[PATCH v2 0/5] Altera PCIe host controller driver with MSI support

[PATCH v2 0/5] Altera PCIe host controller driver with MSI support

[Ley Foon Tan]

This is the 2nd version of patch set to add support for Altera PCIe host
controller with MSI feature on Altera FPGA device families. This patchset
have resolved all comments in v1. Thanks everyone for reviewing.

It is based on patch series from Marc Zyngier "Per-device MSI domain &
platform MSI" [1] to get rid of struct msi_controller.

[1]: https://lkml.org/lkml/2015/7/23/172

v1->v2 changes:
- arm: socfpga: remove ununsed ARCH_SUPPORTS_MSI and PCI from Kconfig
- pci: host: remove depends on OF from Kconfig
- pci: host: remove depends PCIE_ALTERA from Kconfig
- pci: host: change pcie-altera to tristate in Kconfig
- pcie-altera: remove unnecessary newlines and .owner
- pcie-altera: remove altera_pcie_map_irq
- pcie-altera: remove dependency on hw_pci and use pci_scan_root_bus
- pcie-altera-msi: change to use phys_addr_t
- pcie-altera-msi: change to use irq_set_chained_handler
- pcie-altera-msi: add inner_domain
- pcie-altera-msi: add add MSI_FLAG_PCI_MSIX flag
- pcie-altera-msi: return -EINVAL for altera_msi_set_affinity
- pcie-altera-msi: return PTR_ERR() instead of -EADDRNOTAVAIL
- change to use write_relaxed and read_relaxed

History:
-------
[v1]: https://lkml.org/lkml/2015/7/28/395

Ley Foon Tan (5):
  arm: add msi.h to Kbuild
  pci:host: Add Altera PCIe host controller driver
  pci: altera: Add Altera PCIe MSI driver
  Documentation: dt-bindings: pci: altera pcie device tree binding
  MAINTAINERS: Add Altera PCIe and MSI drivers maintainer

 .../devicetree/bindings/pci/altera-pcie-msi.txt    |  27 ++
 .../devicetree/bindings/pci/altera-pcie.txt        |  49 ++
 MAINTAINERS                                        |  16 +
 arch/arm/include/asm/Kbuild                        |   1 +
 drivers/pci/host/Kconfig                           |  15 +
 drivers/pci/host/Makefile                          |   2 +
 drivers/pci/host/pcie-altera-msi.c                 | 324 +++++++++++++
 drivers/pci/host/pcie-altera.c                     | 526 +++++++++++++++++++++
 8 files changed, 960 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/pci/altera-pcie-msi.txt
 create mode 100644 Documentation/devicetree/bindings/pci/altera-pcie.txt
 create mode 100644 drivers/pci/host/pcie-altera-msi.c
 create mode 100644 drivers/pci/host/pcie-altera.c

-- 
1.8.2.1

[PATCH v2 0/5] Altera PCIe host controller driver with MSI support

[Ley Foon Tan]

This is the 2nd version of patch set to add support for Altera PCIe host
controller with MSI feature on Altera FPGA device families. This patchset
have resolved all comments in v1. Thanks everyone for reviewing.

It is based on patch series from Marc Zyngier "Per-device MSI domain &
platform MSI" [1] to get rid of struct msi_controller.

[1]: https://lkml.org/lkml/2015/7/23/172

v1->v2 changes:
- arm: socfpga: remove ununsed ARCH_SUPPORTS_MSI and PCI from Kconfig
- pci: host: remove depends on OF from Kconfig
- pci: host: remove depends PCIE_ALTERA from Kconfig
- pci: host: change pcie-altera to tristate in Kconfig
- pcie-altera: remove unnecessary newlines and .owner
- pcie-altera: remove altera_pcie_map_irq
- pcie-altera: remove dependency on hw_pci and use pci_scan_root_bus
- pcie-altera-msi: change to use phys_addr_t
- pcie-altera-msi: change to use irq_set_chained_handler
- pcie-altera-msi: add inner_domain
- pcie-altera-msi: add add MSI_FLAG_PCI_MSIX flag
- pcie-altera-msi: return -EINVAL for altera_msi_set_affinity
- pcie-altera-msi: return PTR_ERR() instead of -EADDRNOTAVAIL
- change to use write_relaxed and read_relaxed

History:
-------
[v1]: https://lkml.org/lkml/2015/7/28/395

Ley Foon Tan (5):
  arm: add msi.h to Kbuild
  pci:host: Add Altera PCIe host controller driver
  pci: altera: Add Altera PCIe MSI driver
  Documentation: dt-bindings: pci: altera pcie device tree binding
  MAINTAINERS: Add Altera PCIe and MSI drivers maintainer

 .../devicetree/bindings/pci/altera-pcie-msi.txt    |  27 ++
 .../devicetree/bindings/pci/altera-pcie.txt        |  49 ++
 MAINTAINERS                                        |  16 +
 arch/arm/include/asm/Kbuild                        |   1 +
 drivers/pci/host/Kconfig                           |  15 +
 drivers/pci/host/Makefile                          |   2 +
 drivers/pci/host/pcie-altera-msi.c                 | 324 +++++++++++++
 drivers/pci/host/pcie-altera.c                     | 526 +++++++++++++++++++++
 8 files changed, 960 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/pci/altera-pcie-msi.txt
 create mode 100644 Documentation/devicetree/bindings/pci/altera-pcie.txt
 create mode 100644 drivers/pci/host/pcie-altera-msi.c
 create mode 100644 drivers/pci/host/pcie-altera.c

-- 
1.8.2.1

[PATCH v2 0/5] Altera PCIe host controller driver with MSI support

[Ley Foon Tan]

This is the 2nd version of patch set to add support for Altera PCIe host
controller with MSI feature on Altera FPGA device families. This patchset
have resolved all comments in v1. Thanks everyone for reviewing.

It is based on patch series from Marc Zyngier "Per-device MSI domain &
platform MSI" [1] to get rid of struct msi_controller.

[1]: https://lkml.org/lkml/2015/7/23/172

v1->v2 changes:
- arm: socfpga: remove ununsed ARCH_SUPPORTS_MSI and PCI from Kconfig
- pci: host: remove depends on OF from Kconfig
- pci: host: remove depends PCIE_ALTERA from Kconfig
- pci: host: change pcie-altera to tristate in Kconfig
- pcie-altera: remove unnecessary newlines and .owner
- pcie-altera: remove altera_pcie_map_irq
- pcie-altera: remove dependency on hw_pci and use pci_scan_root_bus
- pcie-altera-msi: change to use phys_addr_t
- pcie-altera-msi: change to use irq_set_chained_handler
- pcie-altera-msi: add inner_domain
- pcie-altera-msi: add add MSI_FLAG_PCI_MSIX flag
- pcie-altera-msi: return -EINVAL for altera_msi_set_affinity
- pcie-altera-msi: return PTR_ERR() instead of -EADDRNOTAVAIL
- change to use write_relaxed and read_relaxed

History:
-------
[v1]: https://lkml.org/lkml/2015/7/28/395

Ley Foon Tan (5):
  arm: add msi.h to Kbuild
  pci:host: Add Altera PCIe host controller driver
  pci: altera: Add Altera PCIe MSI driver
  Documentation: dt-bindings: pci: altera pcie device tree binding
  MAINTAINERS: Add Altera PCIe and MSI drivers maintainer

 .../devicetree/bindings/pci/altera-pcie-msi.txt    |  27 ++
 .../devicetree/bindings/pci/altera-pcie.txt        |  49 ++
 MAINTAINERS                                        |  16 +
 arch/arm/include/asm/Kbuild                        |   1 +
 drivers/pci/host/Kconfig                           |  15 +
 drivers/pci/host/Makefile                          |   2 +
 drivers/pci/host/pcie-altera-msi.c                 | 324 +++++++++++++
 drivers/pci/host/pcie-altera.c                     | 526 +++++++++++++++++++++
 8 files changed, 960 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/pci/altera-pcie-msi.txt
 create mode 100644 Documentation/devicetree/bindings/pci/altera-pcie.txt
 create mode 100644 drivers/pci/host/pcie-altera-msi.c
 create mode 100644 drivers/pci/host/pcie-altera.c

-- 
1.8.2.1

[PATCH v2 1/5] arm: add msi.h to Kbuild

[Ley Foon Tan]

Include asm-generic/msi.h to support CONFIG_GENERIC_MSI_IRQ_DOMAIN.
This to fix compilation error:
"include/linux/msi.h:123:21: fatal error: asm/msi.h:
No such file or directory"

Signed-off-by: Ley Foon Tan <lftan@altera.com>
---
 arch/arm/include/asm/Kbuild | 1 +
 1 file changed, 1 insertion(+)

diff --git a/arch/arm/include/asm/Kbuild b/arch/arm/include/asm/Kbuild
index 83c5019..ddc2761 100644
--- a/arch/arm/include/asm/Kbuild
+++ b/arch/arm/include/asm/Kbuild
@@ -14,6 +14,7 @@ generic-y += local.h
 generic-y += local64.h
 generic-y += mcs_spinlock.h
 generic-y += msgbuf.h
+generic-y += msi.h
 generic-y += param.h
 generic-y += parport.h
 generic-y += poll.h
-- 
1.8.2.1

[PATCH v2 1/5] arm: add msi.h to Kbuild

[Ley Foon Tan]

Include asm-generic/msi.h to support CONFIG_GENERIC_MSI_IRQ_DOMAIN.
This to fix compilation error:
"include/linux/msi.h:123:21: fatal error: asm/msi.h:
No such file or directory"

Signed-off-by: Ley Foon Tan <lftan@altera.com>
---
 arch/arm/include/asm/Kbuild | 1 +
 1 file changed, 1 insertion(+)

diff --git a/arch/arm/include/asm/Kbuild b/arch/arm/include/asm/Kbuild
index 83c5019..ddc2761 100644
--- a/arch/arm/include/asm/Kbuild
+++ b/arch/arm/include/asm/Kbuild
@@ -14,6 +14,7 @@ generic-y += local.h
 generic-y += local64.h
 generic-y += mcs_spinlock.h
 generic-y += msgbuf.h
+generic-y += msi.h
 generic-y += param.h
 generic-y += parport.h
 generic-y += poll.h
-- 
1.8.2.1

[PATCH v2 1/5] arm: add msi.h to Kbuild

[Ley Foon Tan]

Include asm-generic/msi.h to support CONFIG_GENERIC_MSI_IRQ_DOMAIN.
This to fix compilation error:
"include/linux/msi.h:123:21: fatal error: asm/msi.h:
No such file or directory"

Signed-off-by: Ley Foon Tan <lftan@altera.com>
---
 arch/arm/include/asm/Kbuild | 1 +
 1 file changed, 1 insertion(+)

diff --git a/arch/arm/include/asm/Kbuild b/arch/arm/include/asm/Kbuild
index 83c5019..ddc2761 100644
--- a/arch/arm/include/asm/Kbuild
+++ b/arch/arm/include/asm/Kbuild
@@ -14,6 +14,7 @@ generic-y += local.h
 generic-y += local64.h
 generic-y += mcs_spinlock.h
 generic-y += msgbuf.h
+generic-y += msi.h
 generic-y += param.h
 generic-y += parport.h
 generic-y += poll.h
-- 
1.8.2.1

[PATCH v2 2/5] pci:host: Add Altera PCIe host controller driver

[Ley Foon Tan]

This patch adds the Altera PCIe host controller driver.

Signed-off-by: Ley Foon Tan <lftan@altera.com>
---
 drivers/pci/host/Kconfig       |   8 +
 drivers/pci/host/Makefile      |   1 +
 drivers/pci/host/pcie-altera.c | 526 +++++++++++++++++++++++++++++++++++++++++
 3 files changed, 535 insertions(+)
 create mode 100644 drivers/pci/host/pcie-altera.c

diff --git a/drivers/pci/host/Kconfig b/drivers/pci/host/Kconfig
index 675c2d1..108500a 100644
--- a/drivers/pci/host/Kconfig
+++ b/drivers/pci/host/Kconfig
@@ -145,4 +145,12 @@ config PCIE_IPROC_BCMA
 	  Say Y here if you want to use the Broadcom iProc PCIe controller
 	  through the BCMA bus interface
 
+config PCIE_ALTERA
+	tristate "Altera PCIe controller"
+	depends on ARCH_SOCFPGA
+	select PCI_MSI_IRQ_DOMAIN if PCI_MSI
+	help
+	  Say Y here if you want to enable PCIe controller support for Altera
+	  SoCFPGA family of SoCs.
+
 endmenu
diff --git a/drivers/pci/host/Makefile b/drivers/pci/host/Makefile
index 140d66f..6954f76 100644
--- a/drivers/pci/host/Makefile
+++ b/drivers/pci/host/Makefile
@@ -17,3 +17,4 @@ obj-$(CONFIG_PCI_VERSATILE) += pci-versatile.o
 obj-$(CONFIG_PCIE_IPROC) += pcie-iproc.o
 obj-$(CONFIG_PCIE_IPROC_PLATFORM) += pcie-iproc-platform.o
 obj-$(CONFIG_PCIE_IPROC_BCMA) += pcie-iproc-bcma.o
+obj-$(CONFIG_PCIE_ALTERA) += pcie-altera.o
diff --git a/drivers/pci/host/pcie-altera.c b/drivers/pci/host/pcie-altera.c
new file mode 100644
index 0000000..3ec51c7
--- /dev/null
+++ b/drivers/pci/host/pcie-altera.c
@@ -0,0 +1,526 @@
+/*
+ * Copyright Altera Corporation (C) 2013-2015. All rights reserved
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+#include <linux/interrupt.h>
+#include <linux/module.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/of_pci.h>
+#include <linux/pci.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+
+#define A2P_ADDR_MAP_LO0		0x1000
+#define A2P_ADDR_MAP_HI0		0x1004
+#define RP_TX_REG0			0x2000
+#define RP_TX_REG1			0x2004
+#define RP_TX_CNTRL			0x2008
+#define RP_TX_EOP			0x2
+#define RP_TX_SOP			0x1
+#define RP_RXCPL_STATUS		0x2010
+#define RP_RXCPL_EOP			0x2
+#define RP_RXCPL_SOP			0x1
+#define RP_RXCPL_REG0			0x2014
+#define RP_RXCPL_REG1			0x2018
+#define P2A_INT_STATUS			0x3060
+#define P2A_INT_STS_ALL			0xF
+#define P2A_INT_ENABLE			0x3070
+#define P2A_INT_ENA_ALL			0xF
+#define RP_LTSSM			0x3C64
+#define LTSSM_L0			0xF
+
+/* TLP configuration type 0 and 1 */
+#define TLP_FMTTYPE_CFGRD0		0x04	/* Configuration Read Type 0 */
+#define TLP_FMTTYPE_CFGWR0		0x44	/* Configuration Write Type 0 */
+#define TLP_FMTTYPE_CFGRD1		0x05	/* Configuration Read Type 1 */
+#define TLP_FMTTYPE_CFGWR1		0x45	/* Configuration Write Type 1 */
+#define TLP_PAYLOAD_SIZE		0x01
+#define TLP_READ_TAG			0x1D
+#define TLP_WRITE_TAG			0x10
+#define TLP_CFG_DW0(fmttype)		(((fmttype) << 24) | TLP_PAYLOAD_SIZE)
+#define TLP_CFG_DW1(reqid, tag)		(((reqid) << 16) | (tag << 8) | 0xF)
+#define TLP_CFG_DW2(bus, devfn, offset)	\
+				(((bus) << 24) | ((devfn) << 16) | (offset))
+#define TLP_REQ_ID(bus, devfn)		(((bus) << 8) | (devfn))
+#define TLP_COMPL_STATUS(hdr)		(((hdr) & 0xE0) >> 13)
+#define TLP_HDR_SIZE			3
+#define TLP_LOOP			10
+
+#define INTX_NUM		4
+
+/* Address translation table entry size */
+#define ATT_ENTRY_SIZE		8
+
+#define DWORD_MASK		3
+
+struct altera_pcie {
+	struct platform_device	*pdev;
+	struct resource		*txs;
+	void __iomem		*cra_base;
+	int			hwirq;
+	u8			root_bus_nr;
+	struct irq_domain		*irq_domain;
+	struct resource		bus_range;
+	struct list_head		resources;
+};
+
+struct tlp_rp_regpair_t {
+	u32 ctrl;
+	u32 reg0;
+	u32 reg1;
+};
+
+static void altera_pcie_retrain(struct pci_dev *dev)
+{
+	u16 linkcap, linkstat;
+
+	/*
+	 * Set the retrain bit if the PCIe rootport support > 2.5GB/s, but
+	 * current speed is 2.5 GB/s.
+	 */
+	pcie_capability_read_word(dev, PCI_EXP_LNKCAP, &linkcap);
+
+	if ((linkcap & PCI_EXP_LNKCAP_SLS) <= PCI_EXP_LNKCAP_SLS_2_5GB)
+		return;
+
+	pcie_capability_read_word(dev, PCI_EXP_LNKSTA, &linkstat);
+	if ((linkstat & PCI_EXP_LNKSTA_CLS) == PCI_EXP_LNKSTA_CLS_2_5GB)
+		pcie_capability_set_word(dev, PCI_EXP_LNKCTL,
+					PCI_EXP_LNKCTL_RL);
+
+}
+DECLARE_PCI_FIXUP_EARLY(PCI_ANY_ID, PCI_ANY_ID, altera_pcie_retrain);
+
+static inline void cra_writel(struct altera_pcie *pcie, u32 value, u32 reg)
+{
+	writel_relaxed(value, pcie->cra_base + reg);
+}
+
+static inline u32 cra_readl(struct altera_pcie *pcie, u32 reg)
+{
+	return readl_relaxed(pcie->cra_base + reg);
+}
+
+static void tlp_read_rx(struct altera_pcie *pcie,
+			struct tlp_rp_regpair_t *tlp_rp_regdata)
+{
+	tlp_rp_regdata->ctrl = cra_readl(pcie, RP_RXCPL_STATUS);
+	tlp_rp_regdata->reg0 = cra_readl(pcie, RP_RXCPL_REG0);
+	tlp_rp_regdata->reg1 = cra_readl(pcie, RP_RXCPL_REG1);
+}
+
+static void tlp_write_tx(struct altera_pcie *pcie,
+			struct tlp_rp_regpair_t *tlp_rp_regdata)
+{
+	cra_writel(pcie, tlp_rp_regdata->reg0, RP_TX_REG0);
+	cra_writel(pcie, tlp_rp_regdata->reg1, RP_TX_REG1);
+	cra_writel(pcie, tlp_rp_regdata->ctrl, RP_TX_CNTRL);
+}
+
+static bool altera_pcie_link_is_up(struct altera_pcie *pcie)
+{
+	return !!(cra_readl(pcie, RP_LTSSM) & LTSSM_L0);
+}
+
+static bool altera_pcie_valid_config(struct altera_pcie *pcie,
+				struct pci_bus *bus, int dev)
+{
+	/* If there is no link, then there is no device */
+	if (bus->number != pcie->root_bus_nr) {
+		if (!altera_pcie_link_is_up(pcie))
+			return false;
+	}
+
+	/* access only one slot on each root port */
+	if (bus->number == pcie->root_bus_nr && dev > 0)
+		return false;
+
+	/*
+	 * Do not read more than one device on the bus directly attached
+	 * to RC.
+	 */
+	if (bus->primary == pcie->root_bus_nr && dev > 0)
+		return false;
+
+	 return true;
+}
+
+static int tlp_read_packet(struct altera_pcie *pcie, u32 *value)
+{
+	u8 loop;
+	struct tlp_rp_regpair_t tlp_rp_regdata;
+
+	for (loop = TLP_LOOP; loop > 0; loop--) {
+		tlp_read_rx(pcie, &tlp_rp_regdata);
+		if (tlp_rp_regdata.ctrl & RP_RXCPL_EOP) {
+			if (value)
+				*value = tlp_rp_regdata.reg0;
+			return PCIBIOS_SUCCESSFUL;
+		}
+	}
+
+	return -ENOENT;
+}
+
+static void tlp_write_packet(struct altera_pcie *pcie, u32 *headers, u32 data)
+{
+	struct tlp_rp_regpair_t tlp_rp_regdata;
+
+	tlp_rp_regdata.reg0 = headers[0];
+	tlp_rp_regdata.reg1 = headers[1];
+	tlp_rp_regdata.ctrl = RP_TX_SOP;
+	tlp_write_tx(pcie, &tlp_rp_regdata);
+
+	tlp_rp_regdata.reg0 = headers[2];
+	tlp_rp_regdata.reg1 = data;
+	tlp_rp_regdata.ctrl = RP_TX_EOP;
+	tlp_write_tx(pcie, &tlp_rp_regdata);
+}
+
+static int tlp_cfg_dword_read(struct altera_pcie *pcie, u8 bus, u32 devfn,
+			      int where, u32 *value)
+{
+	int ret;
+	u32 headers[TLP_HDR_SIZE];
+
+	if (bus == pcie->root_bus_nr)
+		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGRD0);
+	else
+		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGRD1);
+
+	headers[1] = TLP_CFG_DW1(TLP_REQ_ID(pcie->root_bus_nr, devfn),
+					TLP_READ_TAG);
+	headers[2] = TLP_CFG_DW2(bus, devfn, where);
+
+	tlp_write_packet(pcie, headers, 0);
+
+	ret = tlp_read_packet(pcie, value);
+	if (ret)
+		*value = ~0UL;	/* return 0xFFFFFFFF if error */
+
+	return ret;
+}
+
+static int tlp_cfg_dword_write(struct altera_pcie *pcie, u8 bus, u32 devfn,
+			       int where, u32 value)
+{
+	u32 headers[TLP_HDR_SIZE];
+
+	if (bus == pcie->root_bus_nr)
+		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGWR0);
+	else
+		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGWR1);
+
+	headers[1] = TLP_CFG_DW1(TLP_REQ_ID(pcie->root_bus_nr, devfn),
+					TLP_WRITE_TAG);
+	headers[2] = TLP_CFG_DW2(bus, devfn, where);
+
+	tlp_write_packet(pcie, headers, value);
+
+	tlp_read_packet(pcie, NULL);
+
+	/* Keep an eye out for changes to the root bus number */
+	if ((bus == pcie->root_bus_nr) && (where == PCI_PRIMARY_BUS))
+		pcie->root_bus_nr = (u8)(value);
+
+	return PCIBIOS_SUCCESSFUL;
+}
+
+static int altera_pcie_cfg_read(struct pci_bus *bus, unsigned int devfn,
+			       int where, int size, u32 *value)
+{
+	struct altera_pcie *pcie = bus->sysdata;
+	int ret;
+
+	if (!altera_pcie_valid_config(pcie, bus, PCI_SLOT(devfn))) {
+		*value = ~0UL;
+		return PCIBIOS_DEVICE_NOT_FOUND;
+	}
+
+	ret = tlp_cfg_dword_read(pcie, bus->number, devfn,
+		(where & ~DWORD_MASK), value);
+	if (ret != PCIBIOS_SUCCESSFUL)
+		return ret;
+
+	if (size == 1)
+		*value = (*value >> (8 * (where & 0x3))) & 0xff;
+	else if (size == 2)
+		*value = (*value >> (8 * (where & 0x2))) & 0xffff;
+
+	return PCIBIOS_SUCCESSFUL;
+}
+
+static int altera_pcie_cfg_write(struct pci_bus *bus, unsigned int devfn,
+				int where, int size, u32 value)
+{
+	struct altera_pcie *pcie = bus->sysdata;
+	u32 data32 = value;
+	u32 shift = 8 * (where & 3);
+	int ret;
+
+	if (!altera_pcie_valid_config(pcie, bus, PCI_SLOT(devfn)))
+		return PCIBIOS_DEVICE_NOT_FOUND;
+
+	/* write partial */
+	if (size != sizeof(u32)) {
+		ret = tlp_cfg_dword_read(pcie, bus->number, devfn,
+			where & ~DWORD_MASK, &data32);
+		if (ret)
+			return ret;
+
+		if (size == 2)
+			data32 = (data32 & ~(0xffff << shift)) |
+					((value & 0xffff) << shift);
+		 else if (size == 1)
+			data32 = (data32 & ~(0xff << shift)) |
+					((value & 0xff) << shift);
+	}
+
+	return tlp_cfg_dword_write(pcie, bus->number, devfn,
+		(where & ~DWORD_MASK), data32);
+}
+
+static struct pci_ops altera_pcie_ops = {
+	.read = altera_pcie_cfg_read,
+	.write = altera_pcie_cfg_write,
+};
+
+static int altera_pcie_intx_map(struct irq_domain *domain, unsigned int irq,
+				irq_hw_number_t hwirq)
+{
+	irq_set_chip_and_handler(irq, &dummy_irq_chip, handle_simple_irq);
+	irq_set_chip_data(irq, domain->host_data);
+	set_irq_flags(irq, IRQF_VALID);
+
+	return 0;
+}
+
+static const struct irq_domain_ops intx_domain_ops = {
+	.map = altera_pcie_intx_map,
+};
+
+static irqreturn_t altera_pcie_isr(int irq, void *arg)
+{
+	struct altera_pcie *pcie = arg;
+	u32 status, i;
+
+	status = cra_readl(pcie, P2A_INT_STATUS) & P2A_INT_STS_ALL;
+
+	/* clear interrupts */
+	cra_writel(pcie, status, P2A_INT_STATUS);
+
+	for (i = 0; i < INTX_NUM; i++) {
+		if (status & (1 << i))
+			generic_handle_irq(irq_find_mapping(pcie->irq_domain,
+							i + 1));
+	}
+
+	return IRQ_HANDLED;
+}
+
+static void altera_pcie_release_of_pci_ranges(struct altera_pcie *pcie)
+{
+	pci_free_resource_list(&pcie->resources);
+}
+
+static int altera_pcie_parse_request_of_pci_ranges(struct altera_pcie *pcie)
+{
+	int err, res_valid = 0;
+	struct device *dev = &pcie->pdev->dev;
+	struct device_node *np = dev->of_node;
+	resource_size_t iobase;
+	struct resource_entry *win;
+	int offset = 0;
+
+	err = of_pci_get_host_bridge_resources(np, 0, 0xff, &pcie->resources,
+					       &iobase);
+	if (err)
+		return err;
+
+	resource_list_for_each_entry(win, &pcie->resources) {
+		struct resource *parent, *res = win->res;
+
+		switch (resource_type(res)) {
+		case IORESOURCE_MEM:
+			parent = &iomem_resource;
+			res_valid |= !(res->flags & IORESOURCE_PREFETCH);
+			cra_writel(pcie, res->start,
+				A2P_ADDR_MAP_LO0 + offset);
+			cra_writel(pcie, 0,
+				A2P_ADDR_MAP_HI0 + offset);
+			offset += ATT_ENTRY_SIZE;
+			break;
+		default:
+			continue;
+		}
+
+		err = devm_request_resource(dev, parent, res);
+		if (err)
+			goto out_release_res;
+	}
+
+	if (!res_valid) {
+		dev_err(dev, "non-prefetchable memory resource required\n");
+		err = -EINVAL;
+		goto out_release_res;
+	}
+
+	return 0;
+
+out_release_res:
+	altera_pcie_release_of_pci_ranges(pcie);
+	return err;
+}
+
+static void altera_pcie_free_irq_domain(struct altera_pcie *pcie)
+{
+	int i;
+	u32 irq;
+
+	for (i = 0; i < INTX_NUM; i++) {
+		irq = irq_find_mapping(pcie->irq_domain, i);
+		if (irq > 0)
+			irq_dispose_mapping(irq);
+	}
+
+	irq_domain_remove(pcie->irq_domain);
+}
+
+static int altera_pcie_init_irq_domain(struct altera_pcie *pcie)
+{
+	struct device *dev = &pcie->pdev->dev;
+	struct device_node *node = dev->of_node;
+
+	/* Setup INTx */
+	pcie->irq_domain = irq_domain_add_linear(node, INTX_NUM,
+					&intx_domain_ops, pcie);
+	if (!pcie->irq_domain) {
+		dev_err(dev, "Failed to get a INTx IRQ domain\n");
+		return PTR_ERR(pcie->irq_domain);
+	}
+
+	return 0;
+}
+
+static int altera_pcie_parse_dt(struct altera_pcie *pcie)
+{
+	struct resource *cra;
+	int ret;
+	struct platform_device *pdev = pcie->pdev;
+
+	cra = platform_get_resource_byname(pdev, IORESOURCE_MEM, "Cra");
+	pcie->cra_base = devm_ioremap_resource(&pdev->dev, cra);
+	if (IS_ERR(pcie->cra_base)) {
+		dev_err(&pdev->dev, "get Cra resource failed\n");
+		return PTR_ERR(pcie->cra_base);
+	}
+
+	/* setup IRQ */
+	pcie->hwirq = platform_get_irq(pdev, 0);
+	if (pcie->hwirq <= 0) {
+		dev_err(&pdev->dev, "failed to get IRQ: %d\n", pcie->hwirq);
+		return -EINVAL;
+	}
+	ret = devm_request_irq(&pdev->dev, pcie->hwirq, altera_pcie_isr,
+			IRQF_SHARED, pdev->name, pcie);
+
+	if (ret) {
+		dev_err(&pdev->dev, "failed to request irq %d\n", pcie->hwirq);
+		return ret;
+	}
+
+	return 0;
+}
+
+static int altera_pcie_probe(struct platform_device *pdev)
+{
+	struct altera_pcie *pcie;
+	struct pci_bus *bus;
+	int ret;
+
+	pcie = devm_kzalloc(&pdev->dev, sizeof(*pcie), GFP_KERNEL);
+	if (!pcie)
+		return -ENOMEM;
+
+	pcie->pdev = pdev;
+
+	ret = altera_pcie_parse_dt(pcie);
+	if (ret) {
+		dev_err(&pdev->dev, "Parsing DT failed\n");
+		return ret;
+	}
+
+	INIT_LIST_HEAD(&pcie->resources);
+
+	ret = altera_pcie_parse_request_of_pci_ranges(pcie);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed add resources\n");
+		return ret;
+	}
+
+	ret = altera_pcie_init_irq_domain(pcie);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed creating IRQ Domain\n");
+		return ret;
+	}
+
+	pcie->root_bus_nr = 0;
+
+	/* clear all interrupts */
+	cra_writel(pcie, P2A_INT_STS_ALL, P2A_INT_STATUS);
+	/* enable all interrupts */
+	cra_writel(pcie, P2A_INT_ENA_ALL, P2A_INT_ENABLE);
+
+	bus = pci_scan_root_bus(&pdev->dev, pcie->root_bus_nr, &altera_pcie_ops,
+					pcie, &pcie->resources);
+	if (!bus)
+		return -ENOMEM;
+
+	pci_assign_unassigned_bus_resources(bus);
+	pci_bus_add_devices(bus);
+
+	platform_set_drvdata(pdev, pcie);
+	return ret;
+}
+
+static int __exit altera_pcie_remove(struct platform_device *pdev)
+{
+	struct altera_pcie *pcie = platform_get_drvdata(pdev);
+
+	altera_pcie_free_irq_domain(pcie);
+	platform_set_drvdata(pdev, NULL);
+	return 0;
+}
+
+static const struct of_device_id altera_pcie_of_match[] = {
+	{ .compatible = "altr,pcie-root-port-1.0", },
+	{},
+};
+MODULE_DEVICE_TABLE(of, altera_pcie_of_match);
+
+static struct platform_driver altera_pcie_driver = {
+	.probe		= altera_pcie_probe,
+	.remove		= altera_pcie_remove,
+	.driver = {
+		.name	= "altera-pcie",
+		.of_match_table = altera_pcie_of_match,
+	},
+};
+
+module_platform_driver(altera_pcie_driver);
+
+MODULE_AUTHOR("Ley Foon Tan <lftan@altera.com>");
+MODULE_DESCRIPTION("Altera PCIe host controller driver");
+MODULE_LICENSE("GPL v2");
-- 
1.8.2.1

[PATCH v2 2/5] pci:host: Add Altera PCIe host controller driver

[Ley Foon Tan]

This patch adds the Altera PCIe host controller driver.

Signed-off-by: Ley Foon Tan <lftan@altera.com>
---
 drivers/pci/host/Kconfig       |   8 +
 drivers/pci/host/Makefile      |   1 +
 drivers/pci/host/pcie-altera.c | 526 +++++++++++++++++++++++++++++++++++++++++
 3 files changed, 535 insertions(+)
 create mode 100644 drivers/pci/host/pcie-altera.c

diff --git a/drivers/pci/host/Kconfig b/drivers/pci/host/Kconfig
index 675c2d1..108500a 100644
--- a/drivers/pci/host/Kconfig
+++ b/drivers/pci/host/Kconfig
@@ -145,4 +145,12 @@ config PCIE_IPROC_BCMA
 	  Say Y here if you want to use the Broadcom iProc PCIe controller
 	  through the BCMA bus interface
 
+config PCIE_ALTERA
+	tristate "Altera PCIe controller"
+	depends on ARCH_SOCFPGA
+	select PCI_MSI_IRQ_DOMAIN if PCI_MSI
+	help
+	  Say Y here if you want to enable PCIe controller support for Altera
+	  SoCFPGA family of SoCs.
+
 endmenu
diff --git a/drivers/pci/host/Makefile b/drivers/pci/host/Makefile
index 140d66f..6954f76 100644
--- a/drivers/pci/host/Makefile
+++ b/drivers/pci/host/Makefile
@@ -17,3 +17,4 @@ obj-$(CONFIG_PCI_VERSATILE) += pci-versatile.o
 obj-$(CONFIG_PCIE_IPROC) += pcie-iproc.o
 obj-$(CONFIG_PCIE_IPROC_PLATFORM) += pcie-iproc-platform.o
 obj-$(CONFIG_PCIE_IPROC_BCMA) += pcie-iproc-bcma.o
+obj-$(CONFIG_PCIE_ALTERA) += pcie-altera.o
diff --git a/drivers/pci/host/pcie-altera.c b/drivers/pci/host/pcie-altera.c
new file mode 100644
index 0000000..3ec51c7
--- /dev/null
+++ b/drivers/pci/host/pcie-altera.c
@@ -0,0 +1,526 @@
+/*
+ * Copyright Altera Corporation (C) 2013-2015. All rights reserved
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+#include <linux/interrupt.h>
+#include <linux/module.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/of_pci.h>
+#include <linux/pci.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+
+#define A2P_ADDR_MAP_LO0		0x1000
+#define A2P_ADDR_MAP_HI0		0x1004
+#define RP_TX_REG0			0x2000
+#define RP_TX_REG1			0x2004
+#define RP_TX_CNTRL			0x2008
+#define RP_TX_EOP			0x2
+#define RP_TX_SOP			0x1
+#define RP_RXCPL_STATUS		0x2010
+#define RP_RXCPL_EOP			0x2
+#define RP_RXCPL_SOP			0x1
+#define RP_RXCPL_REG0			0x2014
+#define RP_RXCPL_REG1			0x2018
+#define P2A_INT_STATUS			0x3060
+#define P2A_INT_STS_ALL			0xF
+#define P2A_INT_ENABLE			0x3070
+#define P2A_INT_ENA_ALL			0xF
+#define RP_LTSSM			0x3C64
+#define LTSSM_L0			0xF
+
+/* TLP configuration type 0 and 1 */
+#define TLP_FMTTYPE_CFGRD0		0x04	/* Configuration Read Type 0 */
+#define TLP_FMTTYPE_CFGWR0		0x44	/* Configuration Write Type 0 */
+#define TLP_FMTTYPE_CFGRD1		0x05	/* Configuration Read Type 1 */
+#define TLP_FMTTYPE_CFGWR1		0x45	/* Configuration Write Type 1 */
+#define TLP_PAYLOAD_SIZE		0x01
+#define TLP_READ_TAG			0x1D
+#define TLP_WRITE_TAG			0x10
+#define TLP_CFG_DW0(fmttype)		(((fmttype) << 24) | TLP_PAYLOAD_SIZE)
+#define TLP_CFG_DW1(reqid, tag)		(((reqid) << 16) | (tag << 8) | 0xF)
+#define TLP_CFG_DW2(bus, devfn, offset)	\
+				(((bus) << 24) | ((devfn) << 16) | (offset))
+#define TLP_REQ_ID(bus, devfn)		(((bus) << 8) | (devfn))
+#define TLP_COMPL_STATUS(hdr)		(((hdr) & 0xE0) >> 13)
+#define TLP_HDR_SIZE			3
+#define TLP_LOOP			10
+
+#define INTX_NUM		4
+
+/* Address translation table entry size */
+#define ATT_ENTRY_SIZE		8
+
+#define DWORD_MASK		3
+
+struct altera_pcie {
+	struct platform_device	*pdev;
+	struct resource		*txs;
+	void __iomem		*cra_base;
+	int			hwirq;
+	u8			root_bus_nr;
+	struct irq_domain		*irq_domain;
+	struct resource		bus_range;
+	struct list_head		resources;
+};
+
+struct tlp_rp_regpair_t {
+	u32 ctrl;
+	u32 reg0;
+	u32 reg1;
+};
+
+static void altera_pcie_retrain(struct pci_dev *dev)
+{
+	u16 linkcap, linkstat;
+
+	/*
+	 * Set the retrain bit if the PCIe rootport support > 2.5GB/s, but
+	 * current speed is 2.5 GB/s.
+	 */
+	pcie_capability_read_word(dev, PCI_EXP_LNKCAP, &linkcap);
+
+	if ((linkcap & PCI_EXP_LNKCAP_SLS) <= PCI_EXP_LNKCAP_SLS_2_5GB)
+		return;
+
+	pcie_capability_read_word(dev, PCI_EXP_LNKSTA, &linkstat);
+	if ((linkstat & PCI_EXP_LNKSTA_CLS) == PCI_EXP_LNKSTA_CLS_2_5GB)
+		pcie_capability_set_word(dev, PCI_EXP_LNKCTL,
+					PCI_EXP_LNKCTL_RL);
+
+}
+DECLARE_PCI_FIXUP_EARLY(PCI_ANY_ID, PCI_ANY_ID, altera_pcie_retrain);
+
+static inline void cra_writel(struct altera_pcie *pcie, u32 value, u32 reg)
+{
+	writel_relaxed(value, pcie->cra_base + reg);
+}
+
+static inline u32 cra_readl(struct altera_pcie *pcie, u32 reg)
+{
+	return readl_relaxed(pcie->cra_base + reg);
+}
+
+static void tlp_read_rx(struct altera_pcie *pcie,
+			struct tlp_rp_regpair_t *tlp_rp_regdata)
+{
+	tlp_rp_regdata->ctrl = cra_readl(pcie, RP_RXCPL_STATUS);
+	tlp_rp_regdata->reg0 = cra_readl(pcie, RP_RXCPL_REG0);
+	tlp_rp_regdata->reg1 = cra_readl(pcie, RP_RXCPL_REG1);
+}
+
+static void tlp_write_tx(struct altera_pcie *pcie,
+			struct tlp_rp_regpair_t *tlp_rp_regdata)
+{
+	cra_writel(pcie, tlp_rp_regdata->reg0, RP_TX_REG0);
+	cra_writel(pcie, tlp_rp_regdata->reg1, RP_TX_REG1);
+	cra_writel(pcie, tlp_rp_regdata->ctrl, RP_TX_CNTRL);
+}
+
+static bool altera_pcie_link_is_up(struct altera_pcie *pcie)
+{
+	return !!(cra_readl(pcie, RP_LTSSM) & LTSSM_L0);
+}
+
+static bool altera_pcie_valid_config(struct altera_pcie *pcie,
+				struct pci_bus *bus, int dev)
+{
+	/* If there is no link, then there is no device */
+	if (bus->number != pcie->root_bus_nr) {
+		if (!altera_pcie_link_is_up(pcie))
+			return false;
+	}
+
+	/* access only one slot on each root port */
+	if (bus->number == pcie->root_bus_nr && dev > 0)
+		return false;
+
+	/*
+	 * Do not read more than one device on the bus directly attached
+	 * to RC.
+	 */
+	if (bus->primary == pcie->root_bus_nr && dev > 0)
+		return false;
+
+	 return true;
+}
+
+static int tlp_read_packet(struct altera_pcie *pcie, u32 *value)
+{
+	u8 loop;
+	struct tlp_rp_regpair_t tlp_rp_regdata;
+
+	for (loop = TLP_LOOP; loop > 0; loop--) {
+		tlp_read_rx(pcie, &tlp_rp_regdata);
+		if (tlp_rp_regdata.ctrl & RP_RXCPL_EOP) {
+			if (value)
+				*value = tlp_rp_regdata.reg0;
+			return PCIBIOS_SUCCESSFUL;
+		}
+	}
+
+	return -ENOENT;
+}
+
+static void tlp_write_packet(struct altera_pcie *pcie, u32 *headers, u32 data)
+{
+	struct tlp_rp_regpair_t tlp_rp_regdata;
+
+	tlp_rp_regdata.reg0 = headers[0];
+	tlp_rp_regdata.reg1 = headers[1];
+	tlp_rp_regdata.ctrl = RP_TX_SOP;
+	tlp_write_tx(pcie, &tlp_rp_regdata);
+
+	tlp_rp_regdata.reg0 = headers[2];
+	tlp_rp_regdata.reg1 = data;
+	tlp_rp_regdata.ctrl = RP_TX_EOP;
+	tlp_write_tx(pcie, &tlp_rp_regdata);
+}
+
+static int tlp_cfg_dword_read(struct altera_pcie *pcie, u8 bus, u32 devfn,
+			      int where, u32 *value)
+{
+	int ret;
+	u32 headers[TLP_HDR_SIZE];
+
+	if (bus == pcie->root_bus_nr)
+		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGRD0);
+	else
+		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGRD1);
+
+	headers[1] = TLP_CFG_DW1(TLP_REQ_ID(pcie->root_bus_nr, devfn),
+					TLP_READ_TAG);
+	headers[2] = TLP_CFG_DW2(bus, devfn, where);
+
+	tlp_write_packet(pcie, headers, 0);
+
+	ret = tlp_read_packet(pcie, value);
+	if (ret)
+		*value = ~0UL;	/* return 0xFFFFFFFF if error */
+
+	return ret;
+}
+
+static int tlp_cfg_dword_write(struct altera_pcie *pcie, u8 bus, u32 devfn,
+			       int where, u32 value)
+{
+	u32 headers[TLP_HDR_SIZE];
+
+	if (bus == pcie->root_bus_nr)
+		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGWR0);
+	else
+		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGWR1);
+
+	headers[1] = TLP_CFG_DW1(TLP_REQ_ID(pcie->root_bus_nr, devfn),
+					TLP_WRITE_TAG);
+	headers[2] = TLP_CFG_DW2(bus, devfn, where);
+
+	tlp_write_packet(pcie, headers, value);
+
+	tlp_read_packet(pcie, NULL);
+
+	/* Keep an eye out for changes to the root bus number */
+	if ((bus == pcie->root_bus_nr) && (where == PCI_PRIMARY_BUS))
+		pcie->root_bus_nr = (u8)(value);
+
+	return PCIBIOS_SUCCESSFUL;
+}
+
+static int altera_pcie_cfg_read(struct pci_bus *bus, unsigned int devfn,
+			       int where, int size, u32 *value)
+{
+	struct altera_pcie *pcie = bus->sysdata;
+	int ret;
+
+	if (!altera_pcie_valid_config(pcie, bus, PCI_SLOT(devfn))) {
+		*value = ~0UL;
+		return PCIBIOS_DEVICE_NOT_FOUND;
+	}
+
+	ret = tlp_cfg_dword_read(pcie, bus->number, devfn,
+		(where & ~DWORD_MASK), value);
+	if (ret != PCIBIOS_SUCCESSFUL)
+		return ret;
+
+	if (size == 1)
+		*value = (*value >> (8 * (where & 0x3))) & 0xff;
+	else if (size == 2)
+		*value = (*value >> (8 * (where & 0x2))) & 0xffff;
+
+	return PCIBIOS_SUCCESSFUL;
+}
+
+static int altera_pcie_cfg_write(struct pci_bus *bus, unsigned int devfn,
+				int where, int size, u32 value)
+{
+	struct altera_pcie *pcie = bus->sysdata;
+	u32 data32 = value;
+	u32 shift = 8 * (where & 3);
+	int ret;
+
+	if (!altera_pcie_valid_config(pcie, bus, PCI_SLOT(devfn)))
+		return PCIBIOS_DEVICE_NOT_FOUND;
+
+	/* write partial */
+	if (size != sizeof(u32)) {
+		ret = tlp_cfg_dword_read(pcie, bus->number, devfn,
+			where & ~DWORD_MASK, &data32);
+		if (ret)
+			return ret;
+
+		if (size == 2)
+			data32 = (data32 & ~(0xffff << shift)) |
+					((value & 0xffff) << shift);
+		 else if (size == 1)
+			data32 = (data32 & ~(0xff << shift)) |
+					((value & 0xff) << shift);
+	}
+
+	return tlp_cfg_dword_write(pcie, bus->number, devfn,
+		(where & ~DWORD_MASK), data32);
+}
+
+static struct pci_ops altera_pcie_ops = {
+	.read = altera_pcie_cfg_read,
+	.write = altera_pcie_cfg_write,
+};
+
+static int altera_pcie_intx_map(struct irq_domain *domain, unsigned int irq,
+				irq_hw_number_t hwirq)
+{
+	irq_set_chip_and_handler(irq, &dummy_irq_chip, handle_simple_irq);
+	irq_set_chip_data(irq, domain->host_data);
+	set_irq_flags(irq, IRQF_VALID);
+
+	return 0;
+}
+
+static const struct irq_domain_ops intx_domain_ops = {
+	.map = altera_pcie_intx_map,
+};
+
+static irqreturn_t altera_pcie_isr(int irq, void *arg)
+{
+	struct altera_pcie *pcie = arg;
+	u32 status, i;
+
+	status = cra_readl(pcie, P2A_INT_STATUS) & P2A_INT_STS_ALL;
+
+	/* clear interrupts */
+	cra_writel(pcie, status, P2A_INT_STATUS);
+
+	for (i = 0; i < INTX_NUM; i++) {
+		if (status & (1 << i))
+			generic_handle_irq(irq_find_mapping(pcie->irq_domain,
+							i + 1));
+	}
+
+	return IRQ_HANDLED;
+}
+
+static void altera_pcie_release_of_pci_ranges(struct altera_pcie *pcie)
+{
+	pci_free_resource_list(&pcie->resources);
+}
+
+static int altera_pcie_parse_request_of_pci_ranges(struct altera_pcie *pcie)
+{
+	int err, res_valid = 0;
+	struct device *dev = &pcie->pdev->dev;
+	struct device_node *np = dev->of_node;
+	resource_size_t iobase;
+	struct resource_entry *win;
+	int offset = 0;
+
+	err = of_pci_get_host_bridge_resources(np, 0, 0xff, &pcie->resources,
+					       &iobase);
+	if (err)
+		return err;
+
+	resource_list_for_each_entry(win, &pcie->resources) {
+		struct resource *parent, *res = win->res;
+
+		switch (resource_type(res)) {
+		case IORESOURCE_MEM:
+			parent = &iomem_resource;
+			res_valid |= !(res->flags & IORESOURCE_PREFETCH);
+			cra_writel(pcie, res->start,
+				A2P_ADDR_MAP_LO0 + offset);
+			cra_writel(pcie, 0,
+				A2P_ADDR_MAP_HI0 + offset);
+			offset += ATT_ENTRY_SIZE;
+			break;
+		default:
+			continue;
+		}
+
+		err = devm_request_resource(dev, parent, res);
+		if (err)
+			goto out_release_res;
+	}
+
+	if (!res_valid) {
+		dev_err(dev, "non-prefetchable memory resource required\n");
+		err = -EINVAL;
+		goto out_release_res;
+	}
+
+	return 0;
+
+out_release_res:
+	altera_pcie_release_of_pci_ranges(pcie);
+	return err;
+}
+
+static void altera_pcie_free_irq_domain(struct altera_pcie *pcie)
+{
+	int i;
+	u32 irq;
+
+	for (i = 0; i < INTX_NUM; i++) {
+		irq = irq_find_mapping(pcie->irq_domain, i);
+		if (irq > 0)
+			irq_dispose_mapping(irq);
+	}
+
+	irq_domain_remove(pcie->irq_domain);
+}
+
+static int altera_pcie_init_irq_domain(struct altera_pcie *pcie)
+{
+	struct device *dev = &pcie->pdev->dev;
+	struct device_node *node = dev->of_node;
+
+	/* Setup INTx */
+	pcie->irq_domain = irq_domain_add_linear(node, INTX_NUM,
+					&intx_domain_ops, pcie);
+	if (!pcie->irq_domain) {
+		dev_err(dev, "Failed to get a INTx IRQ domain\n");
+		return PTR_ERR(pcie->irq_domain);
+	}
+
+	return 0;
+}
+
+static int altera_pcie_parse_dt(struct altera_pcie *pcie)
+{
+	struct resource *cra;
+	int ret;
+	struct platform_device *pdev = pcie->pdev;
+
+	cra = platform_get_resource_byname(pdev, IORESOURCE_MEM, "Cra");
+	pcie->cra_base = devm_ioremap_resource(&pdev->dev, cra);
+	if (IS_ERR(pcie->cra_base)) {
+		dev_err(&pdev->dev, "get Cra resource failed\n");
+		return PTR_ERR(pcie->cra_base);
+	}
+
+	/* setup IRQ */
+	pcie->hwirq = platform_get_irq(pdev, 0);
+	if (pcie->hwirq <= 0) {
+		dev_err(&pdev->dev, "failed to get IRQ: %d\n", pcie->hwirq);
+		return -EINVAL;
+	}
+	ret = devm_request_irq(&pdev->dev, pcie->hwirq, altera_pcie_isr,
+			IRQF_SHARED, pdev->name, pcie);
+
+	if (ret) {
+		dev_err(&pdev->dev, "failed to request irq %d\n", pcie->hwirq);
+		return ret;
+	}
+
+	return 0;
+}
+
+static int altera_pcie_probe(struct platform_device *pdev)
+{
+	struct altera_pcie *pcie;
+	struct pci_bus *bus;
+	int ret;
+
+	pcie = devm_kzalloc(&pdev->dev, sizeof(*pcie), GFP_KERNEL);
+	if (!pcie)
+		return -ENOMEM;
+
+	pcie->pdev = pdev;
+
+	ret = altera_pcie_parse_dt(pcie);
+	if (ret) {
+		dev_err(&pdev->dev, "Parsing DT failed\n");
+		return ret;
+	}
+
+	INIT_LIST_HEAD(&pcie->resources);
+
+	ret = altera_pcie_parse_request_of_pci_ranges(pcie);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed add resources\n");
+		return ret;
+	}
+
+	ret = altera_pcie_init_irq_domain(pcie);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed creating IRQ Domain\n");
+		return ret;
+	}
+
+	pcie->root_bus_nr = 0;
+
+	/* clear all interrupts */
+	cra_writel(pcie, P2A_INT_STS_ALL, P2A_INT_STATUS);
+	/* enable all interrupts */
+	cra_writel(pcie, P2A_INT_ENA_ALL, P2A_INT_ENABLE);
+
+	bus = pci_scan_root_bus(&pdev->dev, pcie->root_bus_nr, &altera_pcie_ops,
+					pcie, &pcie->resources);
+	if (!bus)
+		return -ENOMEM;
+
+	pci_assign_unassigned_bus_resources(bus);
+	pci_bus_add_devices(bus);
+
+	platform_set_drvdata(pdev, pcie);
+	return ret;
+}
+
+static int __exit altera_pcie_remove(struct platform_device *pdev)
+{
+	struct altera_pcie *pcie = platform_get_drvdata(pdev);
+
+	altera_pcie_free_irq_domain(pcie);
+	platform_set_drvdata(pdev, NULL);
+	return 0;
+}
+
+static const struct of_device_id altera_pcie_of_match[] = {
+	{ .compatible = "altr,pcie-root-port-1.0", },
+	{},
+};
+MODULE_DEVICE_TABLE(of, altera_pcie_of_match);
+
+static struct platform_driver altera_pcie_driver = {
+	.probe		= altera_pcie_probe,
+	.remove		= altera_pcie_remove,
+	.driver = {
+		.name	= "altera-pcie",
+		.of_match_table = altera_pcie_of_match,
+	},
+};
+
+module_platform_driver(altera_pcie_driver);
+
+MODULE_AUTHOR("Ley Foon Tan <lftan@altera.com>");
+MODULE_DESCRIPTION("Altera PCIe host controller driver");
+MODULE_LICENSE("GPL v2");
-- 
1.8.2.1

[PATCH v2 2/5] pci:host: Add Altera PCIe host controller driver

[Ley Foon Tan]

This patch adds the Altera PCIe host controller driver.

Signed-off-by: Ley Foon Tan <lftan@altera.com>
---
 drivers/pci/host/Kconfig       |   8 +
 drivers/pci/host/Makefile      |   1 +
 drivers/pci/host/pcie-altera.c | 526 +++++++++++++++++++++++++++++++++++++++++
 3 files changed, 535 insertions(+)
 create mode 100644 drivers/pci/host/pcie-altera.c

diff --git a/drivers/pci/host/Kconfig b/drivers/pci/host/Kconfig
index 675c2d1..108500a 100644
--- a/drivers/pci/host/Kconfig
+++ b/drivers/pci/host/Kconfig
@@ -145,4 +145,12 @@ config PCIE_IPROC_BCMA
 	  Say Y here if you want to use the Broadcom iProc PCIe controller
 	  through the BCMA bus interface
 
+config PCIE_ALTERA
+	tristate "Altera PCIe controller"
+	depends on ARCH_SOCFPGA
+	select PCI_MSI_IRQ_DOMAIN if PCI_MSI
+	help
+	  Say Y here if you want to enable PCIe controller support for Altera
+	  SoCFPGA family of SoCs.
+
 endmenu
diff --git a/drivers/pci/host/Makefile b/drivers/pci/host/Makefile
index 140d66f..6954f76 100644
--- a/drivers/pci/host/Makefile
+++ b/drivers/pci/host/Makefile
@@ -17,3 +17,4 @@ obj-$(CONFIG_PCI_VERSATILE) += pci-versatile.o
 obj-$(CONFIG_PCIE_IPROC) += pcie-iproc.o
 obj-$(CONFIG_PCIE_IPROC_PLATFORM) += pcie-iproc-platform.o
 obj-$(CONFIG_PCIE_IPROC_BCMA) += pcie-iproc-bcma.o
+obj-$(CONFIG_PCIE_ALTERA) += pcie-altera.o
diff --git a/drivers/pci/host/pcie-altera.c b/drivers/pci/host/pcie-altera.c
new file mode 100644
index 0000000..3ec51c7
--- /dev/null
+++ b/drivers/pci/host/pcie-altera.c
@@ -0,0 +1,526 @@
+/*
+ * Copyright Altera Corporation (C) 2013-2015. All rights reserved
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+#include <linux/interrupt.h>
+#include <linux/module.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/of_pci.h>
+#include <linux/pci.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+
+#define A2P_ADDR_MAP_LO0		0x1000
+#define A2P_ADDR_MAP_HI0		0x1004
+#define RP_TX_REG0			0x2000
+#define RP_TX_REG1			0x2004
+#define RP_TX_CNTRL			0x2008
+#define RP_TX_EOP			0x2
+#define RP_TX_SOP			0x1
+#define RP_RXCPL_STATUS		0x2010
+#define RP_RXCPL_EOP			0x2
+#define RP_RXCPL_SOP			0x1
+#define RP_RXCPL_REG0			0x2014
+#define RP_RXCPL_REG1			0x2018
+#define P2A_INT_STATUS			0x3060
+#define P2A_INT_STS_ALL			0xF
+#define P2A_INT_ENABLE			0x3070
+#define P2A_INT_ENA_ALL			0xF
+#define RP_LTSSM			0x3C64
+#define LTSSM_L0			0xF
+
+/* TLP configuration type 0 and 1 */
+#define TLP_FMTTYPE_CFGRD0		0x04	/* Configuration Read Type 0 */
+#define TLP_FMTTYPE_CFGWR0		0x44	/* Configuration Write Type 0 */
+#define TLP_FMTTYPE_CFGRD1		0x05	/* Configuration Read Type 1 */
+#define TLP_FMTTYPE_CFGWR1		0x45	/* Configuration Write Type 1 */
+#define TLP_PAYLOAD_SIZE		0x01
+#define TLP_READ_TAG			0x1D
+#define TLP_WRITE_TAG			0x10
+#define TLP_CFG_DW0(fmttype)		(((fmttype) << 24) | TLP_PAYLOAD_SIZE)
+#define TLP_CFG_DW1(reqid, tag)		(((reqid) << 16) | (tag << 8) | 0xF)
+#define TLP_CFG_DW2(bus, devfn, offset)	\
+				(((bus) << 24) | ((devfn) << 16) | (offset))
+#define TLP_REQ_ID(bus, devfn)		(((bus) << 8) | (devfn))
+#define TLP_COMPL_STATUS(hdr)		(((hdr) & 0xE0) >> 13)
+#define TLP_HDR_SIZE			3
+#define TLP_LOOP			10
+
+#define INTX_NUM		4
+
+/* Address translation table entry size */
+#define ATT_ENTRY_SIZE		8
+
+#define DWORD_MASK		3
+
+struct altera_pcie {
+	struct platform_device	*pdev;
+	struct resource		*txs;
+	void __iomem		*cra_base;
+	int			hwirq;
+	u8			root_bus_nr;
+	struct irq_domain		*irq_domain;
+	struct resource		bus_range;
+	struct list_head		resources;
+};
+
+struct tlp_rp_regpair_t {
+	u32 ctrl;
+	u32 reg0;
+	u32 reg1;
+};
+
+static void altera_pcie_retrain(struct pci_dev *dev)
+{
+	u16 linkcap, linkstat;
+
+	/*
+	 * Set the retrain bit if the PCIe rootport support > 2.5GB/s, but
+	 * current speed is 2.5 GB/s.
+	 */
+	pcie_capability_read_word(dev, PCI_EXP_LNKCAP, &linkcap);
+
+	if ((linkcap & PCI_EXP_LNKCAP_SLS) <= PCI_EXP_LNKCAP_SLS_2_5GB)
+		return;
+
+	pcie_capability_read_word(dev, PCI_EXP_LNKSTA, &linkstat);
+	if ((linkstat & PCI_EXP_LNKSTA_CLS) == PCI_EXP_LNKSTA_CLS_2_5GB)
+		pcie_capability_set_word(dev, PCI_EXP_LNKCTL,
+					PCI_EXP_LNKCTL_RL);
+
+}
+DECLARE_PCI_FIXUP_EARLY(PCI_ANY_ID, PCI_ANY_ID, altera_pcie_retrain);
+
+static inline void cra_writel(struct altera_pcie *pcie, u32 value, u32 reg)
+{
+	writel_relaxed(value, pcie->cra_base + reg);
+}
+
+static inline u32 cra_readl(struct altera_pcie *pcie, u32 reg)
+{
+	return readl_relaxed(pcie->cra_base + reg);
+}
+
+static void tlp_read_rx(struct altera_pcie *pcie,
+			struct tlp_rp_regpair_t *tlp_rp_regdata)
+{
+	tlp_rp_regdata->ctrl = cra_readl(pcie, RP_RXCPL_STATUS);
+	tlp_rp_regdata->reg0 = cra_readl(pcie, RP_RXCPL_REG0);
+	tlp_rp_regdata->reg1 = cra_readl(pcie, RP_RXCPL_REG1);
+}
+
+static void tlp_write_tx(struct altera_pcie *pcie,
+			struct tlp_rp_regpair_t *tlp_rp_regdata)
+{
+	cra_writel(pcie, tlp_rp_regdata->reg0, RP_TX_REG0);
+	cra_writel(pcie, tlp_rp_regdata->reg1, RP_TX_REG1);
+	cra_writel(pcie, tlp_rp_regdata->ctrl, RP_TX_CNTRL);
+}
+
+static bool altera_pcie_link_is_up(struct altera_pcie *pcie)
+{
+	return !!(cra_readl(pcie, RP_LTSSM) & LTSSM_L0);
+}
+
+static bool altera_pcie_valid_config(struct altera_pcie *pcie,
+				struct pci_bus *bus, int dev)
+{
+	/* If there is no link, then there is no device */
+	if (bus->number != pcie->root_bus_nr) {
+		if (!altera_pcie_link_is_up(pcie))
+			return false;
+	}
+
+	/* access only one slot on each root port */
+	if (bus->number == pcie->root_bus_nr && dev > 0)
+		return false;
+
+	/*
+	 * Do not read more than one device on the bus directly attached
+	 * to RC.
+	 */
+	if (bus->primary == pcie->root_bus_nr && dev > 0)
+		return false;
+
+	 return true;
+}
+
+static int tlp_read_packet(struct altera_pcie *pcie, u32 *value)
+{
+	u8 loop;
+	struct tlp_rp_regpair_t tlp_rp_regdata;
+
+	for (loop = TLP_LOOP; loop > 0; loop--) {
+		tlp_read_rx(pcie, &tlp_rp_regdata);
+		if (tlp_rp_regdata.ctrl & RP_RXCPL_EOP) {
+			if (value)
+				*value = tlp_rp_regdata.reg0;
+			return PCIBIOS_SUCCESSFUL;
+		}
+	}
+
+	return -ENOENT;
+}
+
+static void tlp_write_packet(struct altera_pcie *pcie, u32 *headers, u32 data)
+{
+	struct tlp_rp_regpair_t tlp_rp_regdata;
+
+	tlp_rp_regdata.reg0 = headers[0];
+	tlp_rp_regdata.reg1 = headers[1];
+	tlp_rp_regdata.ctrl = RP_TX_SOP;
+	tlp_write_tx(pcie, &tlp_rp_regdata);
+
+	tlp_rp_regdata.reg0 = headers[2];
+	tlp_rp_regdata.reg1 = data;
+	tlp_rp_regdata.ctrl = RP_TX_EOP;
+	tlp_write_tx(pcie, &tlp_rp_regdata);
+}
+
+static int tlp_cfg_dword_read(struct altera_pcie *pcie, u8 bus, u32 devfn,
+			      int where, u32 *value)
+{
+	int ret;
+	u32 headers[TLP_HDR_SIZE];
+
+	if (bus == pcie->root_bus_nr)
+		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGRD0);
+	else
+		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGRD1);
+
+	headers[1] = TLP_CFG_DW1(TLP_REQ_ID(pcie->root_bus_nr, devfn),
+					TLP_READ_TAG);
+	headers[2] = TLP_CFG_DW2(bus, devfn, where);
+
+	tlp_write_packet(pcie, headers, 0);
+
+	ret = tlp_read_packet(pcie, value);
+	if (ret)
+		*value = ~0UL;	/* return 0xFFFFFFFF if error */
+
+	return ret;
+}
+
+static int tlp_cfg_dword_write(struct altera_pcie *pcie, u8 bus, u32 devfn,
+			       int where, u32 value)
+{
+	u32 headers[TLP_HDR_SIZE];
+
+	if (bus == pcie->root_bus_nr)
+		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGWR0);
+	else
+		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGWR1);
+
+	headers[1] = TLP_CFG_DW1(TLP_REQ_ID(pcie->root_bus_nr, devfn),
+					TLP_WRITE_TAG);
+	headers[2] = TLP_CFG_DW2(bus, devfn, where);
+
+	tlp_write_packet(pcie, headers, value);
+
+	tlp_read_packet(pcie, NULL);
+
+	/* Keep an eye out for changes to the root bus number */
+	if ((bus == pcie->root_bus_nr) && (where == PCI_PRIMARY_BUS))
+		pcie->root_bus_nr = (u8)(value);
+
+	return PCIBIOS_SUCCESSFUL;
+}
+
+static int altera_pcie_cfg_read(struct pci_bus *bus, unsigned int devfn,
+			       int where, int size, u32 *value)
+{
+	struct altera_pcie *pcie = bus->sysdata;
+	int ret;
+
+	if (!altera_pcie_valid_config(pcie, bus, PCI_SLOT(devfn))) {
+		*value = ~0UL;
+		return PCIBIOS_DEVICE_NOT_FOUND;
+	}
+
+	ret = tlp_cfg_dword_read(pcie, bus->number, devfn,
+		(where & ~DWORD_MASK), value);
+	if (ret != PCIBIOS_SUCCESSFUL)
+		return ret;
+
+	if (size == 1)
+		*value = (*value >> (8 * (where & 0x3))) & 0xff;
+	else if (size == 2)
+		*value = (*value >> (8 * (where & 0x2))) & 0xffff;
+
+	return PCIBIOS_SUCCESSFUL;
+}
+
+static int altera_pcie_cfg_write(struct pci_bus *bus, unsigned int devfn,
+				int where, int size, u32 value)
+{
+	struct altera_pcie *pcie = bus->sysdata;
+	u32 data32 = value;
+	u32 shift = 8 * (where & 3);
+	int ret;
+
+	if (!altera_pcie_valid_config(pcie, bus, PCI_SLOT(devfn)))
+		return PCIBIOS_DEVICE_NOT_FOUND;
+
+	/* write partial */
+	if (size != sizeof(u32)) {
+		ret = tlp_cfg_dword_read(pcie, bus->number, devfn,
+			where & ~DWORD_MASK, &data32);
+		if (ret)
+			return ret;
+
+		if (size == 2)
+			data32 = (data32 & ~(0xffff << shift)) |
+					((value & 0xffff) << shift);
+		 else if (size == 1)
+			data32 = (data32 & ~(0xff << shift)) |
+					((value & 0xff) << shift);
+	}
+
+	return tlp_cfg_dword_write(pcie, bus->number, devfn,
+		(where & ~DWORD_MASK), data32);
+}
+
+static struct pci_ops altera_pcie_ops = {
+	.read = altera_pcie_cfg_read,
+	.write = altera_pcie_cfg_write,
+};
+
+static int altera_pcie_intx_map(struct irq_domain *domain, unsigned int irq,
+				irq_hw_number_t hwirq)
+{
+	irq_set_chip_and_handler(irq, &dummy_irq_chip, handle_simple_irq);
+	irq_set_chip_data(irq, domain->host_data);
+	set_irq_flags(irq, IRQF_VALID);
+
+	return 0;
+}
+
+static const struct irq_domain_ops intx_domain_ops = {
+	.map = altera_pcie_intx_map,
+};
+
+static irqreturn_t altera_pcie_isr(int irq, void *arg)
+{
+	struct altera_pcie *pcie = arg;
+	u32 status, i;
+
+	status = cra_readl(pcie, P2A_INT_STATUS) & P2A_INT_STS_ALL;
+
+	/* clear interrupts */
+	cra_writel(pcie, status, P2A_INT_STATUS);
+
+	for (i = 0; i < INTX_NUM; i++) {
+		if (status & (1 << i))
+			generic_handle_irq(irq_find_mapping(pcie->irq_domain,
+							i + 1));
+	}
+
+	return IRQ_HANDLED;
+}
+
+static void altera_pcie_release_of_pci_ranges(struct altera_pcie *pcie)
+{
+	pci_free_resource_list(&pcie->resources);
+}
+
+static int altera_pcie_parse_request_of_pci_ranges(struct altera_pcie *pcie)
+{
+	int err, res_valid = 0;
+	struct device *dev = &pcie->pdev->dev;
+	struct device_node *np = dev->of_node;
+	resource_size_t iobase;
+	struct resource_entry *win;
+	int offset = 0;
+
+	err = of_pci_get_host_bridge_resources(np, 0, 0xff, &pcie->resources,
+					       &iobase);
+	if (err)
+		return err;
+
+	resource_list_for_each_entry(win, &pcie->resources) {
+		struct resource *parent, *res = win->res;
+
+		switch (resource_type(res)) {
+		case IORESOURCE_MEM:
+			parent = &iomem_resource;
+			res_valid |= !(res->flags & IORESOURCE_PREFETCH);
+			cra_writel(pcie, res->start,
+				A2P_ADDR_MAP_LO0 + offset);
+			cra_writel(pcie, 0,
+				A2P_ADDR_MAP_HI0 + offset);
+			offset += ATT_ENTRY_SIZE;
+			break;
+		default:
+			continue;
+		}
+
+		err = devm_request_resource(dev, parent, res);
+		if (err)
+			goto out_release_res;
+	}
+
+	if (!res_valid) {
+		dev_err(dev, "non-prefetchable memory resource required\n");
+		err = -EINVAL;
+		goto out_release_res;
+	}
+
+	return 0;
+
+out_release_res:
+	altera_pcie_release_of_pci_ranges(pcie);
+	return err;
+}
+
+static void altera_pcie_free_irq_domain(struct altera_pcie *pcie)
+{
+	int i;
+	u32 irq;
+
+	for (i = 0; i < INTX_NUM; i++) {
+		irq = irq_find_mapping(pcie->irq_domain, i);
+		if (irq > 0)
+			irq_dispose_mapping(irq);
+	}
+
+	irq_domain_remove(pcie->irq_domain);
+}
+
+static int altera_pcie_init_irq_domain(struct altera_pcie *pcie)
+{
+	struct device *dev = &pcie->pdev->dev;
+	struct device_node *node = dev->of_node;
+
+	/* Setup INTx */
+	pcie->irq_domain = irq_domain_add_linear(node, INTX_NUM,
+					&intx_domain_ops, pcie);
+	if (!pcie->irq_domain) {
+		dev_err(dev, "Failed to get a INTx IRQ domain\n");
+		return PTR_ERR(pcie->irq_domain);
+	}
+
+	return 0;
+}
+
+static int altera_pcie_parse_dt(struct altera_pcie *pcie)
+{
+	struct resource *cra;
+	int ret;
+	struct platform_device *pdev = pcie->pdev;
+
+	cra = platform_get_resource_byname(pdev, IORESOURCE_MEM, "Cra");
+	pcie->cra_base = devm_ioremap_resource(&pdev->dev, cra);
+	if (IS_ERR(pcie->cra_base)) {
+		dev_err(&pdev->dev, "get Cra resource failed\n");
+		return PTR_ERR(pcie->cra_base);
+	}
+
+	/* setup IRQ */
+	pcie->hwirq = platform_get_irq(pdev, 0);
+	if (pcie->hwirq <= 0) {
+		dev_err(&pdev->dev, "failed to get IRQ: %d\n", pcie->hwirq);
+		return -EINVAL;
+	}
+	ret = devm_request_irq(&pdev->dev, pcie->hwirq, altera_pcie_isr,
+			IRQF_SHARED, pdev->name, pcie);
+
+	if (ret) {
+		dev_err(&pdev->dev, "failed to request irq %d\n", pcie->hwirq);
+		return ret;
+	}
+
+	return 0;
+}
+
+static int altera_pcie_probe(struct platform_device *pdev)
+{
+	struct altera_pcie *pcie;
+	struct pci_bus *bus;
+	int ret;
+
+	pcie = devm_kzalloc(&pdev->dev, sizeof(*pcie), GFP_KERNEL);
+	if (!pcie)
+		return -ENOMEM;
+
+	pcie->pdev = pdev;
+
+	ret = altera_pcie_parse_dt(pcie);
+	if (ret) {
+		dev_err(&pdev->dev, "Parsing DT failed\n");
+		return ret;
+	}
+
+	INIT_LIST_HEAD(&pcie->resources);
+
+	ret = altera_pcie_parse_request_of_pci_ranges(pcie);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed add resources\n");
+		return ret;
+	}
+
+	ret = altera_pcie_init_irq_domain(pcie);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed creating IRQ Domain\n");
+		return ret;
+	}
+
+	pcie->root_bus_nr = 0;
+
+	/* clear all interrupts */
+	cra_writel(pcie, P2A_INT_STS_ALL, P2A_INT_STATUS);
+	/* enable all interrupts */
+	cra_writel(pcie, P2A_INT_ENA_ALL, P2A_INT_ENABLE);
+
+	bus = pci_scan_root_bus(&pdev->dev, pcie->root_bus_nr, &altera_pcie_ops,
+					pcie, &pcie->resources);
+	if (!bus)
+		return -ENOMEM;
+
+	pci_assign_unassigned_bus_resources(bus);
+	pci_bus_add_devices(bus);
+
+	platform_set_drvdata(pdev, pcie);
+	return ret;
+}
+
+static int __exit altera_pcie_remove(struct platform_device *pdev)
+{
+	struct altera_pcie *pcie = platform_get_drvdata(pdev);
+
+	altera_pcie_free_irq_domain(pcie);
+	platform_set_drvdata(pdev, NULL);
+	return 0;
+}
+
+static const struct of_device_id altera_pcie_of_match[] = {
+	{ .compatible = "altr,pcie-root-port-1.0", },
+	{},
+};
+MODULE_DEVICE_TABLE(of, altera_pcie_of_match);
+
+static struct platform_driver altera_pcie_driver = {
+	.probe		= altera_pcie_probe,
+	.remove		= altera_pcie_remove,
+	.driver = {
+		.name	= "altera-pcie",
+		.of_match_table = altera_pcie_of_match,
+	},
+};
+
+module_platform_driver(altera_pcie_driver);
+
+MODULE_AUTHOR("Ley Foon Tan <lftan@altera.com>");
+MODULE_DESCRIPTION("Altera PCIe host controller driver");
+MODULE_LICENSE("GPL v2");
-- 
1.8.2.1

[PATCH v2 3/5] pci: altera: Add Altera PCIe MSI driver

[Ley Foon Tan]

This patch adds Altera PCIe MSI driver. This soft IP supports configurable
number of vectors, which is a dts parameter.

Signed-off-by: Ley Foon Tan <lftan@altera.com>
---
 drivers/pci/host/Kconfig           |   7 +
 drivers/pci/host/Makefile          |   1 +
 drivers/pci/host/pcie-altera-msi.c | 324 +++++++++++++++++++++++++++++++++++++
 3 files changed, 332 insertions(+)
 create mode 100644 drivers/pci/host/pcie-altera-msi.c

diff --git a/drivers/pci/host/Kconfig b/drivers/pci/host/Kconfig
index 108500a..51d0468 100644
--- a/drivers/pci/host/Kconfig
+++ b/drivers/pci/host/Kconfig
@@ -153,4 +153,11 @@ config PCIE_ALTERA
 	  Say Y here if you want to enable PCIe controller support for Altera
 	  SoCFPGA family of SoCs.
 
+config PCIE_ALTERA_MSI
+	bool "Altera PCIe MSI feature"
+	depends on PCI_MSI
+	help
+	  Say Y here if you want PCIe MSI support for the Altera SocFPGA SoC.
+	  This MSI driver supports Altera MSI to GIC controller IP.
+
 endmenu
diff --git a/drivers/pci/host/Makefile b/drivers/pci/host/Makefile
index 6954f76..6c4913d 100644
--- a/drivers/pci/host/Makefile
+++ b/drivers/pci/host/Makefile
@@ -18,3 +18,4 @@ obj-$(CONFIG_PCIE_IPROC) += pcie-iproc.o
 obj-$(CONFIG_PCIE_IPROC_PLATFORM) += pcie-iproc-platform.o
 obj-$(CONFIG_PCIE_IPROC_BCMA) += pcie-iproc-bcma.o
 obj-$(CONFIG_PCIE_ALTERA) += pcie-altera.o
+obj-$(CONFIG_PCIE_ALTERA_MSI) += pcie-altera-msi.o
diff --git a/drivers/pci/host/pcie-altera-msi.c b/drivers/pci/host/pcie-altera-msi.c
new file mode 100644
index 0000000..6014719
--- /dev/null
+++ b/drivers/pci/host/pcie-altera-msi.c
@@ -0,0 +1,324 @@
+/*
+ * Copyright Altera Corporation (C) 2013-2015. All rights reserved
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+#include <linux/interrupt.h>
+#include <linux/irqchip/chained_irq.h>
+#include <linux/module.h>
+#include <linux/msi.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/of_pci.h>
+#include <linux/pci.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+
+#define MSI_STATUS		0x0
+#define MSI_ERROR		0x4
+#define MSI_INTMASK		0x8
+
+#define MAX_MSI_VECTORS		32
+struct altera_msi {
+	DECLARE_BITMAP(used, MAX_MSI_VECTORS);
+	struct mutex		lock;	/* proctect used variable */
+	struct platform_device	*pdev;
+	struct irq_domain		*msi_domain;
+	struct irq_domain		*inner_domain;
+	void __iomem		*csr_base;
+	void __iomem		*vector_base;
+	phys_addr_t		vector_phy;
+	u32			num_of_vectors;
+	int			irq;
+};
+
+static inline void msi_writel(struct altera_msi *msi, u32 value, u32 reg)
+{
+	writel_relaxed(value, msi->csr_base + reg);
+}
+
+static inline u32 msi_readl(struct altera_msi *msi, u32 reg)
+{
+	return readl_relaxed(msi->csr_base + reg);
+}
+
+static void altera_msi_isr(unsigned int irq, struct irq_desc *desc)
+{
+	struct irq_chip *chip = irq_desc_get_chip(desc);
+	struct altera_msi *msi;
+	unsigned long status;
+	u32 num_of_vectors;
+	u32 bit;
+
+	chained_irq_enter(chip, desc);
+	msi = irq_desc_get_handler_data(desc);
+	num_of_vectors = msi->num_of_vectors;
+
+	do {
+		status = msi_readl(msi, MSI_STATUS);
+		if (!status)
+			break;
+
+		do {
+			bit = find_first_bit(&status, num_of_vectors);
+			/* Dummy read from vector to clear the interrupt */
+			readl_relaxed(msi->vector_base + (bit * sizeof(u32)));
+
+			irq = irq_find_mapping(msi->inner_domain, bit);
+			if (irq) {
+				if (test_bit(bit, msi->used))
+					generic_handle_irq(irq);
+				else
+					dev_info(&msi->pdev->dev, "unhandled MSI\n");
+			} else
+				dev_info(&msi->pdev->dev, "unexpected MSI\n");
+
+			/* Clear the bit from status and repeat without reading
+			 * again status register. */
+			clear_bit(bit, &status);
+		} while (status);
+	} while (1);
+
+	chained_irq_exit(chip, desc);
+}
+
+static struct irq_chip altera_msi_irq_chip = {
+	.name = "Altera 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 struct msi_domain_info altera_msi_domain_info = {
+	.flags	= (MSI_FLAG_USE_DEF_DOM_OPS | MSI_FLAG_USE_DEF_CHIP_OPS |
+		     MSI_FLAG_PCI_MSIX),
+	.chip	= &altera_msi_irq_chip,
+};
+
+static void altera_compose_msi_msg(struct irq_data *data, struct msi_msg *msg)
+{
+	u32 mask;
+	struct altera_msi *msi = irq_data_get_irq_chip_data(data);
+	u64 addr = msi->vector_phy + (data->hwirq * sizeof(u32));
+
+	msg->address_lo = lower_32_bits(addr);
+	msg->address_hi = upper_32_bits(addr);
+	msg->data = data->hwirq;
+
+	mask = msi_readl(msi, MSI_INTMASK);
+	mask |= 1 << data->hwirq;
+	msi_writel(msi, mask, MSI_INTMASK);
+	dev_dbg(&msi->pdev->dev, "msi#%d address_lo 0x%x\n", (int)data->hwirq,
+		msg->address_lo);
+}
+
+static int altera_msi_set_affinity(struct irq_data *irq_data,
+				 const struct cpumask *mask, bool force)
+{
+	 return -EINVAL;
+}
+
+static struct irq_chip altera_msi_bottom_irq_chip = {
+	.name			= "Altera MSI",
+	.irq_compose_msi_msg	= altera_compose_msi_msg,
+	.irq_set_affinity	= altera_msi_set_affinity,
+};
+
+static int altera_irq_domain_alloc(struct irq_domain *domain, unsigned int virq,
+				  unsigned int nr_irqs, void *args)
+{
+	struct altera_msi *msi = domain->host_data;
+	int bit;
+
+	mutex_lock(&msi->lock);
+
+	bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
+	if (bit < msi->num_of_vectors)
+		set_bit(bit, msi->used);
+
+	mutex_unlock(&msi->lock);
+
+	if (bit < 0)
+		return bit;
+	else if (bit >= msi->num_of_vectors)
+		return -ENOSPC;
+
+	irq_domain_set_info(domain, virq, bit, &altera_msi_bottom_irq_chip,
+			    domain->host_data, handle_simple_irq,
+			    NULL, NULL);
+	set_irq_flags(virq, IRQF_VALID);
+
+	return 0;
+}
+
+static void altera_irq_domain_free(struct irq_domain *domain,
+				  unsigned int virq, unsigned int nr_irqs)
+{
+	struct irq_data *d = irq_domain_get_irq_data(domain, virq);
+	struct altera_msi *msi = irq_data_get_irq_chip_data(d);
+	u32 mask;
+
+	mutex_lock(&msi->lock);
+
+	if (!test_bit(d->hwirq, msi->used))
+		dev_err(&msi->pdev->dev, "trying to free unused MSI#%lu\n",
+		d->hwirq);
+	else {
+		clear_bit(d->hwirq, msi->used);
+		mask = msi_readl(msi, MSI_INTMASK);
+		mask &= ~(1 << d->hwirq);
+		msi_writel(msi, mask, MSI_INTMASK);
+	}
+
+	mutex_unlock(&msi->lock);
+}
+
+static const struct irq_domain_ops msi_domain_ops = {
+	.alloc	= altera_irq_domain_alloc,
+	.free	= altera_irq_domain_free,
+};
+
+static int altera_allocate_domains(struct altera_msi *msi)
+{
+	msi->inner_domain = irq_domain_add_linear(NULL, msi->num_of_vectors,
+					     &msi_domain_ops, msi);
+	if (!msi->inner_domain) {
+		dev_err(&msi->pdev->dev, "failed to create IRQ domain\n");
+		return -ENOMEM;
+	}
+
+	msi->msi_domain = pci_msi_create_irq_domain(
+				msi->pdev->dev.of_node,
+				&altera_msi_domain_info, msi->inner_domain);
+	if (!msi->msi_domain) {
+		dev_err(&msi->pdev->dev, "failed to create MSI domain\n");
+		irq_domain_remove(msi->inner_domain);
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+static void altera_free_domains(struct altera_msi *msi)
+{
+	irq_domain_remove(msi->msi_domain);
+	irq_domain_remove(msi->inner_domain);
+}
+
+
+static int altera_msi_remove(struct platform_device *pdev)
+{
+	struct altera_msi *msi = platform_get_drvdata(pdev);
+
+	msi_writel(msi, 0, MSI_INTMASK);
+	irq_set_chained_handler(msi->irq, NULL);
+	irq_set_handler_data(msi->irq, NULL);
+
+	altera_free_domains(msi);
+
+	platform_set_drvdata(pdev, NULL);
+	return 0;
+}
+
+int altera_msi_probe(struct platform_device *pdev)
+{
+	struct altera_msi *msi;
+	struct device_node *np = pdev->dev.of_node;
+	struct resource *res;
+	int ret;
+
+	msi = devm_kzalloc(&pdev->dev, sizeof(struct altera_msi),
+		GFP_KERNEL);
+	if (!msi)
+		return -ENOMEM;
+
+	mutex_init(&msi->lock);
+	msi->pdev = pdev;
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "csr");
+	msi->csr_base = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(msi->csr_base)) {
+		dev_err(&pdev->dev, "get csr resource failed\n");
+		return PTR_ERR(msi->csr_base);
+	}
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
+			"vector_slave");
+	msi->vector_base = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(msi->vector_base)) {
+		dev_err(&pdev->dev, "get vector slave resource failed\n");
+		return PTR_ERR(msi->vector_base);
+	}
+
+	msi->vector_phy = res->start;
+
+	if (of_property_read_u32(np, "num-vectors", &msi->num_of_vectors)) {
+		dev_err(&pdev->dev, "failed to parse the number of vectors\n");
+		return -EINVAL;
+	}
+
+	ret = altera_allocate_domains(msi);
+	if (ret)
+		return ret;
+
+	msi->irq = platform_get_irq(pdev, 0);
+	if (msi->irq <= 0) {
+		dev_err(&pdev->dev, "failed to map IRQ: %d\n", msi->irq);
+		ret = -ENODEV;
+		goto err;
+	}
+
+	irq_set_chained_handler(msi->irq, altera_msi_isr);
+	ret = irq_set_handler_data(msi->irq, msi);
+	if (ret) {
+		dev_err(&pdev->dev, "failed to register IRQ handler: %d\n",
+			ret);
+		goto err;
+	}
+
+	platform_set_drvdata(pdev, msi);
+
+	return 0;
+
+err:
+	altera_msi_remove(pdev);
+	return ret;
+}
+
+static const struct of_device_id altera_msi_of_match[] = {
+	{ .compatible = "altr,msi-1.0", NULL },
+	{ },
+};
+MODULE_DEVICE_TABLE(of, altera_msi_of_match);
+
+static struct platform_driver altera_msi_driver = {
+	.driver = {
+		.name = "altera-msi",
+		.owner = THIS_MODULE,
+		.of_match_table = altera_msi_of_match,
+	},
+	.probe = altera_msi_probe,
+	.remove = altera_msi_remove,
+};
+
+static int __init altera_msi_init(void)
+{
+	return platform_driver_register(&altera_msi_driver);
+}
+
+subsys_initcall(altera_msi_init);
+
+MODULE_AUTHOR("Ley Foon Tan <lftan@altera.com>");
+MODULE_DESCRIPTION("Altera PCIe MSI support");
+MODULE_LICENSE("GPL v2");
-- 
1.8.2.1

[PATCH v2 3/5] pci: altera: Add Altera PCIe MSI driver

[Ley Foon Tan]

This patch adds Altera PCIe MSI driver. This soft IP supports configurable
number of vectors, which is a dts parameter.

Signed-off-by: Ley Foon Tan <lftan-EIB2kfCEclfQT0dZR+AlfA@public.gmane.org>
---
 drivers/pci/host/Kconfig           |   7 +
 drivers/pci/host/Makefile          |   1 +
 drivers/pci/host/pcie-altera-msi.c | 324 +++++++++++++++++++++++++++++++++++++
 3 files changed, 332 insertions(+)
 create mode 100644 drivers/pci/host/pcie-altera-msi.c

diff --git a/drivers/pci/host/Kconfig b/drivers/pci/host/Kconfig
index 108500a..51d0468 100644
--- a/drivers/pci/host/Kconfig
+++ b/drivers/pci/host/Kconfig
@@ -153,4 +153,11 @@ config PCIE_ALTERA
 	  Say Y here if you want to enable PCIe controller support for Altera
 	  SoCFPGA family of SoCs.
 
+config PCIE_ALTERA_MSI
+	bool "Altera PCIe MSI feature"
+	depends on PCI_MSI
+	help
+	  Say Y here if you want PCIe MSI support for the Altera SocFPGA SoC.
+	  This MSI driver supports Altera MSI to GIC controller IP.
+
 endmenu
diff --git a/drivers/pci/host/Makefile b/drivers/pci/host/Makefile
index 6954f76..6c4913d 100644
--- a/drivers/pci/host/Makefile
+++ b/drivers/pci/host/Makefile
@@ -18,3 +18,4 @@ obj-$(CONFIG_PCIE_IPROC) += pcie-iproc.o
 obj-$(CONFIG_PCIE_IPROC_PLATFORM) += pcie-iproc-platform.o
 obj-$(CONFIG_PCIE_IPROC_BCMA) += pcie-iproc-bcma.o
 obj-$(CONFIG_PCIE_ALTERA) += pcie-altera.o
+obj-$(CONFIG_PCIE_ALTERA_MSI) += pcie-altera-msi.o
diff --git a/drivers/pci/host/pcie-altera-msi.c b/drivers/pci/host/pcie-altera-msi.c
new file mode 100644
index 0000000..6014719
--- /dev/null
+++ b/drivers/pci/host/pcie-altera-msi.c
@@ -0,0 +1,324 @@
+/*
+ * Copyright Altera Corporation (C) 2013-2015. All rights reserved
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+#include <linux/interrupt.h>
+#include <linux/irqchip/chained_irq.h>
+#include <linux/module.h>
+#include <linux/msi.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/of_pci.h>
+#include <linux/pci.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+
+#define MSI_STATUS		0x0
+#define MSI_ERROR		0x4
+#define MSI_INTMASK		0x8
+
+#define MAX_MSI_VECTORS		32
+struct altera_msi {
+	DECLARE_BITMAP(used, MAX_MSI_VECTORS);
+	struct mutex		lock;	/* proctect used variable */
+	struct platform_device	*pdev;
+	struct irq_domain		*msi_domain;
+	struct irq_domain		*inner_domain;
+	void __iomem		*csr_base;
+	void __iomem		*vector_base;
+	phys_addr_t		vector_phy;
+	u32			num_of_vectors;
+	int			irq;
+};
+
+static inline void msi_writel(struct altera_msi *msi, u32 value, u32 reg)
+{
+	writel_relaxed(value, msi->csr_base + reg);
+}
+
+static inline u32 msi_readl(struct altera_msi *msi, u32 reg)
+{
+	return readl_relaxed(msi->csr_base + reg);
+}
+
+static void altera_msi_isr(unsigned int irq, struct irq_desc *desc)
+{
+	struct irq_chip *chip = irq_desc_get_chip(desc);
+	struct altera_msi *msi;
+	unsigned long status;
+	u32 num_of_vectors;
+	u32 bit;
+
+	chained_irq_enter(chip, desc);
+	msi = irq_desc_get_handler_data(desc);
+	num_of_vectors = msi->num_of_vectors;
+
+	do {
+		status = msi_readl(msi, MSI_STATUS);
+		if (!status)
+			break;
+
+		do {
+			bit = find_first_bit(&status, num_of_vectors);
+			/* Dummy read from vector to clear the interrupt */
+			readl_relaxed(msi->vector_base + (bit * sizeof(u32)));
+
+			irq = irq_find_mapping(msi->inner_domain, bit);
+			if (irq) {
+				if (test_bit(bit, msi->used))
+					generic_handle_irq(irq);
+				else
+					dev_info(&msi->pdev->dev, "unhandled MSI\n");
+			} else
+				dev_info(&msi->pdev->dev, "unexpected MSI\n");
+
+			/* Clear the bit from status and repeat without reading
+			 * again status register. */
+			clear_bit(bit, &status);
+		} while (status);
+	} while (1);
+
+	chained_irq_exit(chip, desc);
+}
+
+static struct irq_chip altera_msi_irq_chip = {
+	.name = "Altera 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 struct msi_domain_info altera_msi_domain_info = {
+	.flags	= (MSI_FLAG_USE_DEF_DOM_OPS | MSI_FLAG_USE_DEF_CHIP_OPS |
+		     MSI_FLAG_PCI_MSIX),
+	.chip	= &altera_msi_irq_chip,
+};
+
+static void altera_compose_msi_msg(struct irq_data *data, struct msi_msg *msg)
+{
+	u32 mask;
+	struct altera_msi *msi = irq_data_get_irq_chip_data(data);
+	u64 addr = msi->vector_phy + (data->hwirq * sizeof(u32));
+
+	msg->address_lo = lower_32_bits(addr);
+	msg->address_hi = upper_32_bits(addr);
+	msg->data = data->hwirq;
+
+	mask = msi_readl(msi, MSI_INTMASK);
+	mask |= 1 << data->hwirq;
+	msi_writel(msi, mask, MSI_INTMASK);
+	dev_dbg(&msi->pdev->dev, "msi#%d address_lo 0x%x\n", (int)data->hwirq,
+		msg->address_lo);
+}
+
+static int altera_msi_set_affinity(struct irq_data *irq_data,
+				 const struct cpumask *mask, bool force)
+{
+	 return -EINVAL;
+}
+
+static struct irq_chip altera_msi_bottom_irq_chip = {
+	.name			= "Altera MSI",
+	.irq_compose_msi_msg	= altera_compose_msi_msg,
+	.irq_set_affinity	= altera_msi_set_affinity,
+};
+
+static int altera_irq_domain_alloc(struct irq_domain *domain, unsigned int virq,
+				  unsigned int nr_irqs, void *args)
+{
+	struct altera_msi *msi = domain->host_data;
+	int bit;
+
+	mutex_lock(&msi->lock);
+
+	bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
+	if (bit < msi->num_of_vectors)
+		set_bit(bit, msi->used);
+
+	mutex_unlock(&msi->lock);
+
+	if (bit < 0)
+		return bit;
+	else if (bit >= msi->num_of_vectors)
+		return -ENOSPC;
+
+	irq_domain_set_info(domain, virq, bit, &altera_msi_bottom_irq_chip,
+			    domain->host_data, handle_simple_irq,
+			    NULL, NULL);
+	set_irq_flags(virq, IRQF_VALID);
+
+	return 0;
+}
+
+static void altera_irq_domain_free(struct irq_domain *domain,
+				  unsigned int virq, unsigned int nr_irqs)
+{
+	struct irq_data *d = irq_domain_get_irq_data(domain, virq);
+	struct altera_msi *msi = irq_data_get_irq_chip_data(d);
+	u32 mask;
+
+	mutex_lock(&msi->lock);
+
+	if (!test_bit(d->hwirq, msi->used))
+		dev_err(&msi->pdev->dev, "trying to free unused MSI#%lu\n",
+		d->hwirq);
+	else {
+		clear_bit(d->hwirq, msi->used);
+		mask = msi_readl(msi, MSI_INTMASK);
+		mask &= ~(1 << d->hwirq);
+		msi_writel(msi, mask, MSI_INTMASK);
+	}
+
+	mutex_unlock(&msi->lock);
+}
+
+static const struct irq_domain_ops msi_domain_ops = {
+	.alloc	= altera_irq_domain_alloc,
+	.free	= altera_irq_domain_free,
+};
+
+static int altera_allocate_domains(struct altera_msi *msi)
+{
+	msi->inner_domain = irq_domain_add_linear(NULL, msi->num_of_vectors,
+					     &msi_domain_ops, msi);
+	if (!msi->inner_domain) {
+		dev_err(&msi->pdev->dev, "failed to create IRQ domain\n");
+		return -ENOMEM;
+	}
+
+	msi->msi_domain = pci_msi_create_irq_domain(
+				msi->pdev->dev.of_node,
+				&altera_msi_domain_info, msi->inner_domain);
+	if (!msi->msi_domain) {
+		dev_err(&msi->pdev->dev, "failed to create MSI domain\n");
+		irq_domain_remove(msi->inner_domain);
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+static void altera_free_domains(struct altera_msi *msi)
+{
+	irq_domain_remove(msi->msi_domain);
+	irq_domain_remove(msi->inner_domain);
+}
+
+
+static int altera_msi_remove(struct platform_device *pdev)
+{
+	struct altera_msi *msi = platform_get_drvdata(pdev);
+
+	msi_writel(msi, 0, MSI_INTMASK);
+	irq_set_chained_handler(msi->irq, NULL);
+	irq_set_handler_data(msi->irq, NULL);
+
+	altera_free_domains(msi);
+
+	platform_set_drvdata(pdev, NULL);
+	return 0;
+}
+
+int altera_msi_probe(struct platform_device *pdev)
+{
+	struct altera_msi *msi;
+	struct device_node *np = pdev->dev.of_node;
+	struct resource *res;
+	int ret;
+
+	msi = devm_kzalloc(&pdev->dev, sizeof(struct altera_msi),
+		GFP_KERNEL);
+	if (!msi)
+		return -ENOMEM;
+
+	mutex_init(&msi->lock);
+	msi->pdev = pdev;
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "csr");
+	msi->csr_base = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(msi->csr_base)) {
+		dev_err(&pdev->dev, "get csr resource failed\n");
+		return PTR_ERR(msi->csr_base);
+	}
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
+			"vector_slave");
+	msi->vector_base = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(msi->vector_base)) {
+		dev_err(&pdev->dev, "get vector slave resource failed\n");
+		return PTR_ERR(msi->vector_base);
+	}
+
+	msi->vector_phy = res->start;
+
+	if (of_property_read_u32(np, "num-vectors", &msi->num_of_vectors)) {
+		dev_err(&pdev->dev, "failed to parse the number of vectors\n");
+		return -EINVAL;
+	}
+
+	ret = altera_allocate_domains(msi);
+	if (ret)
+		return ret;
+
+	msi->irq = platform_get_irq(pdev, 0);
+	if (msi->irq <= 0) {
+		dev_err(&pdev->dev, "failed to map IRQ: %d\n", msi->irq);
+		ret = -ENODEV;
+		goto err;
+	}
+
+	irq_set_chained_handler(msi->irq, altera_msi_isr);
+	ret = irq_set_handler_data(msi->irq, msi);
+	if (ret) {
+		dev_err(&pdev->dev, "failed to register IRQ handler: %d\n",
+			ret);
+		goto err;
+	}
+
+	platform_set_drvdata(pdev, msi);
+
+	return 0;
+
+err:
+	altera_msi_remove(pdev);
+	return ret;
+}
+
+static const struct of_device_id altera_msi_of_match[] = {
+	{ .compatible = "altr,msi-1.0", NULL },
+	{ },
+};
+MODULE_DEVICE_TABLE(of, altera_msi_of_match);
+
+static struct platform_driver altera_msi_driver = {
+	.driver = {
+		.name = "altera-msi",
+		.owner = THIS_MODULE,
+		.of_match_table = altera_msi_of_match,
+	},
+	.probe = altera_msi_probe,
+	.remove = altera_msi_remove,
+};
+
+static int __init altera_msi_init(void)
+{
+	return platform_driver_register(&altera_msi_driver);
+}
+
+subsys_initcall(altera_msi_init);
+
+MODULE_AUTHOR("Ley Foon Tan <lftan-EIB2kfCEclfQT0dZR+AlfA@public.gmane.org>");
+MODULE_DESCRIPTION("Altera PCIe MSI support");
+MODULE_LICENSE("GPL v2");
-- 
1.8.2.1

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v2 3/5] pci: altera: Add Altera PCIe MSI driver

[Ley Foon Tan]

This patch adds Altera PCIe MSI driver. This soft IP supports configurable
number of vectors, which is a dts parameter.

Signed-off-by: Ley Foon Tan <lftan@altera.com>
---
 drivers/pci/host/Kconfig           |   7 +
 drivers/pci/host/Makefile          |   1 +
 drivers/pci/host/pcie-altera-msi.c | 324 +++++++++++++++++++++++++++++++++++++
 3 files changed, 332 insertions(+)
 create mode 100644 drivers/pci/host/pcie-altera-msi.c

diff --git a/drivers/pci/host/Kconfig b/drivers/pci/host/Kconfig
index 108500a..51d0468 100644
--- a/drivers/pci/host/Kconfig
+++ b/drivers/pci/host/Kconfig
@@ -153,4 +153,11 @@ config PCIE_ALTERA
 	  Say Y here if you want to enable PCIe controller support for Altera
 	  SoCFPGA family of SoCs.
 
+config PCIE_ALTERA_MSI
+	bool "Altera PCIe MSI feature"
+	depends on PCI_MSI
+	help
+	  Say Y here if you want PCIe MSI support for the Altera SocFPGA SoC.
+	  This MSI driver supports Altera MSI to GIC controller IP.
+
 endmenu
diff --git a/drivers/pci/host/Makefile b/drivers/pci/host/Makefile
index 6954f76..6c4913d 100644
--- a/drivers/pci/host/Makefile
+++ b/drivers/pci/host/Makefile
@@ -18,3 +18,4 @@ obj-$(CONFIG_PCIE_IPROC) += pcie-iproc.o
 obj-$(CONFIG_PCIE_IPROC_PLATFORM) += pcie-iproc-platform.o
 obj-$(CONFIG_PCIE_IPROC_BCMA) += pcie-iproc-bcma.o
 obj-$(CONFIG_PCIE_ALTERA) += pcie-altera.o
+obj-$(CONFIG_PCIE_ALTERA_MSI) += pcie-altera-msi.o
diff --git a/drivers/pci/host/pcie-altera-msi.c b/drivers/pci/host/pcie-altera-msi.c
new file mode 100644
index 0000000..6014719
--- /dev/null
+++ b/drivers/pci/host/pcie-altera-msi.c
@@ -0,0 +1,324 @@
+/*
+ * Copyright Altera Corporation (C) 2013-2015. All rights reserved
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+#include <linux/interrupt.h>
+#include <linux/irqchip/chained_irq.h>
+#include <linux/module.h>
+#include <linux/msi.h>
+#include <linux/of_address.h>
+#include <linux/of_irq.h>
+#include <linux/of_pci.h>
+#include <linux/pci.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+
+#define MSI_STATUS		0x0
+#define MSI_ERROR		0x4
+#define MSI_INTMASK		0x8
+
+#define MAX_MSI_VECTORS		32
+struct altera_msi {
+	DECLARE_BITMAP(used, MAX_MSI_VECTORS);
+	struct mutex		lock;	/* proctect used variable */
+	struct platform_device	*pdev;
+	struct irq_domain		*msi_domain;
+	struct irq_domain		*inner_domain;
+	void __iomem		*csr_base;
+	void __iomem		*vector_base;
+	phys_addr_t		vector_phy;
+	u32			num_of_vectors;
+	int			irq;
+};
+
+static inline void msi_writel(struct altera_msi *msi, u32 value, u32 reg)
+{
+	writel_relaxed(value, msi->csr_base + reg);
+}
+
+static inline u32 msi_readl(struct altera_msi *msi, u32 reg)
+{
+	return readl_relaxed(msi->csr_base + reg);
+}
+
+static void altera_msi_isr(unsigned int irq, struct irq_desc *desc)
+{
+	struct irq_chip *chip = irq_desc_get_chip(desc);
+	struct altera_msi *msi;
+	unsigned long status;
+	u32 num_of_vectors;
+	u32 bit;
+
+	chained_irq_enter(chip, desc);
+	msi = irq_desc_get_handler_data(desc);
+	num_of_vectors = msi->num_of_vectors;
+
+	do {
+		status = msi_readl(msi, MSI_STATUS);
+		if (!status)
+			break;
+
+		do {
+			bit = find_first_bit(&status, num_of_vectors);
+			/* Dummy read from vector to clear the interrupt */
+			readl_relaxed(msi->vector_base + (bit * sizeof(u32)));
+
+			irq = irq_find_mapping(msi->inner_domain, bit);
+			if (irq) {
+				if (test_bit(bit, msi->used))
+					generic_handle_irq(irq);
+				else
+					dev_info(&msi->pdev->dev, "unhandled MSI\n");
+			} else
+				dev_info(&msi->pdev->dev, "unexpected MSI\n");
+
+			/* Clear the bit from status and repeat without reading
+			 * again status register. */
+			clear_bit(bit, &status);
+		} while (status);
+	} while (1);
+
+	chained_irq_exit(chip, desc);
+}
+
+static struct irq_chip altera_msi_irq_chip = {
+	.name = "Altera 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 struct msi_domain_info altera_msi_domain_info = {
+	.flags	= (MSI_FLAG_USE_DEF_DOM_OPS | MSI_FLAG_USE_DEF_CHIP_OPS |
+		     MSI_FLAG_PCI_MSIX),
+	.chip	= &altera_msi_irq_chip,
+};
+
+static void altera_compose_msi_msg(struct irq_data *data, struct msi_msg *msg)
+{
+	u32 mask;
+	struct altera_msi *msi = irq_data_get_irq_chip_data(data);
+	u64 addr = msi->vector_phy + (data->hwirq * sizeof(u32));
+
+	msg->address_lo = lower_32_bits(addr);
+	msg->address_hi = upper_32_bits(addr);
+	msg->data = data->hwirq;
+
+	mask = msi_readl(msi, MSI_INTMASK);
+	mask |= 1 << data->hwirq;
+	msi_writel(msi, mask, MSI_INTMASK);
+	dev_dbg(&msi->pdev->dev, "msi#%d address_lo 0x%x\n", (int)data->hwirq,
+		msg->address_lo);
+}
+
+static int altera_msi_set_affinity(struct irq_data *irq_data,
+				 const struct cpumask *mask, bool force)
+{
+	 return -EINVAL;
+}
+
+static struct irq_chip altera_msi_bottom_irq_chip = {
+	.name			= "Altera MSI",
+	.irq_compose_msi_msg	= altera_compose_msi_msg,
+	.irq_set_affinity	= altera_msi_set_affinity,
+};
+
+static int altera_irq_domain_alloc(struct irq_domain *domain, unsigned int virq,
+				  unsigned int nr_irqs, void *args)
+{
+	struct altera_msi *msi = domain->host_data;
+	int bit;
+
+	mutex_lock(&msi->lock);
+
+	bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
+	if (bit < msi->num_of_vectors)
+		set_bit(bit, msi->used);
+
+	mutex_unlock(&msi->lock);
+
+	if (bit < 0)
+		return bit;
+	else if (bit >= msi->num_of_vectors)
+		return -ENOSPC;
+
+	irq_domain_set_info(domain, virq, bit, &altera_msi_bottom_irq_chip,
+			    domain->host_data, handle_simple_irq,
+			    NULL, NULL);
+	set_irq_flags(virq, IRQF_VALID);
+
+	return 0;
+}
+
+static void altera_irq_domain_free(struct irq_domain *domain,
+				  unsigned int virq, unsigned int nr_irqs)
+{
+	struct irq_data *d = irq_domain_get_irq_data(domain, virq);
+	struct altera_msi *msi = irq_data_get_irq_chip_data(d);
+	u32 mask;
+
+	mutex_lock(&msi->lock);
+
+	if (!test_bit(d->hwirq, msi->used))
+		dev_err(&msi->pdev->dev, "trying to free unused MSI#%lu\n",
+		d->hwirq);
+	else {
+		clear_bit(d->hwirq, msi->used);
+		mask = msi_readl(msi, MSI_INTMASK);
+		mask &= ~(1 << d->hwirq);
+		msi_writel(msi, mask, MSI_INTMASK);
+	}
+
+	mutex_unlock(&msi->lock);
+}
+
+static const struct irq_domain_ops msi_domain_ops = {
+	.alloc	= altera_irq_domain_alloc,
+	.free	= altera_irq_domain_free,
+};
+
+static int altera_allocate_domains(struct altera_msi *msi)
+{
+	msi->inner_domain = irq_domain_add_linear(NULL, msi->num_of_vectors,
+					     &msi_domain_ops, msi);
+	if (!msi->inner_domain) {
+		dev_err(&msi->pdev->dev, "failed to create IRQ domain\n");
+		return -ENOMEM;
+	}
+
+	msi->msi_domain = pci_msi_create_irq_domain(
+				msi->pdev->dev.of_node,
+				&altera_msi_domain_info, msi->inner_domain);
+	if (!msi->msi_domain) {
+		dev_err(&msi->pdev->dev, "failed to create MSI domain\n");
+		irq_domain_remove(msi->inner_domain);
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+static void altera_free_domains(struct altera_msi *msi)
+{
+	irq_domain_remove(msi->msi_domain);
+	irq_domain_remove(msi->inner_domain);
+}
+
+
+static int altera_msi_remove(struct platform_device *pdev)
+{
+	struct altera_msi *msi = platform_get_drvdata(pdev);
+
+	msi_writel(msi, 0, MSI_INTMASK);
+	irq_set_chained_handler(msi->irq, NULL);
+	irq_set_handler_data(msi->irq, NULL);
+
+	altera_free_domains(msi);
+
+	platform_set_drvdata(pdev, NULL);
+	return 0;
+}
+
+int altera_msi_probe(struct platform_device *pdev)
+{
+	struct altera_msi *msi;
+	struct device_node *np = pdev->dev.of_node;
+	struct resource *res;
+	int ret;
+
+	msi = devm_kzalloc(&pdev->dev, sizeof(struct altera_msi),
+		GFP_KERNEL);
+	if (!msi)
+		return -ENOMEM;
+
+	mutex_init(&msi->lock);
+	msi->pdev = pdev;
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "csr");
+	msi->csr_base = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(msi->csr_base)) {
+		dev_err(&pdev->dev, "get csr resource failed\n");
+		return PTR_ERR(msi->csr_base);
+	}
+
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
+			"vector_slave");
+	msi->vector_base = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(msi->vector_base)) {
+		dev_err(&pdev->dev, "get vector slave resource failed\n");
+		return PTR_ERR(msi->vector_base);
+	}
+
+	msi->vector_phy = res->start;
+
+	if (of_property_read_u32(np, "num-vectors", &msi->num_of_vectors)) {
+		dev_err(&pdev->dev, "failed to parse the number of vectors\n");
+		return -EINVAL;
+	}
+
+	ret = altera_allocate_domains(msi);
+	if (ret)
+		return ret;
+
+	msi->irq = platform_get_irq(pdev, 0);
+	if (msi->irq <= 0) {
+		dev_err(&pdev->dev, "failed to map IRQ: %d\n", msi->irq);
+		ret = -ENODEV;
+		goto err;
+	}
+
+	irq_set_chained_handler(msi->irq, altera_msi_isr);
+	ret = irq_set_handler_data(msi->irq, msi);
+	if (ret) {
+		dev_err(&pdev->dev, "failed to register IRQ handler: %d\n",
+			ret);
+		goto err;
+	}
+
+	platform_set_drvdata(pdev, msi);
+
+	return 0;
+
+err:
+	altera_msi_remove(pdev);
+	return ret;
+}
+
+static const struct of_device_id altera_msi_of_match[] = {
+	{ .compatible = "altr,msi-1.0", NULL },
+	{ },
+};
+MODULE_DEVICE_TABLE(of, altera_msi_of_match);
+
+static struct platform_driver altera_msi_driver = {
+	.driver = {
+		.name = "altera-msi",
+		.owner = THIS_MODULE,
+		.of_match_table = altera_msi_of_match,
+	},
+	.probe = altera_msi_probe,
+	.remove = altera_msi_remove,
+};
+
+static int __init altera_msi_init(void)
+{
+	return platform_driver_register(&altera_msi_driver);
+}
+
+subsys_initcall(altera_msi_init);
+
+MODULE_AUTHOR("Ley Foon Tan <lftan@altera.com>");
+MODULE_DESCRIPTION("Altera PCIe MSI support");
+MODULE_LICENSE("GPL v2");
-- 
1.8.2.1

[PATCH v2 4/5] Documentation: dt-bindings: pci: altera pcie device tree binding

[Ley Foon Tan]

This patch adds the bindings for Altera PCIe host controller driver and
Altera PCIe MSI driver.

Signed-off-by: Ley Foon Tan <lftan@altera.com>
---
 .../devicetree/bindings/pci/altera-pcie-msi.txt    | 27 ++++++++++++
 .../devicetree/bindings/pci/altera-pcie.txt        | 49 ++++++++++++++++++++++
 2 files changed, 76 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/pci/altera-pcie-msi.txt
 create mode 100644 Documentation/devicetree/bindings/pci/altera-pcie.txt

diff --git a/Documentation/devicetree/bindings/pci/altera-pcie-msi.txt b/Documentation/devicetree/bindings/pci/altera-pcie-msi.txt
new file mode 100644
index 0000000..7f330c9
--- /dev/null
+++ b/Documentation/devicetree/bindings/pci/altera-pcie-msi.txt
@@ -0,0 +1,27 @@
+* Altera PCIe MSI controller
+
+Required properties:
+- compatible:	should contain "altr,msi-1.0"
+- reg:		specifies the physical base address of the controller and
+		the length of the memory mapped region.
+- reg-names:	Must include the following entries:
+		"csr": CSR registers
+		"vector_slave": vectors region
+-interrupts:	specifies the interrupt source of the parent interrupt
+		controller. The format of the interrupt specifier depends on the
+		parent interrupt controller.
+- num-vectors:	Number of vectors, range 1 to 32.
+- msi-controller:	indicates that this is MSI controller node
+
+
+Example
+msi0: msi@0xFF200000 {
+	compatible = "altr,msi-1.0";
+	reg = <0xFF200000 0x00000010
+		0xFF200010 0x00000080>;
+	reg-names = "csr", "vector_slave";
+	interrupt-parent = <&hps_0_arm_gic_0>;
+	interrupts = <0 42 4>;
+	msi-controller = <1>;
+	num-vectors = <32>;
+};
diff --git a/Documentation/devicetree/bindings/pci/altera-pcie.txt b/Documentation/devicetree/bindings/pci/altera-pcie.txt
new file mode 100644
index 0000000..54c45f5
--- /dev/null
+++ b/Documentation/devicetree/bindings/pci/altera-pcie.txt
@@ -0,0 +1,49 @@
+* Altera PCIe controller
+
+Required properties:
+- compatible :	should contain "altr,pcie-root-port-1.0"
+- reg:		A list of physical base address and length for TXS and CRA.
+- reg-names:	Must include the following entries:
+		"Txs": TXS region
+		"Cra": Control register access region
+-interrupts:	specifies the interrupt source of the parent interrupt controller.
+		The format of the interrupt specifier depends on the parent interrupt
+		controller.
+- device_type:	must be "pci"
+- #address-cells:	set to <3>
+- #size-cells:	set to <2>
+- #interrupt-cells:	set to <1>
+- ranges:		Describes the translation of addresses for root ports and standard
+		PCI regions.
+- interrupt-map-mask and interrupt-map: standard PCI properties
+	to define the mapping of the PCIe interface to interrupt
+	numbers.
+
+Optional properties:
+- msi-parent:	Link to the hardware entity that serves as the MSI controller for this PCIe
+		controller.
+- bus-range:	PCI bus numbers covered
+
+Example
+	pcie_0: pcie@0xc00000000 {
+		compatible = "altr,pcie-root-port-1.0";
+		reg = <0xc0000000 0x20000000>,
+			<0xff220000 0x00004000>;
+		reg-names = "Txs", "Cra";
+		interrupt-parent = <&hps_0_arm_gic_0>;
+		interrupts = <0 40 4>;
+		interrupt-controller;
+		#interrupt-cells = <1>;
+		bus-range = <0x0 0xFF>;
+		device_type = "pci";
+		msi-parent = <&msi_to_gic_gen_0>;
+		#address-cells = <3>;
+		#size-cells = <2>;
+		interrupt-map-mask = <0 0 0 7>;
+		interrupt-map = <0 0 0 1 &pcie_0 1>,
+			            <0 0 0 2 &pcie_0 2>,
+			            <0 0 0 3 &pcie_0 3>,
+			            <0 0 0 4 &pcie_0 4>;
+		ranges = <0x82000000 0x00000000 0xc0000000 0xc0000000 0x00000000 0x10000000
+			    0x82000000 0x00000000 0xd0000000 0xd0000000 0x00000000 0x10000000>;
+	};
-- 
1.8.2.1

[PATCH v2 4/5] Documentation: dt-bindings: pci: altera pcie device tree binding

[Ley Foon Tan]

This patch adds the bindings for Altera PCIe host controller driver and
Altera PCIe MSI driver.

Signed-off-by: Ley Foon Tan <lftan@altera.com>
---
 .../devicetree/bindings/pci/altera-pcie-msi.txt    | 27 ++++++++++++
 .../devicetree/bindings/pci/altera-pcie.txt        | 49 ++++++++++++++++++++++
 2 files changed, 76 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/pci/altera-pcie-msi.txt
 create mode 100644 Documentation/devicetree/bindings/pci/altera-pcie.txt

diff --git a/Documentation/devicetree/bindings/pci/altera-pcie-msi.txt b/Documentation/devicetree/bindings/pci/altera-pcie-msi.txt
new file mode 100644
index 0000000..7f330c9
--- /dev/null
+++ b/Documentation/devicetree/bindings/pci/altera-pcie-msi.txt
@@ -0,0 +1,27 @@
+* Altera PCIe MSI controller
+
+Required properties:
+- compatible:	should contain "altr,msi-1.0"
+- reg:		specifies the physical base address of the controller and
+		the length of the memory mapped region.
+- reg-names:	Must include the following entries:
+		"csr": CSR registers
+		"vector_slave": vectors region
+-interrupts:	specifies the interrupt source of the parent interrupt
+		controller. The format of the interrupt specifier depends on the
+		parent interrupt controller.
+- num-vectors:	Number of vectors, range 1 to 32.
+- msi-controller:	indicates that this is MSI controller node
+
+
+Example
+msi0: msi@0xFF200000 {
+	compatible = "altr,msi-1.0";
+	reg = <0xFF200000 0x00000010
+		0xFF200010 0x00000080>;
+	reg-names = "csr", "vector_slave";
+	interrupt-parent = <&hps_0_arm_gic_0>;
+	interrupts = <0 42 4>;
+	msi-controller = <1>;
+	num-vectors = <32>;
+};
diff --git a/Documentation/devicetree/bindings/pci/altera-pcie.txt b/Documentation/devicetree/bindings/pci/altera-pcie.txt
new file mode 100644
index 0000000..54c45f5
--- /dev/null
+++ b/Documentation/devicetree/bindings/pci/altera-pcie.txt
@@ -0,0 +1,49 @@
+* Altera PCIe controller
+
+Required properties:
+- compatible :	should contain "altr,pcie-root-port-1.0"
+- reg:		A list of physical base address and length for TXS and CRA.
+- reg-names:	Must include the following entries:
+		"Txs": TXS region
+		"Cra": Control register access region
+-interrupts:	specifies the interrupt source of the parent interrupt controller.
+		The format of the interrupt specifier depends on the parent interrupt
+		controller.
+- device_type:	must be "pci"
+- #address-cells:	set to <3>
+- #size-cells:	set to <2>
+- #interrupt-cells:	set to <1>
+- ranges:		Describes the translation of addresses for root ports and standard
+		PCI regions.
+- interrupt-map-mask and interrupt-map: standard PCI properties
+	to define the mapping of the PCIe interface to interrupt
+	numbers.
+
+Optional properties:
+- msi-parent:	Link to the hardware entity that serves as the MSI controller for this PCIe
+		controller.
+- bus-range:	PCI bus numbers covered
+
+Example
+	pcie_0: pcie@0xc00000000 {
+		compatible = "altr,pcie-root-port-1.0";
+		reg = <0xc0000000 0x20000000>,
+			<0xff220000 0x00004000>;
+		reg-names = "Txs", "Cra";
+		interrupt-parent = <&hps_0_arm_gic_0>;
+		interrupts = <0 40 4>;
+		interrupt-controller;
+		#interrupt-cells = <1>;
+		bus-range = <0x0 0xFF>;
+		device_type = "pci";
+		msi-parent = <&msi_to_gic_gen_0>;
+		#address-cells = <3>;
+		#size-cells = <2>;
+		interrupt-map-mask = <0 0 0 7>;
+		interrupt-map = <0 0 0 1 &pcie_0 1>,
+			            <0 0 0 2 &pcie_0 2>,
+			            <0 0 0 3 &pcie_0 3>,
+			            <0 0 0 4 &pcie_0 4>;
+		ranges = <0x82000000 0x00000000 0xc0000000 0xc0000000 0x00000000 0x10000000
+			    0x82000000 0x00000000 0xd0000000 0xd0000000 0x00000000 0x10000000>;
+	};
-- 
1.8.2.1

[PATCH v2 4/5] Documentation: dt-bindings: pci: altera pcie device tree binding

[Ley Foon Tan]

This patch adds the bindings for Altera PCIe host controller driver and
Altera PCIe MSI driver.

Signed-off-by: Ley Foon Tan <lftan@altera.com>
---
 .../devicetree/bindings/pci/altera-pcie-msi.txt    | 27 ++++++++++++
 .../devicetree/bindings/pci/altera-pcie.txt        | 49 ++++++++++++++++++++++
 2 files changed, 76 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/pci/altera-pcie-msi.txt
 create mode 100644 Documentation/devicetree/bindings/pci/altera-pcie.txt

diff --git a/Documentation/devicetree/bindings/pci/altera-pcie-msi.txt b/Documentation/devicetree/bindings/pci/altera-pcie-msi.txt
new file mode 100644
index 0000000..7f330c9
--- /dev/null
+++ b/Documentation/devicetree/bindings/pci/altera-pcie-msi.txt
@@ -0,0 +1,27 @@
+* Altera PCIe MSI controller
+
+Required properties:
+- compatible:	should contain "altr,msi-1.0"
+- reg:		specifies the physical base address of the controller and
+		the length of the memory mapped region.
+- reg-names:	Must include the following entries:
+		"csr": CSR registers
+		"vector_slave": vectors region
+-interrupts:	specifies the interrupt source of the parent interrupt
+		controller. The format of the interrupt specifier depends on the
+		parent interrupt controller.
+- num-vectors:	Number of vectors, range 1 to 32.
+- msi-controller:	indicates that this is MSI controller node
+
+
+Example
+msi0: msi at 0xFF200000 {
+	compatible = "altr,msi-1.0";
+	reg = <0xFF200000 0x00000010
+		0xFF200010 0x00000080>;
+	reg-names = "csr", "vector_slave";
+	interrupt-parent = <&hps_0_arm_gic_0>;
+	interrupts = <0 42 4>;
+	msi-controller = <1>;
+	num-vectors = <32>;
+};
diff --git a/Documentation/devicetree/bindings/pci/altera-pcie.txt b/Documentation/devicetree/bindings/pci/altera-pcie.txt
new file mode 100644
index 0000000..54c45f5
--- /dev/null
+++ b/Documentation/devicetree/bindings/pci/altera-pcie.txt
@@ -0,0 +1,49 @@
+* Altera PCIe controller
+
+Required properties:
+- compatible :	should contain "altr,pcie-root-port-1.0"
+- reg:		A list of physical base address and length for TXS and CRA.
+- reg-names:	Must include the following entries:
+		"Txs": TXS region
+		"Cra": Control register access region
+-interrupts:	specifies the interrupt source of the parent interrupt controller.
+		The format of the interrupt specifier depends on the parent interrupt
+		controller.
+- device_type:	must be "pci"
+- #address-cells:	set to <3>
+- #size-cells:	set to <2>
+- #interrupt-cells:	set to <1>
+- ranges:		Describes the translation of addresses for root ports and standard
+		PCI regions.
+- interrupt-map-mask and interrupt-map: standard PCI properties
+	to define the mapping of the PCIe interface to interrupt
+	numbers.
+
+Optional properties:
+- msi-parent:	Link to the hardware entity that serves as the MSI controller for this PCIe
+		controller.
+- bus-range:	PCI bus numbers covered
+
+Example
+	pcie_0: pcie at 0xc00000000 {
+		compatible = "altr,pcie-root-port-1.0";
+		reg = <0xc0000000 0x20000000>,
+			<0xff220000 0x00004000>;
+		reg-names = "Txs", "Cra";
+		interrupt-parent = <&hps_0_arm_gic_0>;
+		interrupts = <0 40 4>;
+		interrupt-controller;
+		#interrupt-cells = <1>;
+		bus-range = <0x0 0xFF>;
+		device_type = "pci";
+		msi-parent = <&msi_to_gic_gen_0>;
+		#address-cells = <3>;
+		#size-cells = <2>;
+		interrupt-map-mask = <0 0 0 7>;
+		interrupt-map = <0 0 0 1 &pcie_0 1>,
+			            <0 0 0 2 &pcie_0 2>,
+			            <0 0 0 3 &pcie_0 3>,
+			            <0 0 0 4 &pcie_0 4>;
+		ranges = <0x82000000 0x00000000 0xc0000000 0xc0000000 0x00000000 0x10000000
+			    0x82000000 0x00000000 0xd0000000 0xd0000000 0x00000000 0x10000000>;
+	};
-- 
1.8.2.1

[PATCH v2 5/5] MAINTAINERS: Add Altera PCIe and MSI drivers maintainer

[Ley Foon Tan]

Signed-off-by: Ley Foon Tan <lftan@altera.com>
---
 MAINTAINERS | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)

diff --git a/MAINTAINERS b/MAINTAINERS
index fd60784..32f5287 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -7759,6 +7759,14 @@ F:	include/linux/pci*
 F:	arch/x86/pci/
 F:	arch/x86/kernel/quirks.c
 
+PCI DRIVER FOR ALTERA PCIE IP
+M:	Ley Foon Tan <lftan@altera.com>
+L:	rfi@lists.rocketboards.org (moderated for non-subscribers)
+L:	linux-pci@vger.kernel.org
+S:	Supported
+F:	Documentation/devicetree/bindings/pci/altera-pcie.txt
+F:	drivers/pci/host/pcie-altera.c
+
 PCI DRIVER FOR ARM VERSATILE PLATFORM
 M:	Rob Herring <robh@kernel.org>
 L:	linux-pci@vger.kernel.org
@@ -7860,6 +7868,14 @@ L:	linux-pci@vger.kernel.org
 S:	Maintained
 F:	drivers/pci/host/*spear*
 
+PCI MSI DRIVER FOR ALTERA MSI IP
+M:	Ley Foon Tan <lftan@altera.com>
+L:	rfi@lists.rocketboards.org (moderated for non-subscribers)
+L:	linux-pci@vger.kernel.org
+S:	Supported
+F:	Documentation/devicetree/bindings/pci/altera-pcie-msi.txt
+F:	drivers/pci/host/pcie-altera-msi.c
+
 PCI MSI DRIVER FOR APPLIEDMICRO XGENE
 M:	Duc Dang <dhdang@apm.com>
 L:	linux-pci@vger.kernel.org
-- 
1.8.2.1

[PATCH v2 5/5] MAINTAINERS: Add Altera PCIe and MSI drivers maintainer

[Ley Foon Tan]

Signed-off-by: Ley Foon Tan <lftan@altera.com>
---
 MAINTAINERS | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)

diff --git a/MAINTAINERS b/MAINTAINERS
index fd60784..32f5287 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -7759,6 +7759,14 @@ F:	include/linux/pci*
 F:	arch/x86/pci/
 F:	arch/x86/kernel/quirks.c
 
+PCI DRIVER FOR ALTERA PCIE IP
+M:	Ley Foon Tan <lftan@altera.com>
+L:	rfi@lists.rocketboards.org (moderated for non-subscribers)
+L:	linux-pci@vger.kernel.org
+S:	Supported
+F:	Documentation/devicetree/bindings/pci/altera-pcie.txt
+F:	drivers/pci/host/pcie-altera.c
+
 PCI DRIVER FOR ARM VERSATILE PLATFORM
 M:	Rob Herring <robh@kernel.org>
 L:	linux-pci@vger.kernel.org
@@ -7860,6 +7868,14 @@ L:	linux-pci@vger.kernel.org
 S:	Maintained
 F:	drivers/pci/host/*spear*
 
+PCI MSI DRIVER FOR ALTERA MSI IP
+M:	Ley Foon Tan <lftan@altera.com>
+L:	rfi@lists.rocketboards.org (moderated for non-subscribers)
+L:	linux-pci@vger.kernel.org
+S:	Supported
+F:	Documentation/devicetree/bindings/pci/altera-pcie-msi.txt
+F:	drivers/pci/host/pcie-altera-msi.c
+
 PCI MSI DRIVER FOR APPLIEDMICRO XGENE
 M:	Duc Dang <dhdang@apm.com>
 L:	linux-pci@vger.kernel.org
-- 
1.8.2.1

[PATCH v2 5/5] MAINTAINERS: Add Altera PCIe and MSI drivers maintainer

[Ley Foon Tan]

Signed-off-by: Ley Foon Tan <lftan@altera.com>
---
 MAINTAINERS | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)

diff --git a/MAINTAINERS b/MAINTAINERS
index fd60784..32f5287 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -7759,6 +7759,14 @@ F:	include/linux/pci*
 F:	arch/x86/pci/
 F:	arch/x86/kernel/quirks.c
 
+PCI DRIVER FOR ALTERA PCIE IP
+M:	Ley Foon Tan <lftan@altera.com>
+L:	rfi at lists.rocketboards.org (moderated for non-subscribers)
+L:	linux-pci at vger.kernel.org
+S:	Supported
+F:	Documentation/devicetree/bindings/pci/altera-pcie.txt
+F:	drivers/pci/host/pcie-altera.c
+
 PCI DRIVER FOR ARM VERSATILE PLATFORM
 M:	Rob Herring <robh@kernel.org>
 L:	linux-pci at vger.kernel.org
@@ -7860,6 +7868,14 @@ L:	linux-pci at vger.kernel.org
 S:	Maintained
 F:	drivers/pci/host/*spear*
 
+PCI MSI DRIVER FOR ALTERA MSI IP
+M:	Ley Foon Tan <lftan@altera.com>
+L:	rfi at lists.rocketboards.org (moderated for non-subscribers)
+L:	linux-pci at vger.kernel.org
+S:	Supported
+F:	Documentation/devicetree/bindings/pci/altera-pcie-msi.txt
+F:	drivers/pci/host/pcie-altera-msi.c
+
 PCI MSI DRIVER FOR APPLIEDMICRO XGENE
 M:	Duc Dang <dhdang@apm.com>
 L:	linux-pci at vger.kernel.org
-- 
1.8.2.1

Re: [PATCH v2 3/5] pci: altera: Add Altera PCIe MSI driver

[Marc Zyngier]

On 31/07/15 11:15, Ley Foon Tan wrote:
> This patch adds Altera PCIe MSI driver. This soft IP supports configurable
> number of vectors, which is a dts parameter.

I've reviewed the initial drop of this code; basic courtesy would be to
keep me CCed on the follow-up series.

> 
> Signed-off-by: Ley Foon Tan <lftan@altera.com>
> ---
>  drivers/pci/host/Kconfig           |   7 +
>  drivers/pci/host/Makefile          |   1 +
>  drivers/pci/host/pcie-altera-msi.c | 324 +++++++++++++++++++++++++++++++++++++
>  3 files changed, 332 insertions(+)
>  create mode 100644 drivers/pci/host/pcie-altera-msi.c
> 
> diff --git a/drivers/pci/host/Kconfig b/drivers/pci/host/Kconfig
> index 108500a..51d0468 100644
> --- a/drivers/pci/host/Kconfig
> +++ b/drivers/pci/host/Kconfig
> @@ -153,4 +153,11 @@ config PCIE_ALTERA
>  	  Say Y here if you want to enable PCIe controller support for Altera
>  	  SoCFPGA family of SoCs.
>  
> +config PCIE_ALTERA_MSI
> +	bool "Altera PCIe MSI feature"
> +	depends on PCI_MSI
> +	help
> +	  Say Y here if you want PCIe MSI support for the Altera SocFPGA SoC.
> +	  This MSI driver supports Altera MSI to GIC controller IP.
> +
>  endmenu
> diff --git a/drivers/pci/host/Makefile b/drivers/pci/host/Makefile
> index 6954f76..6c4913d 100644
> --- a/drivers/pci/host/Makefile
> +++ b/drivers/pci/host/Makefile
> @@ -18,3 +18,4 @@ obj-$(CONFIG_PCIE_IPROC) += pcie-iproc.o
>  obj-$(CONFIG_PCIE_IPROC_PLATFORM) += pcie-iproc-platform.o
>  obj-$(CONFIG_PCIE_IPROC_BCMA) += pcie-iproc-bcma.o
>  obj-$(CONFIG_PCIE_ALTERA) += pcie-altera.o
> +obj-$(CONFIG_PCIE_ALTERA_MSI) += pcie-altera-msi.o
> diff --git a/drivers/pci/host/pcie-altera-msi.c b/drivers/pci/host/pcie-altera-msi.c
> new file mode 100644
> index 0000000..6014719
> --- /dev/null
> +++ b/drivers/pci/host/pcie-altera-msi.c
> @@ -0,0 +1,324 @@
> +/*
> + * Copyright Altera Corporation (C) 2013-2015. All rights reserved
> + *
> + * This program is free software; you can redistribute it and/or modify it
> + * under the terms and conditions of the GNU General Public License,
> + * version 2, as published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope it will be useful, but WITHOUT
> + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
> + * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
> + * more details.
> + *
> + * You should have received a copy of the GNU General Public License along with
> + * this program.  If not, see <http://www.gnu.org/licenses/>.
> + */
> +#include <linux/interrupt.h>
> +#include <linux/irqchip/chained_irq.h>
> +#include <linux/module.h>
> +#include <linux/msi.h>
> +#include <linux/of_address.h>
> +#include <linux/of_irq.h>
> +#include <linux/of_pci.h>
> +#include <linux/pci.h>
> +#include <linux/platform_device.h>
> +#include <linux/slab.h>
> +
> +#define MSI_STATUS		0x0
> +#define MSI_ERROR		0x4
> +#define MSI_INTMASK		0x8
> +
> +#define MAX_MSI_VECTORS		32
> +struct altera_msi {
> +	DECLARE_BITMAP(used, MAX_MSI_VECTORS);
> +	struct mutex		lock;	/* proctect used variable */
> +	struct platform_device	*pdev;
> +	struct irq_domain		*msi_domain;
> +	struct irq_domain		*inner_domain;
> +	void __iomem		*csr_base;
> +	void __iomem		*vector_base;
> +	phys_addr_t		vector_phy;
> +	u32			num_of_vectors;
> +	int			irq;
> +};
> +
> +static inline void msi_writel(struct altera_msi *msi, u32 value, u32 reg)
> +{
> +	writel_relaxed(value, msi->csr_base + reg);
> +}
> +
> +static inline u32 msi_readl(struct altera_msi *msi, u32 reg)
> +{
> +	return readl_relaxed(msi->csr_base + reg);
> +}
> +
> +static void altera_msi_isr(unsigned int irq, struct irq_desc *desc)
> +{
> +	struct irq_chip *chip = irq_desc_get_chip(desc);
> +	struct altera_msi *msi;
> +	unsigned long status;
> +	u32 num_of_vectors;
> +	u32 bit;
> +
> +	chained_irq_enter(chip, desc);
> +	msi = irq_desc_get_handler_data(desc);
> +	num_of_vectors = msi->num_of_vectors;
> +
> +	do {
> +		status = msi_readl(msi, MSI_STATUS);
> +		if (!status)
> +			break;
> +
> +		do {
> +			bit = find_first_bit(&status, num_of_vectors);
> +			/* Dummy read from vector to clear the interrupt */
> +			readl_relaxed(msi->vector_base + (bit * sizeof(u32)));
> +
> +			irq = irq_find_mapping(msi->inner_domain, bit);

Use of irq is a bit confusing, as this is a parameter describing the
chaining interrupt. Consider using a different name for this variable.

> +			if (irq) {
> +				if (test_bit(bit, msi->used))
> +					generic_handle_irq(irq);
> +				else
> +					dev_info(&msi->pdev->dev, "unhandled MSI\n");
> +			} else
> +				dev_info(&msi->pdev->dev, "unexpected MSI\n");

The whole statement could be rewritten as:

if (irq && test_bit(bit, msi->used))
	generic_handle_irq(irq);
else
	dev_err(...);

In all cases, this is a grave error, so dev_info is not appropriate here.

> +
> +			/* Clear the bit from status and repeat without reading
> +			 * again status register. */
> +			clear_bit(bit, &status);

You don't need an atomic operation here, so __clear_bit() should be enough.

> +		} while (status);
> +	} while (1);
> +
> +	chained_irq_exit(chip, desc);
> +}
> +
> +static struct irq_chip altera_msi_irq_chip = {
> +	.name = "Altera 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 struct msi_domain_info altera_msi_domain_info = {
> +	.flags	= (MSI_FLAG_USE_DEF_DOM_OPS | MSI_FLAG_USE_DEF_CHIP_OPS |
> +		     MSI_FLAG_PCI_MSIX),
> +	.chip	= &altera_msi_irq_chip,
> +};
> +
> +static void altera_compose_msi_msg(struct irq_data *data, struct msi_msg *msg)
> +{
> +	u32 mask;
> +	struct altera_msi *msi = irq_data_get_irq_chip_data(data);
> +	u64 addr = msi->vector_phy + (data->hwirq * sizeof(u32));

Use phys_addr_t.

> +
> +	msg->address_lo = lower_32_bits(addr);
> +	msg->address_hi = upper_32_bits(addr);
> +	msg->data = data->hwirq;
> +
> +	mask = msi_readl(msi, MSI_INTMASK);
> +	mask |= 1 << data->hwirq;
> +	msi_writel(msi, mask, MSI_INTMASK);

It feels a bit weird to unmask the interrupt when you compose the
message. I'd expect this to be done in the mask/unmask methods.

> +	dev_dbg(&msi->pdev->dev, "msi#%d address_lo 0x%x\n", (int)data->hwirq,
> +		msg->address_lo);

You may want to print the whole address here.

> +}
> +
> +static int altera_msi_set_affinity(struct irq_data *irq_data,
> +				 const struct cpumask *mask, bool force)
> +{
> +	 return -EINVAL;
> +}
> +
> +static struct irq_chip altera_msi_bottom_irq_chip = {
> +	.name			= "Altera MSI",
> +	.irq_compose_msi_msg	= altera_compose_msi_msg,
> +	.irq_set_affinity	= altera_msi_set_affinity,
> +};
> +
> +static int altera_irq_domain_alloc(struct irq_domain *domain, unsigned int virq,
> +				  unsigned int nr_irqs, void *args)
> +{
> +	struct altera_msi *msi = domain->host_data;
> +	int bit;
> +
> +	mutex_lock(&msi->lock);
> +
> +	bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
> +	if (bit < msi->num_of_vectors)
> +		set_bit(bit, msi->used);
> +
> +	mutex_unlock(&msi->lock);
> +
> +	if (bit < 0)
> +		return bit;

How can "bit" be negative here? find_first_zero_bit returns an unsigned
long... You probably want to change the type of "bit" to reflect that.

> +	else if (bit >= msi->num_of_vectors)

Useless "else" anyway.

> +		return -ENOSPC;
> +
> +	irq_domain_set_info(domain, virq, bit, &altera_msi_bottom_irq_chip,
> +			    domain->host_data, handle_simple_irq,
> +			    NULL, NULL);
> +	set_irq_flags(virq, IRQF_VALID);
> +
> +	return 0;
> +}
> +
> +static void altera_irq_domain_free(struct irq_domain *domain,
> +				  unsigned int virq, unsigned int nr_irqs)
> +{
> +	struct irq_data *d = irq_domain_get_irq_data(domain, virq);
> +	struct altera_msi *msi = irq_data_get_irq_chip_data(d);
> +	u32 mask;
> +
> +	mutex_lock(&msi->lock);
> +
> +	if (!test_bit(d->hwirq, msi->used))
> +		dev_err(&msi->pdev->dev, "trying to free unused MSI#%lu\n",
> +		d->hwirq);

Put this statement in { } so that it matches the "else" construct.

> +	else {
> +		clear_bit(d->hwirq, msi->used);

No need for an atomic operation here, you own the mutex.

> +		mask = msi_readl(msi, MSI_INTMASK);
> +		mask &= ~(1 << d->hwirq);
> +		msi_writel(msi, mask, MSI_INTMASK);
> +	}
> +
> +	mutex_unlock(&msi->lock);
> +}
> +
> +static const struct irq_domain_ops msi_domain_ops = {
> +	.alloc	= altera_irq_domain_alloc,
> +	.free	= altera_irq_domain_free,
> +};
> +
> +static int altera_allocate_domains(struct altera_msi *msi)
> +{
> +	msi->inner_domain = irq_domain_add_linear(NULL, msi->num_of_vectors,
> +					     &msi_domain_ops, msi);
> +	if (!msi->inner_domain) {
> +		dev_err(&msi->pdev->dev, "failed to create IRQ domain\n");
> +		return -ENOMEM;
> +	}
> +
> +	msi->msi_domain = pci_msi_create_irq_domain(
> +				msi->pdev->dev.of_node,

Please put at least one argument to the function on the same line. This
is otherwise very hard to read.

> +				&altera_msi_domain_info, msi->inner_domain);
> +	if (!msi->msi_domain) {
> +		dev_err(&msi->pdev->dev, "failed to create MSI domain\n");
> +		irq_domain_remove(msi->inner_domain);
> +		return -ENOMEM;
> +	}
> +
> +	return 0;
> +}
> +
> +static void altera_free_domains(struct altera_msi *msi)
> +{
> +	irq_domain_remove(msi->msi_domain);
> +	irq_domain_remove(msi->inner_domain);
> +}
> +
> +
> +static int altera_msi_remove(struct platform_device *pdev)
> +{
> +	struct altera_msi *msi = platform_get_drvdata(pdev);
> +
> +	msi_writel(msi, 0, MSI_INTMASK);
> +	irq_set_chained_handler(msi->irq, NULL);
> +	irq_set_handler_data(msi->irq, NULL);
> +
> +	altera_free_domains(msi);
> +
> +	platform_set_drvdata(pdev, NULL);
> +	return 0;
> +}
> +
> +int altera_msi_probe(struct platform_device *pdev)

Make this static.

> +{
> +	struct altera_msi *msi;
> +	struct device_node *np = pdev->dev.of_node;
> +	struct resource *res;
> +	int ret;
> +
> +	msi = devm_kzalloc(&pdev->dev, sizeof(struct altera_msi),
> +		GFP_KERNEL);
> +	if (!msi)
> +		return -ENOMEM;
> +
> +	mutex_init(&msi->lock);
> +	msi->pdev = pdev;
> +
> +	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "csr");
> +	msi->csr_base = devm_ioremap_resource(&pdev->dev, res);
> +	if (IS_ERR(msi->csr_base)) {
> +		dev_err(&pdev->dev, "get csr resource failed\n");
> +		return PTR_ERR(msi->csr_base);
> +	}
> +
> +	res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
> +			"vector_slave");
> +	msi->vector_base = devm_ioremap_resource(&pdev->dev, res);
> +	if (IS_ERR(msi->vector_base)) {
> +		dev_err(&pdev->dev, "get vector slave resource failed\n");
> +		return PTR_ERR(msi->vector_base);
> +	}
> +
> +	msi->vector_phy = res->start;
> +
> +	if (of_property_read_u32(np, "num-vectors", &msi->num_of_vectors)) {
> +		dev_err(&pdev->dev, "failed to parse the number of vectors\n");
> +		return -EINVAL;
> +	}
> +
> +	ret = altera_allocate_domains(msi);
> +	if (ret)
> +		return ret;
> +
> +	msi->irq = platform_get_irq(pdev, 0);
> +	if (msi->irq <= 0) {
> +		dev_err(&pdev->dev, "failed to map IRQ: %d\n", msi->irq);
> +		ret = -ENODEV;
> +		goto err;
> +	}
> +
> +	irq_set_chained_handler(msi->irq, altera_msi_isr);
> +	ret = irq_set_handler_data(msi->irq, msi);

Use irq_set_chained_handler_and_data.

> +	if (ret) {
> +		dev_err(&pdev->dev, "failed to register IRQ handler: %d\n",
> +			ret);
> +		goto err;
> +	}
> +
> +	platform_set_drvdata(pdev, msi);
> +
> +	return 0;
> +
> +err:
> +	altera_msi_remove(pdev);
> +	return ret;
> +}
> +
> +static const struct of_device_id altera_msi_of_match[] = {
> +	{ .compatible = "altr,msi-1.0", NULL },
> +	{ },
> +};
> +MODULE_DEVICE_TABLE(of, altera_msi_of_match);
> +
> +static struct platform_driver altera_msi_driver = {
> +	.driver = {
> +		.name = "altera-msi",
> +		.owner = THIS_MODULE,

Remove this .owner assignment, the core code deals with it.

> +		.of_match_table = altera_msi_of_match,
> +	},
> +	.probe = altera_msi_probe,
> +	.remove = altera_msi_remove,
> +};
> +
> +static int __init altera_msi_init(void)
> +{
> +	return platform_driver_register(&altera_msi_driver);
> +}
> +
> +subsys_initcall(altera_msi_init);

Hmmm. if you're initializing this as part of the core kernel, it will
never be a module. What is the point of the MODULE_DEVICE_TABLE above?

> +
> +MODULE_AUTHOR("Ley Foon Tan <lftan@altera.com>");
> +MODULE_DESCRIPTION("Altera PCIe MSI support");
> +MODULE_LICENSE("GPL v2");
> 

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...

[PATCH v2 3/5] pci: altera: Add Altera PCIe MSI driver

[Marc Zyngier]

On 31/07/15 11:15, Ley Foon Tan wrote:
> This patch adds Altera PCIe MSI driver. This soft IP supports configurable
> number of vectors, which is a dts parameter.

I've reviewed the initial drop of this code; basic courtesy would be to
keep me CCed on the follow-up series.

> 
> Signed-off-by: Ley Foon Tan <lftan@altera.com>
> ---
>  drivers/pci/host/Kconfig           |   7 +
>  drivers/pci/host/Makefile          |   1 +
>  drivers/pci/host/pcie-altera-msi.c | 324 +++++++++++++++++++++++++++++++++++++
>  3 files changed, 332 insertions(+)
>  create mode 100644 drivers/pci/host/pcie-altera-msi.c
> 
> diff --git a/drivers/pci/host/Kconfig b/drivers/pci/host/Kconfig
> index 108500a..51d0468 100644
> --- a/drivers/pci/host/Kconfig
> +++ b/drivers/pci/host/Kconfig
> @@ -153,4 +153,11 @@ config PCIE_ALTERA
>  	  Say Y here if you want to enable PCIe controller support for Altera
>  	  SoCFPGA family of SoCs.
>  
> +config PCIE_ALTERA_MSI
> +	bool "Altera PCIe MSI feature"
> +	depends on PCI_MSI
> +	help
> +	  Say Y here if you want PCIe MSI support for the Altera SocFPGA SoC.
> +	  This MSI driver supports Altera MSI to GIC controller IP.
> +
>  endmenu
> diff --git a/drivers/pci/host/Makefile b/drivers/pci/host/Makefile
> index 6954f76..6c4913d 100644
> --- a/drivers/pci/host/Makefile
> +++ b/drivers/pci/host/Makefile
> @@ -18,3 +18,4 @@ obj-$(CONFIG_PCIE_IPROC) += pcie-iproc.o
>  obj-$(CONFIG_PCIE_IPROC_PLATFORM) += pcie-iproc-platform.o
>  obj-$(CONFIG_PCIE_IPROC_BCMA) += pcie-iproc-bcma.o
>  obj-$(CONFIG_PCIE_ALTERA) += pcie-altera.o
> +obj-$(CONFIG_PCIE_ALTERA_MSI) += pcie-altera-msi.o
> diff --git a/drivers/pci/host/pcie-altera-msi.c b/drivers/pci/host/pcie-altera-msi.c
> new file mode 100644
> index 0000000..6014719
> --- /dev/null
> +++ b/drivers/pci/host/pcie-altera-msi.c
> @@ -0,0 +1,324 @@
> +/*
> + * Copyright Altera Corporation (C) 2013-2015. All rights reserved
> + *
> + * This program is free software; you can redistribute it and/or modify it
> + * under the terms and conditions of the GNU General Public License,
> + * version 2, as published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope it will be useful, but WITHOUT
> + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
> + * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
> + * more details.
> + *
> + * You should have received a copy of the GNU General Public License along with
> + * this program.  If not, see <http://www.gnu.org/licenses/>.
> + */
> +#include <linux/interrupt.h>
> +#include <linux/irqchip/chained_irq.h>
> +#include <linux/module.h>
> +#include <linux/msi.h>
> +#include <linux/of_address.h>
> +#include <linux/of_irq.h>
> +#include <linux/of_pci.h>
> +#include <linux/pci.h>
> +#include <linux/platform_device.h>
> +#include <linux/slab.h>
> +
> +#define MSI_STATUS		0x0
> +#define MSI_ERROR		0x4
> +#define MSI_INTMASK		0x8
> +
> +#define MAX_MSI_VECTORS		32
> +struct altera_msi {
> +	DECLARE_BITMAP(used, MAX_MSI_VECTORS);
> +	struct mutex		lock;	/* proctect used variable */
> +	struct platform_device	*pdev;
> +	struct irq_domain		*msi_domain;
> +	struct irq_domain		*inner_domain;
> +	void __iomem		*csr_base;
> +	void __iomem		*vector_base;
> +	phys_addr_t		vector_phy;
> +	u32			num_of_vectors;
> +	int			irq;
> +};
> +
> +static inline void msi_writel(struct altera_msi *msi, u32 value, u32 reg)
> +{
> +	writel_relaxed(value, msi->csr_base + reg);
> +}
> +
> +static inline u32 msi_readl(struct altera_msi *msi, u32 reg)
> +{
> +	return readl_relaxed(msi->csr_base + reg);
> +}
> +
> +static void altera_msi_isr(unsigned int irq, struct irq_desc *desc)
> +{
> +	struct irq_chip *chip = irq_desc_get_chip(desc);
> +	struct altera_msi *msi;
> +	unsigned long status;
> +	u32 num_of_vectors;
> +	u32 bit;
> +
> +	chained_irq_enter(chip, desc);
> +	msi = irq_desc_get_handler_data(desc);
> +	num_of_vectors = msi->num_of_vectors;
> +
> +	do {
> +		status = msi_readl(msi, MSI_STATUS);
> +		if (!status)
> +			break;
> +
> +		do {
> +			bit = find_first_bit(&status, num_of_vectors);
> +			/* Dummy read from vector to clear the interrupt */
> +			readl_relaxed(msi->vector_base + (bit * sizeof(u32)));
> +
> +			irq = irq_find_mapping(msi->inner_domain, bit);

Use of irq is a bit confusing, as this is a parameter describing the
chaining interrupt. Consider using a different name for this variable.

> +			if (irq) {
> +				if (test_bit(bit, msi->used))
> +					generic_handle_irq(irq);
> +				else
> +					dev_info(&msi->pdev->dev, "unhandled MSI\n");
> +			} else
> +				dev_info(&msi->pdev->dev, "unexpected MSI\n");

The whole statement could be rewritten as:

if (irq && test_bit(bit, msi->used))
	generic_handle_irq(irq);
else
	dev_err(...);

In all cases, this is a grave error, so dev_info is not appropriate here.

> +
> +			/* Clear the bit from status and repeat without reading
> +			 * again status register. */
> +			clear_bit(bit, &status);

You don't need an atomic operation here, so __clear_bit() should be enough.

> +		} while (status);
> +	} while (1);
> +
> +	chained_irq_exit(chip, desc);
> +}
> +
> +static struct irq_chip altera_msi_irq_chip = {
> +	.name = "Altera 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 struct msi_domain_info altera_msi_domain_info = {
> +	.flags	= (MSI_FLAG_USE_DEF_DOM_OPS | MSI_FLAG_USE_DEF_CHIP_OPS |
> +		     MSI_FLAG_PCI_MSIX),
> +	.chip	= &altera_msi_irq_chip,
> +};
> +
> +static void altera_compose_msi_msg(struct irq_data *data, struct msi_msg *msg)
> +{
> +	u32 mask;
> +	struct altera_msi *msi = irq_data_get_irq_chip_data(data);
> +	u64 addr = msi->vector_phy + (data->hwirq * sizeof(u32));

Use phys_addr_t.

> +
> +	msg->address_lo = lower_32_bits(addr);
> +	msg->address_hi = upper_32_bits(addr);
> +	msg->data = data->hwirq;
> +
> +	mask = msi_readl(msi, MSI_INTMASK);
> +	mask |= 1 << data->hwirq;
> +	msi_writel(msi, mask, MSI_INTMASK);

It feels a bit weird to unmask the interrupt when you compose the
message. I'd expect this to be done in the mask/unmask methods.

> +	dev_dbg(&msi->pdev->dev, "msi#%d address_lo 0x%x\n", (int)data->hwirq,
> +		msg->address_lo);

You may want to print the whole address here.

> +}
> +
> +static int altera_msi_set_affinity(struct irq_data *irq_data,
> +				 const struct cpumask *mask, bool force)
> +{
> +	 return -EINVAL;
> +}
> +
> +static struct irq_chip altera_msi_bottom_irq_chip = {
> +	.name			= "Altera MSI",
> +	.irq_compose_msi_msg	= altera_compose_msi_msg,
> +	.irq_set_affinity	= altera_msi_set_affinity,
> +};
> +
> +static int altera_irq_domain_alloc(struct irq_domain *domain, unsigned int virq,
> +				  unsigned int nr_irqs, void *args)
> +{
> +	struct altera_msi *msi = domain->host_data;
> +	int bit;
> +
> +	mutex_lock(&msi->lock);
> +
> +	bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
> +	if (bit < msi->num_of_vectors)
> +		set_bit(bit, msi->used);
> +
> +	mutex_unlock(&msi->lock);
> +
> +	if (bit < 0)
> +		return bit;

How can "bit" be negative here? find_first_zero_bit returns an unsigned
long... You probably want to change the type of "bit" to reflect that.

> +	else if (bit >= msi->num_of_vectors)

Useless "else" anyway.

> +		return -ENOSPC;
> +
> +	irq_domain_set_info(domain, virq, bit, &altera_msi_bottom_irq_chip,
> +			    domain->host_data, handle_simple_irq,
> +			    NULL, NULL);
> +	set_irq_flags(virq, IRQF_VALID);
> +
> +	return 0;
> +}
> +
> +static void altera_irq_domain_free(struct irq_domain *domain,
> +				  unsigned int virq, unsigned int nr_irqs)
> +{
> +	struct irq_data *d = irq_domain_get_irq_data(domain, virq);
> +	struct altera_msi *msi = irq_data_get_irq_chip_data(d);
> +	u32 mask;
> +
> +	mutex_lock(&msi->lock);
> +
> +	if (!test_bit(d->hwirq, msi->used))
> +		dev_err(&msi->pdev->dev, "trying to free unused MSI#%lu\n",
> +		d->hwirq);

Put this statement in { } so that it matches the "else" construct.

> +	else {
> +		clear_bit(d->hwirq, msi->used);

No need for an atomic operation here, you own the mutex.

> +		mask = msi_readl(msi, MSI_INTMASK);
> +		mask &= ~(1 << d->hwirq);
> +		msi_writel(msi, mask, MSI_INTMASK);
> +	}
> +
> +	mutex_unlock(&msi->lock);
> +}
> +
> +static const struct irq_domain_ops msi_domain_ops = {
> +	.alloc	= altera_irq_domain_alloc,
> +	.free	= altera_irq_domain_free,
> +};
> +
> +static int altera_allocate_domains(struct altera_msi *msi)
> +{
> +	msi->inner_domain = irq_domain_add_linear(NULL, msi->num_of_vectors,
> +					     &msi_domain_ops, msi);
> +	if (!msi->inner_domain) {
> +		dev_err(&msi->pdev->dev, "failed to create IRQ domain\n");
> +		return -ENOMEM;
> +	}
> +
> +	msi->msi_domain = pci_msi_create_irq_domain(
> +				msi->pdev->dev.of_node,

Please put at least one argument to the function on the same line. This
is otherwise very hard to read.

> +				&altera_msi_domain_info, msi->inner_domain);
> +	if (!msi->msi_domain) {
> +		dev_err(&msi->pdev->dev, "failed to create MSI domain\n");
> +		irq_domain_remove(msi->inner_domain);
> +		return -ENOMEM;
> +	}
> +
> +	return 0;
> +}
> +
> +static void altera_free_domains(struct altera_msi *msi)
> +{
> +	irq_domain_remove(msi->msi_domain);
> +	irq_domain_remove(msi->inner_domain);
> +}
> +
> +
> +static int altera_msi_remove(struct platform_device *pdev)
> +{
> +	struct altera_msi *msi = platform_get_drvdata(pdev);
> +
> +	msi_writel(msi, 0, MSI_INTMASK);
> +	irq_set_chained_handler(msi->irq, NULL);
> +	irq_set_handler_data(msi->irq, NULL);
> +
> +	altera_free_domains(msi);
> +
> +	platform_set_drvdata(pdev, NULL);
> +	return 0;
> +}
> +
> +int altera_msi_probe(struct platform_device *pdev)

Make this static.

> +{
> +	struct altera_msi *msi;
> +	struct device_node *np = pdev->dev.of_node;
> +	struct resource *res;
> +	int ret;
> +
> +	msi = devm_kzalloc(&pdev->dev, sizeof(struct altera_msi),
> +		GFP_KERNEL);
> +	if (!msi)
> +		return -ENOMEM;
> +
> +	mutex_init(&msi->lock);
> +	msi->pdev = pdev;
> +
> +	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "csr");
> +	msi->csr_base = devm_ioremap_resource(&pdev->dev, res);
> +	if (IS_ERR(msi->csr_base)) {
> +		dev_err(&pdev->dev, "get csr resource failed\n");
> +		return PTR_ERR(msi->csr_base);
> +	}
> +
> +	res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
> +			"vector_slave");
> +	msi->vector_base = devm_ioremap_resource(&pdev->dev, res);
> +	if (IS_ERR(msi->vector_base)) {
> +		dev_err(&pdev->dev, "get vector slave resource failed\n");
> +		return PTR_ERR(msi->vector_base);
> +	}
> +
> +	msi->vector_phy = res->start;
> +
> +	if (of_property_read_u32(np, "num-vectors", &msi->num_of_vectors)) {
> +		dev_err(&pdev->dev, "failed to parse the number of vectors\n");
> +		return -EINVAL;
> +	}
> +
> +	ret = altera_allocate_domains(msi);
> +	if (ret)
> +		return ret;
> +
> +	msi->irq = platform_get_irq(pdev, 0);
> +	if (msi->irq <= 0) {
> +		dev_err(&pdev->dev, "failed to map IRQ: %d\n", msi->irq);
> +		ret = -ENODEV;
> +		goto err;
> +	}
> +
> +	irq_set_chained_handler(msi->irq, altera_msi_isr);
> +	ret = irq_set_handler_data(msi->irq, msi);

Use irq_set_chained_handler_and_data.

> +	if (ret) {
> +		dev_err(&pdev->dev, "failed to register IRQ handler: %d\n",
> +			ret);
> +		goto err;
> +	}
> +
> +	platform_set_drvdata(pdev, msi);
> +
> +	return 0;
> +
> +err:
> +	altera_msi_remove(pdev);
> +	return ret;
> +}
> +
> +static const struct of_device_id altera_msi_of_match[] = {
> +	{ .compatible = "altr,msi-1.0", NULL },
> +	{ },
> +};
> +MODULE_DEVICE_TABLE(of, altera_msi_of_match);
> +
> +static struct platform_driver altera_msi_driver = {
> +	.driver = {
> +		.name = "altera-msi",
> +		.owner = THIS_MODULE,

Remove this .owner assignment, the core code deals with it.

> +		.of_match_table = altera_msi_of_match,
> +	},
> +	.probe = altera_msi_probe,
> +	.remove = altera_msi_remove,
> +};
> +
> +static int __init altera_msi_init(void)
> +{
> +	return platform_driver_register(&altera_msi_driver);
> +}
> +
> +subsys_initcall(altera_msi_init);

Hmmm. if you're initializing this as part of the core kernel, it will
never be a module. What is the point of the MODULE_DEVICE_TABLE above?

> +
> +MODULE_AUTHOR("Ley Foon Tan <lftan@altera.com>");
> +MODULE_DESCRIPTION("Altera PCIe MSI support");
> +MODULE_LICENSE("GPL v2");
> 

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...

Re: [PATCH v2 2/5] pci:host: Add Altera PCIe host controller driver

[Marc Zyngier]

On 31/07/15 11:15, Ley Foon Tan wrote:
> This patch adds the Altera PCIe host controller driver.
> 
> Signed-off-by: Ley Foon Tan <lftan@altera.com>
> ---
>  drivers/pci/host/Kconfig       |   8 +
>  drivers/pci/host/Makefile      |   1 +
>  drivers/pci/host/pcie-altera.c | 526 +++++++++++++++++++++++++++++++++++++++++
>  3 files changed, 535 insertions(+)
>  create mode 100644 drivers/pci/host/pcie-altera.c
> 
> diff --git a/drivers/pci/host/Kconfig b/drivers/pci/host/Kconfig
> index 675c2d1..108500a 100644
> --- a/drivers/pci/host/Kconfig
> +++ b/drivers/pci/host/Kconfig
> @@ -145,4 +145,12 @@ config PCIE_IPROC_BCMA
>  	  Say Y here if you want to use the Broadcom iProc PCIe controller
>  	  through the BCMA bus interface
>  
> +config PCIE_ALTERA
> +	tristate "Altera PCIe controller"
> +	depends on ARCH_SOCFPGA
> +	select PCI_MSI_IRQ_DOMAIN if PCI_MSI

Why would you select this here? MSI support is in another driver, so I
don't see the point.

> +	help
> +	  Say Y here if you want to enable PCIe controller support for Altera
> +	  SoCFPGA family of SoCs.
> +
>  endmenu
> diff --git a/drivers/pci/host/Makefile b/drivers/pci/host/Makefile
> index 140d66f..6954f76 100644
> --- a/drivers/pci/host/Makefile
> +++ b/drivers/pci/host/Makefile
> @@ -17,3 +17,4 @@ obj-$(CONFIG_PCI_VERSATILE) += pci-versatile.o
>  obj-$(CONFIG_PCIE_IPROC) += pcie-iproc.o
>  obj-$(CONFIG_PCIE_IPROC_PLATFORM) += pcie-iproc-platform.o
>  obj-$(CONFIG_PCIE_IPROC_BCMA) += pcie-iproc-bcma.o
> +obj-$(CONFIG_PCIE_ALTERA) += pcie-altera.o
> diff --git a/drivers/pci/host/pcie-altera.c b/drivers/pci/host/pcie-altera.c
> new file mode 100644
> index 0000000..3ec51c7
> --- /dev/null
> +++ b/drivers/pci/host/pcie-altera.c
> @@ -0,0 +1,526 @@
> +/*
> + * Copyright Altera Corporation (C) 2013-2015. All rights reserved
> + *
> + * This program is free software; you can redistribute it and/or modify it
> + * under the terms and conditions of the GNU General Public License,
> + * version 2, as published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope it will be useful, but WITHOUT
> + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
> + * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
> + * more details.
> + *
> + * You should have received a copy of the GNU General Public License along with
> + * this program.  If not, see <http://www.gnu.org/licenses/>.
> + */
> +#include <linux/interrupt.h>
> +#include <linux/module.h>
> +#include <linux/of_address.h>
> +#include <linux/of_irq.h>
> +#include <linux/of_pci.h>
> +#include <linux/pci.h>
> +#include <linux/platform_device.h>
> +#include <linux/slab.h>
> +
> +#define A2P_ADDR_MAP_LO0		0x1000
> +#define A2P_ADDR_MAP_HI0		0x1004
> +#define RP_TX_REG0			0x2000
> +#define RP_TX_REG1			0x2004
> +#define RP_TX_CNTRL			0x2008
> +#define RP_TX_EOP			0x2
> +#define RP_TX_SOP			0x1
> +#define RP_RXCPL_STATUS		0x2010
> +#define RP_RXCPL_EOP			0x2
> +#define RP_RXCPL_SOP			0x1
> +#define RP_RXCPL_REG0			0x2014
> +#define RP_RXCPL_REG1			0x2018
> +#define P2A_INT_STATUS			0x3060
> +#define P2A_INT_STS_ALL			0xF
> +#define P2A_INT_ENABLE			0x3070
> +#define P2A_INT_ENA_ALL			0xF
> +#define RP_LTSSM			0x3C64
> +#define LTSSM_L0			0xF
> +
> +/* TLP configuration type 0 and 1 */
> +#define TLP_FMTTYPE_CFGRD0		0x04	/* Configuration Read Type 0 */
> +#define TLP_FMTTYPE_CFGWR0		0x44	/* Configuration Write Type 0 */
> +#define TLP_FMTTYPE_CFGRD1		0x05	/* Configuration Read Type 1 */
> +#define TLP_FMTTYPE_CFGWR1		0x45	/* Configuration Write Type 1 */
> +#define TLP_PAYLOAD_SIZE		0x01
> +#define TLP_READ_TAG			0x1D
> +#define TLP_WRITE_TAG			0x10
> +#define TLP_CFG_DW0(fmttype)		(((fmttype) << 24) | TLP_PAYLOAD_SIZE)
> +#define TLP_CFG_DW1(reqid, tag)		(((reqid) << 16) | (tag << 8) | 0xF)
> +#define TLP_CFG_DW2(bus, devfn, offset)	\
> +				(((bus) << 24) | ((devfn) << 16) | (offset))
> +#define TLP_REQ_ID(bus, devfn)		(((bus) << 8) | (devfn))
> +#define TLP_COMPL_STATUS(hdr)		(((hdr) & 0xE0) >> 13)
> +#define TLP_HDR_SIZE			3
> +#define TLP_LOOP			10
> +
> +#define INTX_NUM		4
> +
> +/* Address translation table entry size */
> +#define ATT_ENTRY_SIZE		8
> +
> +#define DWORD_MASK		3
> +
> +struct altera_pcie {
> +	struct platform_device	*pdev;
> +	struct resource		*txs;
> +	void __iomem		*cra_base;
> +	int			hwirq;
> +	u8			root_bus_nr;
> +	struct irq_domain		*irq_domain;
> +	struct resource		bus_range;
> +	struct list_head		resources;
> +};
> +
> +struct tlp_rp_regpair_t {
> +	u32 ctrl;
> +	u32 reg0;
> +	u32 reg1;
> +};
> +
> +static void altera_pcie_retrain(struct pci_dev *dev)
> +{
> +	u16 linkcap, linkstat;
> +
> +	/*
> +	 * Set the retrain bit if the PCIe rootport support > 2.5GB/s, but
> +	 * current speed is 2.5 GB/s.
> +	 */
> +	pcie_capability_read_word(dev, PCI_EXP_LNKCAP, &linkcap);
> +
> +	if ((linkcap & PCI_EXP_LNKCAP_SLS) <= PCI_EXP_LNKCAP_SLS_2_5GB)
> +		return;
> +
> +	pcie_capability_read_word(dev, PCI_EXP_LNKSTA, &linkstat);
> +	if ((linkstat & PCI_EXP_LNKSTA_CLS) == PCI_EXP_LNKSTA_CLS_2_5GB)
> +		pcie_capability_set_word(dev, PCI_EXP_LNKCTL,
> +					PCI_EXP_LNKCTL_RL);
> +
> +}
> +DECLARE_PCI_FIXUP_EARLY(PCI_ANY_ID, PCI_ANY_ID, altera_pcie_retrain);
> +
> +static inline void cra_writel(struct altera_pcie *pcie, u32 value, u32 reg)
> +{
> +	writel_relaxed(value, pcie->cra_base + reg);
> +}
> +
> +static inline u32 cra_readl(struct altera_pcie *pcie, u32 reg)
> +{
> +	return readl_relaxed(pcie->cra_base + reg);
> +}
> +
> +static void tlp_read_rx(struct altera_pcie *pcie,
> +			struct tlp_rp_regpair_t *tlp_rp_regdata)
> +{
> +	tlp_rp_regdata->ctrl = cra_readl(pcie, RP_RXCPL_STATUS);
> +	tlp_rp_regdata->reg0 = cra_readl(pcie, RP_RXCPL_REG0);
> +	tlp_rp_regdata->reg1 = cra_readl(pcie, RP_RXCPL_REG1);
> +}
> +
> +static void tlp_write_tx(struct altera_pcie *pcie,
> +			struct tlp_rp_regpair_t *tlp_rp_regdata)
> +{
> +	cra_writel(pcie, tlp_rp_regdata->reg0, RP_TX_REG0);
> +	cra_writel(pcie, tlp_rp_regdata->reg1, RP_TX_REG1);
> +	cra_writel(pcie, tlp_rp_regdata->ctrl, RP_TX_CNTRL);
> +}
> +
> +static bool altera_pcie_link_is_up(struct altera_pcie *pcie)
> +{
> +	return !!(cra_readl(pcie, RP_LTSSM) & LTSSM_L0);
> +}
> +
> +static bool altera_pcie_valid_config(struct altera_pcie *pcie,
> +				struct pci_bus *bus, int dev)
> +{
> +	/* If there is no link, then there is no device */
> +	if (bus->number != pcie->root_bus_nr) {
> +		if (!altera_pcie_link_is_up(pcie))
> +			return false;
> +	}
> +
> +	/* access only one slot on each root port */
> +	if (bus->number == pcie->root_bus_nr && dev > 0)
> +		return false;
> +
> +	/*
> +	 * Do not read more than one device on the bus directly attached
> +	 * to RC.
> +	 */
> +	if (bus->primary == pcie->root_bus_nr && dev > 0)
> +		return false;
> +
> +	 return true;
> +}
> +
> +static int tlp_read_packet(struct altera_pcie *pcie, u32 *value)
> +{
> +	u8 loop;
> +	struct tlp_rp_regpair_t tlp_rp_regdata;
> +
> +	for (loop = TLP_LOOP; loop > 0; loop--) {
> +		tlp_read_rx(pcie, &tlp_rp_regdata);
> +		if (tlp_rp_regdata.ctrl & RP_RXCPL_EOP) {
> +			if (value)
> +				*value = tlp_rp_regdata.reg0;
> +			return PCIBIOS_SUCCESSFUL;
> +		}
> +	}
> +
> +	return -ENOENT;
> +}
> +
> +static void tlp_write_packet(struct altera_pcie *pcie, u32 *headers, u32 data)
> +{
> +	struct tlp_rp_regpair_t tlp_rp_regdata;
> +
> +	tlp_rp_regdata.reg0 = headers[0];
> +	tlp_rp_regdata.reg1 = headers[1];
> +	tlp_rp_regdata.ctrl = RP_TX_SOP;
> +	tlp_write_tx(pcie, &tlp_rp_regdata);
> +
> +	tlp_rp_regdata.reg0 = headers[2];
> +	tlp_rp_regdata.reg1 = data;
> +	tlp_rp_regdata.ctrl = RP_TX_EOP;
> +	tlp_write_tx(pcie, &tlp_rp_regdata);
> +}
> +
> +static int tlp_cfg_dword_read(struct altera_pcie *pcie, u8 bus, u32 devfn,
> +			      int where, u32 *value)
> +{
> +	int ret;
> +	u32 headers[TLP_HDR_SIZE];
> +
> +	if (bus == pcie->root_bus_nr)
> +		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGRD0);
> +	else
> +		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGRD1);
> +
> +	headers[1] = TLP_CFG_DW1(TLP_REQ_ID(pcie->root_bus_nr, devfn),
> +					TLP_READ_TAG);
> +	headers[2] = TLP_CFG_DW2(bus, devfn, where);
> +
> +	tlp_write_packet(pcie, headers, 0);
> +
> +	ret = tlp_read_packet(pcie, value);
> +	if (ret)
> +		*value = ~0UL;	/* return 0xFFFFFFFF if error */
> +
> +	return ret;
> +}
> +
> +static int tlp_cfg_dword_write(struct altera_pcie *pcie, u8 bus, u32 devfn,
> +			       int where, u32 value)
> +{
> +	u32 headers[TLP_HDR_SIZE];
> +
> +	if (bus == pcie->root_bus_nr)
> +		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGWR0);
> +	else
> +		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGWR1);
> +
> +	headers[1] = TLP_CFG_DW1(TLP_REQ_ID(pcie->root_bus_nr, devfn),
> +					TLP_WRITE_TAG);
> +	headers[2] = TLP_CFG_DW2(bus, devfn, where);
> +
> +	tlp_write_packet(pcie, headers, value);
> +
> +	tlp_read_packet(pcie, NULL);
> +
> +	/* Keep an eye out for changes to the root bus number */
> +	if ((bus == pcie->root_bus_nr) && (where == PCI_PRIMARY_BUS))
> +		pcie->root_bus_nr = (u8)(value);
> +
> +	return PCIBIOS_SUCCESSFUL;
> +}
> +
> +static int altera_pcie_cfg_read(struct pci_bus *bus, unsigned int devfn,
> +			       int where, int size, u32 *value)
> +{
> +	struct altera_pcie *pcie = bus->sysdata;
> +	int ret;
> +
> +	if (!altera_pcie_valid_config(pcie, bus, PCI_SLOT(devfn))) {
> +		*value = ~0UL;
> +		return PCIBIOS_DEVICE_NOT_FOUND;
> +	}
> +
> +	ret = tlp_cfg_dword_read(pcie, bus->number, devfn,
> +		(where & ~DWORD_MASK), value);
> +	if (ret != PCIBIOS_SUCCESSFUL)
> +		return ret;
> +
> +	if (size == 1)
> +		*value = (*value >> (8 * (where & 0x3))) & 0xff;
> +	else if (size == 2)
> +		*value = (*value >> (8 * (where & 0x2))) & 0xffff;
> +
> +	return PCIBIOS_SUCCESSFUL;
> +}
> +
> +static int altera_pcie_cfg_write(struct pci_bus *bus, unsigned int devfn,
> +				int where, int size, u32 value)
> +{
> +	struct altera_pcie *pcie = bus->sysdata;
> +	u32 data32 = value;
> +	u32 shift = 8 * (where & 3);
> +	int ret;
> +
> +	if (!altera_pcie_valid_config(pcie, bus, PCI_SLOT(devfn)))
> +		return PCIBIOS_DEVICE_NOT_FOUND;
> +
> +	/* write partial */
> +	if (size != sizeof(u32)) {
> +		ret = tlp_cfg_dword_read(pcie, bus->number, devfn,
> +			where & ~DWORD_MASK, &data32);
> +		if (ret)
> +			return ret;
> +
> +		if (size == 2)
> +			data32 = (data32 & ~(0xffff << shift)) |
> +					((value & 0xffff) << shift);
> +		 else if (size == 1)
> +			data32 = (data32 & ~(0xff << shift)) |
> +					((value & 0xff) << shift);
> +	}
> +
> +	return tlp_cfg_dword_write(pcie, bus->number, devfn,
> +		(where & ~DWORD_MASK), data32);
> +}
> +
> +static struct pci_ops altera_pcie_ops = {
> +	.read = altera_pcie_cfg_read,
> +	.write = altera_pcie_cfg_write,
> +};
> +
> +static int altera_pcie_intx_map(struct irq_domain *domain, unsigned int irq,
> +				irq_hw_number_t hwirq)
> +{
> +	irq_set_chip_and_handler(irq, &dummy_irq_chip, handle_simple_irq);
> +	irq_set_chip_data(irq, domain->host_data);
> +	set_irq_flags(irq, IRQF_VALID);
> +
> +	return 0;
> +}
> +
> +static const struct irq_domain_ops intx_domain_ops = {
> +	.map = altera_pcie_intx_map,
> +};
> +
> +static irqreturn_t altera_pcie_isr(int irq, void *arg)
> +{
> +	struct altera_pcie *pcie = arg;
> +	u32 status, i;
> +
> +	status = cra_readl(pcie, P2A_INT_STATUS) & P2A_INT_STS_ALL;
> +
> +	/* clear interrupts */
> +	cra_writel(pcie, status, P2A_INT_STATUS);
> +
> +	for (i = 0; i < INTX_NUM; i++) {
> +		if (status & (1 << i))

find_first/next_set_bit?

> +			generic_handle_irq(irq_find_mapping(pcie->irq_domain,
> +							i + 1));

Rule of thumb: if you're calling generic_handle_irq from an interrupt
handler, this is a chained IRQ controller. Please implement it as such.

> +	}
> +
> +	return IRQ_HANDLED;
> +}
> +
> +static void altera_pcie_release_of_pci_ranges(struct altera_pcie *pcie)
> +{
> +	pci_free_resource_list(&pcie->resources);
> +}
> +
> +static int altera_pcie_parse_request_of_pci_ranges(struct altera_pcie *pcie)
> +{
> +	int err, res_valid = 0;
> +	struct device *dev = &pcie->pdev->dev;
> +	struct device_node *np = dev->of_node;
> +	resource_size_t iobase;
> +	struct resource_entry *win;
> +	int offset = 0;
> +
> +	err = of_pci_get_host_bridge_resources(np, 0, 0xff, &pcie->resources,
> +					       &iobase);
> +	if (err)
> +		return err;
> +
> +	resource_list_for_each_entry(win, &pcie->resources) {
> +		struct resource *parent, *res = win->res;
> +
> +		switch (resource_type(res)) {
> +		case IORESOURCE_MEM:
> +			parent = &iomem_resource;
> +			res_valid |= !(res->flags & IORESOURCE_PREFETCH);
> +			cra_writel(pcie, res->start,
> +				A2P_ADDR_MAP_LO0 + offset);
> +			cra_writel(pcie, 0,
> +				A2P_ADDR_MAP_HI0 + offset);
> +			offset += ATT_ENTRY_SIZE;
> +			break;
> +		default:
> +			continue;
> +		}
> +
> +		err = devm_request_resource(dev, parent, res);
> +		if (err)
> +			goto out_release_res;
> +	}
> +
> +	if (!res_valid) {
> +		dev_err(dev, "non-prefetchable memory resource required\n");
> +		err = -EINVAL;
> +		goto out_release_res;
> +	}
> +
> +	return 0;
> +
> +out_release_res:
> +	altera_pcie_release_of_pci_ranges(pcie);
> +	return err;
> +}
> +
> +static void altera_pcie_free_irq_domain(struct altera_pcie *pcie)
> +{
> +	int i;
> +	u32 irq;
> +
> +	for (i = 0; i < INTX_NUM; i++) {
> +		irq = irq_find_mapping(pcie->irq_domain, i);
> +		if (irq > 0)
> +			irq_dispose_mapping(irq);
> +	}
> +
> +	irq_domain_remove(pcie->irq_domain);
> +}
> +
> +static int altera_pcie_init_irq_domain(struct altera_pcie *pcie)
> +{
> +	struct device *dev = &pcie->pdev->dev;
> +	struct device_node *node = dev->of_node;
> +
> +	/* Setup INTx */
> +	pcie->irq_domain = irq_domain_add_linear(node, INTX_NUM,
> +					&intx_domain_ops, pcie);
> +	if (!pcie->irq_domain) {
> +		dev_err(dev, "Failed to get a INTx IRQ domain\n");
> +		return PTR_ERR(pcie->irq_domain);
> +	}
> +
> +	return 0;
> +}
> +
> +static int altera_pcie_parse_dt(struct altera_pcie *pcie)
> +{
> +	struct resource *cra;
> +	int ret;
> +	struct platform_device *pdev = pcie->pdev;
> +
> +	cra = platform_get_resource_byname(pdev, IORESOURCE_MEM, "Cra");
> +	pcie->cra_base = devm_ioremap_resource(&pdev->dev, cra);
> +	if (IS_ERR(pcie->cra_base)) {
> +		dev_err(&pdev->dev, "get Cra resource failed\n");
> +		return PTR_ERR(pcie->cra_base);
> +	}
> +
> +	/* setup IRQ */
> +	pcie->hwirq = platform_get_irq(pdev, 0);
> +	if (pcie->hwirq <= 0) {

This is definitely not a hardware interrupt number. It is completely
virtual. I suggest you give it a better name.

> +		dev_err(&pdev->dev, "failed to get IRQ: %d\n", pcie->hwirq);
> +		return -EINVAL;
> +	}
> +	ret = devm_request_irq(&pdev->dev, pcie->hwirq, altera_pcie_isr,
> +			IRQF_SHARED, pdev->name, pcie);

You're implementing a cascading interrupt controller again. Please fix
it just like you did for your MSI support.

> +
> +	if (ret) {
> +		dev_err(&pdev->dev, "failed to request irq %d\n", pcie->hwirq);
> +		return ret;
> +	}
> +
> +	return 0;
> +}
> +
> +static int altera_pcie_probe(struct platform_device *pdev)
> +{
> +	struct altera_pcie *pcie;
> +	struct pci_bus *bus;
> +	int ret;
> +
> +	pcie = devm_kzalloc(&pdev->dev, sizeof(*pcie), GFP_KERNEL);
> +	if (!pcie)
> +		return -ENOMEM;
> +
> +	pcie->pdev = pdev;
> +
> +	ret = altera_pcie_parse_dt(pcie);
> +	if (ret) {
> +		dev_err(&pdev->dev, "Parsing DT failed\n");
> +		return ret;
> +	}
> +
> +	INIT_LIST_HEAD(&pcie->resources);
> +
> +	ret = altera_pcie_parse_request_of_pci_ranges(pcie);
> +	if (ret) {
> +		dev_err(&pdev->dev, "Failed add resources\n");
> +		return ret;
> +	}
> +
> +	ret = altera_pcie_init_irq_domain(pcie);
> +	if (ret) {
> +		dev_err(&pdev->dev, "Failed creating IRQ Domain\n");
> +		return ret;
> +	}
> +
> +	pcie->root_bus_nr = 0;
> +
> +	/* clear all interrupts */
> +	cra_writel(pcie, P2A_INT_STS_ALL, P2A_INT_STATUS);
> +	/* enable all interrupts */
> +	cra_writel(pcie, P2A_INT_ENA_ALL, P2A_INT_ENABLE);
> +
> +	bus = pci_scan_root_bus(&pdev->dev, pcie->root_bus_nr, &altera_pcie_ops,
> +					pcie, &pcie->resources);
> +	if (!bus)
> +		return -ENOMEM;
> +
> +	pci_assign_unassigned_bus_resources(bus);
> +	pci_bus_add_devices(bus);
> +
> +	platform_set_drvdata(pdev, pcie);
> +	return ret;
> +}
> +
> +static int __exit altera_pcie_remove(struct platform_device *pdev)
> +{
> +	struct altera_pcie *pcie = platform_get_drvdata(pdev);
> +
> +	altera_pcie_free_irq_domain(pcie);
> +	platform_set_drvdata(pdev, NULL);
> +	return 0;
> +}
> +
> +static const struct of_device_id altera_pcie_of_match[] = {
> +	{ .compatible = "altr,pcie-root-port-1.0", },
> +	{},
> +};
> +MODULE_DEVICE_TABLE(of, altera_pcie_of_match);
> +
> +static struct platform_driver altera_pcie_driver = {
> +	.probe		= altera_pcie_probe,
> +	.remove		= altera_pcie_remove,
> +	.driver = {
> +		.name	= "altera-pcie",
> +		.of_match_table = altera_pcie_of_match,
> +	},
> +};
> +
> +module_platform_driver(altera_pcie_driver);
> +
> +MODULE_AUTHOR("Ley Foon Tan <lftan@altera.com>");
> +MODULE_DESCRIPTION("Altera PCIe host controller driver");
> +MODULE_LICENSE("GPL v2");
> 


-- 
Jazz is not dead. It just smells funny...

Re: [PATCH v2 2/5] pci:host: Add Altera PCIe host controller driver

[Marc Zyngier]

On 31/07/15 11:15, Ley Foon Tan wrote:
> This patch adds the Altera PCIe host controller driver.
> 
> Signed-off-by: Ley Foon Tan <lftan-EIB2kfCEclfQT0dZR+AlfA@public.gmane.org>
> ---
>  drivers/pci/host/Kconfig       |   8 +
>  drivers/pci/host/Makefile      |   1 +
>  drivers/pci/host/pcie-altera.c | 526 +++++++++++++++++++++++++++++++++++++++++
>  3 files changed, 535 insertions(+)
>  create mode 100644 drivers/pci/host/pcie-altera.c
> 
> diff --git a/drivers/pci/host/Kconfig b/drivers/pci/host/Kconfig
> index 675c2d1..108500a 100644
> --- a/drivers/pci/host/Kconfig
> +++ b/drivers/pci/host/Kconfig
> @@ -145,4 +145,12 @@ config PCIE_IPROC_BCMA
>  	  Say Y here if you want to use the Broadcom iProc PCIe controller
>  	  through the BCMA bus interface
>  
> +config PCIE_ALTERA
> +	tristate "Altera PCIe controller"
> +	depends on ARCH_SOCFPGA
> +	select PCI_MSI_IRQ_DOMAIN if PCI_MSI

Why would you select this here? MSI support is in another driver, so I
don't see the point.

> +	help
> +	  Say Y here if you want to enable PCIe controller support for Altera
> +	  SoCFPGA family of SoCs.
> +
>  endmenu
> diff --git a/drivers/pci/host/Makefile b/drivers/pci/host/Makefile
> index 140d66f..6954f76 100644
> --- a/drivers/pci/host/Makefile
> +++ b/drivers/pci/host/Makefile
> @@ -17,3 +17,4 @@ obj-$(CONFIG_PCI_VERSATILE) += pci-versatile.o
>  obj-$(CONFIG_PCIE_IPROC) += pcie-iproc.o
>  obj-$(CONFIG_PCIE_IPROC_PLATFORM) += pcie-iproc-platform.o
>  obj-$(CONFIG_PCIE_IPROC_BCMA) += pcie-iproc-bcma.o
> +obj-$(CONFIG_PCIE_ALTERA) += pcie-altera.o
> diff --git a/drivers/pci/host/pcie-altera.c b/drivers/pci/host/pcie-altera.c
> new file mode 100644
> index 0000000..3ec51c7
> --- /dev/null
> +++ b/drivers/pci/host/pcie-altera.c
> @@ -0,0 +1,526 @@
> +/*
> + * Copyright Altera Corporation (C) 2013-2015. All rights reserved
> + *
> + * This program is free software; you can redistribute it and/or modify it
> + * under the terms and conditions of the GNU General Public License,
> + * version 2, as published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope it will be useful, but WITHOUT
> + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
> + * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
> + * more details.
> + *
> + * You should have received a copy of the GNU General Public License along with
> + * this program.  If not, see <http://www.gnu.org/licenses/>.
> + */
> +#include <linux/interrupt.h>
> +#include <linux/module.h>
> +#include <linux/of_address.h>
> +#include <linux/of_irq.h>
> +#include <linux/of_pci.h>
> +#include <linux/pci.h>
> +#include <linux/platform_device.h>
> +#include <linux/slab.h>
> +
> +#define A2P_ADDR_MAP_LO0		0x1000
> +#define A2P_ADDR_MAP_HI0		0x1004
> +#define RP_TX_REG0			0x2000
> +#define RP_TX_REG1			0x2004
> +#define RP_TX_CNTRL			0x2008
> +#define RP_TX_EOP			0x2
> +#define RP_TX_SOP			0x1
> +#define RP_RXCPL_STATUS		0x2010
> +#define RP_RXCPL_EOP			0x2
> +#define RP_RXCPL_SOP			0x1
> +#define RP_RXCPL_REG0			0x2014
> +#define RP_RXCPL_REG1			0x2018
> +#define P2A_INT_STATUS			0x3060
> +#define P2A_INT_STS_ALL			0xF
> +#define P2A_INT_ENABLE			0x3070
> +#define P2A_INT_ENA_ALL			0xF
> +#define RP_LTSSM			0x3C64
> +#define LTSSM_L0			0xF
> +
> +/* TLP configuration type 0 and 1 */
> +#define TLP_FMTTYPE_CFGRD0		0x04	/* Configuration Read Type 0 */
> +#define TLP_FMTTYPE_CFGWR0		0x44	/* Configuration Write Type 0 */
> +#define TLP_FMTTYPE_CFGRD1		0x05	/* Configuration Read Type 1 */
> +#define TLP_FMTTYPE_CFGWR1		0x45	/* Configuration Write Type 1 */
> +#define TLP_PAYLOAD_SIZE		0x01
> +#define TLP_READ_TAG			0x1D
> +#define TLP_WRITE_TAG			0x10
> +#define TLP_CFG_DW0(fmttype)		(((fmttype) << 24) | TLP_PAYLOAD_SIZE)
> +#define TLP_CFG_DW1(reqid, tag)		(((reqid) << 16) | (tag << 8) | 0xF)
> +#define TLP_CFG_DW2(bus, devfn, offset)	\
> +				(((bus) << 24) | ((devfn) << 16) | (offset))
> +#define TLP_REQ_ID(bus, devfn)		(((bus) << 8) | (devfn))
> +#define TLP_COMPL_STATUS(hdr)		(((hdr) & 0xE0) >> 13)
> +#define TLP_HDR_SIZE			3
> +#define TLP_LOOP			10
> +
> +#define INTX_NUM		4
> +
> +/* Address translation table entry size */
> +#define ATT_ENTRY_SIZE		8
> +
> +#define DWORD_MASK		3
> +
> +struct altera_pcie {
> +	struct platform_device	*pdev;
> +	struct resource		*txs;
> +	void __iomem		*cra_base;
> +	int			hwirq;
> +	u8			root_bus_nr;
> +	struct irq_domain		*irq_domain;
> +	struct resource		bus_range;
> +	struct list_head		resources;
> +};
> +
> +struct tlp_rp_regpair_t {
> +	u32 ctrl;
> +	u32 reg0;
> +	u32 reg1;
> +};
> +
> +static void altera_pcie_retrain(struct pci_dev *dev)
> +{
> +	u16 linkcap, linkstat;
> +
> +	/*
> +	 * Set the retrain bit if the PCIe rootport support > 2.5GB/s, but
> +	 * current speed is 2.5 GB/s.
> +	 */
> +	pcie_capability_read_word(dev, PCI_EXP_LNKCAP, &linkcap);
> +
> +	if ((linkcap & PCI_EXP_LNKCAP_SLS) <= PCI_EXP_LNKCAP_SLS_2_5GB)
> +		return;
> +
> +	pcie_capability_read_word(dev, PCI_EXP_LNKSTA, &linkstat);
> +	if ((linkstat & PCI_EXP_LNKSTA_CLS) == PCI_EXP_LNKSTA_CLS_2_5GB)
> +		pcie_capability_set_word(dev, PCI_EXP_LNKCTL,
> +					PCI_EXP_LNKCTL_RL);
> +
> +}
> +DECLARE_PCI_FIXUP_EARLY(PCI_ANY_ID, PCI_ANY_ID, altera_pcie_retrain);
> +
> +static inline void cra_writel(struct altera_pcie *pcie, u32 value, u32 reg)
> +{
> +	writel_relaxed(value, pcie->cra_base + reg);
> +}
> +
> +static inline u32 cra_readl(struct altera_pcie *pcie, u32 reg)
> +{
> +	return readl_relaxed(pcie->cra_base + reg);
> +}
> +
> +static void tlp_read_rx(struct altera_pcie *pcie,
> +			struct tlp_rp_regpair_t *tlp_rp_regdata)
> +{
> +	tlp_rp_regdata->ctrl = cra_readl(pcie, RP_RXCPL_STATUS);
> +	tlp_rp_regdata->reg0 = cra_readl(pcie, RP_RXCPL_REG0);
> +	tlp_rp_regdata->reg1 = cra_readl(pcie, RP_RXCPL_REG1);
> +}
> +
> +static void tlp_write_tx(struct altera_pcie *pcie,
> +			struct tlp_rp_regpair_t *tlp_rp_regdata)
> +{
> +	cra_writel(pcie, tlp_rp_regdata->reg0, RP_TX_REG0);
> +	cra_writel(pcie, tlp_rp_regdata->reg1, RP_TX_REG1);
> +	cra_writel(pcie, tlp_rp_regdata->ctrl, RP_TX_CNTRL);
> +}
> +
> +static bool altera_pcie_link_is_up(struct altera_pcie *pcie)
> +{
> +	return !!(cra_readl(pcie, RP_LTSSM) & LTSSM_L0);
> +}
> +
> +static bool altera_pcie_valid_config(struct altera_pcie *pcie,
> +				struct pci_bus *bus, int dev)
> +{
> +	/* If there is no link, then there is no device */
> +	if (bus->number != pcie->root_bus_nr) {
> +		if (!altera_pcie_link_is_up(pcie))
> +			return false;
> +	}
> +
> +	/* access only one slot on each root port */
> +	if (bus->number == pcie->root_bus_nr && dev > 0)
> +		return false;
> +
> +	/*
> +	 * Do not read more than one device on the bus directly attached
> +	 * to RC.
> +	 */
> +	if (bus->primary == pcie->root_bus_nr && dev > 0)
> +		return false;
> +
> +	 return true;
> +}
> +
> +static int tlp_read_packet(struct altera_pcie *pcie, u32 *value)
> +{
> +	u8 loop;
> +	struct tlp_rp_regpair_t tlp_rp_regdata;
> +
> +	for (loop = TLP_LOOP; loop > 0; loop--) {
> +		tlp_read_rx(pcie, &tlp_rp_regdata);
> +		if (tlp_rp_regdata.ctrl & RP_RXCPL_EOP) {
> +			if (value)
> +				*value = tlp_rp_regdata.reg0;
> +			return PCIBIOS_SUCCESSFUL;
> +		}
> +	}
> +
> +	return -ENOENT;
> +}
> +
> +static void tlp_write_packet(struct altera_pcie *pcie, u32 *headers, u32 data)
> +{
> +	struct tlp_rp_regpair_t tlp_rp_regdata;
> +
> +	tlp_rp_regdata.reg0 = headers[0];
> +	tlp_rp_regdata.reg1 = headers[1];
> +	tlp_rp_regdata.ctrl = RP_TX_SOP;
> +	tlp_write_tx(pcie, &tlp_rp_regdata);
> +
> +	tlp_rp_regdata.reg0 = headers[2];
> +	tlp_rp_regdata.reg1 = data;
> +	tlp_rp_regdata.ctrl = RP_TX_EOP;
> +	tlp_write_tx(pcie, &tlp_rp_regdata);
> +}
> +
> +static int tlp_cfg_dword_read(struct altera_pcie *pcie, u8 bus, u32 devfn,
> +			      int where, u32 *value)
> +{
> +	int ret;
> +	u32 headers[TLP_HDR_SIZE];
> +
> +	if (bus == pcie->root_bus_nr)
> +		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGRD0);
> +	else
> +		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGRD1);
> +
> +	headers[1] = TLP_CFG_DW1(TLP_REQ_ID(pcie->root_bus_nr, devfn),
> +					TLP_READ_TAG);
> +	headers[2] = TLP_CFG_DW2(bus, devfn, where);
> +
> +	tlp_write_packet(pcie, headers, 0);
> +
> +	ret = tlp_read_packet(pcie, value);
> +	if (ret)
> +		*value = ~0UL;	/* return 0xFFFFFFFF if error */
> +
> +	return ret;
> +}
> +
> +static int tlp_cfg_dword_write(struct altera_pcie *pcie, u8 bus, u32 devfn,
> +			       int where, u32 value)
> +{
> +	u32 headers[TLP_HDR_SIZE];
> +
> +	if (bus == pcie->root_bus_nr)
> +		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGWR0);
> +	else
> +		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGWR1);
> +
> +	headers[1] = TLP_CFG_DW1(TLP_REQ_ID(pcie->root_bus_nr, devfn),
> +					TLP_WRITE_TAG);
> +	headers[2] = TLP_CFG_DW2(bus, devfn, where);
> +
> +	tlp_write_packet(pcie, headers, value);
> +
> +	tlp_read_packet(pcie, NULL);
> +
> +	/* Keep an eye out for changes to the root bus number */
> +	if ((bus == pcie->root_bus_nr) && (where == PCI_PRIMARY_BUS))
> +		pcie->root_bus_nr = (u8)(value);
> +
> +	return PCIBIOS_SUCCESSFUL;
> +}
> +
> +static int altera_pcie_cfg_read(struct pci_bus *bus, unsigned int devfn,
> +			       int where, int size, u32 *value)
> +{
> +	struct altera_pcie *pcie = bus->sysdata;
> +	int ret;
> +
> +	if (!altera_pcie_valid_config(pcie, bus, PCI_SLOT(devfn))) {
> +		*value = ~0UL;
> +		return PCIBIOS_DEVICE_NOT_FOUND;
> +	}
> +
> +	ret = tlp_cfg_dword_read(pcie, bus->number, devfn,
> +		(where & ~DWORD_MASK), value);
> +	if (ret != PCIBIOS_SUCCESSFUL)
> +		return ret;
> +
> +	if (size == 1)
> +		*value = (*value >> (8 * (where & 0x3))) & 0xff;
> +	else if (size == 2)
> +		*value = (*value >> (8 * (where & 0x2))) & 0xffff;
> +
> +	return PCIBIOS_SUCCESSFUL;
> +}
> +
> +static int altera_pcie_cfg_write(struct pci_bus *bus, unsigned int devfn,
> +				int where, int size, u32 value)
> +{
> +	struct altera_pcie *pcie = bus->sysdata;
> +	u32 data32 = value;
> +	u32 shift = 8 * (where & 3);
> +	int ret;
> +
> +	if (!altera_pcie_valid_config(pcie, bus, PCI_SLOT(devfn)))
> +		return PCIBIOS_DEVICE_NOT_FOUND;
> +
> +	/* write partial */
> +	if (size != sizeof(u32)) {
> +		ret = tlp_cfg_dword_read(pcie, bus->number, devfn,
> +			where & ~DWORD_MASK, &data32);
> +		if (ret)
> +			return ret;
> +
> +		if (size == 2)
> +			data32 = (data32 & ~(0xffff << shift)) |
> +					((value & 0xffff) << shift);
> +		 else if (size == 1)
> +			data32 = (data32 & ~(0xff << shift)) |
> +					((value & 0xff) << shift);
> +	}
> +
> +	return tlp_cfg_dword_write(pcie, bus->number, devfn,
> +		(where & ~DWORD_MASK), data32);
> +}
> +
> +static struct pci_ops altera_pcie_ops = {
> +	.read = altera_pcie_cfg_read,
> +	.write = altera_pcie_cfg_write,
> +};
> +
> +static int altera_pcie_intx_map(struct irq_domain *domain, unsigned int irq,
> +				irq_hw_number_t hwirq)
> +{
> +	irq_set_chip_and_handler(irq, &dummy_irq_chip, handle_simple_irq);
> +	irq_set_chip_data(irq, domain->host_data);
> +	set_irq_flags(irq, IRQF_VALID);
> +
> +	return 0;
> +}
> +
> +static const struct irq_domain_ops intx_domain_ops = {
> +	.map = altera_pcie_intx_map,
> +};
> +
> +static irqreturn_t altera_pcie_isr(int irq, void *arg)
> +{
> +	struct altera_pcie *pcie = arg;
> +	u32 status, i;
> +
> +	status = cra_readl(pcie, P2A_INT_STATUS) & P2A_INT_STS_ALL;
> +
> +	/* clear interrupts */
> +	cra_writel(pcie, status, P2A_INT_STATUS);
> +
> +	for (i = 0; i < INTX_NUM; i++) {
> +		if (status & (1 << i))

find_first/next_set_bit?

> +			generic_handle_irq(irq_find_mapping(pcie->irq_domain,
> +							i + 1));

Rule of thumb: if you're calling generic_handle_irq from an interrupt
handler, this is a chained IRQ controller. Please implement it as such.

> +	}
> +
> +	return IRQ_HANDLED;
> +}
> +
> +static void altera_pcie_release_of_pci_ranges(struct altera_pcie *pcie)
> +{
> +	pci_free_resource_list(&pcie->resources);
> +}
> +
> +static int altera_pcie_parse_request_of_pci_ranges(struct altera_pcie *pcie)
> +{
> +	int err, res_valid = 0;
> +	struct device *dev = &pcie->pdev->dev;
> +	struct device_node *np = dev->of_node;
> +	resource_size_t iobase;
> +	struct resource_entry *win;
> +	int offset = 0;
> +
> +	err = of_pci_get_host_bridge_resources(np, 0, 0xff, &pcie->resources,
> +					       &iobase);
> +	if (err)
> +		return err;
> +
> +	resource_list_for_each_entry(win, &pcie->resources) {
> +		struct resource *parent, *res = win->res;
> +
> +		switch (resource_type(res)) {
> +		case IORESOURCE_MEM:
> +			parent = &iomem_resource;
> +			res_valid |= !(res->flags & IORESOURCE_PREFETCH);
> +			cra_writel(pcie, res->start,
> +				A2P_ADDR_MAP_LO0 + offset);
> +			cra_writel(pcie, 0,
> +				A2P_ADDR_MAP_HI0 + offset);
> +			offset += ATT_ENTRY_SIZE;
> +			break;
> +		default:
> +			continue;
> +		}
> +
> +		err = devm_request_resource(dev, parent, res);
> +		if (err)
> +			goto out_release_res;
> +	}
> +
> +	if (!res_valid) {
> +		dev_err(dev, "non-prefetchable memory resource required\n");
> +		err = -EINVAL;
> +		goto out_release_res;
> +	}
> +
> +	return 0;
> +
> +out_release_res:
> +	altera_pcie_release_of_pci_ranges(pcie);
> +	return err;
> +}
> +
> +static void altera_pcie_free_irq_domain(struct altera_pcie *pcie)
> +{
> +	int i;
> +	u32 irq;
> +
> +	for (i = 0; i < INTX_NUM; i++) {
> +		irq = irq_find_mapping(pcie->irq_domain, i);
> +		if (irq > 0)
> +			irq_dispose_mapping(irq);
> +	}
> +
> +	irq_domain_remove(pcie->irq_domain);
> +}
> +
> +static int altera_pcie_init_irq_domain(struct altera_pcie *pcie)
> +{
> +	struct device *dev = &pcie->pdev->dev;
> +	struct device_node *node = dev->of_node;
> +
> +	/* Setup INTx */
> +	pcie->irq_domain = irq_domain_add_linear(node, INTX_NUM,
> +					&intx_domain_ops, pcie);
> +	if (!pcie->irq_domain) {
> +		dev_err(dev, "Failed to get a INTx IRQ domain\n");
> +		return PTR_ERR(pcie->irq_domain);
> +	}
> +
> +	return 0;
> +}
> +
> +static int altera_pcie_parse_dt(struct altera_pcie *pcie)
> +{
> +	struct resource *cra;
> +	int ret;
> +	struct platform_device *pdev = pcie->pdev;
> +
> +	cra = platform_get_resource_byname(pdev, IORESOURCE_MEM, "Cra");
> +	pcie->cra_base = devm_ioremap_resource(&pdev->dev, cra);
> +	if (IS_ERR(pcie->cra_base)) {
> +		dev_err(&pdev->dev, "get Cra resource failed\n");
> +		return PTR_ERR(pcie->cra_base);
> +	}
> +
> +	/* setup IRQ */
> +	pcie->hwirq = platform_get_irq(pdev, 0);
> +	if (pcie->hwirq <= 0) {

This is definitely not a hardware interrupt number. It is completely
virtual. I suggest you give it a better name.

> +		dev_err(&pdev->dev, "failed to get IRQ: %d\n", pcie->hwirq);
> +		return -EINVAL;
> +	}
> +	ret = devm_request_irq(&pdev->dev, pcie->hwirq, altera_pcie_isr,
> +			IRQF_SHARED, pdev->name, pcie);

You're implementing a cascading interrupt controller again. Please fix
it just like you did for your MSI support.

> +
> +	if (ret) {
> +		dev_err(&pdev->dev, "failed to request irq %d\n", pcie->hwirq);
> +		return ret;
> +	}
> +
> +	return 0;
> +}
> +
> +static int altera_pcie_probe(struct platform_device *pdev)
> +{
> +	struct altera_pcie *pcie;
> +	struct pci_bus *bus;
> +	int ret;
> +
> +	pcie = devm_kzalloc(&pdev->dev, sizeof(*pcie), GFP_KERNEL);
> +	if (!pcie)
> +		return -ENOMEM;
> +
> +	pcie->pdev = pdev;
> +
> +	ret = altera_pcie_parse_dt(pcie);
> +	if (ret) {
> +		dev_err(&pdev->dev, "Parsing DT failed\n");
> +		return ret;
> +	}
> +
> +	INIT_LIST_HEAD(&pcie->resources);
> +
> +	ret = altera_pcie_parse_request_of_pci_ranges(pcie);
> +	if (ret) {
> +		dev_err(&pdev->dev, "Failed add resources\n");
> +		return ret;
> +	}
> +
> +	ret = altera_pcie_init_irq_domain(pcie);
> +	if (ret) {
> +		dev_err(&pdev->dev, "Failed creating IRQ Domain\n");
> +		return ret;
> +	}
> +
> +	pcie->root_bus_nr = 0;
> +
> +	/* clear all interrupts */
> +	cra_writel(pcie, P2A_INT_STS_ALL, P2A_INT_STATUS);
> +	/* enable all interrupts */
> +	cra_writel(pcie, P2A_INT_ENA_ALL, P2A_INT_ENABLE);
> +
> +	bus = pci_scan_root_bus(&pdev->dev, pcie->root_bus_nr, &altera_pcie_ops,
> +					pcie, &pcie->resources);
> +	if (!bus)
> +		return -ENOMEM;
> +
> +	pci_assign_unassigned_bus_resources(bus);
> +	pci_bus_add_devices(bus);
> +
> +	platform_set_drvdata(pdev, pcie);
> +	return ret;
> +}
> +
> +static int __exit altera_pcie_remove(struct platform_device *pdev)
> +{
> +	struct altera_pcie *pcie = platform_get_drvdata(pdev);
> +
> +	altera_pcie_free_irq_domain(pcie);
> +	platform_set_drvdata(pdev, NULL);
> +	return 0;
> +}
> +
> +static const struct of_device_id altera_pcie_of_match[] = {
> +	{ .compatible = "altr,pcie-root-port-1.0", },
> +	{},
> +};
> +MODULE_DEVICE_TABLE(of, altera_pcie_of_match);
> +
> +static struct platform_driver altera_pcie_driver = {
> +	.probe		= altera_pcie_probe,
> +	.remove		= altera_pcie_remove,
> +	.driver = {
> +		.name	= "altera-pcie",
> +		.of_match_table = altera_pcie_of_match,
> +	},
> +};
> +
> +module_platform_driver(altera_pcie_driver);
> +
> +MODULE_AUTHOR("Ley Foon Tan <lftan-EIB2kfCEclfQT0dZR+AlfA@public.gmane.org>");
> +MODULE_DESCRIPTION("Altera PCIe host controller driver");
> +MODULE_LICENSE("GPL v2");
> 


-- 
Jazz is not dead. It just smells funny...
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v2 2/5] pci:host: Add Altera PCIe host controller driver

[Marc Zyngier]

On 31/07/15 11:15, Ley Foon Tan wrote:
> This patch adds the Altera PCIe host controller driver.
> 
> Signed-off-by: Ley Foon Tan <lftan@altera.com>
> ---
>  drivers/pci/host/Kconfig       |   8 +
>  drivers/pci/host/Makefile      |   1 +
>  drivers/pci/host/pcie-altera.c | 526 +++++++++++++++++++++++++++++++++++++++++
>  3 files changed, 535 insertions(+)
>  create mode 100644 drivers/pci/host/pcie-altera.c
> 
> diff --git a/drivers/pci/host/Kconfig b/drivers/pci/host/Kconfig
> index 675c2d1..108500a 100644
> --- a/drivers/pci/host/Kconfig
> +++ b/drivers/pci/host/Kconfig
> @@ -145,4 +145,12 @@ config PCIE_IPROC_BCMA
>  	  Say Y here if you want to use the Broadcom iProc PCIe controller
>  	  through the BCMA bus interface
>  
> +config PCIE_ALTERA
> +	tristate "Altera PCIe controller"
> +	depends on ARCH_SOCFPGA
> +	select PCI_MSI_IRQ_DOMAIN if PCI_MSI

Why would you select this here? MSI support is in another driver, so I
don't see the point.

> +	help
> +	  Say Y here if you want to enable PCIe controller support for Altera
> +	  SoCFPGA family of SoCs.
> +
>  endmenu
> diff --git a/drivers/pci/host/Makefile b/drivers/pci/host/Makefile
> index 140d66f..6954f76 100644
> --- a/drivers/pci/host/Makefile
> +++ b/drivers/pci/host/Makefile
> @@ -17,3 +17,4 @@ obj-$(CONFIG_PCI_VERSATILE) += pci-versatile.o
>  obj-$(CONFIG_PCIE_IPROC) += pcie-iproc.o
>  obj-$(CONFIG_PCIE_IPROC_PLATFORM) += pcie-iproc-platform.o
>  obj-$(CONFIG_PCIE_IPROC_BCMA) += pcie-iproc-bcma.o
> +obj-$(CONFIG_PCIE_ALTERA) += pcie-altera.o
> diff --git a/drivers/pci/host/pcie-altera.c b/drivers/pci/host/pcie-altera.c
> new file mode 100644
> index 0000000..3ec51c7
> --- /dev/null
> +++ b/drivers/pci/host/pcie-altera.c
> @@ -0,0 +1,526 @@
> +/*
> + * Copyright Altera Corporation (C) 2013-2015. All rights reserved
> + *
> + * This program is free software; you can redistribute it and/or modify it
> + * under the terms and conditions of the GNU General Public License,
> + * version 2, as published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope it will be useful, but WITHOUT
> + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
> + * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
> + * more details.
> + *
> + * You should have received a copy of the GNU General Public License along with
> + * this program.  If not, see <http://www.gnu.org/licenses/>.
> + */
> +#include <linux/interrupt.h>
> +#include <linux/module.h>
> +#include <linux/of_address.h>
> +#include <linux/of_irq.h>
> +#include <linux/of_pci.h>
> +#include <linux/pci.h>
> +#include <linux/platform_device.h>
> +#include <linux/slab.h>
> +
> +#define A2P_ADDR_MAP_LO0		0x1000
> +#define A2P_ADDR_MAP_HI0		0x1004
> +#define RP_TX_REG0			0x2000
> +#define RP_TX_REG1			0x2004
> +#define RP_TX_CNTRL			0x2008
> +#define RP_TX_EOP			0x2
> +#define RP_TX_SOP			0x1
> +#define RP_RXCPL_STATUS		0x2010
> +#define RP_RXCPL_EOP			0x2
> +#define RP_RXCPL_SOP			0x1
> +#define RP_RXCPL_REG0			0x2014
> +#define RP_RXCPL_REG1			0x2018
> +#define P2A_INT_STATUS			0x3060
> +#define P2A_INT_STS_ALL			0xF
> +#define P2A_INT_ENABLE			0x3070
> +#define P2A_INT_ENA_ALL			0xF
> +#define RP_LTSSM			0x3C64
> +#define LTSSM_L0			0xF
> +
> +/* TLP configuration type 0 and 1 */
> +#define TLP_FMTTYPE_CFGRD0		0x04	/* Configuration Read Type 0 */
> +#define TLP_FMTTYPE_CFGWR0		0x44	/* Configuration Write Type 0 */
> +#define TLP_FMTTYPE_CFGRD1		0x05	/* Configuration Read Type 1 */
> +#define TLP_FMTTYPE_CFGWR1		0x45	/* Configuration Write Type 1 */
> +#define TLP_PAYLOAD_SIZE		0x01
> +#define TLP_READ_TAG			0x1D
> +#define TLP_WRITE_TAG			0x10
> +#define TLP_CFG_DW0(fmttype)		(((fmttype) << 24) | TLP_PAYLOAD_SIZE)
> +#define TLP_CFG_DW1(reqid, tag)		(((reqid) << 16) | (tag << 8) | 0xF)
> +#define TLP_CFG_DW2(bus, devfn, offset)	\
> +				(((bus) << 24) | ((devfn) << 16) | (offset))
> +#define TLP_REQ_ID(bus, devfn)		(((bus) << 8) | (devfn))
> +#define TLP_COMPL_STATUS(hdr)		(((hdr) & 0xE0) >> 13)
> +#define TLP_HDR_SIZE			3
> +#define TLP_LOOP			10
> +
> +#define INTX_NUM		4
> +
> +/* Address translation table entry size */
> +#define ATT_ENTRY_SIZE		8
> +
> +#define DWORD_MASK		3
> +
> +struct altera_pcie {
> +	struct platform_device	*pdev;
> +	struct resource		*txs;
> +	void __iomem		*cra_base;
> +	int			hwirq;
> +	u8			root_bus_nr;
> +	struct irq_domain		*irq_domain;
> +	struct resource		bus_range;
> +	struct list_head		resources;
> +};
> +
> +struct tlp_rp_regpair_t {
> +	u32 ctrl;
> +	u32 reg0;
> +	u32 reg1;
> +};
> +
> +static void altera_pcie_retrain(struct pci_dev *dev)
> +{
> +	u16 linkcap, linkstat;
> +
> +	/*
> +	 * Set the retrain bit if the PCIe rootport support > 2.5GB/s, but
> +	 * current speed is 2.5 GB/s.
> +	 */
> +	pcie_capability_read_word(dev, PCI_EXP_LNKCAP, &linkcap);
> +
> +	if ((linkcap & PCI_EXP_LNKCAP_SLS) <= PCI_EXP_LNKCAP_SLS_2_5GB)
> +		return;
> +
> +	pcie_capability_read_word(dev, PCI_EXP_LNKSTA, &linkstat);
> +	if ((linkstat & PCI_EXP_LNKSTA_CLS) == PCI_EXP_LNKSTA_CLS_2_5GB)
> +		pcie_capability_set_word(dev, PCI_EXP_LNKCTL,
> +					PCI_EXP_LNKCTL_RL);
> +
> +}
> +DECLARE_PCI_FIXUP_EARLY(PCI_ANY_ID, PCI_ANY_ID, altera_pcie_retrain);
> +
> +static inline void cra_writel(struct altera_pcie *pcie, u32 value, u32 reg)
> +{
> +	writel_relaxed(value, pcie->cra_base + reg);
> +}
> +
> +static inline u32 cra_readl(struct altera_pcie *pcie, u32 reg)
> +{
> +	return readl_relaxed(pcie->cra_base + reg);
> +}
> +
> +static void tlp_read_rx(struct altera_pcie *pcie,
> +			struct tlp_rp_regpair_t *tlp_rp_regdata)
> +{
> +	tlp_rp_regdata->ctrl = cra_readl(pcie, RP_RXCPL_STATUS);
> +	tlp_rp_regdata->reg0 = cra_readl(pcie, RP_RXCPL_REG0);
> +	tlp_rp_regdata->reg1 = cra_readl(pcie, RP_RXCPL_REG1);
> +}
> +
> +static void tlp_write_tx(struct altera_pcie *pcie,
> +			struct tlp_rp_regpair_t *tlp_rp_regdata)
> +{
> +	cra_writel(pcie, tlp_rp_regdata->reg0, RP_TX_REG0);
> +	cra_writel(pcie, tlp_rp_regdata->reg1, RP_TX_REG1);
> +	cra_writel(pcie, tlp_rp_regdata->ctrl, RP_TX_CNTRL);
> +}
> +
> +static bool altera_pcie_link_is_up(struct altera_pcie *pcie)
> +{
> +	return !!(cra_readl(pcie, RP_LTSSM) & LTSSM_L0);
> +}
> +
> +static bool altera_pcie_valid_config(struct altera_pcie *pcie,
> +				struct pci_bus *bus, int dev)
> +{
> +	/* If there is no link, then there is no device */
> +	if (bus->number != pcie->root_bus_nr) {
> +		if (!altera_pcie_link_is_up(pcie))
> +			return false;
> +	}
> +
> +	/* access only one slot on each root port */
> +	if (bus->number == pcie->root_bus_nr && dev > 0)
> +		return false;
> +
> +	/*
> +	 * Do not read more than one device on the bus directly attached
> +	 * to RC.
> +	 */
> +	if (bus->primary == pcie->root_bus_nr && dev > 0)
> +		return false;
> +
> +	 return true;
> +}
> +
> +static int tlp_read_packet(struct altera_pcie *pcie, u32 *value)
> +{
> +	u8 loop;
> +	struct tlp_rp_regpair_t tlp_rp_regdata;
> +
> +	for (loop = TLP_LOOP; loop > 0; loop--) {
> +		tlp_read_rx(pcie, &tlp_rp_regdata);
> +		if (tlp_rp_regdata.ctrl & RP_RXCPL_EOP) {
> +			if (value)
> +				*value = tlp_rp_regdata.reg0;
> +			return PCIBIOS_SUCCESSFUL;
> +		}
> +	}
> +
> +	return -ENOENT;
> +}
> +
> +static void tlp_write_packet(struct altera_pcie *pcie, u32 *headers, u32 data)
> +{
> +	struct tlp_rp_regpair_t tlp_rp_regdata;
> +
> +	tlp_rp_regdata.reg0 = headers[0];
> +	tlp_rp_regdata.reg1 = headers[1];
> +	tlp_rp_regdata.ctrl = RP_TX_SOP;
> +	tlp_write_tx(pcie, &tlp_rp_regdata);
> +
> +	tlp_rp_regdata.reg0 = headers[2];
> +	tlp_rp_regdata.reg1 = data;
> +	tlp_rp_regdata.ctrl = RP_TX_EOP;
> +	tlp_write_tx(pcie, &tlp_rp_regdata);
> +}
> +
> +static int tlp_cfg_dword_read(struct altera_pcie *pcie, u8 bus, u32 devfn,
> +			      int where, u32 *value)
> +{
> +	int ret;
> +	u32 headers[TLP_HDR_SIZE];
> +
> +	if (bus == pcie->root_bus_nr)
> +		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGRD0);
> +	else
> +		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGRD1);
> +
> +	headers[1] = TLP_CFG_DW1(TLP_REQ_ID(pcie->root_bus_nr, devfn),
> +					TLP_READ_TAG);
> +	headers[2] = TLP_CFG_DW2(bus, devfn, where);
> +
> +	tlp_write_packet(pcie, headers, 0);
> +
> +	ret = tlp_read_packet(pcie, value);
> +	if (ret)
> +		*value = ~0UL;	/* return 0xFFFFFFFF if error */
> +
> +	return ret;
> +}
> +
> +static int tlp_cfg_dword_write(struct altera_pcie *pcie, u8 bus, u32 devfn,
> +			       int where, u32 value)
> +{
> +	u32 headers[TLP_HDR_SIZE];
> +
> +	if (bus == pcie->root_bus_nr)
> +		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGWR0);
> +	else
> +		headers[0] = TLP_CFG_DW0(TLP_FMTTYPE_CFGWR1);
> +
> +	headers[1] = TLP_CFG_DW1(TLP_REQ_ID(pcie->root_bus_nr, devfn),
> +					TLP_WRITE_TAG);
> +	headers[2] = TLP_CFG_DW2(bus, devfn, where);
> +
> +	tlp_write_packet(pcie, headers, value);
> +
> +	tlp_read_packet(pcie, NULL);
> +
> +	/* Keep an eye out for changes to the root bus number */
> +	if ((bus == pcie->root_bus_nr) && (where == PCI_PRIMARY_BUS))
> +		pcie->root_bus_nr = (u8)(value);
> +
> +	return PCIBIOS_SUCCESSFUL;
> +}
> +
> +static int altera_pcie_cfg_read(struct pci_bus *bus, unsigned int devfn,
> +			       int where, int size, u32 *value)
> +{
> +	struct altera_pcie *pcie = bus->sysdata;
> +	int ret;
> +
> +	if (!altera_pcie_valid_config(pcie, bus, PCI_SLOT(devfn))) {
> +		*value = ~0UL;
> +		return PCIBIOS_DEVICE_NOT_FOUND;
> +	}
> +
> +	ret = tlp_cfg_dword_read(pcie, bus->number, devfn,
> +		(where & ~DWORD_MASK), value);
> +	if (ret != PCIBIOS_SUCCESSFUL)
> +		return ret;
> +
> +	if (size == 1)
> +		*value = (*value >> (8 * (where & 0x3))) & 0xff;
> +	else if (size == 2)
> +		*value = (*value >> (8 * (where & 0x2))) & 0xffff;
> +
> +	return PCIBIOS_SUCCESSFUL;
> +}
> +
> +static int altera_pcie_cfg_write(struct pci_bus *bus, unsigned int devfn,
> +				int where, int size, u32 value)
> +{
> +	struct altera_pcie *pcie = bus->sysdata;
> +	u32 data32 = value;
> +	u32 shift = 8 * (where & 3);
> +	int ret;
> +
> +	if (!altera_pcie_valid_config(pcie, bus, PCI_SLOT(devfn)))
> +		return PCIBIOS_DEVICE_NOT_FOUND;
> +
> +	/* write partial */
> +	if (size != sizeof(u32)) {
> +		ret = tlp_cfg_dword_read(pcie, bus->number, devfn,
> +			where & ~DWORD_MASK, &data32);
> +		if (ret)
> +			return ret;
> +
> +		if (size == 2)
> +			data32 = (data32 & ~(0xffff << shift)) |
> +					((value & 0xffff) << shift);
> +		 else if (size == 1)
> +			data32 = (data32 & ~(0xff << shift)) |
> +					((value & 0xff) << shift);
> +	}
> +
> +	return tlp_cfg_dword_write(pcie, bus->number, devfn,
> +		(where & ~DWORD_MASK), data32);
> +}
> +
> +static struct pci_ops altera_pcie_ops = {
> +	.read = altera_pcie_cfg_read,
> +	.write = altera_pcie_cfg_write,
> +};
> +
> +static int altera_pcie_intx_map(struct irq_domain *domain, unsigned int irq,
> +				irq_hw_number_t hwirq)
> +{
> +	irq_set_chip_and_handler(irq, &dummy_irq_chip, handle_simple_irq);
> +	irq_set_chip_data(irq, domain->host_data);
> +	set_irq_flags(irq, IRQF_VALID);
> +
> +	return 0;
> +}
> +
> +static const struct irq_domain_ops intx_domain_ops = {
> +	.map = altera_pcie_intx_map,
> +};
> +
> +static irqreturn_t altera_pcie_isr(int irq, void *arg)
> +{
> +	struct altera_pcie *pcie = arg;
> +	u32 status, i;
> +
> +	status = cra_readl(pcie, P2A_INT_STATUS) & P2A_INT_STS_ALL;
> +
> +	/* clear interrupts */
> +	cra_writel(pcie, status, P2A_INT_STATUS);
> +
> +	for (i = 0; i < INTX_NUM; i++) {
> +		if (status & (1 << i))

find_first/next_set_bit?

> +			generic_handle_irq(irq_find_mapping(pcie->irq_domain,
> +							i + 1));

Rule of thumb: if you're calling generic_handle_irq from an interrupt
handler, this is a chained IRQ controller. Please implement it as such.

> +	}
> +
> +	return IRQ_HANDLED;
> +}
> +
> +static void altera_pcie_release_of_pci_ranges(struct altera_pcie *pcie)
> +{
> +	pci_free_resource_list(&pcie->resources);
> +}
> +
> +static int altera_pcie_parse_request_of_pci_ranges(struct altera_pcie *pcie)
> +{
> +	int err, res_valid = 0;
> +	struct device *dev = &pcie->pdev->dev;
> +	struct device_node *np = dev->of_node;
> +	resource_size_t iobase;
> +	struct resource_entry *win;
> +	int offset = 0;
> +
> +	err = of_pci_get_host_bridge_resources(np, 0, 0xff, &pcie->resources,
> +					       &iobase);
> +	if (err)
> +		return err;
> +
> +	resource_list_for_each_entry(win, &pcie->resources) {
> +		struct resource *parent, *res = win->res;
> +
> +		switch (resource_type(res)) {
> +		case IORESOURCE_MEM:
> +			parent = &iomem_resource;
> +			res_valid |= !(res->flags & IORESOURCE_PREFETCH);
> +			cra_writel(pcie, res->start,
> +				A2P_ADDR_MAP_LO0 + offset);
> +			cra_writel(pcie, 0,
> +				A2P_ADDR_MAP_HI0 + offset);
> +			offset += ATT_ENTRY_SIZE;
> +			break;
> +		default:
> +			continue;
> +		}
> +
> +		err = devm_request_resource(dev, parent, res);
> +		if (err)
> +			goto out_release_res;
> +	}
> +
> +	if (!res_valid) {
> +		dev_err(dev, "non-prefetchable memory resource required\n");
> +		err = -EINVAL;
> +		goto out_release_res;
> +	}
> +
> +	return 0;
> +
> +out_release_res:
> +	altera_pcie_release_of_pci_ranges(pcie);
> +	return err;
> +}
> +
> +static void altera_pcie_free_irq_domain(struct altera_pcie *pcie)
> +{
> +	int i;
> +	u32 irq;
> +
> +	for (i = 0; i < INTX_NUM; i++) {
> +		irq = irq_find_mapping(pcie->irq_domain, i);
> +		if (irq > 0)
> +			irq_dispose_mapping(irq);
> +	}
> +
> +	irq_domain_remove(pcie->irq_domain);
> +}
> +
> +static int altera_pcie_init_irq_domain(struct altera_pcie *pcie)
> +{
> +	struct device *dev = &pcie->pdev->dev;
> +	struct device_node *node = dev->of_node;
> +
> +	/* Setup INTx */
> +	pcie->irq_domain = irq_domain_add_linear(node, INTX_NUM,
> +					&intx_domain_ops, pcie);
> +	if (!pcie->irq_domain) {
> +		dev_err(dev, "Failed to get a INTx IRQ domain\n");
> +		return PTR_ERR(pcie->irq_domain);
> +	}
> +
> +	return 0;
> +}
> +
> +static int altera_pcie_parse_dt(struct altera_pcie *pcie)
> +{
> +	struct resource *cra;
> +	int ret;
> +	struct platform_device *pdev = pcie->pdev;
> +
> +	cra = platform_get_resource_byname(pdev, IORESOURCE_MEM, "Cra");
> +	pcie->cra_base = devm_ioremap_resource(&pdev->dev, cra);
> +	if (IS_ERR(pcie->cra_base)) {
> +		dev_err(&pdev->dev, "get Cra resource failed\n");
> +		return PTR_ERR(pcie->cra_base);
> +	}
> +
> +	/* setup IRQ */
> +	pcie->hwirq = platform_get_irq(pdev, 0);
> +	if (pcie->hwirq <= 0) {

This is definitely not a hardware interrupt number. It is completely
virtual. I suggest you give it a better name.

> +		dev_err(&pdev->dev, "failed to get IRQ: %d\n", pcie->hwirq);
> +		return -EINVAL;
> +	}
> +	ret = devm_request_irq(&pdev->dev, pcie->hwirq, altera_pcie_isr,
> +			IRQF_SHARED, pdev->name, pcie);

You're implementing a cascading interrupt controller again. Please fix
it just like you did for your MSI support.

> +
> +	if (ret) {
> +		dev_err(&pdev->dev, "failed to request irq %d\n", pcie->hwirq);
> +		return ret;
> +	}
> +
> +	return 0;
> +}
> +
> +static int altera_pcie_probe(struct platform_device *pdev)
> +{
> +	struct altera_pcie *pcie;
> +	struct pci_bus *bus;
> +	int ret;
> +
> +	pcie = devm_kzalloc(&pdev->dev, sizeof(*pcie), GFP_KERNEL);
> +	if (!pcie)
> +		return -ENOMEM;
> +
> +	pcie->pdev = pdev;
> +
> +	ret = altera_pcie_parse_dt(pcie);
> +	if (ret) {
> +		dev_err(&pdev->dev, "Parsing DT failed\n");
> +		return ret;
> +	}
> +
> +	INIT_LIST_HEAD(&pcie->resources);
> +
> +	ret = altera_pcie_parse_request_of_pci_ranges(pcie);
> +	if (ret) {
> +		dev_err(&pdev->dev, "Failed add resources\n");
> +		return ret;
> +	}
> +
> +	ret = altera_pcie_init_irq_domain(pcie);
> +	if (ret) {
> +		dev_err(&pdev->dev, "Failed creating IRQ Domain\n");
> +		return ret;
> +	}
> +
> +	pcie->root_bus_nr = 0;
> +
> +	/* clear all interrupts */
> +	cra_writel(pcie, P2A_INT_STS_ALL, P2A_INT_STATUS);
> +	/* enable all interrupts */
> +	cra_writel(pcie, P2A_INT_ENA_ALL, P2A_INT_ENABLE);
> +
> +	bus = pci_scan_root_bus(&pdev->dev, pcie->root_bus_nr, &altera_pcie_ops,
> +					pcie, &pcie->resources);
> +	if (!bus)
> +		return -ENOMEM;
> +
> +	pci_assign_unassigned_bus_resources(bus);
> +	pci_bus_add_devices(bus);
> +
> +	platform_set_drvdata(pdev, pcie);
> +	return ret;
> +}
> +
> +static int __exit altera_pcie_remove(struct platform_device *pdev)
> +{
> +	struct altera_pcie *pcie = platform_get_drvdata(pdev);
> +
> +	altera_pcie_free_irq_domain(pcie);
> +	platform_set_drvdata(pdev, NULL);
> +	return 0;
> +}
> +
> +static const struct of_device_id altera_pcie_of_match[] = {
> +	{ .compatible = "altr,pcie-root-port-1.0", },
> +	{},
> +};
> +MODULE_DEVICE_TABLE(of, altera_pcie_of_match);
> +
> +static struct platform_driver altera_pcie_driver = {
> +	.probe		= altera_pcie_probe,
> +	.remove		= altera_pcie_remove,
> +	.driver = {
> +		.name	= "altera-pcie",
> +		.of_match_table = altera_pcie_of_match,
> +	},
> +};
> +
> +module_platform_driver(altera_pcie_driver);
> +
> +MODULE_AUTHOR("Ley Foon Tan <lftan@altera.com>");
> +MODULE_DESCRIPTION("Altera PCIe host controller driver");
> +MODULE_LICENSE("GPL v2");
> 


-- 
Jazz is not dead. It just smells funny...

Re: [PATCH v2 2/5] pci:host: Add Altera PCIe host controller driver

[Ley Foon Tan]

On Fri, Jul 31, 2015 at 8:34 PM, Marc Zyngier <marc.zyngier@arm.com> wrote:
> On 31/07/15 11:15, Ley Foon Tan wrote:
>> This patch adds the Altera PCIe host controller driver.
>>
>> Signed-off-by: Ley Foon Tan <lftan@altera.com>
>> ---
>>  drivers/pci/host/Kconfig       |   8 +
>>  drivers/pci/host/Makefile      |   1 +
>>  drivers/pci/host/pcie-altera.c | 526 +++++++++++++++++++++++++++++++++++++++++
>>  3 files changed, 535 insertions(+)
>>  create mode 100644 drivers/pci/host/pcie-altera.c
>>
>> diff --git a/drivers/pci/host/Kconfig b/drivers/pci/host/Kconfig
>> index 675c2d1..108500a 100644
>> --- a/drivers/pci/host/Kconfig
>> +++ b/drivers/pci/host/Kconfig
>> @@ -145,4 +145,12 @@ config PCIE_IPROC_BCMA
>>         Say Y here if you want to use the Broadcom iProc PCIe controller
>>         through the BCMA bus interface
>>
>> +config PCIE_ALTERA
>> +     tristate "Altera PCIe controller"
>> +     depends on ARCH_SOCFPGA
>> +     select PCI_MSI_IRQ_DOMAIN if PCI_MSI
>
> Why would you select this here? MSI support is in another driver, so I
> don't see the point.
Will move select PCI_MSI_IRQ_DOMAIN to config  PCIE_ALTERA_MSI section.

>
>> +     help
>> +       Say Y here if you want to enable PCIe controller support for Altera
>> +       SoCFPGA family of SoCs.
>> +
>>  endmenu

>> +static const struct irq_domain_ops intx_domain_ops = {
>> +     .map = altera_pcie_intx_map,
>> +};
>> +
>> +static irqreturn_t altera_pcie_isr(int irq, void *arg)
>> +{
>> +     struct altera_pcie *pcie = arg;
>> +     u32 status, i;
>> +
>> +     status = cra_readl(pcie, P2A_INT_STATUS) & P2A_INT_STS_ALL;
>> +
>> +     /* clear interrupts */
>> +     cra_writel(pcie, status, P2A_INT_STATUS);
>> +
>> +     for (i = 0; i < INTX_NUM; i++) {
>> +             if (status & (1 << i))
>
> find_first/next_set_bit?
Yes, change this to find_first_bit().

>
>> +                     generic_handle_irq(irq_find_mapping(pcie->irq_domain,
>> +                                                     i + 1));
>
> Rule of thumb: if you're calling generic_handle_irq from an interrupt
> handler, this is a chained IRQ controller. Please implement it as such.
Okay, will change this to chained_irq_enter/chained_irq_exit implementation.
>
>> +     }
>> +
>> +     return IRQ_HANDLED;
>> +}
>> +
>> +static void altera_pcie_release_of_pci_ranges(struct altera_pcie *pcie)
>> +{
>> +     pci_free_resource_list(&pcie->resources);
>> +}
>> +
>> +static int altera_pcie_parse_request_of_pci_ranges(struct altera_pcie *pcie)
>> +{
>> +     int err, res_valid = 0;
>> +     struct device *dev = &pcie->pdev->dev;
>> +     struct device_node *np = dev->of_node;
>> +     resource_size_t iobase;
>> +     struct resource_entry *win;
>> +     int offset = 0;
>> +
>> +     err = of_pci_get_host_bridge_resources(np, 0, 0xff, &pcie->resources,
>> +                                            &iobase);
>> +     if (err)
>> +             return err;
>> +
>> +     resource_list_for_each_entry(win, &pcie->resources) {
>> +             struct resource *parent, *res = win->res;
>> +
>> +             switch (resource_type(res)) {
>> +             case IORESOURCE_MEM:
>> +                     parent = &iomem_resource;
>> +                     res_valid |= !(res->flags & IORESOURCE_PREFETCH);
>> +                     cra_writel(pcie, res->start,
>> +                             A2P_ADDR_MAP_LO0 + offset);
>> +                     cra_writel(pcie, 0,
>> +                             A2P_ADDR_MAP_HI0 + offset);
>> +                     offset += ATT_ENTRY_SIZE;
>> +                     break;
>> +             default:
>> +                     continue;
>> +             }
>> +
>> +             err = devm_request_resource(dev, parent, res);
>> +             if (err)
>> +                     goto out_release_res;
>> +     }
>> +
>> +     if (!res_valid) {
>> +             dev_err(dev, "non-prefetchable memory resource required\n");
>> +             err = -EINVAL;
>> +             goto out_release_res;
>> +     }
>> +
>> +     return 0;
>> +
>> +out_release_res:
>> +     altera_pcie_release_of_pci_ranges(pcie);
>> +     return err;
>> +}
>> +
>> +static void altera_pcie_free_irq_domain(struct altera_pcie *pcie)
>> +{
>> +     int i;
>> +     u32 irq;
>> +
>> +     for (i = 0; i < INTX_NUM; i++) {
>> +             irq = irq_find_mapping(pcie->irq_domain, i);
>> +             if (irq > 0)
>> +                     irq_dispose_mapping(irq);
>> +     }
>> +
>> +     irq_domain_remove(pcie->irq_domain);
>> +}
>> +
>> +static int altera_pcie_init_irq_domain(struct altera_pcie *pcie)
>> +{
>> +     struct device *dev = &pcie->pdev->dev;
>> +     struct device_node *node = dev->of_node;
>> +
>> +     /* Setup INTx */
>> +     pcie->irq_domain = irq_domain_add_linear(node, INTX_NUM,
>> +                                     &intx_domain_ops, pcie);
>> +     if (!pcie->irq_domain) {
>> +             dev_err(dev, "Failed to get a INTx IRQ domain\n");
>> +             return PTR_ERR(pcie->irq_domain);
>> +     }
>> +
>> +     return 0;
>> +}
>> +
>> +static int altera_pcie_parse_dt(struct altera_pcie *pcie)
>> +{
>> +     struct resource *cra;
>> +     int ret;
>> +     struct platform_device *pdev = pcie->pdev;
>> +
>> +     cra = platform_get_resource_byname(pdev, IORESOURCE_MEM, "Cra");
>> +     pcie->cra_base = devm_ioremap_resource(&pdev->dev, cra);
>> +     if (IS_ERR(pcie->cra_base)) {
>> +             dev_err(&pdev->dev, "get Cra resource failed\n");
>> +             return PTR_ERR(pcie->cra_base);
>> +     }
>> +
>> +     /* setup IRQ */
>> +     pcie->hwirq = platform_get_irq(pdev, 0);
>> +     if (pcie->hwirq <= 0) {
>
> This is definitely not a hardware interrupt number. It is completely
> virtual. I suggest you give it a better name.
Noted.

>
>> +             dev_err(&pdev->dev, "failed to get IRQ: %d\n", pcie->hwirq);
>> +             return -EINVAL;
>> +     }
>> +     ret = devm_request_irq(&pdev->dev, pcie->hwirq, altera_pcie_isr,
>> +                     IRQF_SHARED, pdev->name, pcie);
>
> You're implementing a cascading interrupt controller again. Please fix
> it just like you did for your MSI support.
Sure, will change to irq_set_chained_handler_and_data here.


Thanks for reviewing.

Regards
Ley Foon

[PATCH v2 2/5] pci:host: Add Altera PCIe host controller driver

[Ley Foon Tan]

On Fri, Jul 31, 2015 at 8:34 PM, Marc Zyngier <marc.zyngier@arm.com> wrote:
> On 31/07/15 11:15, Ley Foon Tan wrote:
>> This patch adds the Altera PCIe host controller driver.
>>
>> Signed-off-by: Ley Foon Tan <lftan@altera.com>
>> ---
>>  drivers/pci/host/Kconfig       |   8 +
>>  drivers/pci/host/Makefile      |   1 +
>>  drivers/pci/host/pcie-altera.c | 526 +++++++++++++++++++++++++++++++++++++++++
>>  3 files changed, 535 insertions(+)
>>  create mode 100644 drivers/pci/host/pcie-altera.c
>>
>> diff --git a/drivers/pci/host/Kconfig b/drivers/pci/host/Kconfig
>> index 675c2d1..108500a 100644
>> --- a/drivers/pci/host/Kconfig
>> +++ b/drivers/pci/host/Kconfig
>> @@ -145,4 +145,12 @@ config PCIE_IPROC_BCMA
>>         Say Y here if you want to use the Broadcom iProc PCIe controller
>>         through the BCMA bus interface
>>
>> +config PCIE_ALTERA
>> +     tristate "Altera PCIe controller"
>> +     depends on ARCH_SOCFPGA
>> +     select PCI_MSI_IRQ_DOMAIN if PCI_MSI
>
> Why would you select this here? MSI support is in another driver, so I
> don't see the point.
Will move select PCI_MSI_IRQ_DOMAIN to config  PCIE_ALTERA_MSI section.

>
>> +     help
>> +       Say Y here if you want to enable PCIe controller support for Altera
>> +       SoCFPGA family of SoCs.
>> +
>>  endmenu

>> +static const struct irq_domain_ops intx_domain_ops = {
>> +     .map = altera_pcie_intx_map,
>> +};
>> +
>> +static irqreturn_t altera_pcie_isr(int irq, void *arg)
>> +{
>> +     struct altera_pcie *pcie = arg;
>> +     u32 status, i;
>> +
>> +     status = cra_readl(pcie, P2A_INT_STATUS) & P2A_INT_STS_ALL;
>> +
>> +     /* clear interrupts */
>> +     cra_writel(pcie, status, P2A_INT_STATUS);
>> +
>> +     for (i = 0; i < INTX_NUM; i++) {
>> +             if (status & (1 << i))
>
> find_first/next_set_bit?
Yes, change this to find_first_bit().

>
>> +                     generic_handle_irq(irq_find_mapping(pcie->irq_domain,
>> +                                                     i + 1));
>
> Rule of thumb: if you're calling generic_handle_irq from an interrupt
> handler, this is a chained IRQ controller. Please implement it as such.
Okay, will change this to chained_irq_enter/chained_irq_exit implementation.
>
>> +     }
>> +
>> +     return IRQ_HANDLED;
>> +}
>> +
>> +static void altera_pcie_release_of_pci_ranges(struct altera_pcie *pcie)
>> +{
>> +     pci_free_resource_list(&pcie->resources);
>> +}
>> +
>> +static int altera_pcie_parse_request_of_pci_ranges(struct altera_pcie *pcie)
>> +{
>> +     int err, res_valid = 0;
>> +     struct device *dev = &pcie->pdev->dev;
>> +     struct device_node *np = dev->of_node;
>> +     resource_size_t iobase;
>> +     struct resource_entry *win;
>> +     int offset = 0;
>> +
>> +     err = of_pci_get_host_bridge_resources(np, 0, 0xff, &pcie->resources,
>> +                                            &iobase);
>> +     if (err)
>> +             return err;
>> +
>> +     resource_list_for_each_entry(win, &pcie->resources) {
>> +             struct resource *parent, *res = win->res;
>> +
>> +             switch (resource_type(res)) {
>> +             case IORESOURCE_MEM:
>> +                     parent = &iomem_resource;
>> +                     res_valid |= !(res->flags & IORESOURCE_PREFETCH);
>> +                     cra_writel(pcie, res->start,
>> +                             A2P_ADDR_MAP_LO0 + offset);
>> +                     cra_writel(pcie, 0,
>> +                             A2P_ADDR_MAP_HI0 + offset);
>> +                     offset += ATT_ENTRY_SIZE;
>> +                     break;
>> +             default:
>> +                     continue;
>> +             }
>> +
>> +             err = devm_request_resource(dev, parent, res);
>> +             if (err)
>> +                     goto out_release_res;
>> +     }
>> +
>> +     if (!res_valid) {
>> +             dev_err(dev, "non-prefetchable memory resource required\n");
>> +             err = -EINVAL;
>> +             goto out_release_res;
>> +     }
>> +
>> +     return 0;
>> +
>> +out_release_res:
>> +     altera_pcie_release_of_pci_ranges(pcie);
>> +     return err;
>> +}
>> +
>> +static void altera_pcie_free_irq_domain(struct altera_pcie *pcie)
>> +{
>> +     int i;
>> +     u32 irq;
>> +
>> +     for (i = 0; i < INTX_NUM; i++) {
>> +             irq = irq_find_mapping(pcie->irq_domain, i);
>> +             if (irq > 0)
>> +                     irq_dispose_mapping(irq);
>> +     }
>> +
>> +     irq_domain_remove(pcie->irq_domain);
>> +}
>> +
>> +static int altera_pcie_init_irq_domain(struct altera_pcie *pcie)
>> +{
>> +     struct device *dev = &pcie->pdev->dev;
>> +     struct device_node *node = dev->of_node;
>> +
>> +     /* Setup INTx */
>> +     pcie->irq_domain = irq_domain_add_linear(node, INTX_NUM,
>> +                                     &intx_domain_ops, pcie);
>> +     if (!pcie->irq_domain) {
>> +             dev_err(dev, "Failed to get a INTx IRQ domain\n");
>> +             return PTR_ERR(pcie->irq_domain);
>> +     }
>> +
>> +     return 0;
>> +}
>> +
>> +static int altera_pcie_parse_dt(struct altera_pcie *pcie)
>> +{
>> +     struct resource *cra;
>> +     int ret;
>> +     struct platform_device *pdev = pcie->pdev;
>> +
>> +     cra = platform_get_resource_byname(pdev, IORESOURCE_MEM, "Cra");
>> +     pcie->cra_base = devm_ioremap_resource(&pdev->dev, cra);
>> +     if (IS_ERR(pcie->cra_base)) {
>> +             dev_err(&pdev->dev, "get Cra resource failed\n");
>> +             return PTR_ERR(pcie->cra_base);
>> +     }
>> +
>> +     /* setup IRQ */
>> +     pcie->hwirq = platform_get_irq(pdev, 0);
>> +     if (pcie->hwirq <= 0) {
>
> This is definitely not a hardware interrupt number. It is completely
> virtual. I suggest you give it a better name.
Noted.

>
>> +             dev_err(&pdev->dev, "failed to get IRQ: %d\n", pcie->hwirq);
>> +             return -EINVAL;
>> +     }
>> +     ret = devm_request_irq(&pdev->dev, pcie->hwirq, altera_pcie_isr,
>> +                     IRQF_SHARED, pdev->name, pcie);
>
> You're implementing a cascading interrupt controller again. Please fix
> it just like you did for your MSI support.
Sure, will change to irq_set_chained_handler_and_data here.


Thanks for reviewing.

Regards
Ley Foon

Re: [PATCH v2 3/5] pci: altera: Add Altera PCIe MSI driver

[Ley Foon Tan]

On Fri, Jul 31, 2015 at 8:12 PM, Marc Zyngier <marc.zyngier@arm.com> wrote:
> On 31/07/15 11:15, Ley Foon Tan wrote:
>> This patch adds Altera PCIe MSI driver. This soft IP supports configurable
>> number of vectors, which is a dts parameter.
>
> I've reviewed the initial drop of this code; basic courtesy would be to
> keep me CCed on the follow-up series.
Will keep you in CC for the following revision. Sorry about this.

>>
>> Signed-off-by: Ley Foon Tan <lftan@altera.com>
>> ---
>>  drivers/pci/host/Kconfig           |   7 +
>>  drivers/pci/host/Makefile          |   1 +
>>  drivers/pci/host/pcie-altera-msi.c | 324 +++++++++++++++++++++++++++++++++++++
>>  3 files changed, 332 insertions(+)
>>  create mode 100644 drivers/pci/host/pcie-altera-msi.c
>>
>> diff --git a/drivers/pci/host/Kconfig b/drivers/pci/host/Kconfig
>> index 108500a..51d0468 100644
>> --- a/drivers/pci/host/Kconfig
>> +++ b/drivers/pci/host/Kconfig
>> @@ -153,4 +153,11 @@ config PCIE_ALTERA
>>         Say Y here if you want to enable PCIe controller support for Altera
>>         SoCFPGA family of SoCs.
>>
>> +config PCIE_ALTERA_MSI
>> +     bool "Altera PCIe MSI feature"
>> +     depends on PCI_MSI
>> +     help
>> +       Say Y here if you want PCIe MSI support for the Altera SocFPGA SoC.
>> +       This MSI driver supports Altera MSI to GIC controller IP.
>> +
>>  endmenu
>> diff --git a/drivers/pci/host/Makefile b/drivers/pci/host/Makefile
>> index 6954f76..6c4913d 100644
>> --- a/drivers/pci/host/Makefile
>> +++ b/drivers/pci/host/Makefile
>> @@ -18,3 +18,4 @@ obj-$(CONFIG_PCIE_IPROC) += pcie-iproc.o
>>  obj-$(CONFIG_PCIE_IPROC_PLATFORM) += pcie-iproc-platform.o
>>  obj-$(CONFIG_PCIE_IPROC_BCMA) += pcie-iproc-bcma.o
>>  obj-$(CONFIG_PCIE_ALTERA) += pcie-altera.o
>> +obj-$(CONFIG_PCIE_ALTERA_MSI) += pcie-altera-msi.o
>> diff --git a/drivers/pci/host/pcie-altera-msi.c b/drivers/pci/host/pcie-altera-msi.c
>> new file mode 100644
>> index 0000000..6014719
>> --- /dev/null
>> +++ b/drivers/pci/host/pcie-altera-msi.c
>> @@ -0,0 +1,324 @@
>> +/*
>> + * Copyright Altera Corporation (C) 2013-2015. All rights reserved
>> + *
>> + * This program is free software; you can redistribute it and/or modify it
>> + * under the terms and conditions of the GNU General Public License,
>> + * version 2, as published by the Free Software Foundation.
>> + *
>> + * This program is distributed in the hope it will be useful, but WITHOUT
>> + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
>> + * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
>> + * more details.
>> + *
>> + * You should have received a copy of the GNU General Public License along with
>> + * this program.  If not, see <http://www.gnu.org/licenses/>.
>> + */
>> +#include <linux/interrupt.h>
>> +#include <linux/irqchip/chained_irq.h>
>> +#include <linux/module.h>
>> +#include <linux/msi.h>
>> +#include <linux/of_address.h>
>> +#include <linux/of_irq.h>
>> +#include <linux/of_pci.h>
>> +#include <linux/pci.h>
>> +#include <linux/platform_device.h>
>> +#include <linux/slab.h>
>> +
>> +#define MSI_STATUS           0x0
>> +#define MSI_ERROR            0x4
>> +#define MSI_INTMASK          0x8
>> +
>> +#define MAX_MSI_VECTORS              32
>> +struct altera_msi {
>> +     DECLARE_BITMAP(used, MAX_MSI_VECTORS);
>> +     struct mutex            lock;   /* proctect used variable */
>> +     struct platform_device  *pdev;
>> +     struct irq_domain               *msi_domain;
>> +     struct irq_domain               *inner_domain;
>> +     void __iomem            *csr_base;
>> +     void __iomem            *vector_base;
>> +     phys_addr_t             vector_phy;
>> +     u32                     num_of_vectors;
>> +     int                     irq;
>> +};
>> +
>> +static inline void msi_writel(struct altera_msi *msi, u32 value, u32 reg)
>> +{
>> +     writel_relaxed(value, msi->csr_base + reg);
>> +}
>> +
>> +static inline u32 msi_readl(struct altera_msi *msi, u32 reg)
>> +{
>> +     return readl_relaxed(msi->csr_base + reg);
>> +}
>> +
>> +static void altera_msi_isr(unsigned int irq, struct irq_desc *desc)
>> +{
>> +     struct irq_chip *chip = irq_desc_get_chip(desc);
>> +     struct altera_msi *msi;
>> +     unsigned long status;
>> +     u32 num_of_vectors;
>> +     u32 bit;
>> +
>> +     chained_irq_enter(chip, desc);
>> +     msi = irq_desc_get_handler_data(desc);
>> +     num_of_vectors = msi->num_of_vectors;
>> +
>> +     do {
>> +             status = msi_readl(msi, MSI_STATUS);
>> +             if (!status)
>> +                     break;
>> +
>> +             do {
>> +                     bit = find_first_bit(&status, num_of_vectors);
>> +                     /* Dummy read from vector to clear the interrupt */
>> +                     readl_relaxed(msi->vector_base + (bit * sizeof(u32)));
>> +
>> +                     irq = irq_find_mapping(msi->inner_domain, bit);
>
> Use of irq is a bit confusing, as this is a parameter describing the
> chaining interrupt. Consider using a different name for this variable.
Okay, will change it to 'virq'.

>
>> +                     if (irq) {
>> +                             if (test_bit(bit, msi->used))
>> +                                     generic_handle_irq(irq);
>> +                             else
>> +                                     dev_info(&msi->pdev->dev, "unhandled MSI\n");
>> +                     } else
>> +                             dev_info(&msi->pdev->dev, "unexpected MSI\n");
>
> The whole statement could be rewritten as:
>
> if (irq && test_bit(bit, msi->used))
>         generic_handle_irq(irq);
> else
>         dev_err(...);
>
> In all cases, this is a grave error, so dev_info is not appropriate here.
Noted.

>
>> +
>> +                     /* Clear the bit from status and repeat without reading
>> +                      * again status register. */
>> +                     clear_bit(bit, &status);
>
> You don't need an atomic operation here, so __clear_bit() should be enough.
Noted.

>
>> +             } while (status);
>> +     } while (1);
>> +
>> +     chained_irq_exit(chip, desc);
>> +}
>> +
>> +static struct irq_chip altera_msi_irq_chip = {
>> +     .name = "Altera 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 struct msi_domain_info altera_msi_domain_info = {
>> +     .flags  = (MSI_FLAG_USE_DEF_DOM_OPS | MSI_FLAG_USE_DEF_CHIP_OPS |
>> +                  MSI_FLAG_PCI_MSIX),
>> +     .chip   = &altera_msi_irq_chip,
>> +};
>> +
>> +static void altera_compose_msi_msg(struct irq_data *data, struct msi_msg *msg)
>> +{
>> +     u32 mask;
>> +     struct altera_msi *msi = irq_data_get_irq_chip_data(data);
>> +     u64 addr = msi->vector_phy + (data->hwirq * sizeof(u32));
>
> Use phys_addr_t.
Noted.

>
>> +
>> +     msg->address_lo = lower_32_bits(addr);
>> +     msg->address_hi = upper_32_bits(addr);
>> +     msg->data = data->hwirq;
>> +
>> +     mask = msi_readl(msi, MSI_INTMASK);
>> +     mask |= 1 << data->hwirq;
>> +     msi_writel(msi, mask, MSI_INTMASK);
>
> It feels a bit weird to unmask the interrupt when you compose the
> message. I'd expect this to be done in the mask/unmask methods.
Do you refer to these 2 callbacks?
.irq_mask = pci_msi_mask_irq,
.irq_unmask = pci_msi_unmask_irq,

How about we move this INTMASK code above to altera_irq_domain_alloc()?
We have unmask code in altera_irq_domain_free() now.

>
>> +     dev_dbg(&msi->pdev->dev, "msi#%d address_lo 0x%x\n", (int)data->hwirq,
>> +             msg->address_lo);
>
> You may want to print the whole address here.
Noted.

>
>> +}
>> +
>> +static int altera_msi_set_affinity(struct irq_data *irq_data,
>> +                              const struct cpumask *mask, bool force)
>> +{
>> +      return -EINVAL;
>> +}
>> +
>> +static struct irq_chip altera_msi_bottom_irq_chip = {
>> +     .name                   = "Altera MSI",
>> +     .irq_compose_msi_msg    = altera_compose_msi_msg,
>> +     .irq_set_affinity       = altera_msi_set_affinity,
>> +};
>> +
>> +static int altera_irq_domain_alloc(struct irq_domain *domain, unsigned int virq,
>> +                               unsigned int nr_irqs, void *args)
>> +{
>> +     struct altera_msi *msi = domain->host_data;
>> +     int bit;
>> +
>> +     mutex_lock(&msi->lock);
>> +
>> +     bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
>> +     if (bit < msi->num_of_vectors)
>> +             set_bit(bit, msi->used);
>> +
>> +     mutex_unlock(&msi->lock);
>> +
>> +     if (bit < 0)
>> +             return bit;
>
> How can "bit" be negative here? find_first_zero_bit returns an unsigned
> long... You probably want to change the type of "bit" to reflect that.
Okay, will change to "unsigned long" type.
>
>> +     else if (bit >= msi->num_of_vectors)
>
> Useless "else" anyway.

I think we still need to check for failing case, if we don't have
available unused bit.
This could be rewritten  as below:

bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
if (bit < msi->num_of_vectors)
      set_bit(bit, msi->used);
else
      return -ENOSPC;

>
>> +             return -ENOSPC;
>> +
>> +     irq_domain_set_info(domain, virq, bit, &altera_msi_bottom_irq_chip,
>> +                         domain->host_data, handle_simple_irq,
>> +                         NULL, NULL);
>> +     set_irq_flags(virq, IRQF_VALID);
>> +
>> +     return 0;
>> +}
>> +
>> +static void altera_irq_domain_free(struct irq_domain *domain,
>> +                               unsigned int virq, unsigned int nr_irqs)
>> +{
>> +     struct irq_data *d = irq_domain_get_irq_data(domain, virq);
>> +     struct altera_msi *msi = irq_data_get_irq_chip_data(d);
>> +     u32 mask;
>> +
>> +     mutex_lock(&msi->lock);
>> +
>> +     if (!test_bit(d->hwirq, msi->used))
>> +             dev_err(&msi->pdev->dev, "trying to free unused MSI#%lu\n",
>> +             d->hwirq);
>
> Put this statement in { } so that it matches the "else" construct.
Noted.

>
>> +     else {
>> +             clear_bit(d->hwirq, msi->used);
>
> No need for an atomic operation here, you own the mutex.
Okay, change to __clear_bit.


>
>> +             mask = msi_readl(msi, MSI_INTMASK);
>> +             mask &= ~(1 << d->hwirq);
>> +             msi_writel(msi, mask, MSI_INTMASK);
>> +     }
>> +
>> +     mutex_unlock(&msi->lock);
>> +}
>> +
>> +static const struct irq_domain_ops msi_domain_ops = {
>> +     .alloc  = altera_irq_domain_alloc,
>> +     .free   = altera_irq_domain_free,
>> +};
>> +
>> +static int altera_allocate_domains(struct altera_msi *msi)
>> +{
>> +     msi->inner_domain = irq_domain_add_linear(NULL, msi->num_of_vectors,
>> +                                          &msi_domain_ops, msi);
>> +     if (!msi->inner_domain) {
>> +             dev_err(&msi->pdev->dev, "failed to create IRQ domain\n");
>> +             return -ENOMEM;
>> +     }
>> +
>> +     msi->msi_domain = pci_msi_create_irq_domain(
>> +                             msi->pdev->dev.of_node,
>
> Please put at least one argument to the function on the same line. This
> is otherwise very hard to read.
Okay.

>
>> +                             &altera_msi_domain_info, msi->inner_domain);
>> +     if (!msi->msi_domain) {
>> +             dev_err(&msi->pdev->dev, "failed to create MSI domain\n");
>> +             irq_domain_remove(msi->inner_domain);
>> +             return -ENOMEM;
>> +     }
>> +
>> +     return 0;
>> +}
>> +
>> +static void altera_free_domains(struct altera_msi *msi)
>> +{
>> +     irq_domain_remove(msi->msi_domain);
>> +     irq_domain_remove(msi->inner_domain);
>> +}
>> +
>> +
>> +static int altera_msi_remove(struct platform_device *pdev)
>> +{
>> +     struct altera_msi *msi = platform_get_drvdata(pdev);
>> +
>> +     msi_writel(msi, 0, MSI_INTMASK);
>> +     irq_set_chained_handler(msi->irq, NULL);
>> +     irq_set_handler_data(msi->irq, NULL);
>> +
>> +     altera_free_domains(msi);
>> +
>> +     platform_set_drvdata(pdev, NULL);
>> +     return 0;
>> +}
>> +
>> +int altera_msi_probe(struct platform_device *pdev)
>
> Make this static.
Okay.

>
>> +{
>> +     struct altera_msi *msi;
>> +     struct device_node *np = pdev->dev.of_node;
>> +     struct resource *res;
>> +     int ret;
>> +
>> +     msi = devm_kzalloc(&pdev->dev, sizeof(struct altera_msi),
>> +             GFP_KERNEL);
>> +     if (!msi)
>> +             return -ENOMEM;
>> +
>> +     mutex_init(&msi->lock);
>> +     msi->pdev = pdev;
>> +
>> +     res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "csr");
>> +     msi->csr_base = devm_ioremap_resource(&pdev->dev, res);
>> +     if (IS_ERR(msi->csr_base)) {
>> +             dev_err(&pdev->dev, "get csr resource failed\n");
>> +             return PTR_ERR(msi->csr_base);
>> +     }
>> +
>> +     res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
>> +                     "vector_slave");
>> +     msi->vector_base = devm_ioremap_resource(&pdev->dev, res);
>> +     if (IS_ERR(msi->vector_base)) {
>> +             dev_err(&pdev->dev, "get vector slave resource failed\n");
>> +             return PTR_ERR(msi->vector_base);
>> +     }
>> +
>> +     msi->vector_phy = res->start;
>> +
>> +     if (of_property_read_u32(np, "num-vectors", &msi->num_of_vectors)) {
>> +             dev_err(&pdev->dev, "failed to parse the number of vectors\n");
>> +             return -EINVAL;
>> +     }
>> +
>> +     ret = altera_allocate_domains(msi);
>> +     if (ret)
>> +             return ret;
>> +
>> +     msi->irq = platform_get_irq(pdev, 0);
>> +     if (msi->irq <= 0) {
>> +             dev_err(&pdev->dev, "failed to map IRQ: %d\n", msi->irq);
>> +             ret = -ENODEV;
>> +             goto err;
>> +     }
>> +
>> +     irq_set_chained_handler(msi->irq, altera_msi_isr);
>> +     ret = irq_set_handler_data(msi->irq, msi);
>
> Use irq_set_chained_handler_and_data.
Okay.

>
>> +     if (ret) {
>> +             dev_err(&pdev->dev, "failed to register IRQ handler: %d\n",
>> +                     ret);
>> +             goto err;
>> +     }
>> +
>> +     platform_set_drvdata(pdev, msi);
>> +
>> +     return 0;
>> +
>> +err:
>> +     altera_msi_remove(pdev);
>> +     return ret;
>> +}
>> +
>> +static const struct of_device_id altera_msi_of_match[] = {
>> +     { .compatible = "altr,msi-1.0", NULL },
>> +     { },
>> +};
>> +MODULE_DEVICE_TABLE(of, altera_msi_of_match);
>> +
>> +static struct platform_driver altera_msi_driver = {
>> +     .driver = {
>> +             .name = "altera-msi",
>> +             .owner = THIS_MODULE,
>
> Remove this .owner assignment, the core code deals with it.
Okay.

>
>> +             .of_match_table = altera_msi_of_match,
>> +     },
>> +     .probe = altera_msi_probe,
>> +     .remove = altera_msi_remove,
>> +};
>> +
>> +static int __init altera_msi_init(void)
>> +{
>> +     return platform_driver_register(&altera_msi_driver);
>> +}
>> +
>> +subsys_initcall(altera_msi_init);
>
> Hmmm. if you're initializing this as part of the core kernel, it will
> never be a module. What is the point of the MODULE_DEVICE_TABLE above?
Okay, will remove MODULE_DEVICE_TABLE.

Thanks for reviewing.

Regards
Ley Foon

[PATCH v2 3/5] pci: altera: Add Altera PCIe MSI driver

[Ley Foon Tan]

On Fri, Jul 31, 2015 at 8:12 PM, Marc Zyngier <marc.zyngier@arm.com> wrote:
> On 31/07/15 11:15, Ley Foon Tan wrote:
>> This patch adds Altera PCIe MSI driver. This soft IP supports configurable
>> number of vectors, which is a dts parameter.
>
> I've reviewed the initial drop of this code; basic courtesy would be to
> keep me CCed on the follow-up series.
Will keep you in CC for the following revision. Sorry about this.

>>
>> Signed-off-by: Ley Foon Tan <lftan@altera.com>
>> ---
>>  drivers/pci/host/Kconfig           |   7 +
>>  drivers/pci/host/Makefile          |   1 +
>>  drivers/pci/host/pcie-altera-msi.c | 324 +++++++++++++++++++++++++++++++++++++
>>  3 files changed, 332 insertions(+)
>>  create mode 100644 drivers/pci/host/pcie-altera-msi.c
>>
>> diff --git a/drivers/pci/host/Kconfig b/drivers/pci/host/Kconfig
>> index 108500a..51d0468 100644
>> --- a/drivers/pci/host/Kconfig
>> +++ b/drivers/pci/host/Kconfig
>> @@ -153,4 +153,11 @@ config PCIE_ALTERA
>>         Say Y here if you want to enable PCIe controller support for Altera
>>         SoCFPGA family of SoCs.
>>
>> +config PCIE_ALTERA_MSI
>> +     bool "Altera PCIe MSI feature"
>> +     depends on PCI_MSI
>> +     help
>> +       Say Y here if you want PCIe MSI support for the Altera SocFPGA SoC.
>> +       This MSI driver supports Altera MSI to GIC controller IP.
>> +
>>  endmenu
>> diff --git a/drivers/pci/host/Makefile b/drivers/pci/host/Makefile
>> index 6954f76..6c4913d 100644
>> --- a/drivers/pci/host/Makefile
>> +++ b/drivers/pci/host/Makefile
>> @@ -18,3 +18,4 @@ obj-$(CONFIG_PCIE_IPROC) += pcie-iproc.o
>>  obj-$(CONFIG_PCIE_IPROC_PLATFORM) += pcie-iproc-platform.o
>>  obj-$(CONFIG_PCIE_IPROC_BCMA) += pcie-iproc-bcma.o
>>  obj-$(CONFIG_PCIE_ALTERA) += pcie-altera.o
>> +obj-$(CONFIG_PCIE_ALTERA_MSI) += pcie-altera-msi.o
>> diff --git a/drivers/pci/host/pcie-altera-msi.c b/drivers/pci/host/pcie-altera-msi.c
>> new file mode 100644
>> index 0000000..6014719
>> --- /dev/null
>> +++ b/drivers/pci/host/pcie-altera-msi.c
>> @@ -0,0 +1,324 @@
>> +/*
>> + * Copyright Altera Corporation (C) 2013-2015. All rights reserved
>> + *
>> + * This program is free software; you can redistribute it and/or modify it
>> + * under the terms and conditions of the GNU General Public License,
>> + * version 2, as published by the Free Software Foundation.
>> + *
>> + * This program is distributed in the hope it will be useful, but WITHOUT
>> + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
>> + * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
>> + * more details.
>> + *
>> + * You should have received a copy of the GNU General Public License along with
>> + * this program.  If not, see <http://www.gnu.org/licenses/>.
>> + */
>> +#include <linux/interrupt.h>
>> +#include <linux/irqchip/chained_irq.h>
>> +#include <linux/module.h>
>> +#include <linux/msi.h>
>> +#include <linux/of_address.h>
>> +#include <linux/of_irq.h>
>> +#include <linux/of_pci.h>
>> +#include <linux/pci.h>
>> +#include <linux/platform_device.h>
>> +#include <linux/slab.h>
>> +
>> +#define MSI_STATUS           0x0
>> +#define MSI_ERROR            0x4
>> +#define MSI_INTMASK          0x8
>> +
>> +#define MAX_MSI_VECTORS              32
>> +struct altera_msi {
>> +     DECLARE_BITMAP(used, MAX_MSI_VECTORS);
>> +     struct mutex            lock;   /* proctect used variable */
>> +     struct platform_device  *pdev;
>> +     struct irq_domain               *msi_domain;
>> +     struct irq_domain               *inner_domain;
>> +     void __iomem            *csr_base;
>> +     void __iomem            *vector_base;
>> +     phys_addr_t             vector_phy;
>> +     u32                     num_of_vectors;
>> +     int                     irq;
>> +};
>> +
>> +static inline void msi_writel(struct altera_msi *msi, u32 value, u32 reg)
>> +{
>> +     writel_relaxed(value, msi->csr_base + reg);
>> +}
>> +
>> +static inline u32 msi_readl(struct altera_msi *msi, u32 reg)
>> +{
>> +     return readl_relaxed(msi->csr_base + reg);
>> +}
>> +
>> +static void altera_msi_isr(unsigned int irq, struct irq_desc *desc)
>> +{
>> +     struct irq_chip *chip = irq_desc_get_chip(desc);
>> +     struct altera_msi *msi;
>> +     unsigned long status;
>> +     u32 num_of_vectors;
>> +     u32 bit;
>> +
>> +     chained_irq_enter(chip, desc);
>> +     msi = irq_desc_get_handler_data(desc);
>> +     num_of_vectors = msi->num_of_vectors;
>> +
>> +     do {
>> +             status = msi_readl(msi, MSI_STATUS);
>> +             if (!status)
>> +                     break;
>> +
>> +             do {
>> +                     bit = find_first_bit(&status, num_of_vectors);
>> +                     /* Dummy read from vector to clear the interrupt */
>> +                     readl_relaxed(msi->vector_base + (bit * sizeof(u32)));
>> +
>> +                     irq = irq_find_mapping(msi->inner_domain, bit);
>
> Use of irq is a bit confusing, as this is a parameter describing the
> chaining interrupt. Consider using a different name for this variable.
Okay, will change it to 'virq'.

>
>> +                     if (irq) {
>> +                             if (test_bit(bit, msi->used))
>> +                                     generic_handle_irq(irq);
>> +                             else
>> +                                     dev_info(&msi->pdev->dev, "unhandled MSI\n");
>> +                     } else
>> +                             dev_info(&msi->pdev->dev, "unexpected MSI\n");
>
> The whole statement could be rewritten as:
>
> if (irq && test_bit(bit, msi->used))
>         generic_handle_irq(irq);
> else
>         dev_err(...);
>
> In all cases, this is a grave error, so dev_info is not appropriate here.
Noted.

>
>> +
>> +                     /* Clear the bit from status and repeat without reading
>> +                      * again status register. */
>> +                     clear_bit(bit, &status);
>
> You don't need an atomic operation here, so __clear_bit() should be enough.
Noted.

>
>> +             } while (status);
>> +     } while (1);
>> +
>> +     chained_irq_exit(chip, desc);
>> +}
>> +
>> +static struct irq_chip altera_msi_irq_chip = {
>> +     .name = "Altera 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 struct msi_domain_info altera_msi_domain_info = {
>> +     .flags  = (MSI_FLAG_USE_DEF_DOM_OPS | MSI_FLAG_USE_DEF_CHIP_OPS |
>> +                  MSI_FLAG_PCI_MSIX),
>> +     .chip   = &altera_msi_irq_chip,
>> +};
>> +
>> +static void altera_compose_msi_msg(struct irq_data *data, struct msi_msg *msg)
>> +{
>> +     u32 mask;
>> +     struct altera_msi *msi = irq_data_get_irq_chip_data(data);
>> +     u64 addr = msi->vector_phy + (data->hwirq * sizeof(u32));
>
> Use phys_addr_t.
Noted.

>
>> +
>> +     msg->address_lo = lower_32_bits(addr);
>> +     msg->address_hi = upper_32_bits(addr);
>> +     msg->data = data->hwirq;
>> +
>> +     mask = msi_readl(msi, MSI_INTMASK);
>> +     mask |= 1 << data->hwirq;
>> +     msi_writel(msi, mask, MSI_INTMASK);
>
> It feels a bit weird to unmask the interrupt when you compose the
> message. I'd expect this to be done in the mask/unmask methods.
Do you refer to these 2 callbacks?
.irq_mask = pci_msi_mask_irq,
.irq_unmask = pci_msi_unmask_irq,

How about we move this INTMASK code above to altera_irq_domain_alloc()?
We have unmask code in altera_irq_domain_free() now.

>
>> +     dev_dbg(&msi->pdev->dev, "msi#%d address_lo 0x%x\n", (int)data->hwirq,
>> +             msg->address_lo);
>
> You may want to print the whole address here.
Noted.

>
>> +}
>> +
>> +static int altera_msi_set_affinity(struct irq_data *irq_data,
>> +                              const struct cpumask *mask, bool force)
>> +{
>> +      return -EINVAL;
>> +}
>> +
>> +static struct irq_chip altera_msi_bottom_irq_chip = {
>> +     .name                   = "Altera MSI",
>> +     .irq_compose_msi_msg    = altera_compose_msi_msg,
>> +     .irq_set_affinity       = altera_msi_set_affinity,
>> +};
>> +
>> +static int altera_irq_domain_alloc(struct irq_domain *domain, unsigned int virq,
>> +                               unsigned int nr_irqs, void *args)
>> +{
>> +     struct altera_msi *msi = domain->host_data;
>> +     int bit;
>> +
>> +     mutex_lock(&msi->lock);
>> +
>> +     bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
>> +     if (bit < msi->num_of_vectors)
>> +             set_bit(bit, msi->used);
>> +
>> +     mutex_unlock(&msi->lock);
>> +
>> +     if (bit < 0)
>> +             return bit;
>
> How can "bit" be negative here? find_first_zero_bit returns an unsigned
> long... You probably want to change the type of "bit" to reflect that.
Okay, will change to "unsigned long" type.
>
>> +     else if (bit >= msi->num_of_vectors)
>
> Useless "else" anyway.

I think we still need to check for failing case, if we don't have
available unused bit.
This could be rewritten  as below:

bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
if (bit < msi->num_of_vectors)
      set_bit(bit, msi->used);
else
      return -ENOSPC;

>
>> +             return -ENOSPC;
>> +
>> +     irq_domain_set_info(domain, virq, bit, &altera_msi_bottom_irq_chip,
>> +                         domain->host_data, handle_simple_irq,
>> +                         NULL, NULL);
>> +     set_irq_flags(virq, IRQF_VALID);
>> +
>> +     return 0;
>> +}
>> +
>> +static void altera_irq_domain_free(struct irq_domain *domain,
>> +                               unsigned int virq, unsigned int nr_irqs)
>> +{
>> +     struct irq_data *d = irq_domain_get_irq_data(domain, virq);
>> +     struct altera_msi *msi = irq_data_get_irq_chip_data(d);
>> +     u32 mask;
>> +
>> +     mutex_lock(&msi->lock);
>> +
>> +     if (!test_bit(d->hwirq, msi->used))
>> +             dev_err(&msi->pdev->dev, "trying to free unused MSI#%lu\n",
>> +             d->hwirq);
>
> Put this statement in { } so that it matches the "else" construct.
Noted.

>
>> +     else {
>> +             clear_bit(d->hwirq, msi->used);
>
> No need for an atomic operation here, you own the mutex.
Okay, change to __clear_bit.


>
>> +             mask = msi_readl(msi, MSI_INTMASK);
>> +             mask &= ~(1 << d->hwirq);
>> +             msi_writel(msi, mask, MSI_INTMASK);
>> +     }
>> +
>> +     mutex_unlock(&msi->lock);
>> +}
>> +
>> +static const struct irq_domain_ops msi_domain_ops = {
>> +     .alloc  = altera_irq_domain_alloc,
>> +     .free   = altera_irq_domain_free,
>> +};
>> +
>> +static int altera_allocate_domains(struct altera_msi *msi)
>> +{
>> +     msi->inner_domain = irq_domain_add_linear(NULL, msi->num_of_vectors,
>> +                                          &msi_domain_ops, msi);
>> +     if (!msi->inner_domain) {
>> +             dev_err(&msi->pdev->dev, "failed to create IRQ domain\n");
>> +             return -ENOMEM;
>> +     }
>> +
>> +     msi->msi_domain = pci_msi_create_irq_domain(
>> +                             msi->pdev->dev.of_node,
>
> Please put at least one argument to the function on the same line. This
> is otherwise very hard to read.
Okay.

>
>> +                             &altera_msi_domain_info, msi->inner_domain);
>> +     if (!msi->msi_domain) {
>> +             dev_err(&msi->pdev->dev, "failed to create MSI domain\n");
>> +             irq_domain_remove(msi->inner_domain);
>> +             return -ENOMEM;
>> +     }
>> +
>> +     return 0;
>> +}
>> +
>> +static void altera_free_domains(struct altera_msi *msi)
>> +{
>> +     irq_domain_remove(msi->msi_domain);
>> +     irq_domain_remove(msi->inner_domain);
>> +}
>> +
>> +
>> +static int altera_msi_remove(struct platform_device *pdev)
>> +{
>> +     struct altera_msi *msi = platform_get_drvdata(pdev);
>> +
>> +     msi_writel(msi, 0, MSI_INTMASK);
>> +     irq_set_chained_handler(msi->irq, NULL);
>> +     irq_set_handler_data(msi->irq, NULL);
>> +
>> +     altera_free_domains(msi);
>> +
>> +     platform_set_drvdata(pdev, NULL);
>> +     return 0;
>> +}
>> +
>> +int altera_msi_probe(struct platform_device *pdev)
>
> Make this static.
Okay.

>
>> +{
>> +     struct altera_msi *msi;
>> +     struct device_node *np = pdev->dev.of_node;
>> +     struct resource *res;
>> +     int ret;
>> +
>> +     msi = devm_kzalloc(&pdev->dev, sizeof(struct altera_msi),
>> +             GFP_KERNEL);
>> +     if (!msi)
>> +             return -ENOMEM;
>> +
>> +     mutex_init(&msi->lock);
>> +     msi->pdev = pdev;
>> +
>> +     res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "csr");
>> +     msi->csr_base = devm_ioremap_resource(&pdev->dev, res);
>> +     if (IS_ERR(msi->csr_base)) {
>> +             dev_err(&pdev->dev, "get csr resource failed\n");
>> +             return PTR_ERR(msi->csr_base);
>> +     }
>> +
>> +     res = platform_get_resource_byname(pdev, IORESOURCE_MEM,
>> +                     "vector_slave");
>> +     msi->vector_base = devm_ioremap_resource(&pdev->dev, res);
>> +     if (IS_ERR(msi->vector_base)) {
>> +             dev_err(&pdev->dev, "get vector slave resource failed\n");
>> +             return PTR_ERR(msi->vector_base);
>> +     }
>> +
>> +     msi->vector_phy = res->start;
>> +
>> +     if (of_property_read_u32(np, "num-vectors", &msi->num_of_vectors)) {
>> +             dev_err(&pdev->dev, "failed to parse the number of vectors\n");
>> +             return -EINVAL;
>> +     }
>> +
>> +     ret = altera_allocate_domains(msi);
>> +     if (ret)
>> +             return ret;
>> +
>> +     msi->irq = platform_get_irq(pdev, 0);
>> +     if (msi->irq <= 0) {
>> +             dev_err(&pdev->dev, "failed to map IRQ: %d\n", msi->irq);
>> +             ret = -ENODEV;
>> +             goto err;
>> +     }
>> +
>> +     irq_set_chained_handler(msi->irq, altera_msi_isr);
>> +     ret = irq_set_handler_data(msi->irq, msi);
>
> Use irq_set_chained_handler_and_data.
Okay.

>
>> +     if (ret) {
>> +             dev_err(&pdev->dev, "failed to register IRQ handler: %d\n",
>> +                     ret);
>> +             goto err;
>> +     }
>> +
>> +     platform_set_drvdata(pdev, msi);
>> +
>> +     return 0;
>> +
>> +err:
>> +     altera_msi_remove(pdev);
>> +     return ret;
>> +}
>> +
>> +static const struct of_device_id altera_msi_of_match[] = {
>> +     { .compatible = "altr,msi-1.0", NULL },
>> +     { },
>> +};
>> +MODULE_DEVICE_TABLE(of, altera_msi_of_match);
>> +
>> +static struct platform_driver altera_msi_driver = {
>> +     .driver = {
>> +             .name = "altera-msi",
>> +             .owner = THIS_MODULE,
>
> Remove this .owner assignment, the core code deals with it.
Okay.

>
>> +             .of_match_table = altera_msi_of_match,
>> +     },
>> +     .probe = altera_msi_probe,
>> +     .remove = altera_msi_remove,
>> +};
>> +
>> +static int __init altera_msi_init(void)
>> +{
>> +     return platform_driver_register(&altera_msi_driver);
>> +}
>> +
>> +subsys_initcall(altera_msi_init);
>
> Hmmm. if you're initializing this as part of the core kernel, it will
> never be a module. What is the point of the MODULE_DEVICE_TABLE above?
Okay, will remove MODULE_DEVICE_TABLE.

Thanks for reviewing.

Regards
Ley Foon

Re: [PATCH v2 3/5] pci: altera: Add Altera PCIe MSI driver

[Marc Zyngier]

On 03/08/15 11:37, Ley Foon Tan wrote:
> On Fri, Jul 31, 2015 at 8:12 PM, Marc Zyngier <marc.zyngier@arm.com> wrote:
>> On 31/07/15 11:15, Ley Foon Tan wrote:
>>> This patch adds Altera PCIe MSI driver. This soft IP supports configurable
>>> number of vectors, which is a dts parameter.
>>
>> I've reviewed the initial drop of this code; basic courtesy would be to
>> keep me CCed on the follow-up series.
> Will keep you in CC for the following revision. Sorry about this.
> 
>>>
>>> Signed-off-by: Ley Foon Tan <lftan@altera.com>
>>> ---
>>>  drivers/pci/host/Kconfig           |   7 +
>>>  drivers/pci/host/Makefile          |   1 +
>>>  drivers/pci/host/pcie-altera-msi.c | 324 +++++++++++++++++++++++++++++++++++++
>>>  3 files changed, 332 insertions(+)
>>>  create mode 100644 drivers/pci/host/pcie-altera-msi.c
>>>

[...]

>>> +
>>> +     msg->address_lo = lower_32_bits(addr);
>>> +     msg->address_hi = upper_32_bits(addr);
>>> +     msg->data = data->hwirq;
>>> +
>>> +     mask = msi_readl(msi, MSI_INTMASK);
>>> +     mask |= 1 << data->hwirq;
>>> +     msi_writel(msi, mask, MSI_INTMASK);
>>
>> It feels a bit weird to unmask the interrupt when you compose the
>> message. I'd expect this to be done in the mask/unmask methods.
> Do you refer to these 2 callbacks?
> .irq_mask = pci_msi_mask_irq,
> .irq_unmask = pci_msi_unmask_irq,
> 
> How about we move this INTMASK code above to altera_irq_domain_alloc()?
> We have unmask code in altera_irq_domain_free() now.

Either you move the mask/unmask code to local irq_mask/irq_unmask
methods (and call pci_msi_(un)mask_irq from there), or you move it
entierely to alloc/free.

Some level of symmetry helps following what is going on in the code.

[...]

>>> +static int altera_irq_domain_alloc(struct irq_domain *domain, unsigned int virq,
>>> +                               unsigned int nr_irqs, void *args)
>>> +{
>>> +     struct altera_msi *msi = domain->host_data;
>>> +     int bit;
>>> +
>>> +     mutex_lock(&msi->lock);
>>> +
>>> +     bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
>>> +     if (bit < msi->num_of_vectors)
>>> +             set_bit(bit, msi->used);
>>> +
>>> +     mutex_unlock(&msi->lock);
>>> +
>>> +     if (bit < 0)
>>> +             return bit;
>>
>> How can "bit" be negative here? find_first_zero_bit returns an unsigned
>> long... You probably want to change the type of "bit" to reflect that.
> Okay, will change to "unsigned long" type.
>>
>>> +     else if (bit >= msi->num_of_vectors)
>>
>> Useless "else" anyway.
> 
> I think we still need to check for failing case, if we don't have
> available unused bit.
> This could be rewritten  as below:
> 
> bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
> if (bit < msi->num_of_vectors)
>       set_bit(bit, msi->used);
> else
>       return -ENOSPC;

The more logical to write this is:

	if (bit >= msi->num_of_vectors)
		return -ENOSPC;

	set_bit(bit, msi->used);

which gets rid of the error case early and streamlines the normal case.

> 
>>
>>> +             return -ENOSPC;
>>> +
>>> +     irq_domain_set_info(domain, virq, bit, &altera_msi_bottom_irq_chip,
>>> +                         domain->host_data, handle_simple_irq,
>>> +                         NULL, NULL);
>>> +     set_irq_flags(virq, IRQF_VALID);
>>> +
>>> +     return 0;
>>> +}

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...

Re: [PATCH v2 3/5] pci: altera: Add Altera PCIe MSI driver

[Marc Zyngier]

On 03/08/15 11:37, Ley Foon Tan wrote:
> On Fri, Jul 31, 2015 at 8:12 PM, Marc Zyngier <marc.zyngier-5wv7dgnIgG8@public.gmane.org> wrote:
>> On 31/07/15 11:15, Ley Foon Tan wrote:
>>> This patch adds Altera PCIe MSI driver. This soft IP supports configurable
>>> number of vectors, which is a dts parameter.
>>
>> I've reviewed the initial drop of this code; basic courtesy would be to
>> keep me CCed on the follow-up series.
> Will keep you in CC for the following revision. Sorry about this.
> 
>>>
>>> Signed-off-by: Ley Foon Tan <lftan-EIB2kfCEclfQT0dZR+AlfA@public.gmane.org>
>>> ---
>>>  drivers/pci/host/Kconfig           |   7 +
>>>  drivers/pci/host/Makefile          |   1 +
>>>  drivers/pci/host/pcie-altera-msi.c | 324 +++++++++++++++++++++++++++++++++++++
>>>  3 files changed, 332 insertions(+)
>>>  create mode 100644 drivers/pci/host/pcie-altera-msi.c
>>>

[...]

>>> +
>>> +     msg->address_lo = lower_32_bits(addr);
>>> +     msg->address_hi = upper_32_bits(addr);
>>> +     msg->data = data->hwirq;
>>> +
>>> +     mask = msi_readl(msi, MSI_INTMASK);
>>> +     mask |= 1 << data->hwirq;
>>> +     msi_writel(msi, mask, MSI_INTMASK);
>>
>> It feels a bit weird to unmask the interrupt when you compose the
>> message. I'd expect this to be done in the mask/unmask methods.
> Do you refer to these 2 callbacks?
> .irq_mask = pci_msi_mask_irq,
> .irq_unmask = pci_msi_unmask_irq,
> 
> How about we move this INTMASK code above to altera_irq_domain_alloc()?
> We have unmask code in altera_irq_domain_free() now.

Either you move the mask/unmask code to local irq_mask/irq_unmask
methods (and call pci_msi_(un)mask_irq from there), or you move it
entierely to alloc/free.

Some level of symmetry helps following what is going on in the code.

[...]

>>> +static int altera_irq_domain_alloc(struct irq_domain *domain, unsigned int virq,
>>> +                               unsigned int nr_irqs, void *args)
>>> +{
>>> +     struct altera_msi *msi = domain->host_data;
>>> +     int bit;
>>> +
>>> +     mutex_lock(&msi->lock);
>>> +
>>> +     bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
>>> +     if (bit < msi->num_of_vectors)
>>> +             set_bit(bit, msi->used);
>>> +
>>> +     mutex_unlock(&msi->lock);
>>> +
>>> +     if (bit < 0)
>>> +             return bit;
>>
>> How can "bit" be negative here? find_first_zero_bit returns an unsigned
>> long... You probably want to change the type of "bit" to reflect that.
> Okay, will change to "unsigned long" type.
>>
>>> +     else if (bit >= msi->num_of_vectors)
>>
>> Useless "else" anyway.
> 
> I think we still need to check for failing case, if we don't have
> available unused bit.
> This could be rewritten  as below:
> 
> bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
> if (bit < msi->num_of_vectors)
>       set_bit(bit, msi->used);
> else
>       return -ENOSPC;

The more logical to write this is:

	if (bit >= msi->num_of_vectors)
		return -ENOSPC;

	set_bit(bit, msi->used);

which gets rid of the error case early and streamlines the normal case.

> 
>>
>>> +             return -ENOSPC;
>>> +
>>> +     irq_domain_set_info(domain, virq, bit, &altera_msi_bottom_irq_chip,
>>> +                         domain->host_data, handle_simple_irq,
>>> +                         NULL, NULL);
>>> +     set_irq_flags(virq, IRQF_VALID);
>>> +
>>> +     return 0;
>>> +}

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v2 3/5] pci: altera: Add Altera PCIe MSI driver

[Marc Zyngier]

On 03/08/15 11:37, Ley Foon Tan wrote:
> On Fri, Jul 31, 2015 at 8:12 PM, Marc Zyngier <marc.zyngier@arm.com> wrote:
>> On 31/07/15 11:15, Ley Foon Tan wrote:
>>> This patch adds Altera PCIe MSI driver. This soft IP supports configurable
>>> number of vectors, which is a dts parameter.
>>
>> I've reviewed the initial drop of this code; basic courtesy would be to
>> keep me CCed on the follow-up series.
> Will keep you in CC for the following revision. Sorry about this.
> 
>>>
>>> Signed-off-by: Ley Foon Tan <lftan@altera.com>
>>> ---
>>>  drivers/pci/host/Kconfig           |   7 +
>>>  drivers/pci/host/Makefile          |   1 +
>>>  drivers/pci/host/pcie-altera-msi.c | 324 +++++++++++++++++++++++++++++++++++++
>>>  3 files changed, 332 insertions(+)
>>>  create mode 100644 drivers/pci/host/pcie-altera-msi.c
>>>

[...]

>>> +
>>> +     msg->address_lo = lower_32_bits(addr);
>>> +     msg->address_hi = upper_32_bits(addr);
>>> +     msg->data = data->hwirq;
>>> +
>>> +     mask = msi_readl(msi, MSI_INTMASK);
>>> +     mask |= 1 << data->hwirq;
>>> +     msi_writel(msi, mask, MSI_INTMASK);
>>
>> It feels a bit weird to unmask the interrupt when you compose the
>> message. I'd expect this to be done in the mask/unmask methods.
> Do you refer to these 2 callbacks?
> .irq_mask = pci_msi_mask_irq,
> .irq_unmask = pci_msi_unmask_irq,
> 
> How about we move this INTMASK code above to altera_irq_domain_alloc()?
> We have unmask code in altera_irq_domain_free() now.

Either you move the mask/unmask code to local irq_mask/irq_unmask
methods (and call pci_msi_(un)mask_irq from there), or you move it
entierely to alloc/free.

Some level of symmetry helps following what is going on in the code.

[...]

>>> +static int altera_irq_domain_alloc(struct irq_domain *domain, unsigned int virq,
>>> +                               unsigned int nr_irqs, void *args)
>>> +{
>>> +     struct altera_msi *msi = domain->host_data;
>>> +     int bit;
>>> +
>>> +     mutex_lock(&msi->lock);
>>> +
>>> +     bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
>>> +     if (bit < msi->num_of_vectors)
>>> +             set_bit(bit, msi->used);
>>> +
>>> +     mutex_unlock(&msi->lock);
>>> +
>>> +     if (bit < 0)
>>> +             return bit;
>>
>> How can "bit" be negative here? find_first_zero_bit returns an unsigned
>> long... You probably want to change the type of "bit" to reflect that.
> Okay, will change to "unsigned long" type.
>>
>>> +     else if (bit >= msi->num_of_vectors)
>>
>> Useless "else" anyway.
> 
> I think we still need to check for failing case, if we don't have
> available unused bit.
> This could be rewritten  as below:
> 
> bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
> if (bit < msi->num_of_vectors)
>       set_bit(bit, msi->used);
> else
>       return -ENOSPC;

The more logical to write this is:

	if (bit >= msi->num_of_vectors)
		return -ENOSPC;

	set_bit(bit, msi->used);

which gets rid of the error case early and streamlines the normal case.

> 
>>
>>> +             return -ENOSPC;
>>> +
>>> +     irq_domain_set_info(domain, virq, bit, &altera_msi_bottom_irq_chip,
>>> +                         domain->host_data, handle_simple_irq,
>>> +                         NULL, NULL);
>>> +     set_irq_flags(virq, IRQF_VALID);
>>> +
>>> +     return 0;
>>> +}

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...

[PATCH v2 3/5] pci: altera: Add Altera PCIe MSI driver

[Marc Zyngier]

On 03/08/15 11:37, Ley Foon Tan wrote:
> On Fri, Jul 31, 2015 at 8:12 PM, Marc Zyngier <marc.zyngier@arm.com> wrote:
>> On 31/07/15 11:15, Ley Foon Tan wrote:
>>> This patch adds Altera PCIe MSI driver. This soft IP supports configurable
>>> number of vectors, which is a dts parameter.
>>
>> I've reviewed the initial drop of this code; basic courtesy would be to
>> keep me CCed on the follow-up series.
> Will keep you in CC for the following revision. Sorry about this.
> 
>>>
>>> Signed-off-by: Ley Foon Tan <lftan@altera.com>
>>> ---
>>>  drivers/pci/host/Kconfig           |   7 +
>>>  drivers/pci/host/Makefile          |   1 +
>>>  drivers/pci/host/pcie-altera-msi.c | 324 +++++++++++++++++++++++++++++++++++++
>>>  3 files changed, 332 insertions(+)
>>>  create mode 100644 drivers/pci/host/pcie-altera-msi.c
>>>

[...]

>>> +
>>> +     msg->address_lo = lower_32_bits(addr);
>>> +     msg->address_hi = upper_32_bits(addr);
>>> +     msg->data = data->hwirq;
>>> +
>>> +     mask = msi_readl(msi, MSI_INTMASK);
>>> +     mask |= 1 << data->hwirq;
>>> +     msi_writel(msi, mask, MSI_INTMASK);
>>
>> It feels a bit weird to unmask the interrupt when you compose the
>> message. I'd expect this to be done in the mask/unmask methods.
> Do you refer to these 2 callbacks?
> .irq_mask = pci_msi_mask_irq,
> .irq_unmask = pci_msi_unmask_irq,
> 
> How about we move this INTMASK code above to altera_irq_domain_alloc()?
> We have unmask code in altera_irq_domain_free() now.

Either you move the mask/unmask code to local irq_mask/irq_unmask
methods (and call pci_msi_(un)mask_irq from there), or you move it
entierely to alloc/free.

Some level of symmetry helps following what is going on in the code.

[...]

>>> +static int altera_irq_domain_alloc(struct irq_domain *domain, unsigned int virq,
>>> +                               unsigned int nr_irqs, void *args)
>>> +{
>>> +     struct altera_msi *msi = domain->host_data;
>>> +     int bit;
>>> +
>>> +     mutex_lock(&msi->lock);
>>> +
>>> +     bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
>>> +     if (bit < msi->num_of_vectors)
>>> +             set_bit(bit, msi->used);
>>> +
>>> +     mutex_unlock(&msi->lock);
>>> +
>>> +     if (bit < 0)
>>> +             return bit;
>>
>> How can "bit" be negative here? find_first_zero_bit returns an unsigned
>> long... You probably want to change the type of "bit" to reflect that.
> Okay, will change to "unsigned long" type.
>>
>>> +     else if (bit >= msi->num_of_vectors)
>>
>> Useless "else" anyway.
> 
> I think we still need to check for failing case, if we don't have
> available unused bit.
> This could be rewritten  as below:
> 
> bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
> if (bit < msi->num_of_vectors)
>       set_bit(bit, msi->used);
> else
>       return -ENOSPC;

The more logical to write this is:

	if (bit >= msi->num_of_vectors)
		return -ENOSPC;

	set_bit(bit, msi->used);

which gets rid of the error case early and streamlines the normal case.

> 
>>
>>> +             return -ENOSPC;
>>> +
>>> +     irq_domain_set_info(domain, virq, bit, &altera_msi_bottom_irq_chip,
>>> +                         domain->host_data, handle_simple_irq,
>>> +                         NULL, NULL);
>>> +     set_irq_flags(virq, IRQF_VALID);
>>> +
>>> +     return 0;
>>> +}

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...

Re: [PATCH v2 3/5] pci: altera: Add Altera PCIe MSI driver

[Ley Foon Tan]

On Mon, Aug 3, 2015 at 6:53 PM, Marc Zyngier <marc.zyngier@arm.com> wrote:
> On 03/08/15 11:37, Ley Foon Tan wrote:

>>>> +     msg->address_lo = lower_32_bits(addr);
>>>> +     msg->address_hi = upper_32_bits(addr);
>>>> +     msg->data = data->hwirq;
>>>> +
>>>> +     mask = msi_readl(msi, MSI_INTMASK);
>>>> +     mask |= 1 << data->hwirq;
>>>> +     msi_writel(msi, mask, MSI_INTMASK);
>>>
>>> It feels a bit weird to unmask the interrupt when you compose the
>>> message. I'd expect this to be done in the mask/unmask methods.
>> Do you refer to these 2 callbacks?
>> .irq_mask = pci_msi_mask_irq,
>> .irq_unmask = pci_msi_unmask_irq,
>>
>> How about we move this INTMASK code above to altera_irq_domain_alloc()?
>> We have unmask code in altera_irq_domain_free() now.
>
> Either you move the mask/unmask code to local irq_mask/irq_unmask
> methods (and call pci_msi_(un)mask_irq from there), or you move it
> entierely to alloc/free.
>
> Some level of symmetry helps following what is going on in the code.
Okay, will move it to alloc/free.
>
> [...]
>
>>>> +static int altera_irq_domain_alloc(struct irq_domain *domain, unsigned int virq,
>>>> +                               unsigned int nr_irqs, void *args)
>>>> +{
>>>> +     struct altera_msi *msi = domain->host_data;
>>>> +     int bit;
>>>> +
>>>> +     mutex_lock(&msi->lock);
>>>> +
>>>> +     bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
>>>> +     if (bit < msi->num_of_vectors)
>>>> +             set_bit(bit, msi->used);
>>>> +
>>>> +     mutex_unlock(&msi->lock);
>>>> +
>>>> +     if (bit < 0)
>>>> +             return bit;
>>>
>>> How can "bit" be negative here? find_first_zero_bit returns an unsigned
>>> long... You probably want to change the type of "bit" to reflect that.
>> Okay, will change to "unsigned long" type.
>>>
>>>> +     else if (bit >= msi->num_of_vectors)
>>>
>>> Useless "else" anyway.
>>
>> I think we still need to check for failing case, if we don't have
>> available unused bit.
>> This could be rewritten  as below:
>>
>> bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
>> if (bit < msi->num_of_vectors)
>>       set_bit(bit, msi->used);
>> else
>>       return -ENOSPC;
>
> The more logical to write this is:
>
>         if (bit >= msi->num_of_vectors)
>                 return -ENOSPC;
>
>         set_bit(bit, msi->used);
>
> which gets rid of the error case early and streamlines the normal case.
Yes, will change to this way.

Thanks.

Regards
Ley Foon

Re: [PATCH v2 3/5] pci: altera: Add Altera PCIe MSI driver

[Ley Foon Tan]

On Mon, Aug 3, 2015 at 6:53 PM, Marc Zyngier <marc.zyngier-5wv7dgnIgG8@public.gmane.org> wrote:
> On 03/08/15 11:37, Ley Foon Tan wrote:

>>>> +     msg->address_lo = lower_32_bits(addr);
>>>> +     msg->address_hi = upper_32_bits(addr);
>>>> +     msg->data = data->hwirq;
>>>> +
>>>> +     mask = msi_readl(msi, MSI_INTMASK);
>>>> +     mask |= 1 << data->hwirq;
>>>> +     msi_writel(msi, mask, MSI_INTMASK);
>>>
>>> It feels a bit weird to unmask the interrupt when you compose the
>>> message. I'd expect this to be done in the mask/unmask methods.
>> Do you refer to these 2 callbacks?
>> .irq_mask = pci_msi_mask_irq,
>> .irq_unmask = pci_msi_unmask_irq,
>>
>> How about we move this INTMASK code above to altera_irq_domain_alloc()?
>> We have unmask code in altera_irq_domain_free() now.
>
> Either you move the mask/unmask code to local irq_mask/irq_unmask
> methods (and call pci_msi_(un)mask_irq from there), or you move it
> entierely to alloc/free.
>
> Some level of symmetry helps following what is going on in the code.
Okay, will move it to alloc/free.
>
> [...]
>
>>>> +static int altera_irq_domain_alloc(struct irq_domain *domain, unsigned int virq,
>>>> +                               unsigned int nr_irqs, void *args)
>>>> +{
>>>> +     struct altera_msi *msi = domain->host_data;
>>>> +     int bit;
>>>> +
>>>> +     mutex_lock(&msi->lock);
>>>> +
>>>> +     bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
>>>> +     if (bit < msi->num_of_vectors)
>>>> +             set_bit(bit, msi->used);
>>>> +
>>>> +     mutex_unlock(&msi->lock);
>>>> +
>>>> +     if (bit < 0)
>>>> +             return bit;
>>>
>>> How can "bit" be negative here? find_first_zero_bit returns an unsigned
>>> long... You probably want to change the type of "bit" to reflect that.
>> Okay, will change to "unsigned long" type.
>>>
>>>> +     else if (bit >= msi->num_of_vectors)
>>>
>>> Useless "else" anyway.
>>
>> I think we still need to check for failing case, if we don't have
>> available unused bit.
>> This could be rewritten  as below:
>>
>> bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
>> if (bit < msi->num_of_vectors)
>>       set_bit(bit, msi->used);
>> else
>>       return -ENOSPC;
>
> The more logical to write this is:
>
>         if (bit >= msi->num_of_vectors)
>                 return -ENOSPC;
>
>         set_bit(bit, msi->used);
>
> which gets rid of the error case early and streamlines the normal case.
Yes, will change to this way.

Thanks.

Regards
Ley Foon
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v2 3/5] pci: altera: Add Altera PCIe MSI driver

[Ley Foon Tan]

On Mon, Aug 3, 2015 at 6:53 PM, Marc Zyngier <marc.zyngier@arm.com> wrote:
> On 03/08/15 11:37, Ley Foon Tan wrote:

>>>> +     msg->address_lo = lower_32_bits(addr);
>>>> +     msg->address_hi = upper_32_bits(addr);
>>>> +     msg->data = data->hwirq;
>>>> +
>>>> +     mask = msi_readl(msi, MSI_INTMASK);
>>>> +     mask |= 1 << data->hwirq;
>>>> +     msi_writel(msi, mask, MSI_INTMASK);
>>>
>>> It feels a bit weird to unmask the interrupt when you compose the
>>> message. I'd expect this to be done in the mask/unmask methods.
>> Do you refer to these 2 callbacks?
>> .irq_mask = pci_msi_mask_irq,
>> .irq_unmask = pci_msi_unmask_irq,
>>
>> How about we move this INTMASK code above to altera_irq_domain_alloc()?
>> We have unmask code in altera_irq_domain_free() now.
>
> Either you move the mask/unmask code to local irq_mask/irq_unmask
> methods (and call pci_msi_(un)mask_irq from there), or you move it
> entierely to alloc/free.
>
> Some level of symmetry helps following what is going on in the code.
Okay, will move it to alloc/free.
>
> [...]
>
>>>> +static int altera_irq_domain_alloc(struct irq_domain *domain, unsigned int virq,
>>>> +                               unsigned int nr_irqs, void *args)
>>>> +{
>>>> +     struct altera_msi *msi = domain->host_data;
>>>> +     int bit;
>>>> +
>>>> +     mutex_lock(&msi->lock);
>>>> +
>>>> +     bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
>>>> +     if (bit < msi->num_of_vectors)
>>>> +             set_bit(bit, msi->used);
>>>> +
>>>> +     mutex_unlock(&msi->lock);
>>>> +
>>>> +     if (bit < 0)
>>>> +             return bit;
>>>
>>> How can "bit" be negative here? find_first_zero_bit returns an unsigned
>>> long... You probably want to change the type of "bit" to reflect that.
>> Okay, will change to "unsigned long" type.
>>>
>>>> +     else if (bit >= msi->num_of_vectors)
>>>
>>> Useless "else" anyway.
>>
>> I think we still need to check for failing case, if we don't have
>> available unused bit.
>> This could be rewritten  as below:
>>
>> bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
>> if (bit < msi->num_of_vectors)
>>       set_bit(bit, msi->used);
>> else
>>       return -ENOSPC;
>
> The more logical to write this is:
>
>         if (bit >= msi->num_of_vectors)
>                 return -ENOSPC;
>
>         set_bit(bit, msi->used);
>
> which gets rid of the error case early and streamlines the normal case.
Yes, will change to this way.

Thanks.

Regards
Ley Foon

[PATCH v2 3/5] pci: altera: Add Altera PCIe MSI driver

[Ley Foon Tan]

On Mon, Aug 3, 2015 at 6:53 PM, Marc Zyngier <marc.zyngier@arm.com> wrote:
> On 03/08/15 11:37, Ley Foon Tan wrote:

>>>> +     msg->address_lo = lower_32_bits(addr);
>>>> +     msg->address_hi = upper_32_bits(addr);
>>>> +     msg->data = data->hwirq;
>>>> +
>>>> +     mask = msi_readl(msi, MSI_INTMASK);
>>>> +     mask |= 1 << data->hwirq;
>>>> +     msi_writel(msi, mask, MSI_INTMASK);
>>>
>>> It feels a bit weird to unmask the interrupt when you compose the
>>> message. I'd expect this to be done in the mask/unmask methods.
>> Do you refer to these 2 callbacks?
>> .irq_mask = pci_msi_mask_irq,
>> .irq_unmask = pci_msi_unmask_irq,
>>
>> How about we move this INTMASK code above to altera_irq_domain_alloc()?
>> We have unmask code in altera_irq_domain_free() now.
>
> Either you move the mask/unmask code to local irq_mask/irq_unmask
> methods (and call pci_msi_(un)mask_irq from there), or you move it
> entierely to alloc/free.
>
> Some level of symmetry helps following what is going on in the code.
Okay, will move it to alloc/free.
>
> [...]
>
>>>> +static int altera_irq_domain_alloc(struct irq_domain *domain, unsigned int virq,
>>>> +                               unsigned int nr_irqs, void *args)
>>>> +{
>>>> +     struct altera_msi *msi = domain->host_data;
>>>> +     int bit;
>>>> +
>>>> +     mutex_lock(&msi->lock);
>>>> +
>>>> +     bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
>>>> +     if (bit < msi->num_of_vectors)
>>>> +             set_bit(bit, msi->used);
>>>> +
>>>> +     mutex_unlock(&msi->lock);
>>>> +
>>>> +     if (bit < 0)
>>>> +             return bit;
>>>
>>> How can "bit" be negative here? find_first_zero_bit returns an unsigned
>>> long... You probably want to change the type of "bit" to reflect that.
>> Okay, will change to "unsigned long" type.
>>>
>>>> +     else if (bit >= msi->num_of_vectors)
>>>
>>> Useless "else" anyway.
>>
>> I think we still need to check for failing case, if we don't have
>> available unused bit.
>> This could be rewritten  as below:
>>
>> bit = find_first_zero_bit(msi->used, msi->num_of_vectors);
>> if (bit < msi->num_of_vectors)
>>       set_bit(bit, msi->used);
>> else
>>       return -ENOSPC;
>
> The more logical to write this is:
>
>         if (bit >= msi->num_of_vectors)
>                 return -ENOSPC;
>
>         set_bit(bit, msi->used);
>
> which gets rid of the error case early and streamlines the normal case.
Yes, will change to this way.

Thanks.

Regards
Ley Foon