rockchip_pcie_ep_probe(struct platform_device *pdev)
>>>        /* Only enable function 0 by default */
>>>        rockchip_pcie_write(rockchip, BIT(0), PCIE_CORE_PHY_FUNC_CFG);
>>>
>>> -     err = pci_epc_mem_init(epc, rockchip->mem_res->start,
>>> -                            resource_size(rockchip->mem_res));
>>> +     mem_window.phys_base = rockchip->mem_res->start;
>>> +     mem_window.size = resource_size(rockchip->mem_res);
>>> +     err = pci_epc_mem_init(epc, &mem_window, 1);
>>>        if (err < 0) {
>>>                dev_err(dev, "failed to initialize the memory space\n");
>>>                goto err_uninit_port;
>>>        }
>>>
>>>        ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
>>> -                                               SZ_128K);
>>> +                                               &window, SZ_128K);
>>>        if (!ep->irq_cpu_addr) {
>>>                dev_err(dev, "failed to reserve memory space for MSI\n");
>>>                err = -ENOMEM;
>>> diff --git a/drivers/pci/endpoint/functions/pci-epf-test.c b/drivers/pci/endpoint/functions/pci-epf-test.c
>>> index 5d74f81ddfe4..475228011703 100644
>>> --- a/drivers/pci/endpoint/functions/pci-epf-test.c
>>> +++ b/drivers/pci/endpoint/functions/pci-epf-test.c
>>> @@ -84,8 +84,10 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>>>        struct pci_epc *epc = epf->epc;
>>>        enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>>>        struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
>>> +     int window;
>>>
>>> -     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr, reg->size);
>>> +     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr,
>>> +                                       &window, reg->size);
>>>        if (!src_addr) {
>>>                dev_err(dev, "Failed to allocate source address\n");
>>>                reg->status = STATUS_SRC_ADDR_INVALID;
>>> @@ -93,15 +95,16 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>>>                goto err;
>>>        }
>>>
>>> -     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, reg->src_addr,
>>> -                            reg->size);
>>> +     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, window,
>>> +                            reg->src_addr, reg->size);
>>>        if (ret) {
>>>                dev_err(dev, "Failed to map source address\n");
>>>                reg->status = STATUS_SRC_ADDR_INVALID;
>>>                goto err_src_addr;
>>>        }
>>>
>>> -     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr, reg->size);
>>> +     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr,
>>> +                                       &window, reg->size);
>>>        if (!dst_addr) {
>>>                dev_err(dev, "Failed to allocate destination address\n");
>>>                reg->status = STATUS_DST_ADDR_INVALID;
>>> @@ -109,8 +112,8 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>>>                goto err_src_map_addr;
>>>        }
>>>
>>> -     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, reg->dst_addr,
>>> -                            reg->size);
>>> +     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, window,
>>> +                            reg->dst_addr, reg->size);
>>>        if (ret) {
>>>                dev_err(dev, "Failed to map destination address\n");
>>>                reg->status = STATUS_DST_ADDR_INVALID;
>>> @@ -146,8 +149,9 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
>>>        struct pci_epc *epc = epf->epc;
>>>        enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>>>        struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
>>> +     int window;
>>>
>>> -     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
>>> +     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
>>>        if (!src_addr) {
>>>                dev_err(dev, "Failed to allocate address\n");
>>>                reg->status = STATUS_SRC_ADDR_INVALID;
>>> @@ -155,8 +159,8 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
>>>                goto err;
>>>        }
>>>
>>> -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->src_addr,
>>> -                            reg->size);
>>> +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
>>> +                            reg->src_addr, reg->size);
>>>        if (ret) {
>>>                dev_err(dev, "Failed to map address\n");
>>>                reg->status = STATUS_SRC_ADDR_INVALID;
>>> @@ -193,13 +197,14 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
>>>        void __iomem *dst_addr;
>>>        void *buf;
>>>        phys_addr_t phys_addr;
>>> +     int window;
>>>        struct pci_epf *epf = epf_test->epf;
>>>        struct device *dev = &epf->dev;
>>>        struct pci_epc *epc = epf->epc;
>>>        enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>>>        struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
>>>
>>> -     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
>>> +     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
>>>        if (!dst_addr) {
>>>                dev_err(dev, "Failed to allocate address\n");
>>>                reg->status = STATUS_DST_ADDR_INVALID;
>>> @@ -207,8 +212,8 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
>>>                goto err;
>>>        }
>>>
>>> -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->dst_addr,
>>> -                            reg->size);
>>> +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
>>> +                            reg->dst_addr, reg->size);
>>>        if (ret) {
>>>                dev_err(dev, "Failed to map address\n");
>>>                reg->status = STATUS_DST_ADDR_INVALID;
>>> diff --git a/drivers/pci/endpoint/pci-epc-core.c b/drivers/pci/endpoint/pci-epc-core.c
>>> index 2091508c1620..289c266c2d90 100644
>>> --- a/drivers/pci/endpoint/pci-epc-core.c
>>> +++ b/drivers/pci/endpoint/pci-epc-core.c
>>> @@ -358,13 +358,15 @@ EXPORT_SYMBOL_GPL(pci_epc_unmap_addr);
>>>    * @epc: the EPC device on which address is allocated
>>>    * @func_no: the endpoint function number in the EPC device
>>>    * @phys_addr: physical address of the local system
>>> + * @window: index to the window region where PCI address will be mapped
>>>    * @pci_addr: PCI address to which the physical address should be mapped
>>>    * @size: the size of the allocation
>>>    *
>>>    * Invoke to map CPU address with PCI address.
>>>    */
>>>   int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
>>> -                  phys_addr_t phys_addr, u64 pci_addr, size_t size)
>>> +                  phys_addr_t phys_addr, int window,
>>> +                  u64 pci_addr, size_t size)
>>>   {
>>>        int ret;
>>>        unsigned long flags;
>>> @@ -376,7 +378,8 @@ int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
>>>                return 0;
>>>
>>>        spin_lock_irqsave(&epc->lock, flags);
>>> -     ret = epc->ops->map_addr(epc, func_no, phys_addr, pci_addr, size);
>>> +     ret = epc->ops->map_addr(epc, func_no, phys_addr,
>>> +                              window, pci_addr, size);
>>>        spin_unlock_irqrestore(&epc->lock, flags);
>>>
>>>        return ret;
>>> diff --git a/drivers/pci/endpoint/pci-epc-mem.c b/drivers/pci/endpoint/pci-epc-mem.c
>>> index d2b174ce15de..c955f2c97944 100644
>>> --- a/drivers/pci/endpoint/pci-epc-mem.c
>>> +++ b/drivers/pci/endpoint/pci-epc-mem.c
>>> @@ -39,56 +39,77 @@ static int pci_epc_mem_get_order(struct pci_epc_mem *mem, size_t size)
>>>    * __pci_epc_mem_init() - initialize the pci_epc_mem structure
>>>    * @epc: the EPC device that invoked pci_epc_mem_init
>>>    * @phys_base: the physical address of the base
>>> - * @size: the size of the address space
>>> + * @num_windows: number of windows device supports
>>
>> struct pci_epc_mem_window is missing here.
> oops my bad will fix that.
> 
>>>    * @page_size: size of each page
>>>    *
>>>    * Invoke to initialize the pci_epc_mem structure used by the
>>>    * endpoint functions to allocate mapped PCI address.
>>>    */
>>> -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_base, size_t size,
>>> -                    size_t page_size)
>>> +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *windows,
>>> +                    int num_windows, size_t page_size)
>>>   {
>>> -     int ret;
>>> -     struct pci_epc_mem *mem;
>>> -     unsigned long *bitmap;
>>> +     struct pci_epc_mem *mem = NULL;
>>> +     unsigned long *bitmap = NULL;
>>>        unsigned int page_shift;
>>> -     int pages;
>>>        int bitmap_size;
>>> +     int pages;
>>> +     int ret;
>>> +     int i;
>>> +
>>> +     epc->mem_windows = 0;
>>> +
>>> +     if (!windows)
>>> +             return -EINVAL;
>>> +
>>> +     if (num_windows <= 0)
>>> +             return -EINVAL;
>>>
>>>        if (page_size < PAGE_SIZE)
>>>                page_size = PAGE_SIZE;
>>>
>>>        page_shift = ilog2(page_size);
>>> -     pages = size >> page_shift;
>>> -     bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
>>>
>>> -     mem = kzalloc(sizeof(*mem), GFP_KERNEL);
>>> -     if (!mem) {
>>> -             ret = -ENOMEM;
>>> -             goto err;
>>> -     }
>>> +     epc->mem = kcalloc(num_windows, sizeof(*mem), GFP_KERNEL);
>>> +     if (!epc->mem)
>>> +             return -EINVAL;
>>>
>>> -     bitmap = kzalloc(bitmap_size, GFP_KERNEL);
>>> -     if (!bitmap) {
>>> -             ret = -ENOMEM;
>>> -             goto err_mem;
>>> -     }
>>> +     for (i = 0; i < num_windows; i++) {
>>> +             pages = windows[i].phys_base >> page_shift;
>>> +             bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
>>>
>>> -     mem->bitmap = bitmap;
>>> -     mem->phys_base = phys_base;
>>> -     mem->page_size = page_size;
>>> -     mem->pages = pages;
>>> -     mem->size = size;
>>> +             mem = kzalloc(sizeof(*mem), GFP_KERNEL);
>>> +             if (!mem) {
>>> +                     ret = -ENOMEM;
>>> +                     goto err_mem;
>>> +             }
>>>
>>> -     epc->mem = mem;
>>> +             bitmap = kzalloc(bitmap_size, GFP_KERNEL);
>>> +             if (!bitmap) {
>>> +                     ret = -ENOMEM;
>>> +                     goto err_mem;
>>> +             }
>>> +
>>> +             mem->bitmap = bitmap;
>>> +             mem->window.phys_base = windows[i].phys_base;
>>> +             mem->page_size = page_size;
>>> +             mem->pages = pages;
>>> +             mem->window.size = windows[i].size;
>>> +             mem->window.map_size = 0;
>>> +
>>> +             epc->mem[i] = mem;
>>> +     }
>>> +     epc->mem_windows = num_windows;
>>>
>>>        return 0;
>>>
>>>   err_mem:
>>> -     kfree(mem);
>>> +     for (; i >= 0; i--) {
>>> +             kfree(mem->bitmap);
>>> +             kfree(epc->mem[i]);
>>> +     }
>>> +     kfree(epc->mem);
>>>
>>> -err:
>>> -return ret;
>>> +     return ret;
>>>   }
>>>   EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
>>>
>>> @@ -101,48 +122,126 @@ EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
>>>    */
>>>   void pci_epc_mem_exit(struct pci_epc *epc)
>>>   {
>>> -     struct pci_epc_mem *mem = epc->mem;
>>> +     struct pci_epc_mem *mem;
>>> +     int i;
>>> +
>>> +     if (!epc->mem_windows)
>>> +             return;
>>> +
>>> +     for (i = 0; i <= epc->mem_windows; i--) {
>>> +             mem = epc->mem[i];
>>> +             kfree(mem->bitmap);
>>> +             kfree(epc->mem[i]);
>>> +     }
>>> +     kfree(epc->mem);
>>>
>>>        epc->mem = NULL;
>>> -     kfree(mem->bitmap);
>>> -     kfree(mem);
>>> +     epc->mem_windows = 0;
>>>   }
>>>   EXPORT_SYMBOL_GPL(pci_epc_mem_exit);
>>>
>>> +static int pci_epc_find_best_fit_window(struct pci_epc *epc, size_t size)
>>> +{
>>> +     size_t window_least_size = 0;
>>> +     int best_fit_window = -1;
>>> +     struct pci_epc_mem *mem;
>>> +     size_t actual_size;
>>> +     int i;
>>> +
>>> +     for (i = 0; i < epc->mem_windows; i++) {
>>> +             mem = epc->mem[i];
>>> +
>>> +             /* if chunk from this region is already used skip it */
>>> +             if (mem->window.map_size)
>>> +                     continue;
>>> +
>>> +             actual_size = ALIGN(size, mem->page_size);
>>> +
>>> +             if (best_fit_window == -1) {
>>> +                     best_fit_window = i;
>>> +                     window_least_size = mem->window.size;
>>> +             } else {
>>> +                     if (actual_size <= mem->window.size &&
>>> +                         mem->window.size < window_least_size) {
>>> +                             best_fit_window = i;
>>> +                             window_least_size = mem->window.size;
>>> +                     }
>>> +             }
>>> +     }
>>> +
>>> +     return best_fit_window;
>>> +}
>>> +
>>>   /**
>>>    * pci_epc_mem_alloc_addr() - allocate memory address from EPC addr space
>>>    * @epc: the EPC device on which memory has to be allocated
>>>    * @phys_addr: populate the allocated physical address here
>>> + * @window: populate the window here which will be used to map PCI address
>>>    * @size: the size of the address space that has to be allocated
>>>    *
>>>    * Invoke to allocate memory address from the EPC address space. This
>>>    * is usually done to map the remote RC address into the local system.
>>>    */
>>>   void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
>>> -                                  phys_addr_t *phys_addr, size_t size)
>>> +                                  phys_addr_t *phys_addr,
>>> +                                  int *window, size_t size)
>>>   {
>>> +     int best_fit = PCI_EPC_DEFAULT_WINDOW;
>>> +     void __iomem *virt_addr = NULL;
>>> +     struct pci_epc_mem *mem;
>>> +     unsigned int page_shift;
>>>        int pageno;
>>> -     void __iomem *virt_addr;
>>> -     struct pci_epc_mem *mem = epc->mem;
>>> -     unsigned int page_shift = ilog2(mem->page_size);
>>>        int order;
>>>
>>> +     if (epc->mem_windows <= 0)
>>> +             return NULL;
>>> +
>>> +     if (epc->mem_windows > 1) {
>>> +             best_fit = pci_epc_find_best_fit_window(epc, size);
>>> +             if (best_fit < 0)
>>> +                     return NULL;
>>> +     }
>>> +
>>> +     mem = epc->mem[best_fit];
>>>        size = ALIGN(size, mem->page_size);
>>> +     if (size > (mem->window.size - mem->window.map_size))
>>> +             return NULL;
>>
>> Assume I have two mem regions, first region is of size 128MB and the second
>> region is of size 4GB. If there is a allocation request for 4GB, will the
>> allocation succeed?
>>
> yes it would succeed, the pci_epc_find_best_fit_window() would get the
> corresponding window.

Can you step through the algo again? pci_epc_find_best_fit_window() will 
return the first region which is of size 128MB.

Thanks
Kishon

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 2/5] pci: endpoint: add support to handle multiple base for mapping outbound memory
@ 2019-12-05 10:22         ` Kishon Vijay Abraham I
  0 siblings, 0 replies; 60+ messages in thread
From: Kishon Vijay Abraham I @ 2019-12-05 10:22 UTC (permalink / raw)
  To: Lad, Prabhakar
  Cc: Mark Rutland, Heiko Stuebner, Geert Uytterhoeven, linux-pci,
	Shawn Lin, Will Deacon, Marek Vasut, Lorenzo Pieralisi,
	Magnus Damm, linux-rockchip, Catalin Marinas,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Chris Paterson, Arnd Bergmann, Lad, Prabhakar, Rob Herring,
	Bjorn Helgaas, LAK, Greg Kroah-Hartman, Yoshihiro Shimoda, LK

Hi,

On 28/11/19 2:51 am, Lad, Prabhakar wrote:
> Hi Kishon,
> 
> Thank you for the review.
> 
> On Wed, Nov 27, 2019 at 5:15 AM Kishon Vijay Abraham I <kishon@ti.com> wrote:
>>
>> Hi Prabhakar,
>>
>> On 07/11/19 1:06 AM, Lad Prabhakar wrote:
>>> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
>>>
>>> rcar pcie controller has support to map multiple memory regions
>>> for mapping the outbound memory in local system, this feature
>>> inspires to add support for handling multiple memory bases in
>>> endpoint framework. In case of multiple memory regions only chunk
>>> or complete region can be mapped and this window needs to be
>>> passed to the controller driver.
>>>
>>> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
>>> Cc: <linux-rockchip@lists.infradead.org>
>>> Cc: Shawn Lin <shawn.lin@rock-chips.com>
>>> Cc: Heiko Stuebner <heiko@sntech.de>
>>> Cc: Tom Joseph <tjoseph@cadence.com>
>>> Cc: Jingoo Han <jingoohan1@gmail.com>
>>> Cc: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
>>> ---
>>>   .../pci/controller/dwc/pcie-designware-ep.c   |  30 ++-
>>>   drivers/pci/controller/pcie-cadence-ep.c      |  11 +-
>>>   drivers/pci/controller/pcie-rockchip-ep.c     |  13 +-
>>>   drivers/pci/endpoint/functions/pci-epf-test.c |  29 +--
>>>   drivers/pci/endpoint/pci-epc-core.c           |   7 +-
>>>   drivers/pci/endpoint/pci-epc-mem.c            | 189 ++++++++++++++----
>>>   include/linux/pci-epc.h                       |  43 ++--
>>>   7 files changed, 234 insertions(+), 88 deletions(-)
>>>
>>> diff --git a/drivers/pci/controller/dwc/pcie-designware-ep.c b/drivers/pci/controller/dwc/pcie-designware-ep.c
>>> index 3dd2e2697294..8d23c20b9afd 100644
>>> --- a/drivers/pci/controller/dwc/pcie-designware-ep.c
>>> +++ b/drivers/pci/controller/dwc/pcie-designware-ep.c
>>> @@ -195,7 +195,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, u8 func_no,
>>>   }
>>>
>>>   static int dw_pcie_ep_map_addr(struct pci_epc *epc, u8 func_no,
>>> -                            phys_addr_t addr,
>>> +                            phys_addr_t addr, int window,
>>>                               u64 pci_addr, size_t size)
>>>   {
>>>        int ret;
>>> @@ -367,6 +367,7 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
>>>        unsigned int aligned_offset;
>>>        u16 msg_ctrl, msg_data;
>>>        u32 msg_addr_lower, msg_addr_upper, reg;
>>> +     int window = PCI_EPC_DEFAULT_WINDOW;
>>>        u64 msg_addr;
>>>        bool has_upper;
>>>        int ret;
>>> @@ -390,11 +391,11 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
>>>                reg = ep->msi_cap + PCI_MSI_DATA_32;
>>>                msg_data = dw_pcie_readw_dbi(pci, reg);
>>>        }
>>> -     aligned_offset = msg_addr_lower & (epc->mem->page_size - 1);
>>> +     aligned_offset = msg_addr_lower & (epc->mem[window]->page_size - 1);
>>>        msg_addr = ((u64)msg_addr_upper) << 32 |
>>>                        (msg_addr_lower & ~aligned_offset);
>>> -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
>>> -                               epc->mem->page_size);
>>> +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
>>> +                               msg_addr, epc->mem[window]->page_size);
>>>        if (ret)
>>>                return ret;
>>>
>>> @@ -416,6 +417,7 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>>>        u32 reg, msg_data, vec_ctrl;
>>>        u64 tbl_addr, msg_addr, reg_u64;
>>>        void __iomem *msix_tbl;
>>> +     int window = PCI_EPC_DEFAULT_WINDOW;
>>>        int ret;
>>>
>>>        reg = ep->msix_cap + PCI_MSIX_TABLE;
>>> @@ -452,8 +454,8 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>>>                return -EPERM;
>>>        }
>>>
>>> -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
>>> -                               epc->mem->page_size);
>>> +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
>>> +                               msg_addr, epc->mem[window]->page_size);
>>>        if (ret)
>>>                return ret;
>>>
>>> @@ -466,10 +468,11 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>>>
>>>   void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
>>>   {
>>> +     int window = PCI_EPC_DEFAULT_WINDOW;
>>>        struct pci_epc *epc = ep->epc;
>>>
>>>        pci_epc_mem_free_addr(epc, ep->msi_mem_phys, ep->msi_mem,
>>> -                           epc->mem->page_size);
>>> +                           epc->mem[window]->page_size);
>>>
>>>        pci_epc_mem_exit(epc);
>>>   }
>>> @@ -499,9 +502,12 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
>>>        u32 reg;
>>>        void *addr;
>>>        u8 hdr_type;
>>> +     int window;
>>>        unsigned int nbars;
>>>        unsigned int offset;
>>>        struct pci_epc *epc;
>>> +     size_t msi_page_size;
>>> +     struct pci_epc_mem_window mem_window;
>>>        struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
>>>        struct device *dev = pci->dev;
>>>        struct device_node *np = dev->of_node;
>>> @@ -574,15 +580,17 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
>>>        if (ret < 0)
>>>                epc->max_functions = 1;
>>>
>>> -     ret = __pci_epc_mem_init(epc, ep->phys_base, ep->addr_size,
>>> -                              ep->page_size);
>>> +     mem_window.phys_base = ep->phys_base;
>>> +     mem_window.size = ep->addr_size;
>>> +     ret = __pci_epc_mem_init(epc, &mem_window, 1, ep->page_size);
>>>        if (ret < 0) {
>>>                dev_err(dev, "Failed to initialize address space\n");
>>>                return ret;
>>>        }
>>>
>>> -     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys,
>>> -                                          epc->mem->page_size);
>>> +     msi_page_size = epc->mem[PCI_EPC_DEFAULT_WINDOW]->page_size;
>>> +     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys, &window,
>>> +                                          msi_page_size);
>>>        if (!ep->msi_mem) {
>>>                dev_err(dev, "Failed to reserve memory for MSI/MSI-X\n");
>>>                return -ENOMEM;
>>> diff --git a/drivers/pci/controller/pcie-cadence-ep.c b/drivers/pci/controller/pcie-cadence-ep.c
>>> index def7820cb824..7991b38a5350 100644
>>> --- a/drivers/pci/controller/pcie-cadence-ep.c
>>> +++ b/drivers/pci/controller/pcie-cadence-ep.c
>>> @@ -172,7 +172,7 @@ static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
>>>   }
>>>
>>>   static int cdns_pcie_ep_map_addr(struct pci_epc *epc, u8 fn, phys_addr_t addr,
>>> -                              u64 pci_addr, size_t size)
>>> +                              int window, u64 pci_addr, size_t size)
>>>   {
>>>        struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>>>        struct cdns_pcie *pcie = &ep->pcie;
>>> @@ -434,12 +434,14 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
>>>   {
>>>        struct device *dev = &pdev->dev;
>>>        struct device_node *np = dev->of_node;
>>> +     struct pci_epc_mem_window mem_window;
>>>        struct cdns_pcie_ep *ep;
>>>        struct cdns_pcie *pcie;
>>>        struct pci_epc *epc;
>>>        struct resource *res;
>>>        int ret;
>>>        int phy_count;
>>> +     int window;
>>>
>>>        ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
>>>        if (!ep)
>>> @@ -502,15 +504,16 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
>>>        if (of_property_read_u8(np, "max-functions", &epc->max_functions) < 0)
>>>                epc->max_functions = 1;
>>>
>>> -     ret = pci_epc_mem_init(epc, pcie->mem_res->start,
>>> -                            resource_size(pcie->mem_res));
>>> +     mem_window.phys_base = pcie->mem_res->start;
>>> +     mem_window.size = resource_size(pcie->mem_res);
>>> +     ret = pci_epc_mem_init(epc, &mem_window, 1);
>>>        if (ret < 0) {
>>>                dev_err(dev, "failed to initialize the memory space\n");
>>>                goto err_init;
>>>        }
>>>
>>>        ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
>>> -                                               SZ_128K);
>>> +                                               &window, SZ_128K);
>>>        if (!ep->irq_cpu_addr) {
>>>                dev_err(dev, "failed to reserve memory space for MSI\n");
>>>                ret = -ENOMEM;
>>> diff --git a/drivers/pci/controller/pcie-rockchip-ep.c b/drivers/pci/controller/pcie-rockchip-ep.c
>>> index d743b0a48988..d59e85c8d319 100644
>>> --- a/drivers/pci/controller/pcie-rockchip-ep.c
>>> +++ b/drivers/pci/controller/pcie-rockchip-ep.c
>>> @@ -256,8 +256,8 @@ static void rockchip_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
>>>   }
>>>
>>>   static int rockchip_pcie_ep_map_addr(struct pci_epc *epc, u8 fn,
>>> -                                  phys_addr_t addr, u64 pci_addr,
>>> -                                  size_t size)
>>> +                                  phys_addr_t addr, int window,
>>> +                                  u64 pci_addr, size_t size)
>>>   {
>>>        struct rockchip_pcie_ep *ep = epc_get_drvdata(epc);
>>>        struct rockchip_pcie *pcie = &ep->rockchip;
>>> @@ -562,11 +562,13 @@ static const struct of_device_id rockchip_pcie_ep_of_match[] = {
>>>
>>>   static int rockchip_pcie_ep_probe(struct platform_device *pdev)
>>>   {
>>> +     struct pci_epc_mem_window mem_window;
>>>        struct device *dev = &pdev->dev;
>>>        struct rockchip_pcie_ep *ep;
>>>        struct rockchip_pcie *rockchip;
>>>        struct pci_epc *epc;
>>>        size_t max_regions;
>>> +     int window;
>>>        int err;
>>>
>>>        ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
>>> @@ -614,15 +616,16 @@ static int rockchip_pcie_ep_probe(struct platform_device *pdev)
>>>        /* Only enable function 0 by default */
>>>        rockchip_pcie_write(rockchip, BIT(0), PCIE_CORE_PHY_FUNC_CFG);
>>>
>>> -     err = pci_epc_mem_init(epc, rockchip->mem_res->start,
>>> -                            resource_size(rockchip->mem_res));
>>> +     mem_window.phys_base = rockchip->mem_res->start;
>>> +     mem_window.size = resource_size(rockchip->mem_res);
>>> +     err = pci_epc_mem_init(epc, &mem_window, 1);
>>>        if (err < 0) {
>>>                dev_err(dev, "failed to initialize the memory space\n");
>>>                goto err_uninit_port;
>>>        }
>>>
>>>        ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
>>> -                                               SZ_128K);
>>> +                                               &window, SZ_128K);
>>>        if (!ep->irq_cpu_addr) {
>>>                dev_err(dev, "failed to reserve memory space for MSI\n");
>>>                err = -ENOMEM;
>>> diff --git a/drivers/pci/endpoint/functions/pci-epf-test.c b/drivers/pci/endpoint/functions/pci-epf-test.c
>>> index 5d74f81ddfe4..475228011703 100644
>>> --- a/drivers/pci/endpoint/functions/pci-epf-test.c
>>> +++ b/drivers/pci/endpoint/functions/pci-epf-test.c
>>> @@ -84,8 +84,10 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>>>        struct pci_epc *epc = epf->epc;
>>>        enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>>>        struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
>>> +     int window;
>>>
>>> -     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr, reg->size);
>>> +     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr,
>>> +                                       &window, reg->size);
>>>        if (!src_addr) {
>>>                dev_err(dev, "Failed to allocate source address\n");
>>>                reg->status = STATUS_SRC_ADDR_INVALID;
>>> @@ -93,15 +95,16 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>>>                goto err;
>>>        }
>>>
>>> -     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, reg->src_addr,
>>> -                            reg->size);
>>> +     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, window,
>>> +                            reg->src_addr, reg->size);
>>>        if (ret) {
>>>                dev_err(dev, "Failed to map source address\n");
>>>                reg->status = STATUS_SRC_ADDR_INVALID;
>>>                goto err_src_addr;
>>>        }
>>>
>>> -     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr, reg->size);
>>> +     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr,
>>> +                                       &window, reg->size);
>>>        if (!dst_addr) {
>>>                dev_err(dev, "Failed to allocate destination address\n");
>>>                reg->status = STATUS_DST_ADDR_INVALID;
>>> @@ -109,8 +112,8 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>>>                goto err_src_map_addr;
>>>        }
>>>
>>> -     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, reg->dst_addr,
>>> -                            reg->size);
>>> +     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, window,
>>> +                            reg->dst_addr, reg->size);
>>>        if (ret) {
>>>                dev_err(dev, "Failed to map destination address\n");
>>>                reg->status = STATUS_DST_ADDR_INVALID;
>>> @@ -146,8 +149,9 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
>>>        struct pci_epc *epc = epf->epc;
>>>        enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>>>        struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
>>> +     int window;
>>>
>>> -     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
>>> +     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
>>>        if (!src_addr) {
>>>                dev_err(dev, "Failed to allocate address\n");
>>>                reg->status = STATUS_SRC_ADDR_INVALID;
>>> @@ -155,8 +159,8 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
>>>                goto err;
>>>        }
>>>
>>> -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->src_addr,
>>> -                            reg->size);
>>> +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
>>> +                            reg->src_addr, reg->size);
>>>        if (ret) {
>>>                dev_err(dev, "Failed to map address\n");
>>>                reg->status = STATUS_SRC_ADDR_INVALID;
>>> @@ -193,13 +197,14 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
>>>        void __iomem *dst_addr;
>>>        void *buf;
>>>        phys_addr_t phys_addr;
>>> +     int window;
>>>        struct pci_epf *epf = epf_test->epf;
>>>        struct device *dev = &epf->dev;
>>>        struct pci_epc *epc = epf->epc;
>>>        enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>>>        struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
>>>
>>> -     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
>>> +     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
>>>        if (!dst_addr) {
>>>                dev_err(dev, "Failed to allocate address\n");
>>>                reg->status = STATUS_DST_ADDR_INVALID;
>>> @@ -207,8 +212,8 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
>>>                goto err;
>>>        }
>>>
>>> -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->dst_addr,
>>> -                            reg->size);
>>> +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
>>> +                            reg->dst_addr, reg->size);
>>>        if (ret) {
>>>                dev_err(dev, "Failed to map address\n");
>>>                reg->status = STATUS_DST_ADDR_INVALID;
>>> diff --git a/drivers/pci/endpoint/pci-epc-core.c b/drivers/pci/endpoint/pci-epc-core.c
>>> index 2091508c1620..289c266c2d90 100644
>>> --- a/drivers/pci/endpoint/pci-epc-core.c
>>> +++ b/drivers/pci/endpoint/pci-epc-core.c
>>> @@ -358,13 +358,15 @@ EXPORT_SYMBOL_GPL(pci_epc_unmap_addr);
>>>    * @epc: the EPC device on which address is allocated
>>>    * @func_no: the endpoint function number in the EPC device
>>>    * @phys_addr: physical address of the local system
>>> + * @window: index to the window region where PCI address will be mapped
>>>    * @pci_addr: PCI address to which the physical address should be mapped
>>>    * @size: the size of the allocation
>>>    *
>>>    * Invoke to map CPU address with PCI address.
>>>    */
>>>   int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
>>> -                  phys_addr_t phys_addr, u64 pci_addr, size_t size)
>>> +                  phys_addr_t phys_addr, int window,
>>> +                  u64 pci_addr, size_t size)
>>>   {
>>>        int ret;
>>>        unsigned long flags;
>>> @@ -376,7 +378,8 @@ int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
>>>                return 0;
>>>
>>>        spin_lock_irqsave(&epc->lock, flags);
>>> -     ret = epc->ops->map_addr(epc, func_no, phys_addr, pci_addr, size);
>>> +     ret = epc->ops->map_addr(epc, func_no, phys_addr,
>>> +                              window, pci_addr, size);
>>>        spin_unlock_irqrestore(&epc->lock, flags);
>>>
>>>        return ret;
>>> diff --git a/drivers/pci/endpoint/pci-epc-mem.c b/drivers/pci/endpoint/pci-epc-mem.c
>>> index d2b174ce15de..c955f2c97944 100644
>>> --- a/drivers/pci/endpoint/pci-epc-mem.c
>>> +++ b/drivers/pci/endpoint/pci-epc-mem.c
>>> @@ -39,56 +39,77 @@ static int pci_epc_mem_get_order(struct pci_epc_mem *mem, size_t size)
>>>    * __pci_epc_mem_init() - initialize the pci_epc_mem structure
>>>    * @epc: the EPC device that invoked pci_epc_mem_init
>>>    * @phys_base: the physical address of the base
>>> - * @size: the size of the address space
>>> + * @num_windows: number of windows device supports
>>
>> struct pci_epc_mem_window is missing here.
> oops my bad will fix that.
> 
>>>    * @page_size: size of each page
>>>    *
>>>    * Invoke to initialize the pci_epc_mem structure used by the
>>>    * endpoint functions to allocate mapped PCI address.
>>>    */
>>> -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_base, size_t size,
>>> -                    size_t page_size)
>>> +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *windows,
>>> +                    int num_windows, size_t page_size)
>>>   {
>>> -     int ret;
>>> -     struct pci_epc_mem *mem;
>>> -     unsigned long *bitmap;
>>> +     struct pci_epc_mem *mem = NULL;
>>> +     unsigned long *bitmap = NULL;
>>>        unsigned int page_shift;
>>> -     int pages;
>>>        int bitmap_size;
>>> +     int pages;
>>> +     int ret;
>>> +     int i;
>>> +
>>> +     epc->mem_windows = 0;
>>> +
>>> +     if (!windows)
>>> +             return -EINVAL;
>>> +
>>> +     if (num_windows <= 0)
>>> +             return -EINVAL;
>>>
>>>        if (page_size < PAGE_SIZE)
>>>                page_size = PAGE_SIZE;
>>>
>>>        page_shift = ilog2(page_size);
>>> -     pages = size >> page_shift;
>>> -     bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
>>>
>>> -     mem = kzalloc(sizeof(*mem), GFP_KERNEL);
>>> -     if (!mem) {
>>> -             ret = -ENOMEM;
>>> -             goto err;
>>> -     }
>>> +     epc->mem = kcalloc(num_windows, sizeof(*mem), GFP_KERNEL);
>>> +     if (!epc->mem)
>>> +             return -EINVAL;
>>>
>>> -     bitmap = kzalloc(bitmap_size, GFP_KERNEL);
>>> -     if (!bitmap) {
>>> -             ret = -ENOMEM;
>>> -             goto err_mem;
>>> -     }
>>> +     for (i = 0; i < num_windows; i++) {
>>> +             pages = windows[i].phys_base >> page_shift;
>>> +             bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
>>>
>>> -     mem->bitmap = bitmap;
>>> -     mem->phys_base = phys_base;
>>> -     mem->page_size = page_size;
>>> -     mem->pages = pages;
>>> -     mem->size = size;
>>> +             mem = kzalloc(sizeof(*mem), GFP_KERNEL);
>>> +             if (!mem) {
>>> +                     ret = -ENOMEM;
>>> +                     goto err_mem;
>>> +             }
>>>
>>> -     epc->mem = mem;
>>> +             bitmap = kzalloc(bitmap_size, GFP_KERNEL);
>>> +             if (!bitmap) {
>>> +                     ret = -ENOMEM;
>>> +                     goto err_mem;
>>> +             }
>>> +
>>> +             mem->bitmap = bitmap;
>>> +             mem->window.phys_base = windows[i].phys_base;
>>> +             mem->page_size = page_size;
>>> +             mem->pages = pages;
>>> +             mem->window.size = windows[i].size;
>>> +             mem->window.map_size = 0;
>>> +
>>> +             epc->mem[i] = mem;
>>> +     }
>>> +     epc->mem_windows = num_windows;
>>>
>>>        return 0;
>>>
>>>   err_mem:
>>> -     kfree(mem);
>>> +     for (; i >= 0; i--) {
>>> +             kfree(mem->bitmap);
>>> +             kfree(epc->mem[i]);
>>> +     }
>>> +     kfree(epc->mem);
>>>
>>> -err:
>>> -return ret;
>>> +     return ret;
>>>   }
>>>   EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
>>>
>>> @@ -101,48 +122,126 @@ EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
>>>    */
>>>   void pci_epc_mem_exit(struct pci_epc *epc)
>>>   {
>>> -     struct pci_epc_mem *mem = epc->mem;
>>> +     struct pci_epc_mem *mem;
>>> +     int i;
>>> +
>>> +     if (!epc->mem_windows)
>>> +             return;
>>> +
>>> +     for (i = 0; i <= epc->mem_windows; i--) {
>>> +             mem = epc->mem[i];
>>> +             kfree(mem->bitmap);
>>> +             kfree(epc->mem[i]);
>>> +     }
>>> +     kfree(epc->mem);
>>>
>>>        epc->mem = NULL;
>>> -     kfree(mem->bitmap);
>>> -     kfree(mem);
>>> +     epc->mem_windows = 0;
>>>   }
>>>   EXPORT_SYMBOL_GPL(pci_epc_mem_exit);
>>>
>>> +static int pci_epc_find_best_fit_window(struct pci_epc *epc, size_t size)
>>> +{
>>> +     size_t window_least_size = 0;
>>> +     int best_fit_window = -1;
>>> +     struct pci_epc_mem *mem;
>>> +     size_t actual_size;
>>> +     int i;
>>> +
>>> +     for (i = 0; i < epc->mem_windows; i++) {
>>> +             mem = epc->mem[i];
>>> +
>>> +             /* if chunk from this region is already used skip it */
>>> +             if (mem->window.map_size)
>>> +                     continue;
>>> +
>>> +             actual_size = ALIGN(size, mem->page_size);
>>> +
>>> +             if (best_fit_window == -1) {
>>> +                     best_fit_window = i;
>>> +                     window_least_size = mem->window.size;
>>> +             } else {
>>> +                     if (actual_size <= mem->window.size &&
>>> +                         mem->window.size < window_least_size) {
>>> +                             best_fit_window = i;
>>> +                             window_least_size = mem->window.size;
>>> +                     }
>>> +             }
>>> +     }
>>> +
>>> +     return best_fit_window;
>>> +}
>>> +
>>>   /**
>>>    * pci_epc_mem_alloc_addr() - allocate memory address from EPC addr space
>>>    * @epc: the EPC device on which memory has to be allocated
>>>    * @phys_addr: populate the allocated physical address here
>>> + * @window: populate the window here which will be used to map PCI address
>>>    * @size: the size of the address space that has to be allocated
>>>    *
>>>    * Invoke to allocate memory address from the EPC address space. This
>>>    * is usually done to map the remote RC address into the local system.
>>>    */
>>>   void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
>>> -                                  phys_addr_t *phys_addr, size_t size)
>>> +                                  phys_addr_t *phys_addr,
>>> +                                  int *window, size_t size)
>>>   {
>>> +     int best_fit = PCI_EPC_DEFAULT_WINDOW;
>>> +     void __iomem *virt_addr = NULL;
>>> +     struct pci_epc_mem *mem;
>>> +     unsigned int page_shift;
>>>        int pageno;
>>> -     void __iomem *virt_addr;
>>> -     struct pci_epc_mem *mem = epc->mem;
>>> -     unsigned int page_shift = ilog2(mem->page_size);
>>>        int order;
>>>
>>> +     if (epc->mem_windows <= 0)
>>> +             return NULL;
>>> +
>>> +     if (epc->mem_windows > 1) {
>>> +             best_fit = pci_epc_find_best_fit_window(epc, size);
>>> +             if (best_fit < 0)
>>> +                     return NULL;
>>> +     }
>>> +
>>> +     mem = epc->mem[best_fit];
>>>        size = ALIGN(size, mem->page_size);
>>> +     if (size > (mem->window.size - mem->window.map_size))
>>> +             return NULL;
>>
>> Assume I have two mem regions, first region is of size 128MB and the second
>> region is of size 4GB. If there is a allocation request for 4GB, will the
>> allocation succeed?
>>
> yes it would succeed, the pci_epc_find_best_fit_window() would get the
> corresponding window.

Can you step through the algo again? pci_epc_find_best_fit_window() will 
return the first region which is of size 128MB.

Thanks
Kishon

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 2/5] pci: endpoint: add support to handle multiple base for mapping outbound memory
@ 2019-12-05 10:22         ` Kishon Vijay Abraham I
  0 siblings, 0 replies; 60+ messages in thread
From: Kishon Vijay Abraham I @ 2019-12-05 10:22 UTC (permalink / raw)
  To: Lad, Prabhakar
  Cc: Mark Rutland, Heiko Stuebner, Geert Uytterhoeven, linux-pci,
	Shawn Lin, Will Deacon, Marek Vasut, Lorenzo Pieralisi,
	Magnus Damm, linux-rockchip, Catalin Marinas,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Chris Paterson, Arnd Bergmann, Lad, Prabhakar, Rob Herring,
	Bjorn Helgaas, LAK, Greg Kroah-Hartman, Yoshihiro Shimoda, LKML,
	Linux-Renesas, Tom Joseph, Jingoo Han, Andrew Murray,
	Gustavo Pimentel

Hi,

On 28/11/19 2:51 am, Lad, Prabhakar wrote:
> Hi Kishon,
> 
> Thank you for the review.
> 
> On Wed, Nov 27, 2019 at 5:15 AM Kishon Vijay Abraham I <kishon@ti.com> wrote:
>>
>> Hi Prabhakar,
>>
>> On 07/11/19 1:06 AM, Lad Prabhakar wrote:
>>> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
>>>
>>> rcar pcie controller has support to map multiple memory regions
>>> for mapping the outbound memory in local system, this feature
>>> inspires to add support for handling multiple memory bases in
>>> endpoint framework. In case of multiple memory regions only chunk
>>> or complete region can be mapped and this window needs to be
>>> passed to the controller driver.
>>>
>>> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
>>> Cc: <linux-rockchip@lists.infradead.org>
>>> Cc: Shawn Lin <shawn.lin@rock-chips.com>
>>> Cc: Heiko Stuebner <heiko@sntech.de>
>>> Cc: Tom Joseph <tjoseph@cadence.com>
>>> Cc: Jingoo Han <jingoohan1@gmail.com>
>>> Cc: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
>>> ---
>>>   .../pci/controller/dwc/pcie-designware-ep.c   |  30 ++-
>>>   drivers/pci/controller/pcie-cadence-ep.c      |  11 +-
>>>   drivers/pci/controller/pcie-rockchip-ep.c     |  13 +-
>>>   drivers/pci/endpoint/functions/pci-epf-test.c |  29 +--
>>>   drivers/pci/endpoint/pci-epc-core.c           |   7 +-
>>>   drivers/pci/endpoint/pci-epc-mem.c            | 189 ++++++++++++++----
>>>   include/linux/pci-epc.h                       |  43 ++--
>>>   7 files changed, 234 insertions(+), 88 deletions(-)
>>>
>>> diff --git a/drivers/pci/controller/dwc/pcie-designware-ep.c b/drivers/pci/controller/dwc/pcie-designware-ep.c
>>> index 3dd2e2697294..8d23c20b9afd 100644
>>> --- a/drivers/pci/controller/dwc/pcie-designware-ep.c
>>> +++ b/drivers/pci/controller/dwc/pcie-designware-ep.c
>>> @@ -195,7 +195,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, u8 func_no,
>>>   }
>>>
>>>   static int dw_pcie_ep_map_addr(struct pci_epc *epc, u8 func_no,
>>> -                            phys_addr_t addr,
>>> +                            phys_addr_t addr, int window,
>>>                               u64 pci_addr, size_t size)
>>>   {
>>>        int ret;
>>> @@ -367,6 +367,7 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
>>>        unsigned int aligned_offset;
>>>        u16 msg_ctrl, msg_data;
>>>        u32 msg_addr_lower, msg_addr_upper, reg;
>>> +     int window = PCI_EPC_DEFAULT_WINDOW;
>>>        u64 msg_addr;
>>>        bool has_upper;
>>>        int ret;
>>> @@ -390,11 +391,11 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
>>>                reg = ep->msi_cap + PCI_MSI_DATA_32;
>>>                msg_data = dw_pcie_readw_dbi(pci, reg);
>>>        }
>>> -     aligned_offset = msg_addr_lower & (epc->mem->page_size - 1);
>>> +     aligned_offset = msg_addr_lower & (epc->mem[window]->page_size - 1);
>>>        msg_addr = ((u64)msg_addr_upper) << 32 |
>>>                        (msg_addr_lower & ~aligned_offset);
>>> -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
>>> -                               epc->mem->page_size);
>>> +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
>>> +                               msg_addr, epc->mem[window]->page_size);
>>>        if (ret)
>>>                return ret;
>>>
>>> @@ -416,6 +417,7 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>>>        u32 reg, msg_data, vec_ctrl;
>>>        u64 tbl_addr, msg_addr, reg_u64;
>>>        void __iomem *msix_tbl;
>>> +     int window = PCI_EPC_DEFAULT_WINDOW;
>>>        int ret;
>>>
>>>        reg = ep->msix_cap + PCI_MSIX_TABLE;
>>> @@ -452,8 +454,8 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>>>                return -EPERM;
>>>        }
>>>
>>> -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
>>> -                               epc->mem->page_size);
>>> +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
>>> +                               msg_addr, epc->mem[window]->page_size);
>>>        if (ret)
>>>                return ret;
>>>
>>> @@ -466,10 +468,11 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
>>>
>>>   void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
>>>   {
>>> +     int window = PCI_EPC_DEFAULT_WINDOW;
>>>        struct pci_epc *epc = ep->epc;
>>>
>>>        pci_epc_mem_free_addr(epc, ep->msi_mem_phys, ep->msi_mem,
>>> -                           epc->mem->page_size);
>>> +                           epc->mem[window]->page_size);
>>>
>>>        pci_epc_mem_exit(epc);
>>>   }
>>> @@ -499,9 +502,12 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
>>>        u32 reg;
>>>        void *addr;
>>>        u8 hdr_type;
>>> +     int window;
>>>        unsigned int nbars;
>>>        unsigned int offset;
>>>        struct pci_epc *epc;
>>> +     size_t msi_page_size;
>>> +     struct pci_epc_mem_window mem_window;
>>>        struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
>>>        struct device *dev = pci->dev;
>>>        struct device_node *np = dev->of_node;
>>> @@ -574,15 +580,17 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
>>>        if (ret < 0)
>>>                epc->max_functions = 1;
>>>
>>> -     ret = __pci_epc_mem_init(epc, ep->phys_base, ep->addr_size,
>>> -                              ep->page_size);
>>> +     mem_window.phys_base = ep->phys_base;
>>> +     mem_window.size = ep->addr_size;
>>> +     ret = __pci_epc_mem_init(epc, &mem_window, 1, ep->page_size);
>>>        if (ret < 0) {
>>>                dev_err(dev, "Failed to initialize address space\n");
>>>                return ret;
>>>        }
>>>
>>> -     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys,
>>> -                                          epc->mem->page_size);
>>> +     msi_page_size = epc->mem[PCI_EPC_DEFAULT_WINDOW]->page_size;
>>> +     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys, &window,
>>> +                                          msi_page_size);
>>>        if (!ep->msi_mem) {
>>>                dev_err(dev, "Failed to reserve memory for MSI/MSI-X\n");
>>>                return -ENOMEM;
>>> diff --git a/drivers/pci/controller/pcie-cadence-ep.c b/drivers/pci/controller/pcie-cadence-ep.c
>>> index def7820cb824..7991b38a5350 100644
>>> --- a/drivers/pci/controller/pcie-cadence-ep.c
>>> +++ b/drivers/pci/controller/pcie-cadence-ep.c
>>> @@ -172,7 +172,7 @@ static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
>>>   }
>>>
>>>   static int cdns_pcie_ep_map_addr(struct pci_epc *epc, u8 fn, phys_addr_t addr,
>>> -                              u64 pci_addr, size_t size)
>>> +                              int window, u64 pci_addr, size_t size)
>>>   {
>>>        struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
>>>        struct cdns_pcie *pcie = &ep->pcie;
>>> @@ -434,12 +434,14 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
>>>   {
>>>        struct device *dev = &pdev->dev;
>>>        struct device_node *np = dev->of_node;
>>> +     struct pci_epc_mem_window mem_window;
>>>        struct cdns_pcie_ep *ep;
>>>        struct cdns_pcie *pcie;
>>>        struct pci_epc *epc;
>>>        struct resource *res;
>>>        int ret;
>>>        int phy_count;
>>> +     int window;
>>>
>>>        ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
>>>        if (!ep)
>>> @@ -502,15 +504,16 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
>>>        if (of_property_read_u8(np, "max-functions", &epc->max_functions) < 0)
>>>                epc->max_functions = 1;
>>>
>>> -     ret = pci_epc_mem_init(epc, pcie->mem_res->start,
>>> -                            resource_size(pcie->mem_res));
>>> +     mem_window.phys_base = pcie->mem_res->start;
>>> +     mem_window.size = resource_size(pcie->mem_res);
>>> +     ret = pci_epc_mem_init(epc, &mem_window, 1);
>>>        if (ret < 0) {
>>>                dev_err(dev, "failed to initialize the memory space\n");
>>>                goto err_init;
>>>        }
>>>
>>>        ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
>>> -                                               SZ_128K);
>>> +                                               &window, SZ_128K);
>>>        if (!ep->irq_cpu_addr) {
>>>                dev_err(dev, "failed to reserve memory space for MSI\n");
>>>                ret = -ENOMEM;
>>> diff --git a/drivers/pci/controller/pcie-rockchip-ep.c b/drivers/pci/controller/pcie-rockchip-ep.c
>>> index d743b0a48988..d59e85c8d319 100644
>>> --- a/drivers/pci/controller/pcie-rockchip-ep.c
>>> +++ b/drivers/pci/controller/pcie-rockchip-ep.c
>>> @@ -256,8 +256,8 @@ static void rockchip_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
>>>   }
>>>
>>>   static int rockchip_pcie_ep_map_addr(struct pci_epc *epc, u8 fn,
>>> -                                  phys_addr_t addr, u64 pci_addr,
>>> -                                  size_t size)
>>> +                                  phys_addr_t addr, int window,
>>> +                                  u64 pci_addr, size_t size)
>>>   {
>>>        struct rockchip_pcie_ep *ep = epc_get_drvdata(epc);
>>>        struct rockchip_pcie *pcie = &ep->rockchip;
>>> @@ -562,11 +562,13 @@ static const struct of_device_id rockchip_pcie_ep_of_match[] = {
>>>
>>>   static int rockchip_pcie_ep_probe(struct platform_device *pdev)
>>>   {
>>> +     struct pci_epc_mem_window mem_window;
>>>        struct device *dev = &pdev->dev;
>>>        struct rockchip_pcie_ep *ep;
>>>        struct rockchip_pcie *rockchip;
>>>        struct pci_epc *epc;
>>>        size_t max_regions;
>>> +     int window;
>>>        int err;
>>>
>>>        ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
>>> @@ -614,15 +616,16 @@ static int rockchip_pcie_ep_probe(struct platform_device *pdev)
>>>        /* Only enable function 0 by default */
>>>        rockchip_pcie_write(rockchip, BIT(0), PCIE_CORE_PHY_FUNC_CFG);
>>>
>>> -     err = pci_epc_mem_init(epc, rockchip->mem_res->start,
>>> -                            resource_size(rockchip->mem_res));
>>> +     mem_window.phys_base = rockchip->mem_res->start;
>>> +     mem_window.size = resource_size(rockchip->mem_res);
>>> +     err = pci_epc_mem_init(epc, &mem_window, 1);
>>>        if (err < 0) {
>>>                dev_err(dev, "failed to initialize the memory space\n");
>>>                goto err_uninit_port;
>>>        }
>>>
>>>        ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
>>> -                                               SZ_128K);
>>> +                                               &window, SZ_128K);
>>>        if (!ep->irq_cpu_addr) {
>>>                dev_err(dev, "failed to reserve memory space for MSI\n");
>>>                err = -ENOMEM;
>>> diff --git a/drivers/pci/endpoint/functions/pci-epf-test.c b/drivers/pci/endpoint/functions/pci-epf-test.c
>>> index 5d74f81ddfe4..475228011703 100644
>>> --- a/drivers/pci/endpoint/functions/pci-epf-test.c
>>> +++ b/drivers/pci/endpoint/functions/pci-epf-test.c
>>> @@ -84,8 +84,10 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>>>        struct pci_epc *epc = epf->epc;
>>>        enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>>>        struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
>>> +     int window;
>>>
>>> -     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr, reg->size);
>>> +     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr,
>>> +                                       &window, reg->size);
>>>        if (!src_addr) {
>>>                dev_err(dev, "Failed to allocate source address\n");
>>>                reg->status = STATUS_SRC_ADDR_INVALID;
>>> @@ -93,15 +95,16 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>>>                goto err;
>>>        }
>>>
>>> -     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, reg->src_addr,
>>> -                            reg->size);
>>> +     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, window,
>>> +                            reg->src_addr, reg->size);
>>>        if (ret) {
>>>                dev_err(dev, "Failed to map source address\n");
>>>                reg->status = STATUS_SRC_ADDR_INVALID;
>>>                goto err_src_addr;
>>>        }
>>>
>>> -     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr, reg->size);
>>> +     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr,
>>> +                                       &window, reg->size);
>>>        if (!dst_addr) {
>>>                dev_err(dev, "Failed to allocate destination address\n");
>>>                reg->status = STATUS_DST_ADDR_INVALID;
>>> @@ -109,8 +112,8 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
>>>                goto err_src_map_addr;
>>>        }
>>>
>>> -     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, reg->dst_addr,
>>> -                            reg->size);
>>> +     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, window,
>>> +                            reg->dst_addr, reg->size);
>>>        if (ret) {
>>>                dev_err(dev, "Failed to map destination address\n");
>>>                reg->status = STATUS_DST_ADDR_INVALID;
>>> @@ -146,8 +149,9 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
>>>        struct pci_epc *epc = epf->epc;
>>>        enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>>>        struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
>>> +     int window;
>>>
>>> -     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
>>> +     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
>>>        if (!src_addr) {
>>>                dev_err(dev, "Failed to allocate address\n");
>>>                reg->status = STATUS_SRC_ADDR_INVALID;
>>> @@ -155,8 +159,8 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
>>>                goto err;
>>>        }
>>>
>>> -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->src_addr,
>>> -                            reg->size);
>>> +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
>>> +                            reg->src_addr, reg->size);
>>>        if (ret) {
>>>                dev_err(dev, "Failed to map address\n");
>>>                reg->status = STATUS_SRC_ADDR_INVALID;
>>> @@ -193,13 +197,14 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
>>>        void __iomem *dst_addr;
>>>        void *buf;
>>>        phys_addr_t phys_addr;
>>> +     int window;
>>>        struct pci_epf *epf = epf_test->epf;
>>>        struct device *dev = &epf->dev;
>>>        struct pci_epc *epc = epf->epc;
>>>        enum pci_barno test_reg_bar = epf_test->test_reg_bar;
>>>        struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
>>>
>>> -     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
>>> +     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
>>>        if (!dst_addr) {
>>>                dev_err(dev, "Failed to allocate address\n");
>>>                reg->status = STATUS_DST_ADDR_INVALID;
>>> @@ -207,8 +212,8 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
>>>                goto err;
>>>        }
>>>
>>> -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->dst_addr,
>>> -                            reg->size);
>>> +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
>>> +                            reg->dst_addr, reg->size);
>>>        if (ret) {
>>>                dev_err(dev, "Failed to map address\n");
>>>                reg->status = STATUS_DST_ADDR_INVALID;
>>> diff --git a/drivers/pci/endpoint/pci-epc-core.c b/drivers/pci/endpoint/pci-epc-core.c
>>> index 2091508c1620..289c266c2d90 100644
>>> --- a/drivers/pci/endpoint/pci-epc-core.c
>>> +++ b/drivers/pci/endpoint/pci-epc-core.c
>>> @@ -358,13 +358,15 @@ EXPORT_SYMBOL_GPL(pci_epc_unmap_addr);
>>>    * @epc: the EPC device on which address is allocated
>>>    * @func_no: the endpoint function number in the EPC device
>>>    * @phys_addr: physical address of the local system
>>> + * @window: index to the window region where PCI address will be mapped
>>>    * @pci_addr: PCI address to which the physical address should be mapped
>>>    * @size: the size of the allocation
>>>    *
>>>    * Invoke to map CPU address with PCI address.
>>>    */
>>>   int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
>>> -                  phys_addr_t phys_addr, u64 pci_addr, size_t size)
>>> +                  phys_addr_t phys_addr, int window,
>>> +                  u64 pci_addr, size_t size)
>>>   {
>>>        int ret;
>>>        unsigned long flags;
>>> @@ -376,7 +378,8 @@ int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
>>>                return 0;
>>>
>>>        spin_lock_irqsave(&epc->lock, flags);
>>> -     ret = epc->ops->map_addr(epc, func_no, phys_addr, pci_addr, size);
>>> +     ret = epc->ops->map_addr(epc, func_no, phys_addr,
>>> +                              window, pci_addr, size);
>>>        spin_unlock_irqrestore(&epc->lock, flags);
>>>
>>>        return ret;
>>> diff --git a/drivers/pci/endpoint/pci-epc-mem.c b/drivers/pci/endpoint/pci-epc-mem.c
>>> index d2b174ce15de..c955f2c97944 100644
>>> --- a/drivers/pci/endpoint/pci-epc-mem.c
>>> +++ b/drivers/pci/endpoint/pci-epc-mem.c
>>> @@ -39,56 +39,77 @@ static int pci_epc_mem_get_order(struct pci_epc_mem *mem, size_t size)
>>>    * __pci_epc_mem_init() - initialize the pci_epc_mem structure
>>>    * @epc: the EPC device that invoked pci_epc_mem_init
>>>    * @phys_base: the physical address of the base
>>> - * @size: the size of the address space
>>> + * @num_windows: number of windows device supports
>>
>> struct pci_epc_mem_window is missing here.
> oops my bad will fix that.
> 
>>>    * @page_size: size of each page
>>>    *
>>>    * Invoke to initialize the pci_epc_mem structure used by the
>>>    * endpoint functions to allocate mapped PCI address.
>>>    */
>>> -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_base, size_t size,
>>> -                    size_t page_size)
>>> +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *windows,
>>> +                    int num_windows, size_t page_size)
>>>   {
>>> -     int ret;
>>> -     struct pci_epc_mem *mem;
>>> -     unsigned long *bitmap;
>>> +     struct pci_epc_mem *mem = NULL;
>>> +     unsigned long *bitmap = NULL;
>>>        unsigned int page_shift;
>>> -     int pages;
>>>        int bitmap_size;
>>> +     int pages;
>>> +     int ret;
>>> +     int i;
>>> +
>>> +     epc->mem_windows = 0;
>>> +
>>> +     if (!windows)
>>> +             return -EINVAL;
>>> +
>>> +     if (num_windows <= 0)
>>> +             return -EINVAL;
>>>
>>>        if (page_size < PAGE_SIZE)
>>>                page_size = PAGE_SIZE;
>>>
>>>        page_shift = ilog2(page_size);
>>> -     pages = size >> page_shift;
>>> -     bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
>>>
>>> -     mem = kzalloc(sizeof(*mem), GFP_KERNEL);
>>> -     if (!mem) {
>>> -             ret = -ENOMEM;
>>> -             goto err;
>>> -     }
>>> +     epc->mem = kcalloc(num_windows, sizeof(*mem), GFP_KERNEL);
>>> +     if (!epc->mem)
>>> +             return -EINVAL;
>>>
>>> -     bitmap = kzalloc(bitmap_size, GFP_KERNEL);
>>> -     if (!bitmap) {
>>> -             ret = -ENOMEM;
>>> -             goto err_mem;
>>> -     }
>>> +     for (i = 0; i < num_windows; i++) {
>>> +             pages = windows[i].phys_base >> page_shift;
>>> +             bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
>>>
>>> -     mem->bitmap = bitmap;
>>> -     mem->phys_base = phys_base;
>>> -     mem->page_size = page_size;
>>> -     mem->pages = pages;
>>> -     mem->size = size;
>>> +             mem = kzalloc(sizeof(*mem), GFP_KERNEL);
>>> +             if (!mem) {
>>> +                     ret = -ENOMEM;
>>> +                     goto err_mem;
>>> +             }
>>>
>>> -     epc->mem = mem;
>>> +             bitmap = kzalloc(bitmap_size, GFP_KERNEL);
>>> +             if (!bitmap) {
>>> +                     ret = -ENOMEM;
>>> +                     goto err_mem;
>>> +             }
>>> +
>>> +             mem->bitmap = bitmap;
>>> +             mem->window.phys_base = windows[i].phys_base;
>>> +             mem->page_size = page_size;
>>> +             mem->pages = pages;
>>> +             mem->window.size = windows[i].size;
>>> +             mem->window.map_size = 0;
>>> +
>>> +             epc->mem[i] = mem;
>>> +     }
>>> +     epc->mem_windows = num_windows;
>>>
>>>        return 0;
>>>
>>>   err_mem:
>>> -     kfree(mem);
>>> +     for (; i >= 0; i--) {
>>> +             kfree(mem->bitmap);
>>> +             kfree(epc->mem[i]);
>>> +     }
>>> +     kfree(epc->mem);
>>>
>>> -err:
>>> -return ret;
>>> +     return ret;
>>>   }
>>>   EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
>>>
>>> @@ -101,48 +122,126 @@ EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
>>>    */
>>>   void pci_epc_mem_exit(struct pci_epc *epc)
>>>   {
>>> -     struct pci_epc_mem *mem = epc->mem;
>>> +     struct pci_epc_mem *mem;
>>> +     int i;
>>> +
>>> +     if (!epc->mem_windows)
>>> +             return;
>>> +
>>> +     for (i = 0; i <= epc->mem_windows; i--) {
>>> +             mem = epc->mem[i];
>>> +             kfree(mem->bitmap);
>>> +             kfree(epc->mem[i]);
>>> +     }
>>> +     kfree(epc->mem);
>>>
>>>        epc->mem = NULL;
>>> -     kfree(mem->bitmap);
>>> -     kfree(mem);
>>> +     epc->mem_windows = 0;
>>>   }
>>>   EXPORT_SYMBOL_GPL(pci_epc_mem_exit);
>>>
>>> +static int pci_epc_find_best_fit_window(struct pci_epc *epc, size_t size)
>>> +{
>>> +     size_t window_least_size = 0;
>>> +     int best_fit_window = -1;
>>> +     struct pci_epc_mem *mem;
>>> +     size_t actual_size;
>>> +     int i;
>>> +
>>> +     for (i = 0; i < epc->mem_windows; i++) {
>>> +             mem = epc->mem[i];
>>> +
>>> +             /* if chunk from this region is already used skip it */
>>> +             if (mem->window.map_size)
>>> +                     continue;
>>> +
>>> +             actual_size = ALIGN(size, mem->page_size);
>>> +
>>> +             if (best_fit_window == -1) {
>>> +                     best_fit_window = i;
>>> +                     window_least_size = mem->window.size;
>>> +             } else {
>>> +                     if (actual_size <= mem->window.size &&
>>> +                         mem->window.size < window_least_size) {
>>> +                             best_fit_window = i;
>>> +                             window_least_size = mem->window.size;
>>> +                     }
>>> +             }
>>> +     }
>>> +
>>> +     return best_fit_window;
>>> +}
>>> +
>>>   /**
>>>    * pci_epc_mem_alloc_addr() - allocate memory address from EPC addr space
>>>    * @epc: the EPC device on which memory has to be allocated
>>>    * @phys_addr: populate the allocated physical address here
>>> + * @window: populate the window here which will be used to map PCI address
>>>    * @size: the size of the address space that has to be allocated
>>>    *
>>>    * Invoke to allocate memory address from the EPC address space. This
>>>    * is usually done to map the remote RC address into the local system.
>>>    */
>>>   void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
>>> -                                  phys_addr_t *phys_addr, size_t size)
>>> +                                  phys_addr_t *phys_addr,
>>> +                                  int *window, size_t size)
>>>   {
>>> +     int best_fit = PCI_EPC_DEFAULT_WINDOW;
>>> +     void __iomem *virt_addr = NULL;
>>> +     struct pci_epc_mem *mem;
>>> +     unsigned int page_shift;
>>>        int pageno;
>>> -     void __iomem *virt_addr;
>>> -     struct pci_epc_mem *mem = epc->mem;
>>> -     unsigned int page_shift = ilog2(mem->page_size);
>>>        int order;
>>>
>>> +     if (epc->mem_windows <= 0)
>>> +             return NULL;
>>> +
>>> +     if (epc->mem_windows > 1) {
>>> +             best_fit = pci_epc_find_best_fit_window(epc, size);
>>> +             if (best_fit < 0)
>>> +                     return NULL;
>>> +     }
>>> +
>>> +     mem = epc->mem[best_fit];
>>>        size = ALIGN(size, mem->page_size);
>>> +     if (size > (mem->window.size - mem->window.map_size))
>>> +             return NULL;
>>
>> Assume I have two mem regions, first region is of size 128MB and the second
>> region is of size 4GB. If there is a allocation request for 4GB, will the
>> allocation succeed?
>>
> yes it would succeed, the pci_epc_find_best_fit_window() would get the
> corresponding window.

Can you step through the algo again? pci_epc_find_best_fit_window() will 
return the first region which is of size 128MB.

Thanks
Kishon

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 2/5] pci: endpoint: add support to handle multiple base for mapping outbound memory
  2019-12-05 10:22         ` Kishon Vijay Abraham I
  (?)
@ 2019-12-06 13:08           ` Lad, Prabhakar
  -1 siblings, 0 replies; 60+ messages in thread
From: Lad, Prabhakar @ 2019-12-06 13:08 UTC (permalink / raw)
  To: Kishon Vijay Abraham I
  Cc: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Marek Vasut, Yoshihiro Shimoda, linux-pci,
	Catalin Marinas, Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS, LKML,
	LAK, Linux-Renesas, Chris Paterson, Lad, Prabhakar,
	linux-rockchip, Shawn Lin, Heiko Stuebner, Tom Joseph,
	Jingoo Han, Gustavo Pimentel

Hi Kishon,

On Thu, Dec 5, 2019 at 10:22 AM Kishon Vijay Abraham I <kishon@ti.com> wrote:
>
> Hi,
>
> On 28/11/19 2:51 am, Lad, Prabhakar wrote:
> > Hi Kishon,
> >
> > Thank you for the review.
> >
> > On Wed, Nov 27, 2019 at 5:15 AM Kishon Vijay Abraham I <kishon@ti.com> wrote:
> >>
> >> Hi Prabhakar,
> >>
> >> On 07/11/19 1:06 AM, Lad Prabhakar wrote:
> >>> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >>>
> >>> rcar pcie controller has support to map multiple memory regions
> >>> for mapping the outbound memory in local system, this feature
> >>> inspires to add support for handling multiple memory bases in
> >>> endpoint framework. In case of multiple memory regions only chunk
> >>> or complete region can be mapped and this window needs to be
> >>> passed to the controller driver.
> >>>
> >>> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >>> Cc: <linux-rockchip@lists.infradead.org>
> >>> Cc: Shawn Lin <shawn.lin@rock-chips.com>
> >>> Cc: Heiko Stuebner <heiko@sntech.de>
> >>> Cc: Tom Joseph <tjoseph@cadence.com>
> >>> Cc: Jingoo Han <jingoohan1@gmail.com>
> >>> Cc: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
> >>> ---
> >>>   .../pci/controller/dwc/pcie-designware-ep.c   |  30 ++-
> >>>   drivers/pci/controller/pcie-cadence-ep.c      |  11 +-
> >>>   drivers/pci/controller/pcie-rockchip-ep.c     |  13 +-
> >>>   drivers/pci/endpoint/functions/pci-epf-test.c |  29 +--
> >>>   drivers/pci/endpoint/pci-epc-core.c           |   7 +-
> >>>   drivers/pci/endpoint/pci-epc-mem.c            | 189 ++++++++++++++----
> >>>   include/linux/pci-epc.h                       |  43 ++--
> >>>   7 files changed, 234 insertions(+), 88 deletions(-)
> >>>
> >>> diff --git a/drivers/pci/controller/dwc/pcie-designware-ep.c b/drivers/pci/controller/dwc/pcie-designware-ep.c
> >>> index 3dd2e2697294..8d23c20b9afd 100644
> >>> --- a/drivers/pci/controller/dwc/pcie-designware-ep.c
> >>> +++ b/drivers/pci/controller/dwc/pcie-designware-ep.c
> >>> @@ -195,7 +195,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, u8 func_no,
> >>>   }
> >>>
> >>>   static int dw_pcie_ep_map_addr(struct pci_epc *epc, u8 func_no,
> >>> -                            phys_addr_t addr,
> >>> +                            phys_addr_t addr, int window,
> >>>                               u64 pci_addr, size_t size)
> >>>   {
> >>>        int ret;
> >>> @@ -367,6 +367,7 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
> >>>        unsigned int aligned_offset;
> >>>        u16 msg_ctrl, msg_data;
> >>>        u32 msg_addr_lower, msg_addr_upper, reg;
> >>> +     int window = PCI_EPC_DEFAULT_WINDOW;
> >>>        u64 msg_addr;
> >>>        bool has_upper;
> >>>        int ret;
> >>> @@ -390,11 +391,11 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
> >>>                reg = ep->msi_cap + PCI_MSI_DATA_32;
> >>>                msg_data = dw_pcie_readw_dbi(pci, reg);
> >>>        }
> >>> -     aligned_offset = msg_addr_lower & (epc->mem->page_size - 1);
> >>> +     aligned_offset = msg_addr_lower & (epc->mem[window]->page_size - 1);
> >>>        msg_addr = ((u64)msg_addr_upper) << 32 |
> >>>                        (msg_addr_lower & ~aligned_offset);
> >>> -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> >>> -                               epc->mem->page_size);
> >>> +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> >>> +                               msg_addr, epc->mem[window]->page_size);
> >>>        if (ret)
> >>>                return ret;
> >>>
> >>> @@ -416,6 +417,7 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >>>        u32 reg, msg_data, vec_ctrl;
> >>>        u64 tbl_addr, msg_addr, reg_u64;
> >>>        void __iomem *msix_tbl;
> >>> +     int window = PCI_EPC_DEFAULT_WINDOW;
> >>>        int ret;
> >>>
> >>>        reg = ep->msix_cap + PCI_MSIX_TABLE;
> >>> @@ -452,8 +454,8 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >>>                return -EPERM;
> >>>        }
> >>>
> >>> -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> >>> -                               epc->mem->page_size);
> >>> +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> >>> +                               msg_addr, epc->mem[window]->page_size);
> >>>        if (ret)
> >>>                return ret;
> >>>
> >>> @@ -466,10 +468,11 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >>>
> >>>   void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
> >>>   {
> >>> +     int window = PCI_EPC_DEFAULT_WINDOW;
> >>>        struct pci_epc *epc = ep->epc;
> >>>
> >>>        pci_epc_mem_free_addr(epc, ep->msi_mem_phys, ep->msi_mem,
> >>> -                           epc->mem->page_size);
> >>> +                           epc->mem[window]->page_size);
> >>>
> >>>        pci_epc_mem_exit(epc);
> >>>   }
> >>> @@ -499,9 +502,12 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
> >>>        u32 reg;
> >>>        void *addr;
> >>>        u8 hdr_type;
> >>> +     int window;
> >>>        unsigned int nbars;
> >>>        unsigned int offset;
> >>>        struct pci_epc *epc;
> >>> +     size_t msi_page_size;
> >>> +     struct pci_epc_mem_window mem_window;
> >>>        struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
> >>>        struct device *dev = pci->dev;
> >>>        struct device_node *np = dev->of_node;
> >>> @@ -574,15 +580,17 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
> >>>        if (ret < 0)
> >>>                epc->max_functions = 1;
> >>>
> >>> -     ret = __pci_epc_mem_init(epc, ep->phys_base, ep->addr_size,
> >>> -                              ep->page_size);
> >>> +     mem_window.phys_base = ep->phys_base;
> >>> +     mem_window.size = ep->addr_size;
> >>> +     ret = __pci_epc_mem_init(epc, &mem_window, 1, ep->page_size);
> >>>        if (ret < 0) {
> >>>                dev_err(dev, "Failed to initialize address space\n");
> >>>                return ret;
> >>>        }
> >>>
> >>> -     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys,
> >>> -                                          epc->mem->page_size);
> >>> +     msi_page_size = epc->mem[PCI_EPC_DEFAULT_WINDOW]->page_size;
> >>> +     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys, &window,
> >>> +                                          msi_page_size);
> >>>        if (!ep->msi_mem) {
> >>>                dev_err(dev, "Failed to reserve memory for MSI/MSI-X\n");
> >>>                return -ENOMEM;
> >>> diff --git a/drivers/pci/controller/pcie-cadence-ep.c b/drivers/pci/controller/pcie-cadence-ep.c
> >>> index def7820cb824..7991b38a5350 100644
> >>> --- a/drivers/pci/controller/pcie-cadence-ep.c
> >>> +++ b/drivers/pci/controller/pcie-cadence-ep.c
> >>> @@ -172,7 +172,7 @@ static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
> >>>   }
> >>>
> >>>   static int cdns_pcie_ep_map_addr(struct pci_epc *epc, u8 fn, phys_addr_t addr,
> >>> -                              u64 pci_addr, size_t size)
> >>> +                              int window, u64 pci_addr, size_t size)
> >>>   {
> >>>        struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
> >>>        struct cdns_pcie *pcie = &ep->pcie;
> >>> @@ -434,12 +434,14 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
> >>>   {
> >>>        struct device *dev = &pdev->dev;
> >>>        struct device_node *np = dev->of_node;
> >>> +     struct pci_epc_mem_window mem_window;
> >>>        struct cdns_pcie_ep *ep;
> >>>        struct cdns_pcie *pcie;
> >>>        struct pci_epc *epc;
> >>>        struct resource *res;
> >>>        int ret;
> >>>        int phy_count;
> >>> +     int window;
> >>>
> >>>        ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> >>>        if (!ep)
> >>> @@ -502,15 +504,16 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
> >>>        if (of_property_read_u8(np, "max-functions", &epc->max_functions) < 0)
> >>>                epc->max_functions = 1;
> >>>
> >>> -     ret = pci_epc_mem_init(epc, pcie->mem_res->start,
> >>> -                            resource_size(pcie->mem_res));
> >>> +     mem_window.phys_base = pcie->mem_res->start;
> >>> +     mem_window.size = resource_size(pcie->mem_res);
> >>> +     ret = pci_epc_mem_init(epc, &mem_window, 1);
> >>>        if (ret < 0) {
> >>>                dev_err(dev, "failed to initialize the memory space\n");
> >>>                goto err_init;
> >>>        }
> >>>
> >>>        ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> >>> -                                               SZ_128K);
> >>> +                                               &window, SZ_128K);
> >>>        if (!ep->irq_cpu_addr) {
> >>>                dev_err(dev, "failed to reserve memory space for MSI\n");
> >>>                ret = -ENOMEM;
> >>> diff --git a/drivers/pci/controller/pcie-rockchip-ep.c b/drivers/pci/controller/pcie-rockchip-ep.c
> >>> index d743b0a48988..d59e85c8d319 100644
> >>> --- a/drivers/pci/controller/pcie-rockchip-ep.c
> >>> +++ b/drivers/pci/controller/pcie-rockchip-ep.c
> >>> @@ -256,8 +256,8 @@ static void rockchip_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
> >>>   }
> >>>
> >>>   static int rockchip_pcie_ep_map_addr(struct pci_epc *epc, u8 fn,
> >>> -                                  phys_addr_t addr, u64 pci_addr,
> >>> -                                  size_t size)
> >>> +                                  phys_addr_t addr, int window,
> >>> +                                  u64 pci_addr, size_t size)
> >>>   {
> >>>        struct rockchip_pcie_ep *ep = epc_get_drvdata(epc);
> >>>        struct rockchip_pcie *pcie = &ep->rockchip;
> >>> @@ -562,11 +562,13 @@ static const struct of_device_id rockchip_pcie_ep_of_match[] = {
> >>>
> >>>   static int rockchip_pcie_ep_probe(struct platform_device *pdev)
> >>>   {
> >>> +     struct pci_epc_mem_window mem_window;
> >>>        struct device *dev = &pdev->dev;
> >>>        struct rockchip_pcie_ep *ep;
> >>>        struct rockchip_pcie *rockchip;
> >>>        struct pci_epc *epc;
> >>>        size_t max_regions;
> >>> +     int window;
> >>>        int err;
> >>>
> >>>        ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> >>> @@ -614,15 +616,16 @@ static int rockchip_pcie_ep_probe(struct platform_device *pdev)
> >>>        /* Only enable function 0 by default */
> >>>        rockchip_pcie_write(rockchip, BIT(0), PCIE_CORE_PHY_FUNC_CFG);
> >>>
> >>> -     err = pci_epc_mem_init(epc, rockchip->mem_res->start,
> >>> -                            resource_size(rockchip->mem_res));
> >>> +     mem_window.phys_base = rockchip->mem_res->start;
> >>> +     mem_window.size = resource_size(rockchip->mem_res);
> >>> +     err = pci_epc_mem_init(epc, &mem_window, 1);
> >>>        if (err < 0) {
> >>>                dev_err(dev, "failed to initialize the memory space\n");
> >>>                goto err_uninit_port;
> >>>        }
> >>>
> >>>        ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> >>> -                                               SZ_128K);
> >>> +                                               &window, SZ_128K);
> >>>        if (!ep->irq_cpu_addr) {
> >>>                dev_err(dev, "failed to reserve memory space for MSI\n");
> >>>                err = -ENOMEM;
> >>> diff --git a/drivers/pci/endpoint/functions/pci-epf-test.c b/drivers/pci/endpoint/functions/pci-epf-test.c
> >>> index 5d74f81ddfe4..475228011703 100644
> >>> --- a/drivers/pci/endpoint/functions/pci-epf-test.c
> >>> +++ b/drivers/pci/endpoint/functions/pci-epf-test.c
> >>> @@ -84,8 +84,10 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >>>        struct pci_epc *epc = epf->epc;
> >>>        enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >>>        struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> >>> +     int window;
> >>>
> >>> -     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr, reg->size);
> >>> +     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr,
> >>> +                                       &window, reg->size);
> >>>        if (!src_addr) {
> >>>                dev_err(dev, "Failed to allocate source address\n");
> >>>                reg->status = STATUS_SRC_ADDR_INVALID;
> >>> @@ -93,15 +95,16 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >>>                goto err;
> >>>        }
> >>>
> >>> -     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, reg->src_addr,
> >>> -                            reg->size);
> >>> +     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, window,
> >>> +                            reg->src_addr, reg->size);
> >>>        if (ret) {
> >>>                dev_err(dev, "Failed to map source address\n");
> >>>                reg->status = STATUS_SRC_ADDR_INVALID;
> >>>                goto err_src_addr;
> >>>        }
> >>>
> >>> -     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr, reg->size);
> >>> +     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr,
> >>> +                                       &window, reg->size);
> >>>        if (!dst_addr) {
> >>>                dev_err(dev, "Failed to allocate destination address\n");
> >>>                reg->status = STATUS_DST_ADDR_INVALID;
> >>> @@ -109,8 +112,8 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >>>                goto err_src_map_addr;
> >>>        }
> >>>
> >>> -     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, reg->dst_addr,
> >>> -                            reg->size);
> >>> +     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, window,
> >>> +                            reg->dst_addr, reg->size);
> >>>        if (ret) {
> >>>                dev_err(dev, "Failed to map destination address\n");
> >>>                reg->status = STATUS_DST_ADDR_INVALID;
> >>> @@ -146,8 +149,9 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
> >>>        struct pci_epc *epc = epf->epc;
> >>>        enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >>>        struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> >>> +     int window;
> >>>
> >>> -     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> >>> +     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
> >>>        if (!src_addr) {
> >>>                dev_err(dev, "Failed to allocate address\n");
> >>>                reg->status = STATUS_SRC_ADDR_INVALID;
> >>> @@ -155,8 +159,8 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
> >>>                goto err;
> >>>        }
> >>>
> >>> -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->src_addr,
> >>> -                            reg->size);
> >>> +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> >>> +                            reg->src_addr, reg->size);
> >>>        if (ret) {
> >>>                dev_err(dev, "Failed to map address\n");
> >>>                reg->status = STATUS_SRC_ADDR_INVALID;
> >>> @@ -193,13 +197,14 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
> >>>        void __iomem *dst_addr;
> >>>        void *buf;
> >>>        phys_addr_t phys_addr;
> >>> +     int window;
> >>>        struct pci_epf *epf = epf_test->epf;
> >>>        struct device *dev = &epf->dev;
> >>>        struct pci_epc *epc = epf->epc;
> >>>        enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >>>        struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> >>>
> >>> -     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> >>> +     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
> >>>        if (!dst_addr) {
> >>>                dev_err(dev, "Failed to allocate address\n");
> >>>                reg->status = STATUS_DST_ADDR_INVALID;
> >>> @@ -207,8 +212,8 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
> >>>                goto err;
> >>>        }
> >>>
> >>> -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->dst_addr,
> >>> -                            reg->size);
> >>> +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> >>> +                            reg->dst_addr, reg->size);
> >>>        if (ret) {
> >>>                dev_err(dev, "Failed to map address\n");
> >>>                reg->status = STATUS_DST_ADDR_INVALID;
> >>> diff --git a/drivers/pci/endpoint/pci-epc-core.c b/drivers/pci/endpoint/pci-epc-core.c
> >>> index 2091508c1620..289c266c2d90 100644
> >>> --- a/drivers/pci/endpoint/pci-epc-core.c
> >>> +++ b/drivers/pci/endpoint/pci-epc-core.c
> >>> @@ -358,13 +358,15 @@ EXPORT_SYMBOL_GPL(pci_epc_unmap_addr);
> >>>    * @epc: the EPC device on which address is allocated
> >>>    * @func_no: the endpoint function number in the EPC device
> >>>    * @phys_addr: physical address of the local system
> >>> + * @window: index to the window region where PCI address will be mapped
> >>>    * @pci_addr: PCI address to which the physical address should be mapped
> >>>    * @size: the size of the allocation
> >>>    *
> >>>    * Invoke to map CPU address with PCI address.
> >>>    */
> >>>   int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> >>> -                  phys_addr_t phys_addr, u64 pci_addr, size_t size)
> >>> +                  phys_addr_t phys_addr, int window,
> >>> +                  u64 pci_addr, size_t size)
> >>>   {
> >>>        int ret;
> >>>        unsigned long flags;
> >>> @@ -376,7 +378,8 @@ int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> >>>                return 0;
> >>>
> >>>        spin_lock_irqsave(&epc->lock, flags);
> >>> -     ret = epc->ops->map_addr(epc, func_no, phys_addr, pci_addr, size);
> >>> +     ret = epc->ops->map_addr(epc, func_no, phys_addr,
> >>> +                              window, pci_addr, size);
> >>>        spin_unlock_irqrestore(&epc->lock, flags);
> >>>
> >>>        return ret;
> >>> diff --git a/drivers/pci/endpoint/pci-epc-mem.c b/drivers/pci/endpoint/pci-epc-mem.c
> >>> index d2b174ce15de..c955f2c97944 100644
> >>> --- a/drivers/pci/endpoint/pci-epc-mem.c
> >>> +++ b/drivers/pci/endpoint/pci-epc-mem.c
> >>> @@ -39,56 +39,77 @@ static int pci_epc_mem_get_order(struct pci_epc_mem *mem, size_t size)
> >>>    * __pci_epc_mem_init() - initialize the pci_epc_mem structure
> >>>    * @epc: the EPC device that invoked pci_epc_mem_init
> >>>    * @phys_base: the physical address of the base
> >>> - * @size: the size of the address space
> >>> + * @num_windows: number of windows device supports
> >>
> >> struct pci_epc_mem_window is missing here.
> > oops my bad will fix that.
> >
> >>>    * @page_size: size of each page
> >>>    *
> >>>    * Invoke to initialize the pci_epc_mem structure used by the
> >>>    * endpoint functions to allocate mapped PCI address.
> >>>    */
> >>> -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_base, size_t size,
> >>> -                    size_t page_size)
> >>> +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *windows,
> >>> +                    int num_windows, size_t page_size)
> >>>   {
> >>> -     int ret;
> >>> -     struct pci_epc_mem *mem;
> >>> -     unsigned long *bitmap;
> >>> +     struct pci_epc_mem *mem = NULL;
> >>> +     unsigned long *bitmap = NULL;
> >>>        unsigned int page_shift;
> >>> -     int pages;
> >>>        int bitmap_size;
> >>> +     int pages;
> >>> +     int ret;
> >>> +     int i;
> >>> +
> >>> +     epc->mem_windows = 0;
> >>> +
> >>> +     if (!windows)
> >>> +             return -EINVAL;
> >>> +
> >>> +     if (num_windows <= 0)
> >>> +             return -EINVAL;
> >>>
> >>>        if (page_size < PAGE_SIZE)
> >>>                page_size = PAGE_SIZE;
> >>>
> >>>        page_shift = ilog2(page_size);
> >>> -     pages = size >> page_shift;
> >>> -     bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
> >>>
> >>> -     mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> >>> -     if (!mem) {
> >>> -             ret = -ENOMEM;
> >>> -             goto err;
> >>> -     }
> >>> +     epc->mem = kcalloc(num_windows, sizeof(*mem), GFP_KERNEL);
> >>> +     if (!epc->mem)
> >>> +             return -EINVAL;
> >>>
> >>> -     bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> >>> -     if (!bitmap) {
> >>> -             ret = -ENOMEM;
> >>> -             goto err_mem;
> >>> -     }
> >>> +     for (i = 0; i < num_windows; i++) {
> >>> +             pages = windows[i].phys_base >> page_shift;
> >>> +             bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
> >>>
> >>> -     mem->bitmap = bitmap;
> >>> -     mem->phys_base = phys_base;
> >>> -     mem->page_size = page_size;
> >>> -     mem->pages = pages;
> >>> -     mem->size = size;
> >>> +             mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> >>> +             if (!mem) {
> >>> +                     ret = -ENOMEM;
> >>> +                     goto err_mem;
> >>> +             }
> >>>
> >>> -     epc->mem = mem;
> >>> +             bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> >>> +             if (!bitmap) {
> >>> +                     ret = -ENOMEM;
> >>> +                     goto err_mem;
> >>> +             }
> >>> +
> >>> +             mem->bitmap = bitmap;
> >>> +             mem->window.phys_base = windows[i].phys_base;
> >>> +             mem->page_size = page_size;
> >>> +             mem->pages = pages;
> >>> +             mem->window.size = windows[i].size;
> >>> +             mem->window.map_size = 0;
> >>> +
> >>> +             epc->mem[i] = mem;
> >>> +     }
> >>> +     epc->mem_windows = num_windows;
> >>>
> >>>        return 0;
> >>>
> >>>   err_mem:
> >>> -     kfree(mem);
> >>> +     for (; i >= 0; i--) {
> >>> +             kfree(mem->bitmap);
> >>> +             kfree(epc->mem[i]);
> >>> +     }
> >>> +     kfree(epc->mem);
> >>>
> >>> -err:
> >>> -return ret;
> >>> +     return ret;
> >>>   }
> >>>   EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
> >>>
> >>> @@ -101,48 +122,126 @@ EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
> >>>    */
> >>>   void pci_epc_mem_exit(struct pci_epc *epc)
> >>>   {
> >>> -     struct pci_epc_mem *mem = epc->mem;
> >>> +     struct pci_epc_mem *mem;
> >>> +     int i;
> >>> +
> >>> +     if (!epc->mem_windows)
> >>> +             return;
> >>> +
> >>> +     for (i = 0; i <= epc->mem_windows; i--) {
> >>> +             mem = epc->mem[i];
> >>> +             kfree(mem->bitmap);
> >>> +             kfree(epc->mem[i]);
> >>> +     }
> >>> +     kfree(epc->mem);
> >>>
> >>>        epc->mem = NULL;
> >>> -     kfree(mem->bitmap);
> >>> -     kfree(mem);
> >>> +     epc->mem_windows = 0;
> >>>   }
> >>>   EXPORT_SYMBOL_GPL(pci_epc_mem_exit);
> >>>
> >>> +static int pci_epc_find_best_fit_window(struct pci_epc *epc, size_t size)
> >>> +{
> >>> +     size_t window_least_size = 0;
> >>> +     int best_fit_window = -1;
> >>> +     struct pci_epc_mem *mem;
> >>> +     size_t actual_size;
> >>> +     int i;
> >>> +
> >>> +     for (i = 0; i < epc->mem_windows; i++) {
^^
here we loop through all the possible windows.

> >>> +             mem = epc->mem[i];
> >>> +
> >>> +             /* if chunk from this region is already used skip it */
> >>> +             if (mem->window.map_size)
> >>> +                     continue;
> >>> +
^^
If a chunk is used in the window we skip it.
but the next version of the patchest  will check for  the flag if the
controller supports multiple allocation of chunks
in same window.

> >>> +             actual_size = ALIGN(size, mem->page_size);
> >>> +
> >>> +             if (best_fit_window == -1) {
> >>> +                     best_fit_window = i;
> >>> +                     window_least_size = mem->window.size;
firstly it assigns a first free region, although the above code doesnt
check if requested size can be fitted in the window,
I'll fix that in next version as below :

if (actual_size <= mem->window.size) {
                    best_fit_window = i;
                    window_least_size = mem->window.size;
}

window_least_size  is temp variable to compare against other window sizes.

> >>> +             } else {
> >>> +                     if (actual_size <= mem->window.size &&
> >>> +                         mem->window.size < window_least_size) {
> >>> +                             best_fit_window = i;
> >>> +                             window_least_size = mem->window.size;
> >>> +                     }

add in above we check if the requested size does fit in the window,
and the windows size is lesser then
the previously assigned one, if the criteria matches we choose this
window. and this is looped through all
the windows.

> >>> +             }
> >>> +     }
> >>> +
> >>> +     return best_fit_window;
> >>> +}
> >>> +
> >>>   /**
> >>>    * pci_epc_mem_alloc_addr() - allocate memory address from EPC addr space
> >>>    * @epc: the EPC device on which memory has to be allocated
> >>>    * @phys_addr: populate the allocated physical address here
> >>> + * @window: populate the window here which will be used to map PCI address
> >>>    * @size: the size of the address space that has to be allocated
> >>>    *
> >>>    * Invoke to allocate memory address from the EPC address space. This
> >>>    * is usually done to map the remote RC address into the local system.
> >>>    */
> >>>   void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
> >>> -                                  phys_addr_t *phys_addr, size_t size)
> >>> +                                  phys_addr_t *phys_addr,
> >>> +                                  int *window, size_t size)
> >>>   {
> >>> +     int best_fit = PCI_EPC_DEFAULT_WINDOW;
> >>> +     void __iomem *virt_addr = NULL;
> >>> +     struct pci_epc_mem *mem;
> >>> +     unsigned int page_shift;
> >>>        int pageno;
> >>> -     void __iomem *virt_addr;
> >>> -     struct pci_epc_mem *mem = epc->mem;
> >>> -     unsigned int page_shift = ilog2(mem->page_size);
> >>>        int order;
> >>>
> >>> +     if (epc->mem_windows <= 0)
> >>> +             return NULL;
> >>> +
> >>> +     if (epc->mem_windows > 1) {
> >>> +             best_fit = pci_epc_find_best_fit_window(epc, size);
> >>> +             if (best_fit < 0)
> >>> +                     return NULL;
> >>> +     }
> >>> +
> >>> +     mem = epc->mem[best_fit];
> >>>        size = ALIGN(size, mem->page_size);
> >>> +     if (size > (mem->window.size - mem->window.map_size))
> >>> +             return NULL;
> >>
> >> Assume I have two mem regions, first region is of size 128MB and the second
> >> region is of size 4GB. If there is a allocation request for 4GB, will the
> >> allocation succeed?
> >>
> > yes it would succeed, the pci_epc_find_best_fit_window() would get the
> > corresponding window.
>
> Can you step through the algo again? pci_epc_find_best_fit_window() will
> return the first region which is of size 128MB.
>
done above as requested.

Cheers,
--Prabhakar

> Thanks
> Kishon

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 2/5] pci: endpoint: add support to handle multiple base for mapping outbound memory
@ 2019-12-06 13:08           ` Lad, Prabhakar
  0 siblings, 0 replies; 60+ messages in thread
From: Lad, Prabhakar @ 2019-12-06 13:08 UTC (permalink / raw)
  To: Kishon Vijay Abraham I
  Cc: Bjorn Helgaas, Rob Herring, Mark Rutland, Geert Uytterhoeven,
	Magnus Damm, Marek Vasut, Yoshihiro Shimoda, linux-pci,
	Catalin Marinas, Will Deacon, Lorenzo Pieralisi, Arnd Bergmann,
	Greg Kroah-Hartman, Andrew Murray,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS, LKML,
	LAK

Hi Kishon,

On Thu, Dec 5, 2019 at 10:22 AM Kishon Vijay Abraham I <kishon@ti.com> wrote:
>
> Hi,
>
> On 28/11/19 2:51 am, Lad, Prabhakar wrote:
> > Hi Kishon,
> >
> > Thank you for the review.
> >
> > On Wed, Nov 27, 2019 at 5:15 AM Kishon Vijay Abraham I <kishon@ti.com> wrote:
> >>
> >> Hi Prabhakar,
> >>
> >> On 07/11/19 1:06 AM, Lad Prabhakar wrote:
> >>> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >>>
> >>> rcar pcie controller has support to map multiple memory regions
> >>> for mapping the outbound memory in local system, this feature
> >>> inspires to add support for handling multiple memory bases in
> >>> endpoint framework. In case of multiple memory regions only chunk
> >>> or complete region can be mapped and this window needs to be
> >>> passed to the controller driver.
> >>>
> >>> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >>> Cc: <linux-rockchip@lists.infradead.org>
> >>> Cc: Shawn Lin <shawn.lin@rock-chips.com>
> >>> Cc: Heiko Stuebner <heiko@sntech.de>
> >>> Cc: Tom Joseph <tjoseph@cadence.com>
> >>> Cc: Jingoo Han <jingoohan1@gmail.com>
> >>> Cc: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
> >>> ---
> >>>   .../pci/controller/dwc/pcie-designware-ep.c   |  30 ++-
> >>>   drivers/pci/controller/pcie-cadence-ep.c      |  11 +-
> >>>   drivers/pci/controller/pcie-rockchip-ep.c     |  13 +-
> >>>   drivers/pci/endpoint/functions/pci-epf-test.c |  29 +--
> >>>   drivers/pci/endpoint/pci-epc-core.c           |   7 +-
> >>>   drivers/pci/endpoint/pci-epc-mem.c            | 189 ++++++++++++++----
> >>>   include/linux/pci-epc.h                       |  43 ++--
> >>>   7 files changed, 234 insertions(+), 88 deletions(-)
> >>>
> >>> diff --git a/drivers/pci/controller/dwc/pcie-designware-ep.c b/drivers/pci/controller/dwc/pcie-designware-ep.c
> >>> index 3dd2e2697294..8d23c20b9afd 100644
> >>> --- a/drivers/pci/controller/dwc/pcie-designware-ep.c
> >>> +++ b/drivers/pci/controller/dwc/pcie-designware-ep.c
> >>> @@ -195,7 +195,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, u8 func_no,
> >>>   }
> >>>
> >>>   static int dw_pcie_ep_map_addr(struct pci_epc *epc, u8 func_no,
> >>> -                            phys_addr_t addr,
> >>> +                            phys_addr_t addr, int window,
> >>>                               u64 pci_addr, size_t size)
> >>>   {
> >>>        int ret;
> >>> @@ -367,6 +367,7 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
> >>>        unsigned int aligned_offset;
> >>>        u16 msg_ctrl, msg_data;
> >>>        u32 msg_addr_lower, msg_addr_upper, reg;
> >>> +     int window = PCI_EPC_DEFAULT_WINDOW;
> >>>        u64 msg_addr;
> >>>        bool has_upper;
> >>>        int ret;
> >>> @@ -390,11 +391,11 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
> >>>                reg = ep->msi_cap + PCI_MSI_DATA_32;
> >>>                msg_data = dw_pcie_readw_dbi(pci, reg);
> >>>        }
> >>> -     aligned_offset = msg_addr_lower & (epc->mem->page_size - 1);
> >>> +     aligned_offset = msg_addr_lower & (epc->mem[window]->page_size - 1);
> >>>        msg_addr = ((u64)msg_addr_upper) << 32 |
> >>>                        (msg_addr_lower & ~aligned_offset);
> >>> -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> >>> -                               epc->mem->page_size);
> >>> +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> >>> +                               msg_addr, epc->mem[window]->page_size);
> >>>        if (ret)
> >>>                return ret;
> >>>
> >>> @@ -416,6 +417,7 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >>>        u32 reg, msg_data, vec_ctrl;
> >>>        u64 tbl_addr, msg_addr, reg_u64;
> >>>        void __iomem *msix_tbl;
> >>> +     int window = PCI_EPC_DEFAULT_WINDOW;
> >>>        int ret;
> >>>
> >>>        reg = ep->msix_cap + PCI_MSIX_TABLE;
> >>> @@ -452,8 +454,8 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >>>                return -EPERM;
> >>>        }
> >>>
> >>> -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> >>> -                               epc->mem->page_size);
> >>> +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> >>> +                               msg_addr, epc->mem[window]->page_size);
> >>>        if (ret)
> >>>                return ret;
> >>>
> >>> @@ -466,10 +468,11 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >>>
> >>>   void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
> >>>   {
> >>> +     int window = PCI_EPC_DEFAULT_WINDOW;
> >>>        struct pci_epc *epc = ep->epc;
> >>>
> >>>        pci_epc_mem_free_addr(epc, ep->msi_mem_phys, ep->msi_mem,
> >>> -                           epc->mem->page_size);
> >>> +                           epc->mem[window]->page_size);
> >>>
> >>>        pci_epc_mem_exit(epc);
> >>>   }
> >>> @@ -499,9 +502,12 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
> >>>        u32 reg;
> >>>        void *addr;
> >>>        u8 hdr_type;
> >>> +     int window;
> >>>        unsigned int nbars;
> >>>        unsigned int offset;
> >>>        struct pci_epc *epc;
> >>> +     size_t msi_page_size;
> >>> +     struct pci_epc_mem_window mem_window;
> >>>        struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
> >>>        struct device *dev = pci->dev;
> >>>        struct device_node *np = dev->of_node;
> >>> @@ -574,15 +580,17 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
> >>>        if (ret < 0)
> >>>                epc->max_functions = 1;
> >>>
> >>> -     ret = __pci_epc_mem_init(epc, ep->phys_base, ep->addr_size,
> >>> -                              ep->page_size);
> >>> +     mem_window.phys_base = ep->phys_base;
> >>> +     mem_window.size = ep->addr_size;
> >>> +     ret = __pci_epc_mem_init(epc, &mem_window, 1, ep->page_size);
> >>>        if (ret < 0) {
> >>>                dev_err(dev, "Failed to initialize address space\n");
> >>>                return ret;
> >>>        }
> >>>
> >>> -     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys,
> >>> -                                          epc->mem->page_size);
> >>> +     msi_page_size = epc->mem[PCI_EPC_DEFAULT_WINDOW]->page_size;
> >>> +     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys, &window,
> >>> +                                          msi_page_size);
> >>>        if (!ep->msi_mem) {
> >>>                dev_err(dev, "Failed to reserve memory for MSI/MSI-X\n");
> >>>                return -ENOMEM;
> >>> diff --git a/drivers/pci/controller/pcie-cadence-ep.c b/drivers/pci/controller/pcie-cadence-ep.c
> >>> index def7820cb824..7991b38a5350 100644
> >>> --- a/drivers/pci/controller/pcie-cadence-ep.c
> >>> +++ b/drivers/pci/controller/pcie-cadence-ep.c
> >>> @@ -172,7 +172,7 @@ static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
> >>>   }
> >>>
> >>>   static int cdns_pcie_ep_map_addr(struct pci_epc *epc, u8 fn, phys_addr_t addr,
> >>> -                              u64 pci_addr, size_t size)
> >>> +                              int window, u64 pci_addr, size_t size)
> >>>   {
> >>>        struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
> >>>        struct cdns_pcie *pcie = &ep->pcie;
> >>> @@ -434,12 +434,14 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
> >>>   {
> >>>        struct device *dev = &pdev->dev;
> >>>        struct device_node *np = dev->of_node;
> >>> +     struct pci_epc_mem_window mem_window;
> >>>        struct cdns_pcie_ep *ep;
> >>>        struct cdns_pcie *pcie;
> >>>        struct pci_epc *epc;
> >>>        struct resource *res;
> >>>        int ret;
> >>>        int phy_count;
> >>> +     int window;
> >>>
> >>>        ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> >>>        if (!ep)
> >>> @@ -502,15 +504,16 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
> >>>        if (of_property_read_u8(np, "max-functions", &epc->max_functions) < 0)
> >>>                epc->max_functions = 1;
> >>>
> >>> -     ret = pci_epc_mem_init(epc, pcie->mem_res->start,
> >>> -                            resource_size(pcie->mem_res));
> >>> +     mem_window.phys_base = pcie->mem_res->start;
> >>> +     mem_window.size = resource_size(pcie->mem_res);
> >>> +     ret = pci_epc_mem_init(epc, &mem_window, 1);
> >>>        if (ret < 0) {
> >>>                dev_err(dev, "failed to initialize the memory space\n");
> >>>                goto err_init;
> >>>        }
> >>>
> >>>        ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> >>> -                                               SZ_128K);
> >>> +                                               &window, SZ_128K);
> >>>        if (!ep->irq_cpu_addr) {
> >>>                dev_err(dev, "failed to reserve memory space for MSI\n");
> >>>                ret = -ENOMEM;
> >>> diff --git a/drivers/pci/controller/pcie-rockchip-ep.c b/drivers/pci/controller/pcie-rockchip-ep.c
> >>> index d743b0a48988..d59e85c8d319 100644
> >>> --- a/drivers/pci/controller/pcie-rockchip-ep.c
> >>> +++ b/drivers/pci/controller/pcie-rockchip-ep.c
> >>> @@ -256,8 +256,8 @@ static void rockchip_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
> >>>   }
> >>>
> >>>   static int rockchip_pcie_ep_map_addr(struct pci_epc *epc, u8 fn,
> >>> -                                  phys_addr_t addr, u64 pci_addr,
> >>> -                                  size_t size)
> >>> +                                  phys_addr_t addr, int window,
> >>> +                                  u64 pci_addr, size_t size)
> >>>   {
> >>>        struct rockchip_pcie_ep *ep = epc_get_drvdata(epc);
> >>>        struct rockchip_pcie *pcie = &ep->rockchip;
> >>> @@ -562,11 +562,13 @@ static const struct of_device_id rockchip_pcie_ep_of_match[] = {
> >>>
> >>>   static int rockchip_pcie_ep_probe(struct platform_device *pdev)
> >>>   {
> >>> +     struct pci_epc_mem_window mem_window;
> >>>        struct device *dev = &pdev->dev;
> >>>        struct rockchip_pcie_ep *ep;
> >>>        struct rockchip_pcie *rockchip;
> >>>        struct pci_epc *epc;
> >>>        size_t max_regions;
> >>> +     int window;
> >>>        int err;
> >>>
> >>>        ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> >>> @@ -614,15 +616,16 @@ static int rockchip_pcie_ep_probe(struct platform_device *pdev)
> >>>        /* Only enable function 0 by default */
> >>>        rockchip_pcie_write(rockchip, BIT(0), PCIE_CORE_PHY_FUNC_CFG);
> >>>
> >>> -     err = pci_epc_mem_init(epc, rockchip->mem_res->start,
> >>> -                            resource_size(rockchip->mem_res));
> >>> +     mem_window.phys_base = rockchip->mem_res->start;
> >>> +     mem_window.size = resource_size(rockchip->mem_res);
> >>> +     err = pci_epc_mem_init(epc, &mem_window, 1);
> >>>        if (err < 0) {
> >>>                dev_err(dev, "failed to initialize the memory space\n");
> >>>                goto err_uninit_port;
> >>>        }
> >>>
> >>>        ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> >>> -                                               SZ_128K);
> >>> +                                               &window, SZ_128K);
> >>>        if (!ep->irq_cpu_addr) {
> >>>                dev_err(dev, "failed to reserve memory space for MSI\n");
> >>>                err = -ENOMEM;
> >>> diff --git a/drivers/pci/endpoint/functions/pci-epf-test.c b/drivers/pci/endpoint/functions/pci-epf-test.c
> >>> index 5d74f81ddfe4..475228011703 100644
> >>> --- a/drivers/pci/endpoint/functions/pci-epf-test.c
> >>> +++ b/drivers/pci/endpoint/functions/pci-epf-test.c
> >>> @@ -84,8 +84,10 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >>>        struct pci_epc *epc = epf->epc;
> >>>        enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >>>        struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> >>> +     int window;
> >>>
> >>> -     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr, reg->size);
> >>> +     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr,
> >>> +                                       &window, reg->size);
> >>>        if (!src_addr) {
> >>>                dev_err(dev, "Failed to allocate source address\n");
> >>>                reg->status = STATUS_SRC_ADDR_INVALID;
> >>> @@ -93,15 +95,16 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >>>                goto err;
> >>>        }
> >>>
> >>> -     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, reg->src_addr,
> >>> -                            reg->size);
> >>> +     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, window,
> >>> +                            reg->src_addr, reg->size);
> >>>        if (ret) {
> >>>                dev_err(dev, "Failed to map source address\n");
> >>>                reg->status = STATUS_SRC_ADDR_INVALID;
> >>>                goto err_src_addr;
> >>>        }
> >>>
> >>> -     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr, reg->size);
> >>> +     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr,
> >>> +                                       &window, reg->size);
> >>>        if (!dst_addr) {
> >>>                dev_err(dev, "Failed to allocate destination address\n");
> >>>                reg->status = STATUS_DST_ADDR_INVALID;
> >>> @@ -109,8 +112,8 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >>>                goto err_src_map_addr;
> >>>        }
> >>>
> >>> -     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, reg->dst_addr,
> >>> -                            reg->size);
> >>> +     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, window,
> >>> +                            reg->dst_addr, reg->size);
> >>>        if (ret) {
> >>>                dev_err(dev, "Failed to map destination address\n");
> >>>                reg->status = STATUS_DST_ADDR_INVALID;
> >>> @@ -146,8 +149,9 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
> >>>        struct pci_epc *epc = epf->epc;
> >>>        enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >>>        struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> >>> +     int window;
> >>>
> >>> -     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> >>> +     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
> >>>        if (!src_addr) {
> >>>                dev_err(dev, "Failed to allocate address\n");
> >>>                reg->status = STATUS_SRC_ADDR_INVALID;
> >>> @@ -155,8 +159,8 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
> >>>                goto err;
> >>>        }
> >>>
> >>> -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->src_addr,
> >>> -                            reg->size);
> >>> +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> >>> +                            reg->src_addr, reg->size);
> >>>        if (ret) {
> >>>                dev_err(dev, "Failed to map address\n");
> >>>                reg->status = STATUS_SRC_ADDR_INVALID;
> >>> @@ -193,13 +197,14 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
> >>>        void __iomem *dst_addr;
> >>>        void *buf;
> >>>        phys_addr_t phys_addr;
> >>> +     int window;
> >>>        struct pci_epf *epf = epf_test->epf;
> >>>        struct device *dev = &epf->dev;
> >>>        struct pci_epc *epc = epf->epc;
> >>>        enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >>>        struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> >>>
> >>> -     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> >>> +     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
> >>>        if (!dst_addr) {
> >>>                dev_err(dev, "Failed to allocate address\n");
> >>>                reg->status = STATUS_DST_ADDR_INVALID;
> >>> @@ -207,8 +212,8 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
> >>>                goto err;
> >>>        }
> >>>
> >>> -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->dst_addr,
> >>> -                            reg->size);
> >>> +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> >>> +                            reg->dst_addr, reg->size);
> >>>        if (ret) {
> >>>                dev_err(dev, "Failed to map address\n");
> >>>                reg->status = STATUS_DST_ADDR_INVALID;
> >>> diff --git a/drivers/pci/endpoint/pci-epc-core.c b/drivers/pci/endpoint/pci-epc-core.c
> >>> index 2091508c1620..289c266c2d90 100644
> >>> --- a/drivers/pci/endpoint/pci-epc-core.c
> >>> +++ b/drivers/pci/endpoint/pci-epc-core.c
> >>> @@ -358,13 +358,15 @@ EXPORT_SYMBOL_GPL(pci_epc_unmap_addr);
> >>>    * @epc: the EPC device on which address is allocated
> >>>    * @func_no: the endpoint function number in the EPC device
> >>>    * @phys_addr: physical address of the local system
> >>> + * @window: index to the window region where PCI address will be mapped
> >>>    * @pci_addr: PCI address to which the physical address should be mapped
> >>>    * @size: the size of the allocation
> >>>    *
> >>>    * Invoke to map CPU address with PCI address.
> >>>    */
> >>>   int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> >>> -                  phys_addr_t phys_addr, u64 pci_addr, size_t size)
> >>> +                  phys_addr_t phys_addr, int window,
> >>> +                  u64 pci_addr, size_t size)
> >>>   {
> >>>        int ret;
> >>>        unsigned long flags;
> >>> @@ -376,7 +378,8 @@ int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> >>>                return 0;
> >>>
> >>>        spin_lock_irqsave(&epc->lock, flags);
> >>> -     ret = epc->ops->map_addr(epc, func_no, phys_addr, pci_addr, size);
> >>> +     ret = epc->ops->map_addr(epc, func_no, phys_addr,
> >>> +                              window, pci_addr, size);
> >>>        spin_unlock_irqrestore(&epc->lock, flags);
> >>>
> >>>        return ret;
> >>> diff --git a/drivers/pci/endpoint/pci-epc-mem.c b/drivers/pci/endpoint/pci-epc-mem.c
> >>> index d2b174ce15de..c955f2c97944 100644
> >>> --- a/drivers/pci/endpoint/pci-epc-mem.c
> >>> +++ b/drivers/pci/endpoint/pci-epc-mem.c
> >>> @@ -39,56 +39,77 @@ static int pci_epc_mem_get_order(struct pci_epc_mem *mem, size_t size)
> >>>    * __pci_epc_mem_init() - initialize the pci_epc_mem structure
> >>>    * @epc: the EPC device that invoked pci_epc_mem_init
> >>>    * @phys_base: the physical address of the base
> >>> - * @size: the size of the address space
> >>> + * @num_windows: number of windows device supports
> >>
> >> struct pci_epc_mem_window is missing here.
> > oops my bad will fix that.
> >
> >>>    * @page_size: size of each page
> >>>    *
> >>>    * Invoke to initialize the pci_epc_mem structure used by the
> >>>    * endpoint functions to allocate mapped PCI address.
> >>>    */
> >>> -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_base, size_t size,
> >>> -                    size_t page_size)
> >>> +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *windows,
> >>> +                    int num_windows, size_t page_size)
> >>>   {
> >>> -     int ret;
> >>> -     struct pci_epc_mem *mem;
> >>> -     unsigned long *bitmap;
> >>> +     struct pci_epc_mem *mem = NULL;
> >>> +     unsigned long *bitmap = NULL;
> >>>        unsigned int page_shift;
> >>> -     int pages;
> >>>        int bitmap_size;
> >>> +     int pages;
> >>> +     int ret;
> >>> +     int i;
> >>> +
> >>> +     epc->mem_windows = 0;
> >>> +
> >>> +     if (!windows)
> >>> +             return -EINVAL;
> >>> +
> >>> +     if (num_windows <= 0)
> >>> +             return -EINVAL;
> >>>
> >>>        if (page_size < PAGE_SIZE)
> >>>                page_size = PAGE_SIZE;
> >>>
> >>>        page_shift = ilog2(page_size);
> >>> -     pages = size >> page_shift;
> >>> -     bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
> >>>
> >>> -     mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> >>> -     if (!mem) {
> >>> -             ret = -ENOMEM;
> >>> -             goto err;
> >>> -     }
> >>> +     epc->mem = kcalloc(num_windows, sizeof(*mem), GFP_KERNEL);
> >>> +     if (!epc->mem)
> >>> +             return -EINVAL;
> >>>
> >>> -     bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> >>> -     if (!bitmap) {
> >>> -             ret = -ENOMEM;
> >>> -             goto err_mem;
> >>> -     }
> >>> +     for (i = 0; i < num_windows; i++) {
> >>> +             pages = windows[i].phys_base >> page_shift;
> >>> +             bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
> >>>
> >>> -     mem->bitmap = bitmap;
> >>> -     mem->phys_base = phys_base;
> >>> -     mem->page_size = page_size;
> >>> -     mem->pages = pages;
> >>> -     mem->size = size;
> >>> +             mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> >>> +             if (!mem) {
> >>> +                     ret = -ENOMEM;
> >>> +                     goto err_mem;
> >>> +             }
> >>>
> >>> -     epc->mem = mem;
> >>> +             bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> >>> +             if (!bitmap) {
> >>> +                     ret = -ENOMEM;
> >>> +                     goto err_mem;
> >>> +             }
> >>> +
> >>> +             mem->bitmap = bitmap;
> >>> +             mem->window.phys_base = windows[i].phys_base;
> >>> +             mem->page_size = page_size;
> >>> +             mem->pages = pages;
> >>> +             mem->window.size = windows[i].size;
> >>> +             mem->window.map_size = 0;
> >>> +
> >>> +             epc->mem[i] = mem;
> >>> +     }
> >>> +     epc->mem_windows = num_windows;
> >>>
> >>>        return 0;
> >>>
> >>>   err_mem:
> >>> -     kfree(mem);
> >>> +     for (; i >= 0; i--) {
> >>> +             kfree(mem->bitmap);
> >>> +             kfree(epc->mem[i]);
> >>> +     }
> >>> +     kfree(epc->mem);
> >>>
> >>> -err:
> >>> -return ret;
> >>> +     return ret;
> >>>   }
> >>>   EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
> >>>
> >>> @@ -101,48 +122,126 @@ EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
> >>>    */
> >>>   void pci_epc_mem_exit(struct pci_epc *epc)
> >>>   {
> >>> -     struct pci_epc_mem *mem = epc->mem;
> >>> +     struct pci_epc_mem *mem;
> >>> +     int i;
> >>> +
> >>> +     if (!epc->mem_windows)
> >>> +             return;
> >>> +
> >>> +     for (i = 0; i <= epc->mem_windows; i--) {
> >>> +             mem = epc->mem[i];
> >>> +             kfree(mem->bitmap);
> >>> +             kfree(epc->mem[i]);
> >>> +     }
> >>> +     kfree(epc->mem);
> >>>
> >>>        epc->mem = NULL;
> >>> -     kfree(mem->bitmap);
> >>> -     kfree(mem);
> >>> +     epc->mem_windows = 0;
> >>>   }
> >>>   EXPORT_SYMBOL_GPL(pci_epc_mem_exit);
> >>>
> >>> +static int pci_epc_find_best_fit_window(struct pci_epc *epc, size_t size)
> >>> +{
> >>> +     size_t window_least_size = 0;
> >>> +     int best_fit_window = -1;
> >>> +     struct pci_epc_mem *mem;
> >>> +     size_t actual_size;
> >>> +     int i;
> >>> +
> >>> +     for (i = 0; i < epc->mem_windows; i++) {
^^
here we loop through all the possible windows.

> >>> +             mem = epc->mem[i];
> >>> +
> >>> +             /* if chunk from this region is already used skip it */
> >>> +             if (mem->window.map_size)
> >>> +                     continue;
> >>> +
^^
If a chunk is used in the window we skip it.
but the next version of the patchest  will check for  the flag if the
controller supports multiple allocation of chunks
in same window.

> >>> +             actual_size = ALIGN(size, mem->page_size);
> >>> +
> >>> +             if (best_fit_window == -1) {
> >>> +                     best_fit_window = i;
> >>> +                     window_least_size = mem->window.size;
firstly it assigns a first free region, although the above code doesnt
check if requested size can be fitted in the window,
I'll fix that in next version as below :

if (actual_size <= mem->window.size) {
                    best_fit_window = i;
                    window_least_size = mem->window.size;
}

window_least_size  is temp variable to compare against other window sizes.

> >>> +             } else {
> >>> +                     if (actual_size <= mem->window.size &&
> >>> +                         mem->window.size < window_least_size) {
> >>> +                             best_fit_window = i;
> >>> +                             window_least_size = mem->window.size;
> >>> +                     }

add in above we check if the requested size does fit in the window,
and the windows size is lesser then
the previously assigned one, if the criteria matches we choose this
window. and this is looped through all
the windows.

> >>> +             }
> >>> +     }
> >>> +
> >>> +     return best_fit_window;
> >>> +}
> >>> +
> >>>   /**
> >>>    * pci_epc_mem_alloc_addr() - allocate memory address from EPC addr space
> >>>    * @epc: the EPC device on which memory has to be allocated
> >>>    * @phys_addr: populate the allocated physical address here
> >>> + * @window: populate the window here which will be used to map PCI address
> >>>    * @size: the size of the address space that has to be allocated
> >>>    *
> >>>    * Invoke to allocate memory address from the EPC address space. This
> >>>    * is usually done to map the remote RC address into the local system.
> >>>    */
> >>>   void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
> >>> -                                  phys_addr_t *phys_addr, size_t size)
> >>> +                                  phys_addr_t *phys_addr,
> >>> +                                  int *window, size_t size)
> >>>   {
> >>> +     int best_fit = PCI_EPC_DEFAULT_WINDOW;
> >>> +     void __iomem *virt_addr = NULL;
> >>> +     struct pci_epc_mem *mem;
> >>> +     unsigned int page_shift;
> >>>        int pageno;
> >>> -     void __iomem *virt_addr;
> >>> -     struct pci_epc_mem *mem = epc->mem;
> >>> -     unsigned int page_shift = ilog2(mem->page_size);
> >>>        int order;
> >>>
> >>> +     if (epc->mem_windows <= 0)
> >>> +             return NULL;
> >>> +
> >>> +     if (epc->mem_windows > 1) {
> >>> +             best_fit = pci_epc_find_best_fit_window(epc, size);
> >>> +             if (best_fit < 0)
> >>> +                     return NULL;
> >>> +     }
> >>> +
> >>> +     mem = epc->mem[best_fit];
> >>>        size = ALIGN(size, mem->page_size);
> >>> +     if (size > (mem->window.size - mem->window.map_size))
> >>> +             return NULL;
> >>
> >> Assume I have two mem regions, first region is of size 128MB and the second
> >> region is of size 4GB. If there is a allocation request for 4GB, will the
> >> allocation succeed?
> >>
> > yes it would succeed, the pci_epc_find_best_fit_window() would get the
> > corresponding window.
>
> Can you step through the algo again? pci_epc_find_best_fit_window() will
> return the first region which is of size 128MB.
>
done above as requested.

Cheers,
--Prabhakar

> Thanks
> Kishon

^ permalink raw reply	[flat|nested] 60+ messages in thread

* Re: [PATCH 2/5] pci: endpoint: add support to handle multiple base for mapping outbound memory
@ 2019-12-06 13:08           ` Lad, Prabhakar
  0 siblings, 0 replies; 60+ messages in thread
From: Lad, Prabhakar @ 2019-12-06 13:08 UTC (permalink / raw)
  To: Kishon Vijay Abraham I
  Cc: Mark Rutland, Heiko Stuebner, Geert Uytterhoeven, linux-pci,
	Shawn Lin, Will Deacon, Marek Vasut, Lorenzo Pieralisi,
	Magnus Damm, linux-rockchip, Catalin Marinas,
	open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS,
	Chris Paterson, Arnd Bergmann, Lad, Prabhakar, Rob Herring,
	Bjorn Helgaas, LAK, Greg Kroah-Hartman, Yoshihiro Shimoda, LKML,
	Linux-Renesas, Tom Joseph, Jingoo Han, Andrew Murray,
	Gustavo Pimentel

Hi Kishon,

On Thu, Dec 5, 2019 at 10:22 AM Kishon Vijay Abraham I <kishon@ti.com> wrote:
>
> Hi,
>
> On 28/11/19 2:51 am, Lad, Prabhakar wrote:
> > Hi Kishon,
> >
> > Thank you for the review.
> >
> > On Wed, Nov 27, 2019 at 5:15 AM Kishon Vijay Abraham I <kishon@ti.com> wrote:
> >>
> >> Hi Prabhakar,
> >>
> >> On 07/11/19 1:06 AM, Lad Prabhakar wrote:
> >>> From: "Lad, Prabhakar" <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >>>
> >>> rcar pcie controller has support to map multiple memory regions
> >>> for mapping the outbound memory in local system, this feature
> >>> inspires to add support for handling multiple memory bases in
> >>> endpoint framework. In case of multiple memory regions only chunk
> >>> or complete region can be mapped and this window needs to be
> >>> passed to the controller driver.
> >>>
> >>> Signed-off-by: Lad, Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
> >>> Cc: <linux-rockchip@lists.infradead.org>
> >>> Cc: Shawn Lin <shawn.lin@rock-chips.com>
> >>> Cc: Heiko Stuebner <heiko@sntech.de>
> >>> Cc: Tom Joseph <tjoseph@cadence.com>
> >>> Cc: Jingoo Han <jingoohan1@gmail.com>
> >>> Cc: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
> >>> ---
> >>>   .../pci/controller/dwc/pcie-designware-ep.c   |  30 ++-
> >>>   drivers/pci/controller/pcie-cadence-ep.c      |  11 +-
> >>>   drivers/pci/controller/pcie-rockchip-ep.c     |  13 +-
> >>>   drivers/pci/endpoint/functions/pci-epf-test.c |  29 +--
> >>>   drivers/pci/endpoint/pci-epc-core.c           |   7 +-
> >>>   drivers/pci/endpoint/pci-epc-mem.c            | 189 ++++++++++++++----
> >>>   include/linux/pci-epc.h                       |  43 ++--
> >>>   7 files changed, 234 insertions(+), 88 deletions(-)
> >>>
> >>> diff --git a/drivers/pci/controller/dwc/pcie-designware-ep.c b/drivers/pci/controller/dwc/pcie-designware-ep.c
> >>> index 3dd2e2697294..8d23c20b9afd 100644
> >>> --- a/drivers/pci/controller/dwc/pcie-designware-ep.c
> >>> +++ b/drivers/pci/controller/dwc/pcie-designware-ep.c
> >>> @@ -195,7 +195,7 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, u8 func_no,
> >>>   }
> >>>
> >>>   static int dw_pcie_ep_map_addr(struct pci_epc *epc, u8 func_no,
> >>> -                            phys_addr_t addr,
> >>> +                            phys_addr_t addr, int window,
> >>>                               u64 pci_addr, size_t size)
> >>>   {
> >>>        int ret;
> >>> @@ -367,6 +367,7 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
> >>>        unsigned int aligned_offset;
> >>>        u16 msg_ctrl, msg_data;
> >>>        u32 msg_addr_lower, msg_addr_upper, reg;
> >>> +     int window = PCI_EPC_DEFAULT_WINDOW;
> >>>        u64 msg_addr;
> >>>        bool has_upper;
> >>>        int ret;
> >>> @@ -390,11 +391,11 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
> >>>                reg = ep->msi_cap + PCI_MSI_DATA_32;
> >>>                msg_data = dw_pcie_readw_dbi(pci, reg);
> >>>        }
> >>> -     aligned_offset = msg_addr_lower & (epc->mem->page_size - 1);
> >>> +     aligned_offset = msg_addr_lower & (epc->mem[window]->page_size - 1);
> >>>        msg_addr = ((u64)msg_addr_upper) << 32 |
> >>>                        (msg_addr_lower & ~aligned_offset);
> >>> -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> >>> -                               epc->mem->page_size);
> >>> +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> >>> +                               msg_addr, epc->mem[window]->page_size);
> >>>        if (ret)
> >>>                return ret;
> >>>
> >>> @@ -416,6 +417,7 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >>>        u32 reg, msg_data, vec_ctrl;
> >>>        u64 tbl_addr, msg_addr, reg_u64;
> >>>        void __iomem *msix_tbl;
> >>> +     int window = PCI_EPC_DEFAULT_WINDOW;
> >>>        int ret;
> >>>
> >>>        reg = ep->msix_cap + PCI_MSIX_TABLE;
> >>> @@ -452,8 +454,8 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >>>                return -EPERM;
> >>>        }
> >>>
> >>> -     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, msg_addr,
> >>> -                               epc->mem->page_size);
> >>> +     ret = dw_pcie_ep_map_addr(epc, func_no, ep->msi_mem_phys, window,
> >>> +                               msg_addr, epc->mem[window]->page_size);
> >>>        if (ret)
> >>>                return ret;
> >>>
> >>> @@ -466,10 +468,11 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no,
> >>>
> >>>   void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
> >>>   {
> >>> +     int window = PCI_EPC_DEFAULT_WINDOW;
> >>>        struct pci_epc *epc = ep->epc;
> >>>
> >>>        pci_epc_mem_free_addr(epc, ep->msi_mem_phys, ep->msi_mem,
> >>> -                           epc->mem->page_size);
> >>> +                           epc->mem[window]->page_size);
> >>>
> >>>        pci_epc_mem_exit(epc);
> >>>   }
> >>> @@ -499,9 +502,12 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
> >>>        u32 reg;
> >>>        void *addr;
> >>>        u8 hdr_type;
> >>> +     int window;
> >>>        unsigned int nbars;
> >>>        unsigned int offset;
> >>>        struct pci_epc *epc;
> >>> +     size_t msi_page_size;
> >>> +     struct pci_epc_mem_window mem_window;
> >>>        struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
> >>>        struct device *dev = pci->dev;
> >>>        struct device_node *np = dev->of_node;
> >>> @@ -574,15 +580,17 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
> >>>        if (ret < 0)
> >>>                epc->max_functions = 1;
> >>>
> >>> -     ret = __pci_epc_mem_init(epc, ep->phys_base, ep->addr_size,
> >>> -                              ep->page_size);
> >>> +     mem_window.phys_base = ep->phys_base;
> >>> +     mem_window.size = ep->addr_size;
> >>> +     ret = __pci_epc_mem_init(epc, &mem_window, 1, ep->page_size);
> >>>        if (ret < 0) {
> >>>                dev_err(dev, "Failed to initialize address space\n");
> >>>                return ret;
> >>>        }
> >>>
> >>> -     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys,
> >>> -                                          epc->mem->page_size);
> >>> +     msi_page_size = epc->mem[PCI_EPC_DEFAULT_WINDOW]->page_size;
> >>> +     ep->msi_mem = pci_epc_mem_alloc_addr(epc, &ep->msi_mem_phys, &window,
> >>> +                                          msi_page_size);
> >>>        if (!ep->msi_mem) {
> >>>                dev_err(dev, "Failed to reserve memory for MSI/MSI-X\n");
> >>>                return -ENOMEM;
> >>> diff --git a/drivers/pci/controller/pcie-cadence-ep.c b/drivers/pci/controller/pcie-cadence-ep.c
> >>> index def7820cb824..7991b38a5350 100644
> >>> --- a/drivers/pci/controller/pcie-cadence-ep.c
> >>> +++ b/drivers/pci/controller/pcie-cadence-ep.c
> >>> @@ -172,7 +172,7 @@ static void cdns_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
> >>>   }
> >>>
> >>>   static int cdns_pcie_ep_map_addr(struct pci_epc *epc, u8 fn, phys_addr_t addr,
> >>> -                              u64 pci_addr, size_t size)
> >>> +                              int window, u64 pci_addr, size_t size)
> >>>   {
> >>>        struct cdns_pcie_ep *ep = epc_get_drvdata(epc);
> >>>        struct cdns_pcie *pcie = &ep->pcie;
> >>> @@ -434,12 +434,14 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
> >>>   {
> >>>        struct device *dev = &pdev->dev;
> >>>        struct device_node *np = dev->of_node;
> >>> +     struct pci_epc_mem_window mem_window;
> >>>        struct cdns_pcie_ep *ep;
> >>>        struct cdns_pcie *pcie;
> >>>        struct pci_epc *epc;
> >>>        struct resource *res;
> >>>        int ret;
> >>>        int phy_count;
> >>> +     int window;
> >>>
> >>>        ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> >>>        if (!ep)
> >>> @@ -502,15 +504,16 @@ static int cdns_pcie_ep_probe(struct platform_device *pdev)
> >>>        if (of_property_read_u8(np, "max-functions", &epc->max_functions) < 0)
> >>>                epc->max_functions = 1;
> >>>
> >>> -     ret = pci_epc_mem_init(epc, pcie->mem_res->start,
> >>> -                            resource_size(pcie->mem_res));
> >>> +     mem_window.phys_base = pcie->mem_res->start;
> >>> +     mem_window.size = resource_size(pcie->mem_res);
> >>> +     ret = pci_epc_mem_init(epc, &mem_window, 1);
> >>>        if (ret < 0) {
> >>>                dev_err(dev, "failed to initialize the memory space\n");
> >>>                goto err_init;
> >>>        }
> >>>
> >>>        ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> >>> -                                               SZ_128K);
> >>> +                                               &window, SZ_128K);
> >>>        if (!ep->irq_cpu_addr) {
> >>>                dev_err(dev, "failed to reserve memory space for MSI\n");
> >>>                ret = -ENOMEM;
> >>> diff --git a/drivers/pci/controller/pcie-rockchip-ep.c b/drivers/pci/controller/pcie-rockchip-ep.c
> >>> index d743b0a48988..d59e85c8d319 100644
> >>> --- a/drivers/pci/controller/pcie-rockchip-ep.c
> >>> +++ b/drivers/pci/controller/pcie-rockchip-ep.c
> >>> @@ -256,8 +256,8 @@ static void rockchip_pcie_ep_clear_bar(struct pci_epc *epc, u8 fn,
> >>>   }
> >>>
> >>>   static int rockchip_pcie_ep_map_addr(struct pci_epc *epc, u8 fn,
> >>> -                                  phys_addr_t addr, u64 pci_addr,
> >>> -                                  size_t size)
> >>> +                                  phys_addr_t addr, int window,
> >>> +                                  u64 pci_addr, size_t size)
> >>>   {
> >>>        struct rockchip_pcie_ep *ep = epc_get_drvdata(epc);
> >>>        struct rockchip_pcie *pcie = &ep->rockchip;
> >>> @@ -562,11 +562,13 @@ static const struct of_device_id rockchip_pcie_ep_of_match[] = {
> >>>
> >>>   static int rockchip_pcie_ep_probe(struct platform_device *pdev)
> >>>   {
> >>> +     struct pci_epc_mem_window mem_window;
> >>>        struct device *dev = &pdev->dev;
> >>>        struct rockchip_pcie_ep *ep;
> >>>        struct rockchip_pcie *rockchip;
> >>>        struct pci_epc *epc;
> >>>        size_t max_regions;
> >>> +     int window;
> >>>        int err;
> >>>
> >>>        ep = devm_kzalloc(dev, sizeof(*ep), GFP_KERNEL);
> >>> @@ -614,15 +616,16 @@ static int rockchip_pcie_ep_probe(struct platform_device *pdev)
> >>>        /* Only enable function 0 by default */
> >>>        rockchip_pcie_write(rockchip, BIT(0), PCIE_CORE_PHY_FUNC_CFG);
> >>>
> >>> -     err = pci_epc_mem_init(epc, rockchip->mem_res->start,
> >>> -                            resource_size(rockchip->mem_res));
> >>> +     mem_window.phys_base = rockchip->mem_res->start;
> >>> +     mem_window.size = resource_size(rockchip->mem_res);
> >>> +     err = pci_epc_mem_init(epc, &mem_window, 1);
> >>>        if (err < 0) {
> >>>                dev_err(dev, "failed to initialize the memory space\n");
> >>>                goto err_uninit_port;
> >>>        }
> >>>
> >>>        ep->irq_cpu_addr = pci_epc_mem_alloc_addr(epc, &ep->irq_phys_addr,
> >>> -                                               SZ_128K);
> >>> +                                               &window, SZ_128K);
> >>>        if (!ep->irq_cpu_addr) {
> >>>                dev_err(dev, "failed to reserve memory space for MSI\n");
> >>>                err = -ENOMEM;
> >>> diff --git a/drivers/pci/endpoint/functions/pci-epf-test.c b/drivers/pci/endpoint/functions/pci-epf-test.c
> >>> index 5d74f81ddfe4..475228011703 100644
> >>> --- a/drivers/pci/endpoint/functions/pci-epf-test.c
> >>> +++ b/drivers/pci/endpoint/functions/pci-epf-test.c
> >>> @@ -84,8 +84,10 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >>>        struct pci_epc *epc = epf->epc;
> >>>        enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >>>        struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> >>> +     int window;
> >>>
> >>> -     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr, reg->size);
> >>> +     src_addr = pci_epc_mem_alloc_addr(epc, &src_phys_addr,
> >>> +                                       &window, reg->size);
> >>>        if (!src_addr) {
> >>>                dev_err(dev, "Failed to allocate source address\n");
> >>>                reg->status = STATUS_SRC_ADDR_INVALID;
> >>> @@ -93,15 +95,16 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >>>                goto err;
> >>>        }
> >>>
> >>> -     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, reg->src_addr,
> >>> -                            reg->size);
> >>> +     ret = pci_epc_map_addr(epc, epf->func_no, src_phys_addr, window,
> >>> +                            reg->src_addr, reg->size);
> >>>        if (ret) {
> >>>                dev_err(dev, "Failed to map source address\n");
> >>>                reg->status = STATUS_SRC_ADDR_INVALID;
> >>>                goto err_src_addr;
> >>>        }
> >>>
> >>> -     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr, reg->size);
> >>> +     dst_addr = pci_epc_mem_alloc_addr(epc, &dst_phys_addr,
> >>> +                                       &window, reg->size);
> >>>        if (!dst_addr) {
> >>>                dev_err(dev, "Failed to allocate destination address\n");
> >>>                reg->status = STATUS_DST_ADDR_INVALID;
> >>> @@ -109,8 +112,8 @@ static int pci_epf_test_copy(struct pci_epf_test *epf_test)
> >>>                goto err_src_map_addr;
> >>>        }
> >>>
> >>> -     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, reg->dst_addr,
> >>> -                            reg->size);
> >>> +     ret = pci_epc_map_addr(epc, epf->func_no, dst_phys_addr, window,
> >>> +                            reg->dst_addr, reg->size);
> >>>        if (ret) {
> >>>                dev_err(dev, "Failed to map destination address\n");
> >>>                reg->status = STATUS_DST_ADDR_INVALID;
> >>> @@ -146,8 +149,9 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
> >>>        struct pci_epc *epc = epf->epc;
> >>>        enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >>>        struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> >>> +     int window;
> >>>
> >>> -     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> >>> +     src_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
> >>>        if (!src_addr) {
> >>>                dev_err(dev, "Failed to allocate address\n");
> >>>                reg->status = STATUS_SRC_ADDR_INVALID;
> >>> @@ -155,8 +159,8 @@ static int pci_epf_test_read(struct pci_epf_test *epf_test)
> >>>                goto err;
> >>>        }
> >>>
> >>> -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->src_addr,
> >>> -                            reg->size);
> >>> +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> >>> +                            reg->src_addr, reg->size);
> >>>        if (ret) {
> >>>                dev_err(dev, "Failed to map address\n");
> >>>                reg->status = STATUS_SRC_ADDR_INVALID;
> >>> @@ -193,13 +197,14 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
> >>>        void __iomem *dst_addr;
> >>>        void *buf;
> >>>        phys_addr_t phys_addr;
> >>> +     int window;
> >>>        struct pci_epf *epf = epf_test->epf;
> >>>        struct device *dev = &epf->dev;
> >>>        struct pci_epc *epc = epf->epc;
> >>>        enum pci_barno test_reg_bar = epf_test->test_reg_bar;
> >>>        struct pci_epf_test_reg *reg = epf_test->reg[test_reg_bar];
> >>>
> >>> -     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, reg->size);
> >>> +     dst_addr = pci_epc_mem_alloc_addr(epc, &phys_addr, &window, reg->size);
> >>>        if (!dst_addr) {
> >>>                dev_err(dev, "Failed to allocate address\n");
> >>>                reg->status = STATUS_DST_ADDR_INVALID;
> >>> @@ -207,8 +212,8 @@ static int pci_epf_test_write(struct pci_epf_test *epf_test)
> >>>                goto err;
> >>>        }
> >>>
> >>> -     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, reg->dst_addr,
> >>> -                            reg->size);
> >>> +     ret = pci_epc_map_addr(epc, epf->func_no, phys_addr, window,
> >>> +                            reg->dst_addr, reg->size);
> >>>        if (ret) {
> >>>                dev_err(dev, "Failed to map address\n");
> >>>                reg->status = STATUS_DST_ADDR_INVALID;
> >>> diff --git a/drivers/pci/endpoint/pci-epc-core.c b/drivers/pci/endpoint/pci-epc-core.c
> >>> index 2091508c1620..289c266c2d90 100644
> >>> --- a/drivers/pci/endpoint/pci-epc-core.c
> >>> +++ b/drivers/pci/endpoint/pci-epc-core.c
> >>> @@ -358,13 +358,15 @@ EXPORT_SYMBOL_GPL(pci_epc_unmap_addr);
> >>>    * @epc: the EPC device on which address is allocated
> >>>    * @func_no: the endpoint function number in the EPC device
> >>>    * @phys_addr: physical address of the local system
> >>> + * @window: index to the window region where PCI address will be mapped
> >>>    * @pci_addr: PCI address to which the physical address should be mapped
> >>>    * @size: the size of the allocation
> >>>    *
> >>>    * Invoke to map CPU address with PCI address.
> >>>    */
> >>>   int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> >>> -                  phys_addr_t phys_addr, u64 pci_addr, size_t size)
> >>> +                  phys_addr_t phys_addr, int window,
> >>> +                  u64 pci_addr, size_t size)
> >>>   {
> >>>        int ret;
> >>>        unsigned long flags;
> >>> @@ -376,7 +378,8 @@ int pci_epc_map_addr(struct pci_epc *epc, u8 func_no,
> >>>                return 0;
> >>>
> >>>        spin_lock_irqsave(&epc->lock, flags);
> >>> -     ret = epc->ops->map_addr(epc, func_no, phys_addr, pci_addr, size);
> >>> +     ret = epc->ops->map_addr(epc, func_no, phys_addr,
> >>> +                              window, pci_addr, size);
> >>>        spin_unlock_irqrestore(&epc->lock, flags);
> >>>
> >>>        return ret;
> >>> diff --git a/drivers/pci/endpoint/pci-epc-mem.c b/drivers/pci/endpoint/pci-epc-mem.c
> >>> index d2b174ce15de..c955f2c97944 100644
> >>> --- a/drivers/pci/endpoint/pci-epc-mem.c
> >>> +++ b/drivers/pci/endpoint/pci-epc-mem.c
> >>> @@ -39,56 +39,77 @@ static int pci_epc_mem_get_order(struct pci_epc_mem *mem, size_t size)
> >>>    * __pci_epc_mem_init() - initialize the pci_epc_mem structure
> >>>    * @epc: the EPC device that invoked pci_epc_mem_init
> >>>    * @phys_base: the physical address of the base
> >>> - * @size: the size of the address space
> >>> + * @num_windows: number of windows device supports
> >>
> >> struct pci_epc_mem_window is missing here.
> > oops my bad will fix that.
> >
> >>>    * @page_size: size of each page
> >>>    *
> >>>    * Invoke to initialize the pci_epc_mem structure used by the
> >>>    * endpoint functions to allocate mapped PCI address.
> >>>    */
> >>> -int __pci_epc_mem_init(struct pci_epc *epc, phys_addr_t phys_base, size_t size,
> >>> -                    size_t page_size)
> >>> +int __pci_epc_mem_init(struct pci_epc *epc, struct pci_epc_mem_window *windows,
> >>> +                    int num_windows, size_t page_size)
> >>>   {
> >>> -     int ret;
> >>> -     struct pci_epc_mem *mem;
> >>> -     unsigned long *bitmap;
> >>> +     struct pci_epc_mem *mem = NULL;
> >>> +     unsigned long *bitmap = NULL;
> >>>        unsigned int page_shift;
> >>> -     int pages;
> >>>        int bitmap_size;
> >>> +     int pages;
> >>> +     int ret;
> >>> +     int i;
> >>> +
> >>> +     epc->mem_windows = 0;
> >>> +
> >>> +     if (!windows)
> >>> +             return -EINVAL;
> >>> +
> >>> +     if (num_windows <= 0)
> >>> +             return -EINVAL;
> >>>
> >>>        if (page_size < PAGE_SIZE)
> >>>                page_size = PAGE_SIZE;
> >>>
> >>>        page_shift = ilog2(page_size);
> >>> -     pages = size >> page_shift;
> >>> -     bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
> >>>
> >>> -     mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> >>> -     if (!mem) {
> >>> -             ret = -ENOMEM;
> >>> -             goto err;
> >>> -     }
> >>> +     epc->mem = kcalloc(num_windows, sizeof(*mem), GFP_KERNEL);
> >>> +     if (!epc->mem)
> >>> +             return -EINVAL;
> >>>
> >>> -     bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> >>> -     if (!bitmap) {
> >>> -             ret = -ENOMEM;
> >>> -             goto err_mem;
> >>> -     }
> >>> +     for (i = 0; i < num_windows; i++) {
> >>> +             pages = windows[i].phys_base >> page_shift;
> >>> +             bitmap_size = BITS_TO_LONGS(pages) * sizeof(long);
> >>>
> >>> -     mem->bitmap = bitmap;
> >>> -     mem->phys_base = phys_base;
> >>> -     mem->page_size = page_size;
> >>> -     mem->pages = pages;
> >>> -     mem->size = size;
> >>> +             mem = kzalloc(sizeof(*mem), GFP_KERNEL);
> >>> +             if (!mem) {
> >>> +                     ret = -ENOMEM;
> >>> +                     goto err_mem;
> >>> +             }
> >>>
> >>> -     epc->mem = mem;
> >>> +             bitmap = kzalloc(bitmap_size, GFP_KERNEL);
> >>> +             if (!bitmap) {
> >>> +                     ret = -ENOMEM;
> >>> +                     goto err_mem;
> >>> +             }
> >>> +
> >>> +             mem->bitmap = bitmap;
> >>> +             mem->window.phys_base = windows[i].phys_base;
> >>> +             mem->page_size = page_size;
> >>> +             mem->pages = pages;
> >>> +             mem->window.size = windows[i].size;
> >>> +             mem->window.map_size = 0;
> >>> +
> >>> +             epc->mem[i] = mem;
> >>> +     }
> >>> +     epc->mem_windows = num_windows;
> >>>
> >>>        return 0;
> >>>
> >>>   err_mem:
> >>> -     kfree(mem);
> >>> +     for (; i >= 0; i--) {
> >>> +             kfree(mem->bitmap);
> >>> +             kfree(epc->mem[i]);
> >>> +     }
> >>> +     kfree(epc->mem);
> >>>
> >>> -err:
> >>> -return ret;
> >>> +     return ret;
> >>>   }
> >>>   EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
> >>>
> >>> @@ -101,48 +122,126 @@ EXPORT_SYMBOL_GPL(__pci_epc_mem_init);
> >>>    */
> >>>   void pci_epc_mem_exit(struct pci_epc *epc)
> >>>   {
> >>> -     struct pci_epc_mem *mem = epc->mem;
> >>> +     struct pci_epc_mem *mem;
> >>> +     int i;
> >>> +
> >>> +     if (!epc->mem_windows)
> >>> +             return;
> >>> +
> >>> +     for (i = 0; i <= epc->mem_windows; i--) {
> >>> +             mem = epc->mem[i];
> >>> +             kfree(mem->bitmap);
> >>> +             kfree(epc->mem[i]);
> >>> +     }
> >>> +     kfree(epc->mem);
> >>>
> >>>        epc->mem = NULL;
> >>> -     kfree(mem->bitmap);
> >>> -     kfree(mem);
> >>> +     epc->mem_windows = 0;
> >>>   }
> >>>   EXPORT_SYMBOL_GPL(pci_epc_mem_exit);
> >>>
> >>> +static int pci_epc_find_best_fit_window(struct pci_epc *epc, size_t size)
> >>> +{
> >>> +     size_t window_least_size = 0;
> >>> +     int best_fit_window = -1;
> >>> +     struct pci_epc_mem *mem;
> >>> +     size_t actual_size;
> >>> +     int i;
> >>> +
> >>> +     for (i = 0; i < epc->mem_windows; i++) {
^^
here we loop through all the possible windows.

> >>> +             mem = epc->mem[i];
> >>> +
> >>> +             /* if chunk from this region is already used skip it */
> >>> +             if (mem->window.map_size)
> >>> +                     continue;
> >>> +
^^
If a chunk is used in the window we skip it.
but the next version of the patchest  will check for  the flag if the
controller supports multiple allocation of chunks
in same window.

> >>> +             actual_size = ALIGN(size, mem->page_size);
> >>> +
> >>> +             if (best_fit_window == -1) {
> >>> +                     best_fit_window = i;
> >>> +                     window_least_size = mem->window.size;
firstly it assigns a first free region, although the above code doesnt
check if requested size can be fitted in the window,
I'll fix that in next version as below :

if (actual_size <= mem->window.size) {
                    best_fit_window = i;
                    window_least_size = mem->window.size;
}

window_least_size  is temp variable to compare against other window sizes.

> >>> +             } else {
> >>> +                     if (actual_size <= mem->window.size &&
> >>> +                         mem->window.size < window_least_size) {
> >>> +                             best_fit_window = i;
> >>> +                             window_least_size = mem->window.size;
> >>> +                     }

add in above we check if the requested size does fit in the window,
and the windows size is lesser then
the previously assigned one, if the criteria matches we choose this
window. and this is looped through all
the windows.

> >>> +             }
> >>> +     }
> >>> +
> >>> +     return best_fit_window;
> >>> +}
> >>> +
> >>>   /**
> >>>    * pci_epc_mem_alloc_addr() - allocate memory address from EPC addr space
> >>>    * @epc: the EPC device on which memory has to be allocated
> >>>    * @phys_addr: populate the allocated physical address here
> >>> + * @window: populate the window here which will be used to map PCI address
> >>>    * @size: the size of the address space that has to be allocated
> >>>    *
> >>>    * Invoke to allocate memory address from the EPC address space. This
> >>>    * is usually done to map the remote RC address into the local system.
> >>>    */
> >>>   void __iomem *pci_epc_mem_alloc_addr(struct pci_epc *epc,
> >>> -                                  phys_addr_t *phys_addr, size_t size)
> >>> +                                  phys_addr_t *phys_addr,
> >>> +                                  int *window, size_t size)
> >>>   {
> >>> +     int best_fit = PCI_EPC_DEFAULT_WINDOW;
> >>> +     void __iomem *virt_addr = NULL;
> >>> +     struct pci_epc_mem *mem;
> >>> +     unsigned int page_shift;
> >>>        int pageno;
> >>> -     void __iomem *virt_addr;
> >>> -     struct pci_epc_mem *mem = epc->mem;
> >>> -     unsigned int page_shift = ilog2(mem->page_size);
> >>>        int order;
> >>>
> >>> +     if (epc->mem_windows <= 0)
> >>> +             return NULL;
> >>> +
> >>> +     if (epc->mem_windows > 1) {
> >>> +             best_fit = pci_epc_find_best_fit_window(epc, size);
> >>> +             if (best_fit < 0)
> >>> +                     return NULL;
> >>> +     }
> >>> +
> >>> +     mem = epc->mem[best_fit];
> >>>        size = ALIGN(size, mem->page_size);
> >>> +     if (size > (mem->window.size - mem->window.map_size))
> >>> +             return NULL;
> >>
> >> Assume I have two mem regions, first region is of size 128MB and the second
> >> region is of size 4GB. If there is a allocation request for 4GB, will the
> >> allocation succeed?
> >>
> > yes it would succeed, the pci_epc_find_best_fit_window() would get the
> > corresponding window.
>
> Can you step through the algo again? pci_epc_find_best_fit_window() will
> return the first region which is of size 128MB.
>
done above as requested.

Cheers,
--Prabhakar

> Thanks
> Kishon

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 60+ messages in thread

end of thread, other threads:[~2019-12-06 13:09 UTC | newest]

Thread overview: 60+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-11-06 19:36 [PATCH 0/5] Add support for PCIe controller to work in endpoint mode on R-Car SoCs Lad Prabhakar
2019-11-06 19:36 ` Lad Prabhakar
2019-11-06 19:36 ` [PATCH 1/5] pci: pcie-rcar: preparation for adding endpoint support Lad Prabhakar
2019-11-06 19:36   ` Lad Prabhakar
2019-11-27  4:55   ` Kishon Vijay Abraham I
2019-11-27  4:55     ` Kishon Vijay Abraham I
2019-11-27 20:51     ` Lad, Prabhakar
2019-11-27 20:51       ` Lad, Prabhakar
2019-11-06 19:36 ` [PATCH 2/5] pci: endpoint: add support to handle multiple base for mapping outbound memory Lad Prabhakar
2019-11-06 19:36   ` Lad Prabhakar
2019-11-27  5:14   ` Kishon Vijay Abraham I
2019-11-27  5:14     ` Kishon Vijay Abraham I
2019-11-27  5:14     ` Kishon Vijay Abraham I
2019-11-27 21:21     ` Lad, Prabhakar
2019-11-27 21:21       ` Lad, Prabhakar
2019-11-27 21:21       ` Lad, Prabhakar
2019-12-05 10:22       ` Kishon Vijay Abraham I
2019-12-05 10:22         ` Kishon Vijay Abraham I
2019-12-05 10:22         ` Kishon Vijay Abraham I
2019-12-06 13:08         ` Lad, Prabhakar
2019-12-06 13:08           ` Lad, Prabhakar
2019-12-06 13:08           ` Lad, Prabhakar
2019-11-27  6:05   ` Kishon Vijay Abraham I
2019-11-27  6:05     ` Kishon Vijay Abraham I
2019-11-27  6:05     ` Kishon Vijay Abraham I
2019-11-27 21:28     ` Lad, Prabhakar
2019-11-27 21:28       ` Lad, Prabhakar
2019-11-27 21:28       ` Lad, Prabhakar
2019-11-06 19:36 ` [PATCH 3/5] PCI: rcar: Add R-Car PCIe endpoint device tree bindings Lad Prabhakar
2019-11-06 19:36   ` Lad Prabhakar
2019-11-07  7:39   ` Biju Das
2019-11-07  7:39     ` Biju Das
2019-11-07  8:10     ` Lad, Prabhakar
2019-11-07  8:10       ` Lad, Prabhakar
2019-11-07  8:44   ` Geert Uytterhoeven
2019-11-07  8:44     ` Geert Uytterhoeven
2019-11-07  9:25     ` Lad, Prabhakar
2019-11-07  9:25       ` Lad, Prabhakar
2019-11-07 20:08       ` Geert Uytterhoeven
2019-11-07 20:08         ` Geert Uytterhoeven
2019-11-07 22:46         ` Lad, Prabhakar
2019-11-07 22:46           ` Lad, Prabhakar
2019-11-08  8:36           ` Geert Uytterhoeven
2019-11-08  8:36             ` Geert Uytterhoeven
2019-11-13  4:08         ` Rob Herring
2019-11-13  4:08           ` Rob Herring
2019-11-27  5:44           ` Kishon Vijay Abraham I
2019-11-27  5:44             ` Kishon Vijay Abraham I
2019-11-27 21:00             ` Lad, Prabhakar
2019-11-27 21:00               ` Lad, Prabhakar
2019-11-06 19:36 ` [PATCH 4/5] pci: rcar: add support for rcar pcie controller in endpoint mode Lad Prabhakar
2019-11-06 19:36   ` Lad Prabhakar
2019-11-06 19:36 ` [PATCH 5/5] misc: pci_endpoint_test: add device-id for RZ/G2 pcie controller Lad Prabhakar
2019-11-06 19:36   ` Lad Prabhakar
2019-11-27  5:45   ` Kishon Vijay Abraham I
2019-11-27  5:45     ` Kishon Vijay Abraham I
2019-11-26 14:33 ` [PATCH 0/5] Add support for PCIe controller to work in endpoint mode on R-Car SoCs Lad, Prabhakar
2019-11-26 14:33   ` Lad, Prabhakar
2019-11-27  6:07   ` Kishon Vijay Abraham I
2019-11-27  6:07     ` Kishon Vijay Abraham I

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.