[Patch v7 00/18] Enable support of IOAPIC hotplug on x86 platforms

[Patch v7 00/18] Enable support of IOAPIC hotplug on x86 platforms

[Jiang Liu]

This patch set enhances IOAPIC core and ACPI drivers to support IOAPIC
hotplug on x86 platforms. It's based on v3.18-rc2 at
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git

You may pull it from 
https://github.com/jiangliu/linux.git ioapic/hotplug_v7

We have pick up several patches from Yinghai's original IOAPIC hotplug
patch set and reimplemented IOAPIC driver as an ACPI driver instead of
a PCI driver.

It has been tested on a 4-socket Intel SDV with socket hot-addition
capability. Any suggestions are welcomed!

Patch 1-3 are bugfixes against v3.17 and should target v3.18. Patch 1
has been merged into tip/x86/urgent.
Patch 4-7 are bugfixes and enhancements to ACPI subsystem
Patch 8 killes PCI IOAPIC driver
Patch 9-17 enhances IOAPIC core to support IOAPIC hotplug
Patch 18 reimplements ACPI IOAPIC driver and enables IOAPIC hotplug

V6->V7:
1) Rebase to v3.18-rc2
2) Three bugfixes for v3.18
V5->V6:
1) Rebase to the latest v3.17-16
2) Minor fixes for comments
V4->V5:
1) Fix a building error
2) Don't rename processor_core.c as apic_id.c
3) Improve comments
4) Reorder patches to get rid of some temporary code
V3->V4:
1) Fix a bug in manage IOAPIC reference count
2) Rebase to v3.17-rc2
3) Refine commit messages
V2->V3:
1) Refine ACPI resource walk functions for PCI root bus and IOAPIC
2) Improve commit messages
3) Reorder patch order for better maintenence

Jiang Liu (15):
  ACPI, irq: fix regression casued by 6b9fb7082409
  x86, intel-mid: Create IRQs for APB timers and RTC timers
  ACPI, irq, x86: Return IRQ instead of GSI in mp_register_gsi()
  x86, PCI, ACPI: Kill private function resource_to_addr() in
    arch/x86/pci/acpi.c
  ACPI: Correct return value of acpi_dev_resource_address_space()
  ACPI: Fix minor syntax issues in processor_core.c
  PCI: Remove PCI ioapic driver
  x86, irq: Remove __init marker for functions will be used by IOAPIC
    hotplug
  x86, irq: Keep balance of IOAPIC pin reference count
  x86, irq: Refine mp_register_ioapic() to prepare for IOAPIC hotplug
  x86, irq, ACPI: Introduce a rwsem to protect IOAPIC operations from
    hotplug
  x86, irq, ACPI: Implement interface to support ACPI based IOAPIC
    hot-addition
  x86, irq, ACPI: Implement interfaces to support ACPI based IOAPIC
    hot-removal
  x86, irq: Introduce helper to check whether an IOAPIC has been
    registered
  x86, irq, ACPI: Implement ACPI driver to support IOAPIC hotplug

Yinghai Lu (3):
  ACPI: Add interfaces to parse IOAPIC ID for IOAPIC hotplug
  x86, irq: Split out alloc_ioapic_save_registers()
  x86, irq: Prefer assigned ID in APIC ID register for x86_64

 arch/x86/include/asm/io_apic.h    |    6 +-
 arch/x86/kernel/acpi/boot.c       |  102 ++++++++++++++--
 arch/x86/kernel/apb_timer.c       |    2 -
 arch/x86/kernel/apic/io_apic.c    |  234 +++++++++++++++++++++++++++++-------
 arch/x86/pci/acpi.c               |  144 +++++++++--------------
 arch/x86/pci/intel_mid_pci.c      |   10 +-
 arch/x86/pci/irq.c                |    7 +-
 arch/x86/platform/intel-mid/sfi.c |    2 +
 drivers/acpi/Kconfig              |    6 +
 drivers/acpi/Makefile             |    1 +
 drivers/acpi/internal.h           |    7 ++
 drivers/acpi/ioapic.c             |  235 +++++++++++++++++++++++++++++++++++++
 drivers/acpi/pci_irq.c            |   11 +-
 drivers/acpi/pci_root.c           |    3 +
 drivers/acpi/processor_core.c     |  131 ++++++++++++++++++---
 drivers/acpi/resource.c           |    2 +-
 drivers/pci/Kconfig               |    7 --
 drivers/pci/Makefile              |    2 -
 drivers/pci/ioapic.c              |  121 -------------------
 include/linux/acpi.h              |    5 +
 include/linux/pci.h               |    1 +
 21 files changed, 739 insertions(+), 300 deletions(-)
 create mode 100644 drivers/acpi/ioapic.c
 delete mode 100644 drivers/pci/ioapic.c

-- 
1.7.10.4

[Patch v7 01/18] ACPI, irq: fix regression casued by 6b9fb7082409

[Jiang Liu]

When IOAPIC is disabled, acpi_gsi_to_irq() should return gsi directly
instead of calling mp_map_gsi_to_irq() to translate gsi to IRQ by IOAPIC.
It fixes https://bugzilla.kernel.org/show_bug.cgi?id=84381.

Reported-by: Thomas Richter <thor@math.tu-berlin.de>
Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
Cc: rui.zhang@intel.com
Cc: <stable@vger.kernel.org> # 3.17
---
 arch/x86/kernel/acpi/boot.c |   13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)

diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c
index b436fc735aa4..eceba9d9e116 100644
--- a/arch/x86/kernel/acpi/boot.c
+++ b/arch/x86/kernel/acpi/boot.c
@@ -604,14 +604,19 @@ void __init acpi_pic_sci_set_trigger(unsigned int irq, u16 trigger)
 
 int acpi_gsi_to_irq(u32 gsi, unsigned int *irqp)
 {
-	int irq = mp_map_gsi_to_irq(gsi, IOAPIC_MAP_ALLOC | IOAPIC_MAP_CHECK);
+	int irq;
 
-	if (irq >= 0) {
+	if (acpi_irq_model == ACPI_IRQ_MODEL_PIC) {
+		*irqp = gsi;
+	} else {
+		irq = mp_map_gsi_to_irq(gsi,
+					IOAPIC_MAP_ALLOC | IOAPIC_MAP_CHECK);
+		if (irq < 0)
+			return -1;
 		*irqp = irq;
-		return 0;
 	}
 
-	return -1;
+	return 0;
 }
 EXPORT_SYMBOL_GPL(acpi_gsi_to_irq);
 
-- 
1.7.10.4

[Patch v7 02/18] x86, intel-mid: Create IRQs for APB timers and RTC timers

[Jiang Liu]

Intel MID platforms has no legacy interrupts, so no IRQ descriptors
preallocated. We need to call mp_map_gsi_to_irq() to create IRQ
descriptors for APB timers and RTC timers, otherwise it may cause
invalid memory access as:
[    0.116839] BUG: unable to handle kernel NULL pointer dereference at
0000003a
[    0.123803] IP: [<c1071c0e>] setup_irq+0xf/0x4d

Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
Tested-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: <stable@vger.kernel.org> # 3.17
---
 arch/x86/kernel/apb_timer.c       |    2 --
 arch/x86/platform/intel-mid/sfi.c |    2 ++
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/arch/x86/kernel/apb_timer.c b/arch/x86/kernel/apb_timer.c
index 5972b108f15a..b708738d016e 100644
--- a/arch/x86/kernel/apb_timer.c
+++ b/arch/x86/kernel/apb_timer.c
@@ -185,8 +185,6 @@ static void apbt_setup_irq(struct apbt_dev *adev)
 
 	irq_modify_status(adev->irq, 0, IRQ_MOVE_PCNTXT);
 	irq_set_affinity(adev->irq, cpumask_of(adev->cpu));
-	/* APB timer irqs are set up as mp_irqs, timer is edge type */
-	__irq_set_handler(adev->irq, handle_edge_irq, 0, "edge");
 }
 
 /* Should be called with per cpu */
diff --git a/arch/x86/platform/intel-mid/sfi.c b/arch/x86/platform/intel-mid/sfi.c
index 3c53a90fdb18..c14ad34776c4 100644
--- a/arch/x86/platform/intel-mid/sfi.c
+++ b/arch/x86/platform/intel-mid/sfi.c
@@ -106,6 +106,7 @@ int __init sfi_parse_mtmr(struct sfi_table_header *table)
 			mp_irq.dstapic = MP_APIC_ALL;
 			mp_irq.dstirq = pentry->irq;
 			mp_save_irq(&mp_irq);
+			mp_map_gsi_to_irq(pentry->irq, IOAPIC_MAP_ALLOC);
 	}
 
 	return 0;
@@ -176,6 +177,7 @@ int __init sfi_parse_mrtc(struct sfi_table_header *table)
 		mp_irq.dstapic = MP_APIC_ALL;
 		mp_irq.dstirq = pentry->irq;
 		mp_save_irq(&mp_irq);
+		mp_map_gsi_to_irq(pentry->irq, IOAPIC_MAP_ALLOC);
 	}
 	return 0;
 }
-- 
1.7.10.4

[Patch v7 03/18] ACPI, irq, x86: Return IRQ instead of GSI in mp_register_gsi()

[Jiang Liu]

The GSI for ACPI SCI may be shared with other devices. For example,

Function mp_register_gsi() should return IRQ number, so fix a regression
by returning mp_map_gsi_to_irq(gsi, 0) instead of gsi.

The regression was caused by commit 84245af7297ced9e8fe "x86, irq, ACPI:
Change __acpi_register_gsi to return IRQ number instead of GSI" and
exposed on a SuperMicro system, which shares one GSI between ACPI SCI
and PCI device, with following failure:

http://sourceforge.net/p/linux1394/mailman/linux1394-user/?viewmonth=201410
[    0.000000] ACPI: INT_SRC_OVR (bus 0 bus_irq 9 global_irq 20 low
level)
[    2.699224] firewire_ohci 0000:06:00.0: failed to allocate interrupt
20

Reported-and-Tested-by: Daniel Robbins <drobbins@funtoo.org>
Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
Cc: <stable@vger.kernel.org> # 3.17
---
 arch/x86/kernel/acpi/boot.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c
index eceba9d9e116..e077c080a519 100644
--- a/arch/x86/kernel/acpi/boot.c
+++ b/arch/x86/kernel/acpi/boot.c
@@ -397,7 +397,7 @@ static int mp_register_gsi(struct device *dev, u32 gsi, int trigger,
 
 	/* Don't set up the ACPI SCI because it's already set up */
 	if (acpi_gbl_FADT.sci_interrupt == gsi)
-		return gsi;
+		return mp_map_gsi_to_irq(gsi, IOAPIC_MAP_ALLOC);
 
 	trigger = trigger == ACPI_EDGE_SENSITIVE ? 0 : 1;
 	polarity = polarity == ACPI_ACTIVE_HIGH ? 0 : 1;
-- 
1.7.10.4

[Patch v7 04/18] x86, PCI, ACPI: Kill private function resource_to_addr() in arch/x86/pci/acpi.c

[Jiang Liu]

Private function resource_to_addr() is used to parse ACPI resources
for PCI host bridge. There are public interfaces available for that
purpose, so replace resource_to_addr() with public interfaces.

Reviewed-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
---
 arch/x86/pci/acpi.c |  144 +++++++++++++++++++--------------------------------
 1 file changed, 53 insertions(+), 91 deletions(-)

diff --git a/arch/x86/pci/acpi.c b/arch/x86/pci/acpi.c
index cfd1b132b8e3..3f72d934dc16 100644
--- a/arch/x86/pci/acpi.c
+++ b/arch/x86/pci/acpi.c
@@ -218,114 +218,76 @@ static void teardown_mcfg_map(struct pci_root_info *info)
 }
 #endif
 
-static acpi_status resource_to_addr(struct acpi_resource *resource,
-				    struct acpi_resource_address64 *addr)
-{
-	acpi_status status;
-	struct acpi_resource_memory24 *memory24;
-	struct acpi_resource_memory32 *memory32;
-	struct acpi_resource_fixed_memory32 *fixed_memory32;
-
-	memset(addr, 0, sizeof(*addr));
-	switch (resource->type) {
-	case ACPI_RESOURCE_TYPE_MEMORY24:
-		memory24 = &resource->data.memory24;
-		addr->resource_type = ACPI_MEMORY_RANGE;
-		addr->minimum = memory24->minimum;
-		addr->address_length = memory24->address_length;
-		addr->maximum = addr->minimum + addr->address_length - 1;
-		return AE_OK;
-	case ACPI_RESOURCE_TYPE_MEMORY32:
-		memory32 = &resource->data.memory32;
-		addr->resource_type = ACPI_MEMORY_RANGE;
-		addr->minimum = memory32->minimum;
-		addr->address_length = memory32->address_length;
-		addr->maximum = addr->minimum + addr->address_length - 1;
-		return AE_OK;
-	case ACPI_RESOURCE_TYPE_FIXED_MEMORY32:
-		fixed_memory32 = &resource->data.fixed_memory32;
-		addr->resource_type = ACPI_MEMORY_RANGE;
-		addr->minimum = fixed_memory32->address;
-		addr->address_length = fixed_memory32->address_length;
-		addr->maximum = addr->minimum + addr->address_length - 1;
-		return AE_OK;
-	case ACPI_RESOURCE_TYPE_ADDRESS16:
-	case ACPI_RESOURCE_TYPE_ADDRESS32:
-	case ACPI_RESOURCE_TYPE_ADDRESS64:
-		status = acpi_resource_to_address64(resource, addr);
-		if (ACPI_SUCCESS(status) &&
-		    (addr->resource_type == ACPI_MEMORY_RANGE ||
-		    addr->resource_type == ACPI_IO_RANGE) &&
-		    addr->address_length > 0) {
-			return AE_OK;
-		}
-		break;
-	}
-	return AE_ERROR;
-}
-
 static acpi_status count_resource(struct acpi_resource *acpi_res, void *data)
 {
 	struct pci_root_info *info = data;
-	struct acpi_resource_address64 addr;
-	acpi_status status;
+	struct resource r = {
+		.flags = 0
+	};
 
-	status = resource_to_addr(acpi_res, &addr);
-	if (ACPI_SUCCESS(status))
+	if (!acpi_dev_resource_memory(acpi_res, &r) &&
+	    !acpi_dev_resource_address_space(acpi_res, &r))
+		return AE_OK;
+
+	if ((r.flags & (IORESOURCE_IO | IORESOURCE_MEM)) && resource_size(&r))
 		info->res_num++;
+
 	return AE_OK;
 }
 
 static acpi_status setup_resource(struct acpi_resource *acpi_res, void *data)
 {
 	struct pci_root_info *info = data;
-	struct resource *res;
-	struct acpi_resource_address64 addr;
-	acpi_status status;
-	unsigned long flags;
-	u64 start, orig_end, end;
+	u64 translation_offset = 0;
+	struct resource r = {
+		.flags = 0
+	};
+
+	if (acpi_dev_resource_memory(acpi_res, &r)) {
+		r.flags &= IORESOURCE_MEM | IORESOURCE_IO;
+	} else if (acpi_dev_resource_address_space(acpi_res, &r)) {
+		u64 orig_end;
+		struct acpi_resource_address64 addr;
+
+		r.flags &= IORESOURCE_MEM | IORESOURCE_IO;
+		if (r.flags == 0)
+			return AE_OK;
 
-	status = resource_to_addr(acpi_res, &addr);
-	if (!ACPI_SUCCESS(status))
-		return AE_OK;
+		if (ACPI_FAILURE(acpi_resource_to_address64(acpi_res, &addr)))
+			return AE_OK;
 
-	if (addr.resource_type == ACPI_MEMORY_RANGE) {
-		flags = IORESOURCE_MEM;
-		if (addr.info.mem.caching == ACPI_PREFETCHABLE_MEMORY)
-			flags |= IORESOURCE_PREFETCH;
-	} else if (addr.resource_type == ACPI_IO_RANGE) {
-		flags = IORESOURCE_IO;
-	} else
-		return AE_OK;
+		if (addr.resource_type == ACPI_MEMORY_RANGE &&
+		    addr.info.mem.caching == ACPI_PREFETCHABLE_MEMORY)
+			r.flags |= IORESOURCE_PREFETCH;
 
-	start = addr.minimum + addr.translation_offset;
-	orig_end = end = addr.maximum + addr.translation_offset;
+		translation_offset = addr.translation_offset;
+		orig_end = r.end;
+		r.start += translation_offset;
+		r.end += translation_offset;
 
-	/* Exclude non-addressable range or non-addressable portion of range */
-	end = min(end, (u64)iomem_resource.end);
-	if (end <= start) {
-		dev_info(&info->bridge->dev,
-			"host bridge window [%#llx-%#llx] "
-			"(ignored, not CPU addressable)\n", start, orig_end);
-		return AE_OK;
-	} else if (orig_end != end) {
-		dev_info(&info->bridge->dev,
-			"host bridge window [%#llx-%#llx] "
-			"([%#llx-%#llx] ignored, not CPU addressable)\n", 
-			start, orig_end, end + 1, orig_end);
+		/* Exclude non-addressable range or non-addressable portion of range */
+		r.end = min(r.end, iomem_resource.end);
+		if (r.end <= r.start) {
+			dev_info(&info->bridge->dev,
+				"host bridge window [%#llx-%#llx] (ignored, not CPU addressable)\n",
+				 r.start, orig_end);
+			return AE_OK;
+		} else if (orig_end != r.end) {
+			dev_info(&info->bridge->dev,
+				"host bridge window [%#llx-%#llx] ([%#llx-%#llx] ignored, not CPU addressable)\n",
+				r.start, orig_end, r.end + 1, orig_end);
+		}
 	}
 
-	res = &info->res[info->res_num];
-	res->name = info->name;
-	res->flags = flags;
-	res->start = start;
-	res->end = end;
-	info->res_offset[info->res_num] = addr.translation_offset;
-	info->res_num++;
-
-	if (!pci_use_crs)
-		dev_printk(KERN_DEBUG, &info->bridge->dev,
-			   "host bridge window %pR (ignored)\n", res);
+	if (r.flags && resource_size(&r)) {
+		r.name = info->name;
+		info->res[info->res_num] = r;
+		info->res_offset[info->res_num] = translation_offset;
+		info->res_num++;
+		if (!pci_use_crs)
+			dev_printk(KERN_DEBUG, &info->bridge->dev,
+				   "host bridge window %pR (ignored)\n", &r);
+	}
 
 	return AE_OK;
 }
-- 
1.7.10.4

[Patch v7 05/18] ACPI: Correct return value of acpi_dev_resource_address_space()

[Jiang Liu]

Change acpi_dev_resource_address_space() to return failure if the
acpi_resource structure can't be converted to an ACPI address64
structure, so caller could correctly detect failure.

Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
---
 drivers/acpi/resource.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/acpi/resource.c b/drivers/acpi/resource.c
index 2ba8f02ced36..782a0d15c25f 100644
--- a/drivers/acpi/resource.c
+++ b/drivers/acpi/resource.c
@@ -200,7 +200,7 @@ bool acpi_dev_resource_address_space(struct acpi_resource *ares,
 
 	status = acpi_resource_to_address64(ares, &addr);
 	if (ACPI_FAILURE(status))
-		return true;
+		return false;
 
 	res->start = addr.minimum;
 	res->end = addr.maximum;
-- 
1.7.10.4

[Patch v7 06/18] ACPI: Fix minor syntax issues in processor_core.c

[Jiang Liu]

Fix minor syntax issues in processor_core.c to follow coding styles.

Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
---
 drivers/acpi/processor_core.c |    9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

diff --git a/drivers/acpi/processor_core.c b/drivers/acpi/processor_core.c
index ef58f46c8442..342942f90a10 100644
--- a/drivers/acpi/processor_core.c
+++ b/drivers/acpi/processor_core.c
@@ -125,13 +125,12 @@ static int map_mat_entry(acpi_handle handle, int type, u32 acpi_id)
 	}
 
 	header = (struct acpi_subtable_header *)obj->buffer.pointer;
-	if (header->type == ACPI_MADT_TYPE_LOCAL_APIC) {
+	if (header->type == ACPI_MADT_TYPE_LOCAL_APIC)
 		map_lapic_id(header, acpi_id, &apic_id);
-	} else if (header->type == ACPI_MADT_TYPE_LOCAL_SAPIC) {
+	else if (header->type == ACPI_MADT_TYPE_LOCAL_SAPIC)
 		map_lsapic_id(header, type, acpi_id, &apic_id);
-	} else if (header->type == ACPI_MADT_TYPE_LOCAL_X2APIC) {
+	else if (header->type == ACPI_MADT_TYPE_LOCAL_X2APIC)
 		map_x2apic_id(header, type, acpi_id, &apic_id);
-	}
 
 exit:
 	kfree(buffer.pointer);
@@ -164,7 +163,7 @@ int acpi_map_cpuid(int apic_id, u32 acpi_id)
 		 * For example,
 		 *
 		 * Scope (_PR)
-                 * {
+		 * {
 		 *     Processor (CPU0, 0x00, 0x00000410, 0x06) {}
 		 *     Processor (CPU1, 0x01, 0x00000410, 0x06) {}
 		 *     Processor (CPU2, 0x02, 0x00000410, 0x06) {}
-- 
1.7.10.4

[Patch v7 07/18] ACPI: Add interfaces to parse IOAPIC ID for IOAPIC hotplug

[Jiang Liu]

From: Yinghai Lu <yinghai@kernel.org>

We need to parse APIC ID for IOAPIC registration for IOAPIC hotplug.
ACPI _MAT method and MADT table are used to figure out IOAPIC ID, just
like parsing CPU APIC ID for CPU hotplug.

Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
---
 drivers/acpi/processor_core.c |  122 ++++++++++++++++++++++++++++++++++++++---
 include/linux/acpi.h          |    4 ++
 2 files changed, 117 insertions(+), 9 deletions(-)

diff --git a/drivers/acpi/processor_core.c b/drivers/acpi/processor_core.c
index 342942f90a10..9ff12905aed9 100644
--- a/drivers/acpi/processor_core.c
+++ b/drivers/acpi/processor_core.c
@@ -4,6 +4,10 @@
  *
  *	Alex Chiang <achiang@hp.com>
  *	- Unified x86/ia64 implementations
+ *
+ * I/O APIC hotplug support
+ *	Yinghai Lu <yinghai@kernel.org>
+ *	Jiang Liu <jiang.liu@intel.com>
  */
 #include <linux/export.h>
 #include <linux/acpi.h>
@@ -12,6 +16,21 @@
 #define _COMPONENT		ACPI_PROCESSOR_COMPONENT
 ACPI_MODULE_NAME("processor_core");
 
+static struct acpi_table_madt *get_madt_table(void)
+{
+	static struct acpi_table_madt *madt;
+	static int read_madt;
+
+	if (!read_madt) {
+		if (ACPI_FAILURE(acpi_get_table(ACPI_SIG_MADT, 0,
+					(struct acpi_table_header **)&madt)))
+			madt = NULL;
+		read_madt++;
+	}
+
+	return madt;
+}
+
 static int map_lapic_id(struct acpi_subtable_header *entry,
 		 u32 acpi_id, int *apic_id)
 {
@@ -67,17 +86,10 @@ static int map_lsapic_id(struct acpi_subtable_header *entry,
 static int map_madt_entry(int type, u32 acpi_id)
 {
 	unsigned long madt_end, entry;
-	static struct acpi_table_madt *madt;
-	static int read_madt;
 	int apic_id = -1;
+	struct acpi_table_madt *madt;
 
-	if (!read_madt) {
-		if (ACPI_FAILURE(acpi_get_table(ACPI_SIG_MADT, 0,
-					(struct acpi_table_header **)&madt)))
-			madt = NULL;
-		read_madt++;
-	}
-
+	madt = get_madt_table();
 	if (!madt)
 		return apic_id;
 
@@ -203,3 +215,95 @@ int acpi_get_cpuid(acpi_handle handle, int type, u32 acpi_id)
 	return acpi_map_cpuid(apic_id, acpi_id);
 }
 EXPORT_SYMBOL_GPL(acpi_get_cpuid);
+
+#ifdef CONFIG_ACPI_HOTPLUG_IOAPIC
+static int get_ioapic_id(struct acpi_subtable_header *entry, u32 gsi_base,
+			 u64 *phys_addr, int *ioapic_id)
+{
+	struct acpi_madt_io_apic *ioapic = (struct acpi_madt_io_apic *)entry;
+
+	if (ioapic->global_irq_base != gsi_base)
+		return 0;
+
+	*phys_addr = ioapic->address;
+	*ioapic_id = ioapic->id;
+	return 1;
+}
+
+static int parse_madt_ioapic_entry(u32 gsi_base, u64 *phys_addr)
+{
+	struct acpi_subtable_header *hdr;
+	unsigned long madt_end, entry;
+	int apic_id = -1;
+	struct acpi_table_madt *madt;
+
+	madt = get_madt_table();
+	if (!madt)
+		return apic_id;
+
+	entry = (unsigned long)madt;
+	madt_end = entry + madt->header.length;
+
+	/* Parse all entries looking for a match. */
+	entry += sizeof(struct acpi_table_madt);
+	while (entry + sizeof(struct acpi_subtable_header) < madt_end) {
+		hdr = (struct acpi_subtable_header *)entry;
+		if (hdr->type == ACPI_MADT_TYPE_IO_APIC &&
+		    get_ioapic_id(hdr, gsi_base, phys_addr, &apic_id))
+			break;
+		else
+			entry += hdr->length;
+	}
+
+	return apic_id;
+}
+
+static int parse_mat_ioapic_entry(acpi_handle handle, u32 gsi_base,
+				  u64 *phys_addr)
+{
+	struct acpi_buffer buffer = { ACPI_ALLOCATE_BUFFER, NULL };
+	union acpi_object *obj;
+	struct acpi_subtable_header *header;
+	int apic_id = -1;
+
+	if (ACPI_FAILURE(acpi_evaluate_object(handle, "_MAT", NULL, &buffer)))
+		goto exit;
+
+	if (!buffer.length || !buffer.pointer)
+		goto exit;
+
+	obj = buffer.pointer;
+	if (obj->type != ACPI_TYPE_BUFFER ||
+	    obj->buffer.length < sizeof(struct acpi_subtable_header))
+		goto exit;
+
+	header = (struct acpi_subtable_header *)obj->buffer.pointer;
+	if (header->type == ACPI_MADT_TYPE_IO_APIC)
+		get_ioapic_id(header, gsi_base, phys_addr, &apic_id);
+
+exit:
+	kfree(buffer.pointer);
+	return apic_id;
+}
+
+/**
+ * acpi_get_ioapic_id - Get IOAPIC ID and physical address matching @gsi_base
+ * @handle:	ACPI object for IOAPIC device
+ * @gsi_base:	GSI base to match with
+ * @phys_add:	hold physical address of matching IOAPIC record
+ *
+ * Walk resources returned by ACPI_MAT method, then ACPI MADT table, to search
+ * for an ACPI IOAPIC record matching @gsi_base.
+ * Return IOAPIC id and physical address if found a match, otherwise return <0.
+ */
+int acpi_get_ioapic_id(acpi_handle handle, u32 gsi_base, u64 *phys_addr)
+{
+	int apic_id;
+
+	apic_id = parse_mat_ioapic_entry(handle, gsi_base, phys_addr);
+	if (apic_id == -1)
+		apic_id = parse_madt_ioapic_entry(gsi_base, phys_addr);
+
+	return apic_id;
+}
+#endif /* CONFIG_ACPI_HOTPLUG_IOAPIC */
diff --git a/include/linux/acpi.h b/include/linux/acpi.h
index 407a12f663eb..77cecd0350e7 100644
--- a/include/linux/acpi.h
+++ b/include/linux/acpi.h
@@ -146,6 +146,10 @@ int acpi_map_lsapic(acpi_handle handle, int physid, int *pcpu);
 int acpi_unmap_lsapic(int cpu);
 #endif /* CONFIG_ACPI_HOTPLUG_CPU */
 
+#ifdef CONFIG_ACPI_HOTPLUG_IOAPIC
+int acpi_get_ioapic_id(acpi_handle handle, u32 gsi_base, u64 *phys_addr);
+#endif
+
 int acpi_register_ioapic(acpi_handle handle, u64 phys_addr, u32 gsi_base);
 int acpi_unregister_ioapic(acpi_handle handle, u32 gsi_base);
 void acpi_irq_stats_init(void);
-- 
1.7.10.4

[Patch v7 08/18] PCI: Remove PCI ioapic driver

[Jiang Liu]

To support IOAPIC hotplug on x86 and IA64 platforms, OS needs to figure
out global interrupt source number(GSI) and IOAPIC enumeration ID
through ACPI interfaces. So BIOS must implement an ACPI IOAPIC device
with _GSB/_UID or _MAT method to support IOAPIC hotplug. OS also needs
to figure out base physical address to access IOAPIC registers. OS may
get the base physical address through PCI BARs if IOAPIC device is
visible in PCI domain, otherwise OS may get the address by ACPI _CRS
method if IOAPIC device is hidden from PCI domain by BIOS.

When adding a PCI subtree, we need to add IOAPIC devices before enabling
all other PCI devices because other PCI devices may use the IOAPIC to
allocate PCI interrupts.

So we plan to reimplement IOAPIC driver as an ACPI instead of PCI driver
due to:
1) hot-pluggable IOAPIC devices are always visible in ACPI domain,
   but may or may not be visible in PCI domain.
2) we could explicitly control the order between IOAPIC and other PCI
   devices.

We also have another choice to use a PCI driver to manage IOAPIC device
if it's visible in PCI domain and use an ACPI driver if it's only
visible in ACPI domain. But this solution is a little complex.

It shouldn't cause serious backward compatibility issues because:
1) IOAPIC hotplug is never supported on x86 yet because it hasn't
   implemented the required acpi_register_ioapic() and
   acpi_unregister_ioapic().
2) Currently only ACPI based IOAPIC hotplug is possible on x86 and
   IA64, we don't know other specifications and interfaces to support
   IOAPIC hotplug yet.
3) We will reimplement an ACPI IOAPICｔdriver support IOAPIC hotplug.

This change also helps to get rid of the false alarm on all current
Linux distributions:
[    6.952497] ioapic: probe of 0000:00:05.4 failed with error -22
[    6.959542] ioapic: probe of 0000:80:05.4 failed with error -22

Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
---
 drivers/pci/Kconfig  |    7 ---
 drivers/pci/Makefile |    2 -
 drivers/pci/ioapic.c |  121 --------------------------------------------------
 3 files changed, 130 deletions(-)
 delete mode 100644 drivers/pci/ioapic.c

diff --git a/drivers/pci/Kconfig b/drivers/pci/Kconfig
index 893503fa1782..39866d18004e 100644
--- a/drivers/pci/Kconfig
+++ b/drivers/pci/Kconfig
@@ -104,13 +104,6 @@ config PCI_PASID
 
 	  If unsure, say N.
 
-config PCI_IOAPIC
-	bool "PCI IO-APIC hotplug support" if X86
-	depends on PCI
-	depends on ACPI
-	depends on X86_IO_APIC
-	default !X86
-
 config PCI_LABEL
 	def_bool y if (DMI || ACPI)
 	select NLS
diff --git a/drivers/pci/Makefile b/drivers/pci/Makefile
index e04fe2d9df3b..73e4af400a5a 100644
--- a/drivers/pci/Makefile
+++ b/drivers/pci/Makefile
@@ -13,8 +13,6 @@ obj-$(CONFIG_PCI_QUIRKS) += quirks.o
 # Build PCI Express stuff if needed
 obj-$(CONFIG_PCIEPORTBUS) += pcie/
 
-obj-$(CONFIG_PCI_IOAPIC) += ioapic.o
-
 # Build the PCI Hotplug drivers if we were asked to
 obj-$(CONFIG_HOTPLUG_PCI) += hotplug/
 ifdef CONFIG_HOTPLUG_PCI
diff --git a/drivers/pci/ioapic.c b/drivers/pci/ioapic.c
deleted file mode 100644
index f6219d36227f..000000000000
--- a/drivers/pci/ioapic.c
+++ /dev/null
@@ -1,121 +0,0 @@
-/*
- * IOAPIC/IOxAPIC/IOSAPIC driver
- *
- * Copyright (C) 2009 Fujitsu Limited.
- * (c) Copyright 2009 Hewlett-Packard Development Company, L.P.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 as
- * published by the Free Software Foundation.
- */
-
-/*
- * This driver manages PCI I/O APICs added by hotplug after boot.  We try to
- * claim all I/O APIC PCI devices, but those present at boot were registered
- * when we parsed the ACPI MADT, so we'll fail when we try to re-register
- * them.
- */
-
-#include <linux/pci.h>
-#include <linux/module.h>
-#include <linux/acpi.h>
-#include <linux/slab.h>
-
-struct ioapic {
-	acpi_handle	handle;
-	u32		gsi_base;
-};
-
-static int ioapic_probe(struct pci_dev *dev, const struct pci_device_id *ent)
-{
-	acpi_handle handle;
-	acpi_status status;
-	unsigned long long gsb;
-	struct ioapic *ioapic;
-	int ret;
-	char *type;
-	struct resource *res;
-
-	handle = ACPI_HANDLE(&dev->dev);
-	if (!handle)
-		return -EINVAL;
-
-	status = acpi_evaluate_integer(handle, "_GSB", NULL, &gsb);
-	if (ACPI_FAILURE(status))
-		return -EINVAL;
-
-	/*
-	 * The previous code in acpiphp evaluated _MAT if _GSB failed, but
-	 * ACPI spec 4.0 sec 6.2.2 requires _GSB for hot-pluggable I/O APICs.
-	 */
-
-	ioapic = kzalloc(sizeof(*ioapic), GFP_KERNEL);
-	if (!ioapic)
-		return -ENOMEM;
-
-	ioapic->handle = handle;
-	ioapic->gsi_base = (u32) gsb;
-
-	if (dev->class == PCI_CLASS_SYSTEM_PIC_IOAPIC)
-		type = "IOAPIC";
-	else
-		type = "IOxAPIC";
-
-	ret = pci_enable_device(dev);
-	if (ret < 0)
-		goto exit_free;
-
-	pci_set_master(dev);
-
-	if (pci_request_region(dev, 0, type))
-		goto exit_disable;
-
-	res = &dev->resource[0];
-	if (acpi_register_ioapic(ioapic->handle, res->start, ioapic->gsi_base))
-		goto exit_release;
-
-	pci_set_drvdata(dev, ioapic);
-	dev_info(&dev->dev, "%s at %pR, GSI %u\n", type, res, ioapic->gsi_base);
-	return 0;
-
-exit_release:
-	pci_release_region(dev, 0);
-exit_disable:
-	pci_disable_device(dev);
-exit_free:
-	kfree(ioapic);
-	return -ENODEV;
-}
-
-static void ioapic_remove(struct pci_dev *dev)
-{
-	struct ioapic *ioapic = pci_get_drvdata(dev);
-
-	acpi_unregister_ioapic(ioapic->handle, ioapic->gsi_base);
-	pci_release_region(dev, 0);
-	pci_disable_device(dev);
-	kfree(ioapic);
-}
-
-
-static const struct pci_device_id ioapic_devices[] = {
-	{ PCI_DEVICE_CLASS(PCI_CLASS_SYSTEM_PIC_IOAPIC, ~0) },
-	{ PCI_DEVICE_CLASS(PCI_CLASS_SYSTEM_PIC_IOXAPIC, ~0) },
-	{ }
-};
-MODULE_DEVICE_TABLE(pci, ioapic_devices);
-
-static struct pci_driver ioapic_driver = {
-	.name		= "ioapic",
-	.id_table	= ioapic_devices,
-	.probe		= ioapic_probe,
-	.remove		= ioapic_remove,
-};
-
-static int __init ioapic_init(void)
-{
-	return pci_register_driver(&ioapic_driver);
-}
-module_init(ioapic_init);
-
-MODULE_LICENSE("GPL");
-- 
1.7.10.4

[Patch v7 09/18] x86, irq: Split out alloc_ioapic_save_registers()

[Jiang Liu]

From: Yinghai Lu <yinghai@kernel.org>

Split out alloc_ioapic_save_registers() from early_irq_init(),
so it could be used for ioapic hotplug later.

Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
Cc: Joerg Roedel <joro@8bytes.org>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Sebastian Andrzej Siewior <sebastian@breakpoint.cc>
---
 arch/x86/kernel/apic/io_apic.c |   22 +++++++++++++++-------
 1 file changed, 15 insertions(+), 7 deletions(-)

diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
index 1183d545da1e..acaee90fc9cc 100644
--- a/arch/x86/kernel/apic/io_apic.c
+++ b/arch/x86/kernel/apic/io_apic.c
@@ -237,6 +237,19 @@ static struct irq_pin_list *alloc_irq_pin_list(int node)
 	return kzalloc_node(sizeof(struct irq_pin_list), GFP_KERNEL, node);
 }
 
+static void alloc_ioapic_saved_registers(int idx)
+{
+	size_t size;
+
+	if (ioapics[idx].saved_registers)
+		return;
+
+	size = sizeof(struct IO_APIC_route_entry) * ioapics[idx].nr_registers;
+	ioapics[idx].saved_registers = kzalloc(size, GFP_KERNEL);
+	if (!ioapics[idx].saved_registers)
+		pr_err("IOAPIC %d: suspend/resume impossible!\n", idx);
+}
+
 int __init arch_early_irq_init(void)
 {
 	struct irq_cfg *cfg;
@@ -245,13 +258,8 @@ int __init arch_early_irq_init(void)
 	if (!nr_legacy_irqs())
 		io_apic_irqs = ~0UL;
 
-	for_each_ioapic(i) {
-		ioapics[i].saved_registers =
-			kzalloc(sizeof(struct IO_APIC_route_entry) *
-				ioapics[i].nr_registers, GFP_KERNEL);
-		if (!ioapics[i].saved_registers)
-			pr_err("IOAPIC %d: suspend/resume impossible!\n", i);
-	}
+	for_each_ioapic(i)
+		alloc_ioapic_saved_registers(i);
 
 	/*
 	 * For legacy IRQ's, start with assigning irq0 to irq15 to
-- 
1.7.10.4

[Patch v7 10/18] x86, irq: Prefer assigned ID in APIC ID register for x86_64

[Jiang Liu]

From: Yinghai Lu <yinghai@kernel.org>

Perfer the assigned ID in APIC ID register for x86_64 if it's still
available.

Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Cc: Joerg Roedel <joro@8bytes.org>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Sebastian Andrzej Siewior <sebastian@breakpoint.cc>
Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
---
 arch/x86/kernel/apic/io_apic.c |   35 ++++++++++++++++++++++++++++++-----
 1 file changed, 30 insertions(+), 5 deletions(-)

diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
index acaee90fc9cc..eaa7ea385600 100644
--- a/arch/x86/kernel/apic/io_apic.c
+++ b/arch/x86/kernel/apic/io_apic.c
@@ -3580,26 +3580,51 @@ static int __init io_apic_get_unique_id(int ioapic, int apic_id)
 	return apic_id;
 }
 
-static u8 __init io_apic_unique_id(u8 id)
+static u8 __init io_apic_unique_id(int idx, u8 id)
 {
 	if ((boot_cpu_data.x86_vendor == X86_VENDOR_INTEL) &&
 	    !APIC_XAPIC(apic_version[boot_cpu_physical_apicid]))
-		return io_apic_get_unique_id(nr_ioapics, id);
+		return io_apic_get_unique_id(idx, id);
 	else
 		return id;
 }
 #else
-static u8 __init io_apic_unique_id(u8 id)
+static u8 __init io_apic_unique_id(int idx, u8 id)
 {
 	int i;
+	u8 new_id;
+	unsigned long flags;
 	DECLARE_BITMAP(used, 256);
+	union IO_APIC_reg_00 reg_00;
 
 	bitmap_zero(used, 256);
 	for_each_ioapic(i)
 		__set_bit(mpc_ioapic_id(i), used);
 	if (!test_bit(id, used))
 		return id;
-	return find_first_zero_bit(used, 256);
+
+	/* check register at first */
+	raw_spin_lock_irqsave(&ioapic_lock, flags);
+	reg_00.raw = io_apic_read(idx, 0);
+	raw_spin_unlock_irqrestore(&ioapic_lock, flags);
+	new_id = reg_00.bits.ID;
+	if (!test_bit(new_id, used)) {
+		apic_printk(APIC_VERBOSE, KERN_INFO
+			"IOAPIC[%d]: Using reg apic_id %d instead of %d\n",
+			 idx, new_id, id);
+		return new_id;
+	}
+
+	new_id = find_first_zero_bit(used, 256);
+	reg_00.bits.ID = new_id;
+	raw_spin_lock_irqsave(&ioapic_lock, flags);
+	io_apic_write(idx, 0, reg_00.raw);
+	reg_00.raw = io_apic_read(idx, 0);
+	raw_spin_unlock_irqrestore(&ioapic_lock, flags);
+	/* Sanity check */
+	BUG_ON(reg_00.bits.ID != new_id);
+
+	return new_id;
 }
 #endif
 
@@ -3865,7 +3890,7 @@ void __init mp_register_ioapic(int id, u32 address, u32 gsi_base,
 		return;
 	}
 
-	ioapics[idx].mp_config.apicid = io_apic_unique_id(id);
+	ioapics[idx].mp_config.apicid = io_apic_unique_id(idx, id);
 	ioapics[idx].mp_config.apicver = io_apic_get_version(idx);
 
 	/*
-- 
1.7.10.4

[Patch v7 11/18] x86, irq: Remove __init marker for functions will be used by IOAPIC hotplug

[Jiang Liu]

Remove __init marker for functions which will be used by IOAPIC hotplug
at runtime.

Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
---
 arch/x86/include/asm/io_apic.h |    4 ++--
 arch/x86/kernel/apic/io_apic.c |   18 +++++++++---------
 2 files changed, 11 insertions(+), 11 deletions(-)

diff --git a/arch/x86/include/asm/io_apic.h b/arch/x86/include/asm/io_apic.h
index 1733ab49ac5e..a666dcdeea67 100644
--- a/arch/x86/include/asm/io_apic.h
+++ b/arch/x86/include/asm/io_apic.h
@@ -188,8 +188,8 @@ extern int mp_find_ioapic_pin(int ioapic, u32 gsi);
 extern u32 mp_pin_to_gsi(int ioapic, int pin);
 extern int mp_map_gsi_to_irq(u32 gsi, unsigned int flags);
 extern void mp_unmap_irq(int irq);
-extern void __init mp_register_ioapic(int id, u32 address, u32 gsi_base,
-				      struct ioapic_domain_cfg *cfg);
+extern void mp_register_ioapic(int id, u32 address, u32 gsi_base,
+			       struct ioapic_domain_cfg *cfg);
 extern int mp_irqdomain_map(struct irq_domain *domain, unsigned int virq,
 			    irq_hw_number_t hwirq);
 extern void mp_irqdomain_unmap(struct irq_domain *domain, unsigned int virq);
diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
index eaa7ea385600..581e51728871 100644
--- a/arch/x86/kernel/apic/io_apic.c
+++ b/arch/x86/kernel/apic/io_apic.c
@@ -3459,7 +3459,7 @@ io_apic_setup_irq_pin(unsigned int irq, int node, struct io_apic_irq_attr *attr)
 	return ret;
 }
 
-static int __init io_apic_get_redir_entries(int ioapic)
+static int io_apic_get_redir_entries(int ioapic)
 {
 	union IO_APIC_reg_01	reg_01;
 	unsigned long flags;
@@ -3505,7 +3505,7 @@ int __init arch_probe_nr_irqs(void)
 }
 
 #ifdef CONFIG_X86_32
-static int __init io_apic_get_unique_id(int ioapic, int apic_id)
+static int io_apic_get_unique_id(int ioapic, int apic_id)
 {
 	union IO_APIC_reg_00 reg_00;
 	static physid_mask_t apic_id_map = PHYSID_MASK_NONE;
@@ -3580,7 +3580,7 @@ static int __init io_apic_get_unique_id(int ioapic, int apic_id)
 	return apic_id;
 }
 
-static u8 __init io_apic_unique_id(int idx, u8 id)
+static u8 io_apic_unique_id(int idx, u8 id)
 {
 	if ((boot_cpu_data.x86_vendor == X86_VENDOR_INTEL) &&
 	    !APIC_XAPIC(apic_version[boot_cpu_physical_apicid]))
@@ -3589,7 +3589,7 @@ static u8 __init io_apic_unique_id(int idx, u8 id)
 		return id;
 }
 #else
-static u8 __init io_apic_unique_id(int idx, u8 id)
+static u8 io_apic_unique_id(int idx, u8 id)
 {
 	int i;
 	u8 new_id;
@@ -3628,7 +3628,7 @@ static u8 __init io_apic_unique_id(int idx, u8 id)
 }
 #endif
 
-static int __init io_apic_get_version(int ioapic)
+static int io_apic_get_version(int ioapic)
 {
 	union IO_APIC_reg_01	reg_01;
 	unsigned long flags;
@@ -3832,7 +3832,7 @@ int mp_find_ioapic_pin(int ioapic, u32 gsi)
 	return gsi - gsi_cfg->gsi_base;
 }
 
-static __init int bad_ioapic(unsigned long address)
+static int bad_ioapic(unsigned long address)
 {
 	if (nr_ioapics >= MAX_IO_APICS) {
 		pr_warn("WARNING: Max # of I/O APICs (%d) exceeded (found %d), skipping\n",
@@ -3846,7 +3846,7 @@ static __init int bad_ioapic(unsigned long address)
 	return 0;
 }
 
-static __init int bad_ioapic_register(int idx)
+static int bad_ioapic_register(int idx)
 {
 	union IO_APIC_reg_00 reg_00;
 	union IO_APIC_reg_01 reg_01;
@@ -3865,8 +3865,8 @@ static __init int bad_ioapic_register(int idx)
 	return 0;
 }
 
-void __init mp_register_ioapic(int id, u32 address, u32 gsi_base,
-			       struct ioapic_domain_cfg *cfg)
+void mp_register_ioapic(int id, u32 address, u32 gsi_base,
+			struct ioapic_domain_cfg *cfg)
 {
 	int idx = 0;
 	int entries;
-- 
1.7.10.4

[Patch v7 12/18] x86, irq: Keep balance of IOAPIC pin reference count

[Jiang Liu]

To keep balance of IOAPIC pin reference count, we need to protect
pirq_enable_irq(), acpi_pci_irq_enable() and intel_mid_pci_irq_enable()
from reentrance. There are two cases which will cause reentrance.

The first case is caused by suspend/hibernation. If pcibios_disable_irq
is called during suspending/hibernating, we don't release the assigned
IRQ number, otherwise it may break the suspend/hibernation. So late when
pcibios_enable_irq is called during resume, we shouldn't allocate IRQ
number again.

The second case is that function acpi_pci_irq_enable() may be called
twice for PCI devices present at boot time as below:
1) pci_acpi_init()
	--> acpi_pci_irq_enable() if pci_routeirq is true
2) pci_enable_device()
	--> pcibios_enable_device()
		--> acpi_pci_irq_enable()
We can't kill kernel parameter pci_routeirq yet because it's still
needed for debugging purpose.

So flag irq_managed is introduced to track whether IRQ number is
assigned by OS and to protect pirq_enable_irq(), acpi_pci_irq_enable()
and intel_mid_pci_irq_enable() from reentrance.

Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
---
 arch/x86/pci/intel_mid_pci.c |   10 +++++++++-
 arch/x86/pci/irq.c           |    7 ++++++-
 drivers/acpi/pci_irq.c       |   11 +++++++++--
 include/linux/pci.h          |    1 +
 4 files changed, 25 insertions(+), 4 deletions(-)

diff --git a/arch/x86/pci/intel_mid_pci.c b/arch/x86/pci/intel_mid_pci.c
index b9958c364075..44b9271580b5 100644
--- a/arch/x86/pci/intel_mid_pci.c
+++ b/arch/x86/pci/intel_mid_pci.c
@@ -210,6 +210,9 @@ static int intel_mid_pci_irq_enable(struct pci_dev *dev)
 {
 	int polarity;
 
+	if (dev->irq_managed && dev->irq > 0)
+		return 0;
+
 	if (intel_mid_identify_cpu() == INTEL_MID_CPU_CHIP_TANGIER)
 		polarity = 0; /* active high */
 	else
@@ -224,13 +227,18 @@ static int intel_mid_pci_irq_enable(struct pci_dev *dev)
 	if (mp_map_gsi_to_irq(dev->irq, IOAPIC_MAP_ALLOC) < 0)
 		return -EBUSY;
 
+	dev->irq_managed = 1;
+
 	return 0;
 }
 
 static void intel_mid_pci_irq_disable(struct pci_dev *dev)
 {
-	if (!mp_should_keep_irq(&dev->dev) && dev->irq > 0)
+	if (!mp_should_keep_irq(&dev->dev) && dev->irq_managed &&
+	    dev->irq > 0) {
 		mp_unmap_irq(dev->irq);
+		dev->irq_managed = 0;
+	}
 }
 
 struct pci_ops intel_mid_pci_ops = {
diff --git a/arch/x86/pci/irq.c b/arch/x86/pci/irq.c
index eb500c2592ad..a47e2dea0972 100644
--- a/arch/x86/pci/irq.c
+++ b/arch/x86/pci/irq.c
@@ -1202,6 +1202,9 @@ static int pirq_enable_irq(struct pci_dev *dev)
 			int irq;
 			struct io_apic_irq_attr irq_attr;
 
+			if (dev->irq_managed && dev->irq > 0)
+				return 0;
+
 			irq = IO_APIC_get_PCI_irq_vector(dev->bus->number,
 						PCI_SLOT(dev->devfn),
 						pin - 1, &irq_attr);
@@ -1228,6 +1231,7 @@ static int pirq_enable_irq(struct pci_dev *dev)
 			}
 			dev = temp_dev;
 			if (irq >= 0) {
+				dev->irq_managed = 1;
 				dev->irq = irq;
 				dev_info(&dev->dev, "PCI->APIC IRQ transform: "
 					 "INT %c -> IRQ %d\n", 'A' + pin - 1, irq);
@@ -1257,8 +1261,9 @@ static int pirq_enable_irq(struct pci_dev *dev)
 static void pirq_disable_irq(struct pci_dev *dev)
 {
 	if (io_apic_assign_pci_irqs && !mp_should_keep_irq(&dev->dev) &&
-	    dev->irq) {
+	    dev->irq_managed && dev->irq) {
 		mp_unmap_irq(dev->irq);
 		dev->irq = 0;
+		dev->irq_managed = 0;
 	}
 }
diff --git a/drivers/acpi/pci_irq.c b/drivers/acpi/pci_irq.c
index 6e6b80eb0bba..5f1fdca65e5f 100644
--- a/drivers/acpi/pci_irq.c
+++ b/drivers/acpi/pci_irq.c
@@ -413,6 +413,9 @@ int acpi_pci_irq_enable(struct pci_dev *dev)
 		return 0;
 	}
 
+	if (dev->irq_managed && dev->irq > 0)
+		return 0;
+
 	entry = acpi_pci_irq_lookup(dev, pin);
 	if (!entry) {
 		/*
@@ -456,6 +459,7 @@ int acpi_pci_irq_enable(struct pci_dev *dev)
 		return rc;
 	}
 	dev->irq = rc;
+	dev->irq_managed = 1;
 
 	if (link)
 		snprintf(link_desc, sizeof(link_desc), " -> Link[%s]", link);
@@ -478,7 +482,7 @@ void acpi_pci_irq_disable(struct pci_dev *dev)
 	u8 pin;
 
 	pin = dev->pin;
-	if (!pin)
+	if (!pin || !dev->irq_managed || dev->irq <= 0)
 		return;
 
 	/* Keep IOAPIC pin configuration when suspending */
@@ -506,6 +510,9 @@ void acpi_pci_irq_disable(struct pci_dev *dev)
 	 */
 
 	dev_dbg(&dev->dev, "PCI INT %c disabled\n", pin_name(pin));
-	if (gsi >= 0 && dev->irq > 0)
+	if (gsi >= 0) {
 		acpi_unregister_gsi(gsi);
+		dev->irq = 0;
+		dev->irq_managed = 0;
+	}
 }
diff --git a/include/linux/pci.h b/include/linux/pci.h
index 5be8db45e368..2a8c4055ae9d 100644
--- a/include/linux/pci.h
+++ b/include/linux/pci.h
@@ -348,6 +348,7 @@ struct pci_dev {
 	unsigned int	__aer_firmware_first:1;
 	unsigned int	broken_intx_masking:1;
 	unsigned int	io_window_1k:1;	/* Intel P2P bridge 1K I/O windows */
+	unsigned int	irq_managed:1;
 	pci_dev_flags_t dev_flags;
 	atomic_t	enable_cnt;	/* pci_enable_device has been called */
 
-- 
1.7.10.4

[Patch v7 13/18] x86, irq: Refine mp_register_ioapic() to prepare for IOAPIC hotplug

[Jiang Liu]

Refine mp_register_ioapic() to prepare for IOAPIC hotplug by:
1) change return value from void to int.
2) check for gsi range conflicts
3) check for IOAPIC physical address conflicts
4) enhance the way to allocate IOAPIC index

Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
---
 arch/x86/include/asm/io_apic.h |    4 +-
 arch/x86/kernel/apic/io_apic.c |   87 ++++++++++++++++++++++++++--------------
 2 files changed, 58 insertions(+), 33 deletions(-)

diff --git a/arch/x86/include/asm/io_apic.h b/arch/x86/include/asm/io_apic.h
index a666dcdeea67..8741c1ea33c3 100644
--- a/arch/x86/include/asm/io_apic.h
+++ b/arch/x86/include/asm/io_apic.h
@@ -188,8 +188,8 @@ extern int mp_find_ioapic_pin(int ioapic, u32 gsi);
 extern u32 mp_pin_to_gsi(int ioapic, int pin);
 extern int mp_map_gsi_to_irq(u32 gsi, unsigned int flags);
 extern void mp_unmap_irq(int irq);
-extern void mp_register_ioapic(int id, u32 address, u32 gsi_base,
-			       struct ioapic_domain_cfg *cfg);
+extern int mp_register_ioapic(int id, u32 address, u32 gsi_base,
+			      struct ioapic_domain_cfg *cfg);
 extern int mp_irqdomain_map(struct irq_domain *domain, unsigned int virq,
 			    irq_hw_number_t hwirq);
 extern void mp_irqdomain_unmap(struct irq_domain *domain, unsigned int virq);
diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
index 581e51728871..59267659e5b1 100644
--- a/arch/x86/kernel/apic/io_apic.c
+++ b/arch/x86/kernel/apic/io_apic.c
@@ -3832,20 +3832,6 @@ int mp_find_ioapic_pin(int ioapic, u32 gsi)
 	return gsi - gsi_cfg->gsi_base;
 }
 
-static int bad_ioapic(unsigned long address)
-{
-	if (nr_ioapics >= MAX_IO_APICS) {
-		pr_warn("WARNING: Max # of I/O APICs (%d) exceeded (found %d), skipping\n",
-			MAX_IO_APICS, nr_ioapics);
-		return 1;
-	}
-	if (!address) {
-		pr_warn("WARNING: Bogus (zero) I/O APIC address found in table, skipping!\n");
-		return 1;
-	}
-	return 0;
-}
-
 static int bad_ioapic_register(int idx)
 {
 	union IO_APIC_reg_00 reg_00;
@@ -3865,29 +3851,51 @@ static int bad_ioapic_register(int idx)
 	return 0;
 }
 
-void mp_register_ioapic(int id, u32 address, u32 gsi_base,
-			struct ioapic_domain_cfg *cfg)
+static int find_free_ioapic_entry(void)
 {
-	int idx = 0;
-	int entries;
+	return nr_ioapics;
+}
+
+/**
+ * mp_register_ioapic - Register an IOAPIC device
+ * @id:		hardware IOAPIC ID
+ * @address:	physical address of IOAPIC register area
+ * @gsi_base:	base of GSI associated with the IOAPIC
+ * @cfg:	configuration information for the IOAPIC
+ */
+int mp_register_ioapic(int id, u32 address, u32 gsi_base,
+		       struct ioapic_domain_cfg *cfg)
+{
+	u32 gsi_end;
+	int idx, ioapic, entries;
 	struct mp_ioapic_gsi *gsi_cfg;
 
-	if (bad_ioapic(address))
-		return;
+	if (!address) {
+		pr_warn("Bogus (zero) I/O APIC address found, skipping!\n");
+		return -EINVAL;
+	}
+	for_each_ioapic(ioapic)
+		if (ioapics[ioapic].mp_config.apicaddr == address) {
+			pr_warn("address 0x%x conflicts with IOAPIC%d\n",
+				address, ioapic);
+			return -EEXIST;
+		}
 
-	idx = nr_ioapics;
+	idx = find_free_ioapic_entry();
+	if (idx >= MAX_IO_APICS) {
+		pr_warn("Max # of I/O APICs (%d) exceeded (found %d), skipping\n",
+			MAX_IO_APICS, idx);
+		return -ENOSPC;
+	}
 
 	ioapics[idx].mp_config.type = MP_IOAPIC;
 	ioapics[idx].mp_config.flags = MPC_APIC_USABLE;
 	ioapics[idx].mp_config.apicaddr = address;
-	ioapics[idx].irqdomain = NULL;
-	ioapics[idx].irqdomain_cfg = *cfg;
 
 	set_fixmap_nocache(FIX_IO_APIC_BASE_0 + idx, address);
-
 	if (bad_ioapic_register(idx)) {
 		clear_fixmap(FIX_IO_APIC_BASE_0 + idx);
-		return;
+		return -ENODEV;
 	}
 
 	ioapics[idx].mp_config.apicid = io_apic_unique_id(idx, id);
@@ -3898,24 +3906,41 @@ void mp_register_ioapic(int id, u32 address, u32 gsi_base,
 	 * and to prevent reprogramming of IOAPIC pins (PCI GSIs).
 	 */
 	entries = io_apic_get_redir_entries(idx);
+	gsi_end = gsi_base + entries - 1;
+	for_each_ioapic(ioapic) {
+		gsi_cfg = mp_ioapic_gsi_routing(ioapic);
+		if ((gsi_base >= gsi_cfg->gsi_base &&
+		     gsi_base <= gsi_cfg->gsi_end) ||
+		    (gsi_end >= gsi_cfg->gsi_base &&
+		     gsi_end <= gsi_cfg->gsi_end)) {
+			pr_warn("GSI range [%u-%u] for new IOAPIC conflicts with GSI[%u-%u]\n",
+				gsi_base, gsi_end,
+				gsi_cfg->gsi_base, gsi_cfg->gsi_end);
+			clear_fixmap(FIX_IO_APIC_BASE_0 + idx);
+			return -ENOSPC;
+		}
+	}
 	gsi_cfg = mp_ioapic_gsi_routing(idx);
 	gsi_cfg->gsi_base = gsi_base;
-	gsi_cfg->gsi_end = gsi_base + entries - 1;
+	gsi_cfg->gsi_end = gsi_end;
 
-	/*
-	 * The number of IO-APIC IRQ registers (== #pins):
-	 */
-	ioapics[idx].nr_registers = entries;
+	ioapics[idx].irqdomain = NULL;
+	ioapics[idx].irqdomain_cfg = *cfg;
 
 	if (gsi_cfg->gsi_end >= gsi_top)
 		gsi_top = gsi_cfg->gsi_end + 1;
+	if (nr_ioapics <= idx)
+		nr_ioapics = idx + 1;
+
+	/* Set nr_registers to mark entry present */
+	ioapics[idx].nr_registers = entries;
 
 	pr_info("IOAPIC[%d]: apic_id %d, version %d, address 0x%x, GSI %d-%d\n",
 		idx, mpc_ioapic_id(idx),
 		mpc_ioapic_ver(idx), mpc_ioapic_addr(idx),
 		gsi_cfg->gsi_base, gsi_cfg->gsi_end);
 
-	nr_ioapics++;
+	return 0;
 }
 
 int mp_irqdomain_map(struct irq_domain *domain, unsigned int virq,
-- 
1.7.10.4

[Patch v7 14/18] x86, irq, ACPI: Introduce a rwsem to protect IOAPIC operations from hotplug

[Jiang Liu]

We are going to support ACPI based IOAPIC hotplug, so introduce a rwsem
to protect IOAPIC data structures from IOAPIC hotplug. We choose to
serialize in ACPI instead of in the IOAPIC core because:
1) currently we are only plan to support ACPI based IOAPIC hotplug
2) it's much more cleaner and easier
3) It does't affect IOAPIC discovered by devicetree, SFI and mpparse.

Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
---
 arch/x86/kernel/acpi/boot.c |   21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)

diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c
index e077c080a519..3f6b665f5aa6 100644
--- a/arch/x86/kernel/acpi/boot.c
+++ b/arch/x86/kernel/acpi/boot.c
@@ -76,6 +76,19 @@ int acpi_fix_pin2_polarity __initdata;
 static u64 acpi_lapic_addr __initdata = APIC_DEFAULT_PHYS_BASE;
 #endif
 
+/*
+ * Locks related to IOAPIC hotplug
+ * Hotplug side:
+ * 	->lock_device_hotplug()	//device_hotplug_lock
+ *		->acpi_ioapic_rwsem
+ *			->ioapic_lock
+ * Interrupt mapping side:
+ *	->acpi_ioapic_rwsem
+ *		->ioapic_mutex
+ *			->ioapic_lock
+ */
+static DECLARE_RWSEM(acpi_ioapic_rwsem);
+
 /* --------------------------------------------------------------------------
                               Boot-time Configuration
    -------------------------------------------------------------------------- */
@@ -609,8 +622,10 @@ int acpi_gsi_to_irq(u32 gsi, unsigned int *irqp)
 	if (acpi_irq_model == ACPI_IRQ_MODEL_PIC) {
 		*irqp = gsi;
 	} else {
+		down_read(&acpi_ioapic_rwsem);
 		irq = mp_map_gsi_to_irq(gsi,
 					IOAPIC_MAP_ALLOC | IOAPIC_MAP_CHECK);
+		up_read(&acpi_ioapic_rwsem);
 		if (irq < 0)
 			return -1;
 		*irqp = irq;
@@ -651,7 +666,9 @@ static int acpi_register_gsi_ioapic(struct device *dev, u32 gsi,
 	int irq = gsi;
 
 #ifdef CONFIG_X86_IO_APIC
+	down_read(&acpi_ioapic_rwsem);
 	irq = mp_register_gsi(dev, gsi, trigger, polarity);
+	up_read(&acpi_ioapic_rwsem);
 #endif
 
 	return irq;
@@ -660,7 +677,9 @@ static int acpi_register_gsi_ioapic(struct device *dev, u32 gsi,
 static void acpi_unregister_gsi_ioapic(u32 gsi)
 {
 #ifdef CONFIG_X86_IO_APIC
+	down_read(&acpi_ioapic_rwsem);
 	mp_unregister_gsi(gsi);
+	up_read(&acpi_ioapic_rwsem);
 #endif
 }
 
@@ -1186,7 +1205,9 @@ static void __init acpi_process_madt(void)
 			/*
 			 * Parse MADT IO-APIC entries
 			 */
+			down_write(&acpi_ioapic_rwsem);
 			error = acpi_parse_madt_ioapic_entries();
+			up_write(&acpi_ioapic_rwsem);
 			if (!error) {
 				acpi_set_irq_model_ioapic();
 
-- 
1.7.10.4

[Patch v7 15/18] x86, irq, ACPI: Implement interface to support ACPI based IOAPIC hot-addition

[Jiang Liu]

Implement acpi_register_ioapic() and enhance mp_register_ioapic()
to support ACPI based IOAPIC hot-addition.

Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
---
 arch/x86/kernel/acpi/boot.c    |   31 +++++++++++++++++++++++++++++--
 arch/x86/kernel/apic/io_apic.c |   22 +++++++++++++++++++++-
 2 files changed, 50 insertions(+), 3 deletions(-)

diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c
index 3f6b665f5aa6..8205f6381d77 100644
--- a/arch/x86/kernel/acpi/boot.c
+++ b/arch/x86/kernel/acpi/boot.c
@@ -779,8 +779,35 @@ EXPORT_SYMBOL(acpi_unmap_lsapic);
 
 int acpi_register_ioapic(acpi_handle handle, u64 phys_addr, u32 gsi_base)
 {
-	/* TBD */
-	return -EINVAL;
+	int ret = -ENOSYS;
+#ifdef CONFIG_ACPI_HOTPLUG_IOAPIC
+	int ioapic_id;
+	u64 addr;
+	struct ioapic_domain_cfg cfg = {
+		.type = IOAPIC_DOMAIN_DYNAMIC,
+		.ops = &acpi_irqdomain_ops,
+	};
+
+	ioapic_id = acpi_get_ioapic_id(handle, gsi_base, &addr);
+	if (ioapic_id < 0) {
+		unsigned long long uid;
+		acpi_status status;
+
+		status = acpi_evaluate_integer(handle, METHOD_NAME__UID,
+					       NULL, &uid);
+		if (ACPI_FAILURE(status)) {
+			acpi_handle_warn(handle, "failed to get IOAPIC ID.\n");
+			return -EINVAL;
+		}
+		ioapic_id = (int)uid;
+	}
+
+	down_write(&acpi_ioapic_rwsem);
+	ret  = mp_register_ioapic(ioapic_id, phys_addr, gsi_base, &cfg);
+	up_write(&acpi_ioapic_rwsem);
+#endif
+
+	return ret;
 }
 
 EXPORT_SYMBOL(acpi_register_ioapic);
diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
index 59267659e5b1..4d9d87cc3259 100644
--- a/arch/x86/kernel/apic/io_apic.c
+++ b/arch/x86/kernel/apic/io_apic.c
@@ -3853,7 +3853,13 @@ static int bad_ioapic_register(int idx)
 
 static int find_free_ioapic_entry(void)
 {
-	return nr_ioapics;
+	int idx;
+
+	for (idx = 0; idx < MAX_IO_APICS; idx++)
+		if (ioapics[idx].nr_registers == 0)
+			return idx;
+
+	return MAX_IO_APICS;
 }
 
 /**
@@ -3869,6 +3875,7 @@ int mp_register_ioapic(int id, u32 address, u32 gsi_base,
 	u32 gsi_end;
 	int idx, ioapic, entries;
 	struct mp_ioapic_gsi *gsi_cfg;
+	bool hotplug = !!ioapic_initialized;
 
 	if (!address) {
 		pr_warn("Bogus (zero) I/O APIC address found, skipping!\n");
@@ -3927,6 +3934,19 @@ int mp_register_ioapic(int id, u32 address, u32 gsi_base,
 	ioapics[idx].irqdomain = NULL;
 	ioapics[idx].irqdomain_cfg = *cfg;
 
+	/*
+	 * If mp_register_ioapic() is called during early boot stage when
+	 * walking ACPI/SFI/DT tables, it's too early to create irqdomain,
+	 * we are still using bootmem allocator. So delay it to setup_IO_APIC().
+	 */
+	if (hotplug) {
+		if (mp_irqdomain_create(idx)) {
+			clear_fixmap(FIX_IO_APIC_BASE_0 + idx);
+			return -ENOMEM;
+		}
+		alloc_ioapic_saved_registers(idx);
+	}
+
 	if (gsi_cfg->gsi_end >= gsi_top)
 		gsi_top = gsi_cfg->gsi_end + 1;
 	if (nr_ioapics <= idx)
-- 
1.7.10.4

[Patch v7 16/18] x86, irq, ACPI: Implement interfaces to support ACPI based IOAPIC hot-removal

[Jiang Liu]

Implement acpi_unregister_ioapic() to support ACPI based IOAPIC hot-removal.
An IOAPIC could only be removed when all its pins are unused.

Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
---
 arch/x86/include/asm/io_apic.h |    1 +
 arch/x86/kernel/acpi/boot.c    |   13 +++++++---
 arch/x86/kernel/apic/io_apic.c |   55 ++++++++++++++++++++++++++++++++++++++++
 3 files changed, 65 insertions(+), 4 deletions(-)

diff --git a/arch/x86/include/asm/io_apic.h b/arch/x86/include/asm/io_apic.h
index 8741c1ea33c3..7c04b8ec22a4 100644
--- a/arch/x86/include/asm/io_apic.h
+++ b/arch/x86/include/asm/io_apic.h
@@ -190,6 +190,7 @@ extern int mp_map_gsi_to_irq(u32 gsi, unsigned int flags);
 extern void mp_unmap_irq(int irq);
 extern int mp_register_ioapic(int id, u32 address, u32 gsi_base,
 			      struct ioapic_domain_cfg *cfg);
+extern int mp_unregister_ioapic(u32 gsi_base);
 extern int mp_irqdomain_map(struct irq_domain *domain, unsigned int virq,
 			    irq_hw_number_t hwirq);
 extern void mp_irqdomain_unmap(struct irq_domain *domain, unsigned int virq);
diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c
index 8205f6381d77..bc61a5c0e7b6 100644
--- a/arch/x86/kernel/acpi/boot.c
+++ b/arch/x86/kernel/acpi/boot.c
@@ -809,15 +809,20 @@ int acpi_register_ioapic(acpi_handle handle, u64 phys_addr, u32 gsi_base)
 
 	return ret;
 }
-
 EXPORT_SYMBOL(acpi_register_ioapic);
 
 int acpi_unregister_ioapic(acpi_handle handle, u32 gsi_base)
 {
-	/* TBD */
-	return -EINVAL;
-}
+	int ret = -ENOSYS;
+
+#ifdef CONFIG_ACPI_HOTPLUG_IOAPIC
+	down_write(&acpi_ioapic_rwsem);
+	ret  = mp_unregister_ioapic(gsi_base);
+	up_write(&acpi_ioapic_rwsem);
+#endif
 
+	return ret;
+}
 EXPORT_SYMBOL(acpi_unregister_ioapic);
 
 static int __init acpi_parse_sbf(struct acpi_table_header *table)
diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
index 4d9d87cc3259..9e5432ee2070 100644
--- a/arch/x86/kernel/apic/io_apic.c
+++ b/arch/x86/kernel/apic/io_apic.c
@@ -112,6 +112,7 @@ static struct ioapic {
 	struct ioapic_domain_cfg irqdomain_cfg;
 	struct irq_domain *irqdomain;
 	struct mp_pin_info *pin_info;
+	struct resource *iomem_res;
 } ioapics[MAX_IO_APICS];
 
 #define mpc_ioapic_ver(ioapic_idx)	ioapics[ioapic_idx].mp_config.apicver
@@ -250,6 +251,12 @@ static void alloc_ioapic_saved_registers(int idx)
 		pr_err("IOAPIC %d: suspend/resume impossible!\n", idx);
 }
 
+static void free_ioapic_saved_registers(int idx)
+{
+	kfree(ioapics[idx].saved_registers);
+	ioapics[idx].saved_registers = NULL;
+}
+
 int __init arch_early_irq_init(void)
 {
 	struct irq_cfg *cfg;
@@ -2973,6 +2980,16 @@ static int mp_irqdomain_create(int ioapic)
 	return 0;
 }
 
+static void ioapic_destroy_irqdomain(int idx)
+{
+	if (ioapics[idx].irqdomain) {
+		irq_domain_remove(ioapics[idx].irqdomain);
+		ioapics[idx].irqdomain = NULL;
+	}
+	kfree(ioapics[idx].pin_info);
+	ioapics[idx].pin_info = NULL;
+}
+
 void __init setup_IO_APIC(void)
 {
 	int ioapic;
@@ -3735,6 +3752,7 @@ static struct resource * __init ioapic_setup_resources(void)
 		snprintf(mem, IOAPIC_RESOURCE_NAME_SIZE, "IOAPIC %u", i);
 		mem += IOAPIC_RESOURCE_NAME_SIZE;
 		num++;
+		ioapics[i].iomem_res = res;
 	}
 
 	ioapic_resources = res;
@@ -3963,6 +3981,43 @@ int mp_register_ioapic(int id, u32 address, u32 gsi_base,
 	return 0;
 }
 
+int mp_unregister_ioapic(u32 gsi_base)
+{
+	int ioapic, pin;
+	int found = 0;
+	struct mp_pin_info *pin_info;
+
+	for_each_ioapic(ioapic)
+		if (ioapics[ioapic].gsi_config.gsi_base == gsi_base) {
+			found = 1;
+			break;
+		}
+	if (!found) {
+		pr_warn("can't find IOAPIC for GSI %d\n", gsi_base);
+		return -ENODEV;
+	}
+
+	for_each_pin(ioapic, pin) {
+		pin_info = mp_pin_info(ioapic, pin);
+		if (pin_info->count) {
+			pr_warn("pin%d on IOAPIC%d is still in use.\n",
+				pin, ioapic);
+			return -EBUSY;
+		}
+	}
+
+	/* Mark entry not present */
+	ioapics[ioapic].nr_registers  = 0;
+	ioapic_destroy_irqdomain(ioapic);
+	free_ioapic_saved_registers(ioapic);
+	if (ioapics[ioapic].iomem_res)
+		release_resource(ioapics[ioapic].iomem_res);
+	clear_fixmap(FIX_IO_APIC_BASE_0 + ioapic);
+	memset(&ioapics[ioapic], 0, sizeof(ioapics[ioapic]));
+
+	return 0;
+}
+
 int mp_irqdomain_map(struct irq_domain *domain, unsigned int virq,
 		     irq_hw_number_t hwirq)
 {
-- 
1.7.10.4

[Patch v7 17/18] x86, irq: Introduce helper to check whether an IOAPIC has been registered

[Jiang Liu]

Introduce acpi_ioapic_registered() to check whether an IOAPIC has already
been registered, it will be used when enabling IOAPIC hotplug.

Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
---
 arch/x86/include/asm/io_apic.h |    1 +
 arch/x86/kernel/acpi/boot.c    |   22 ++++++++++++++++++++++
 arch/x86/kernel/apic/io_apic.c |   11 +++++++++++
 include/linux/acpi.h           |    1 +
 4 files changed, 35 insertions(+)

diff --git a/arch/x86/include/asm/io_apic.h b/arch/x86/include/asm/io_apic.h
index 7c04b8ec22a4..ca742d5249db 100644
--- a/arch/x86/include/asm/io_apic.h
+++ b/arch/x86/include/asm/io_apic.h
@@ -191,6 +191,7 @@ extern void mp_unmap_irq(int irq);
 extern int mp_register_ioapic(int id, u32 address, u32 gsi_base,
 			      struct ioapic_domain_cfg *cfg);
 extern int mp_unregister_ioapic(u32 gsi_base);
+extern int mp_ioapic_registered(u32 gsi_base);
 extern int mp_irqdomain_map(struct irq_domain *domain, unsigned int virq,
 			    irq_hw_number_t hwirq);
 extern void mp_irqdomain_unmap(struct irq_domain *domain, unsigned int virq);
diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c
index bc61a5c0e7b6..f27b33194e06 100644
--- a/arch/x86/kernel/acpi/boot.c
+++ b/arch/x86/kernel/acpi/boot.c
@@ -825,6 +825,28 @@ int acpi_unregister_ioapic(acpi_handle handle, u32 gsi_base)
 }
 EXPORT_SYMBOL(acpi_unregister_ioapic);
 
+/**
+ * acpi_ioapic_registered - Check whether IOAPIC assoicatied with @gsi_base
+ *			    has been registered
+ * @handle:	ACPI handle of the IOAPIC deivce
+ * @gsi_base:	GSI base associated with the IOAPIC
+ *
+ * Assume caller holds some type of lock to serialize acpi_ioapic_registered()
+ * with acpi_register_ioapic()/acpi_unregister_ioapic().
+ */
+int acpi_ioapic_registered(acpi_handle handle, u32 gsi_base)
+{
+	int ret = 0;
+
+#ifdef CONFIG_ACPI_HOTPLUG_IOAPIC
+	down_read(&acpi_ioapic_rwsem);
+	ret  = mp_ioapic_registered(gsi_base);
+	up_read(&acpi_ioapic_rwsem);
+#endif
+
+	return ret;
+}
+
 static int __init acpi_parse_sbf(struct acpi_table_header *table)
 {
 	struct acpi_table_boot *sb;
diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
index 9e5432ee2070..df3858de3031 100644
--- a/arch/x86/kernel/apic/io_apic.c
+++ b/arch/x86/kernel/apic/io_apic.c
@@ -4018,6 +4018,17 @@ int mp_unregister_ioapic(u32 gsi_base)
 	return 0;
 }
 
+int mp_ioapic_registered(u32 gsi_base)
+{
+	int ioapic;
+
+	for_each_ioapic(ioapic)
+		if (ioapics[ioapic].gsi_config.gsi_base == gsi_base)
+			return 1;
+
+	return 0;
+}
+
 int mp_irqdomain_map(struct irq_domain *domain, unsigned int virq,
 		     irq_hw_number_t hwirq)
 {
diff --git a/include/linux/acpi.h b/include/linux/acpi.h
index 77cecd0350e7..be6be0f91337 100644
--- a/include/linux/acpi.h
+++ b/include/linux/acpi.h
@@ -152,6 +152,7 @@ int acpi_get_ioapic_id(acpi_handle handle, u32 gsi_base, u64 *phys_addr);
 
 int acpi_register_ioapic(acpi_handle handle, u64 phys_addr, u32 gsi_base);
 int acpi_unregister_ioapic(acpi_handle handle, u32 gsi_base);
+int acpi_ioapic_registered(acpi_handle handle, u32 gsi_base);
 void acpi_irq_stats_init(void);
 extern u32 acpi_irq_handled;
 extern u32 acpi_irq_not_handled;
-- 
1.7.10.4

[Patch v7 18/18] x86, irq, ACPI: Implement ACPI driver to support IOAPIC hotplug

[Jiang Liu]

Enable support of IOAPIC hotplug by:
1) reintroducing ACPI based IOAPIC driver
2) enhance pci_root driver to hook hotplug events

The ACPI IOAPIC driver is always enabled if all of ACPI, PCI and IOAPIC
are enabled.

Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
---
 drivers/acpi/Kconfig    |    6 ++
 drivers/acpi/Makefile   |    1 +
 drivers/acpi/internal.h |    7 ++
 drivers/acpi/ioapic.c   |  235 +++++++++++++++++++++++++++++++++++++++++++++++
 drivers/acpi/pci_root.c |    3 +
 5 files changed, 252 insertions(+)
 create mode 100644 drivers/acpi/ioapic.c

diff --git a/drivers/acpi/Kconfig b/drivers/acpi/Kconfig
index b23fe37f67c0..127bc2fc43be 100644
--- a/drivers/acpi/Kconfig
+++ b/drivers/acpi/Kconfig
@@ -315,6 +315,12 @@ config ACPI_HOTPLUG_MEMORY
 	  To compile this driver as a module, choose M here:
 	  the module will be called acpi_memhotplug.
 
+config ACPI_HOTPLUG_IOAPIC
+	bool
+	depends on PCI
+	depends on X86_IO_APIC
+	default y
+
 config ACPI_SBS
 	tristate "Smart Battery System"
 	depends on X86
diff --git a/drivers/acpi/Makefile b/drivers/acpi/Makefile
index c3b2fcb729f3..3b283a232f7a 100644
--- a/drivers/acpi/Makefile
+++ b/drivers/acpi/Makefile
@@ -69,6 +69,7 @@ obj-$(CONFIG_ACPI_PROCESSOR)	+= processor.o
 obj-y				+= container.o
 obj-$(CONFIG_ACPI_THERMAL)	+= thermal.o
 obj-y				+= acpi_memhotplug.o
+obj-$(CONFIG_ACPI_HOTPLUG_IOAPIC) += ioapic.o
 obj-$(CONFIG_ACPI_BATTERY)	+= battery.o
 obj-$(CONFIG_ACPI_SBS)		+= sbshc.o
 obj-$(CONFIG_ACPI_SBS)		+= sbs.o
diff --git a/drivers/acpi/internal.h b/drivers/acpi/internal.h
index 447f6d679b29..73efdcb8d8ad 100644
--- a/drivers/acpi/internal.h
+++ b/drivers/acpi/internal.h
@@ -35,6 +35,13 @@ void acpi_int340x_thermal_init(void);
 int acpi_sysfs_init(void);
 void acpi_container_init(void);
 void acpi_memory_hotplug_init(void);
+#ifdef	CONFIG_ACPI_HOTPLUG_IOAPIC
+int acpi_ioapic_add(struct acpi_pci_root *root);
+int acpi_ioapic_remove(struct acpi_pci_root *root);
+#else
+static inline int acpi_ioapic_add(struct acpi_pci_root *root) { return 0; }
+static inline int acpi_ioapic_remove(struct acpi_pci_root *root) { return 0; }
+#endif
 #ifdef CONFIG_ACPI_DOCK
 void register_dock_dependent_device(struct acpi_device *adev,
 				    acpi_handle dshandle);
diff --git a/drivers/acpi/ioapic.c b/drivers/acpi/ioapic.c
new file mode 100644
index 000000000000..9a631a61e270
--- /dev/null
+++ b/drivers/acpi/ioapic.c
@@ -0,0 +1,235 @@
+/*
+ * IOAPIC/IOxAPIC/IOSAPIC driver
+ *
+ * Copyright (C) 2009 Fujitsu Limited.
+ * (c) Copyright 2009 Hewlett-Packard Development Company, L.P.
+ *
+ * Copyright (C) 2014 Intel Corporation
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * Based on original drivers/pci/ioapic.c
+ *	Yinghai Lu <yinghai@kernel.org>
+ *	Jiang Liu <jiang.liu@intel.com>
+ */
+
+/*
+ * This driver manages I/O APICs added by hotplug after boot.
+ * We try to claim all I/O APIC devices, but those present at boot were
+ * registered when we parsed the ACPI MADT.
+ */
+
+#define pr_fmt(fmt) "ACPI : IOAPIC: " fmt
+
+#include <linux/slab.h>
+#include <linux/acpi.h>
+#include <linux/pci.h>
+#include <acpi/acpi.h>
+
+struct acpi_pci_ioapic {
+	acpi_handle	root_handle;
+	acpi_handle	handle;
+	u32		gsi_base;
+	struct resource	res;
+	struct pci_dev	*pdev;
+	struct list_head list;
+};
+
+static LIST_HEAD(ioapic_list);
+static DEFINE_MUTEX(ioapic_list_lock);
+
+static acpi_status setup_res(struct acpi_resource *acpi_res, void *data)
+{
+	struct resource *res = data;
+
+	memset(res, 0, sizeof(*res));
+	if (acpi_dev_resource_memory(acpi_res, res)) {
+		res->flags &= IORESOURCE_MEM;
+		if (res->flags)
+			return AE_OK;
+	} else if (acpi_dev_resource_address_space(acpi_res, res)) {
+		struct acpi_resource_address64 addr;
+
+		res->flags &= IORESOURCE_MEM;
+		if (res->flags &&
+		    ACPI_SUCCESS(acpi_resource_to_address64(acpi_res, &addr)) &&
+		    addr.info.mem.caching != ACPI_PREFETCHABLE_MEMORY) {
+			res->start += addr.translation_offset;
+			res->end += addr.translation_offset;
+			return AE_OK;
+		}
+	}
+	res->flags = 0;
+
+	return AE_OK;
+}
+
+static bool acpi_is_ioapic(acpi_handle handle, char **type)
+{
+	acpi_status status;
+	struct acpi_device_info *info;
+	char *hid = NULL;
+	bool match = false;
+
+	if (!acpi_has_method(handle, "_GSB"))
+		return false;
+
+	status = acpi_get_object_info(handle, &info);
+	if (ACPI_SUCCESS(status)) {
+		if (info->valid & ACPI_VALID_HID)
+			hid = info->hardware_id.string;
+		if (hid) {
+			if (strcmp(hid, "ACPI0009") == 0) {
+				*type = "IOxAPIC";
+				match = true;
+			} else if (strcmp(hid, "ACPI000A") == 0) {
+				*type = "IOAPIC";
+				match = true;
+			}
+		}
+		kfree(info);
+	}
+
+	return match;
+}
+
+static acpi_status handle_ioapic_add(acpi_handle handle, u32 lvl,
+				     void *context, void **rv)
+{
+	acpi_status status;
+	unsigned long long gsi_base;
+	struct acpi_pci_ioapic *ioapic;
+	struct pci_dev *dev = NULL;
+	struct resource *res = NULL;
+	char *type = NULL;
+
+	if (!acpi_is_ioapic(handle, &type))
+		return AE_OK;
+
+	mutex_lock(&ioapic_list_lock);
+	list_for_each_entry(ioapic, &ioapic_list, list)
+		if (ioapic->handle == handle) {
+			mutex_unlock(&ioapic_list_lock);
+			return AE_OK;
+		}
+
+	status = acpi_evaluate_integer(handle, "_GSB", NULL, &gsi_base);
+	if (ACPI_FAILURE(status)) {
+		acpi_handle_warn(handle, "failed to evaluate _GSB method\n");
+		goto exit;
+	}
+
+	ioapic = kzalloc(sizeof(*ioapic), GFP_KERNEL);
+	if (!ioapic) {
+		pr_err("cannot allocate memory for new IOAPIC\n");
+		goto exit;
+	} else {
+		ioapic->root_handle = (acpi_handle)context;
+		ioapic->handle = handle;
+		ioapic->gsi_base = (u32)gsi_base;
+		ioapic->res.flags = IORESOURCE_UNSET;
+	}
+
+	if (acpi_ioapic_registered(handle, (u32)gsi_base))
+		goto done;
+
+	dev = acpi_get_pci_dev(handle);
+	if (dev && pci_resource_len(dev, 0)) {
+		if (pci_enable_device(dev) < 0)
+			goto exit_put;
+		pci_set_master(dev);
+		if (pci_request_region(dev, 0, type))
+			goto exit_disable;
+		res = &dev->resource[0];
+		ioapic->pdev = dev;
+	} else {
+		pci_dev_put(dev);
+		dev = NULL;
+
+		res = &ioapic->res;
+		acpi_walk_resources(handle, METHOD_NAME__CRS, setup_res, res);
+		if (res->flags == IORESOURCE_UNSET) {
+			acpi_handle_warn(handle, "failed to get resource\n");
+			goto exit_free;
+		} else if (request_resource(&iomem_resource, res)) {
+			acpi_handle_warn(handle, "failed to insert resource\n");
+			goto exit_free;
+		}
+	}
+
+	if (acpi_register_ioapic(handle, res->start, (u32)gsi_base)) {
+		acpi_handle_warn(handle, "failed to register IOAPIC\n");
+		goto exit_release;
+	}
+done:
+	list_add(&ioapic->list, &ioapic_list);
+	mutex_unlock(&ioapic_list_lock);
+
+	if (dev)
+		dev_info(&dev->dev, "%s at %pR, GSI %u\n",
+			 type, res, (u32)gsi_base);
+	else
+		acpi_handle_info(handle, "%s at %pR, GSI %u\n",
+				 type, res, (u32)gsi_base);
+
+	return AE_OK;
+
+exit_release:
+	if (dev)
+		pci_release_region(dev, 0);
+	else
+		release_resource(res);
+exit_disable:
+	if (dev)
+		pci_disable_device(dev);
+exit_put:
+	if (dev)
+		pci_dev_put(dev);
+exit_free:
+	kfree(ioapic);
+exit:
+	mutex_unlock(&ioapic_list_lock);
+	*(acpi_status *)rv = AE_ERROR;
+	return AE_OK;
+}
+
+int acpi_ioapic_add(struct acpi_pci_root *root)
+{
+	acpi_status status, retval = AE_OK;
+
+	status = acpi_walk_namespace(ACPI_TYPE_DEVICE, root->device->handle,
+				     UINT_MAX, handle_ioapic_add, NULL,
+				     root->device->handle, (void **)&retval);
+
+	return ACPI_SUCCESS(status) && ACPI_SUCCESS(retval) ? 0 : -ENODEV;
+}
+
+int acpi_ioapic_remove(struct acpi_pci_root *root)
+{
+	int retval = 0;
+	struct acpi_pci_ioapic *ioapic, *tmp;
+
+	mutex_lock(&ioapic_list_lock);
+	list_for_each_entry_safe(ioapic, tmp, &ioapic_list, list) {
+		if (root->device->handle != ioapic->root_handle)
+			continue;
+
+		if (acpi_unregister_ioapic(ioapic->handle, ioapic->gsi_base))
+			retval = -EBUSY;
+
+		if (ioapic->pdev) {
+			pci_release_region(ioapic->pdev, 0);
+			pci_disable_device(ioapic->pdev);
+			pci_dev_put(ioapic->pdev);
+		} else if (ioapic->res.flags != IORESOURCE_UNSET) {
+			release_resource(&ioapic->res);
+		}
+		list_del(&ioapic->list);
+		kfree(ioapic);
+	}
+	mutex_unlock(&ioapic_list_lock);
+
+	return retval;
+}
diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c
index cd4de7e038ea..f3f77689bafc 100644
--- a/drivers/acpi/pci_root.c
+++ b/drivers/acpi/pci_root.c
@@ -614,6 +614,7 @@ static int acpi_pci_root_add(struct acpi_device *device,
 	if (system_state != SYSTEM_BOOTING) {
 		pcibios_resource_survey_bus(root->bus);
 		pci_assign_unassigned_root_bus_resources(root->bus);
+		acpi_ioapic_add(root);
 	}
 
 	pci_lock_rescan_remove();
@@ -634,6 +635,8 @@ static void acpi_pci_root_remove(struct acpi_device *device)
 
 	pci_stop_root_bus(root->bus);
 
+	WARN_ON(acpi_ioapic_remove(root));
+
 	device_set_run_wake(root->bus->bridge, false);
 	pci_acpi_remove_bus_pm_notifier(device);
 
-- 
1.7.10.4

Re: [Patch v7 05/18] ACPI: Correct return value of acpi_dev_resource_address_space()

[Rafael J. Wysocki]

On Monday, October 27, 2014 01:21:35 PM Jiang Liu wrote:
> Change acpi_dev_resource_address_space() to return failure if the
> acpi_resource structure can't be converted to an ACPI address64
> structure, so caller could correctly detect failure.
> 
> Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>

ACK

> ---
>  drivers/acpi/resource.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/acpi/resource.c b/drivers/acpi/resource.c
> index 2ba8f02ced36..782a0d15c25f 100644
> --- a/drivers/acpi/resource.c
> +++ b/drivers/acpi/resource.c
> @@ -200,7 +200,7 @@ bool acpi_dev_resource_address_space(struct acpi_resource *ares,
>  
>  	status = acpi_resource_to_address64(ares, &addr);
>  	if (ACPI_FAILURE(status))
> -		return true;
> +		return false;
>  
>  	res->start = addr.minimum;
>  	res->end = addr.maximum;
> 

-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.

Re: [Patch v7 05/18] ACPI: Correct return value of acpi_dev_resource_address_space()

[Bjorn Helgaas]

On Sun, Oct 26, 2014 at 11:21 PM, Jiang Liu <jiang.liu@linux.intel.com> wrote:
> Change acpi_dev_resource_address_space() to return failure if the
> acpi_resource structure can't be converted to an ACPI address64
> structure, so caller could correctly detect failure.

Fixes: 046d9ce6820e ("ACPI: Move device resources interpretation code
from PNP to ACPI core")

Possible stable candidate?  046d9ce6820e appeared in v3.8.

> Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
> ---
>  drivers/acpi/resource.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/acpi/resource.c b/drivers/acpi/resource.c
> index 2ba8f02ced36..782a0d15c25f 100644
> --- a/drivers/acpi/resource.c
> +++ b/drivers/acpi/resource.c
> @@ -200,7 +200,7 @@ bool acpi_dev_resource_address_space(struct acpi_resource *ares,
>
>         status = acpi_resource_to_address64(ares, &addr);
>         if (ACPI_FAILURE(status))
> -               return true;
> +               return false;
>
>         res->start = addr.minimum;
>         res->end = addr.maximum;
> --
> 1.7.10.4
>

Re: [Patch v7 05/18] ACPI: Correct return value of acpi_dev_resource_address_space()

[Rafael J. Wysocki]

On Monday, October 27, 2014 04:49:06 PM Bjorn Helgaas wrote:
> On Sun, Oct 26, 2014 at 11:21 PM, Jiang Liu <jiang.liu@linux.intel.com> wrote:
> > Change acpi_dev_resource_address_space() to return failure if the
> > acpi_resource structure can't be converted to an ACPI address64
> > structure, so caller could correctly detect failure.
> 
> Fixes: 046d9ce6820e ("ACPI: Move device resources interpretation code
> from PNP to ACPI core")
> 
> Possible stable candidate?  046d9ce6820e appeared in v3.8.

Well, are we seeing problems with it?

> > Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
> > ---
> >  drivers/acpi/resource.c |    2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/acpi/resource.c b/drivers/acpi/resource.c
> > index 2ba8f02ced36..782a0d15c25f 100644
> > --- a/drivers/acpi/resource.c
> > +++ b/drivers/acpi/resource.c
> > @@ -200,7 +200,7 @@ bool acpi_dev_resource_address_space(struct acpi_resource *ares,
> >
> >         status = acpi_resource_to_address64(ares, &addr);
> >         if (ACPI_FAILURE(status))
> > -               return true;
> > +               return false;
> >
> >         res->start = addr.minimum;
> >         res->end = addr.maximum;
> > --
> > 1.7.10.4
> >
> --
> To unsubscribe from this list: send the line "unsubscribe linux-pci" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.

Re: [Patch v7 05/18] ACPI: Correct return value of acpi_dev_resource_address_space()

[Bjorn Helgaas]

On Mon, Oct 27, 2014 at 5:11 PM, Rafael J. Wysocki <rjw@rjwysocki.net> wrote:
> On Monday, October 27, 2014 04:49:06 PM Bjorn Helgaas wrote:
>> On Sun, Oct 26, 2014 at 11:21 PM, Jiang Liu <jiang.liu@linux.intel.com> wrote:
>> > Change acpi_dev_resource_address_space() to return failure if the
>> > acpi_resource structure can't be converted to an ACPI address64
>> > structure, so caller could correctly detect failure.
>>
>> Fixes: 046d9ce6820e ("ACPI: Move device resources interpretation code
>> from PNP to ACPI core")
>>
>> Possible stable candidate?  046d9ce6820e appeared in v3.8.
>
> Well, are we seeing problems with it?

Not to my knowledge.  I only raised the possibility because it seems
like we are seeing more and more things backported to stable.  But if
we we're not seeing a problem, I agree we probably shouldn't push it.

>> > Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
>> > ---
>> >  drivers/acpi/resource.c |    2 +-
>> >  1 file changed, 1 insertion(+), 1 deletion(-)
>> >
>> > diff --git a/drivers/acpi/resource.c b/drivers/acpi/resource.c
>> > index 2ba8f02ced36..782a0d15c25f 100644
>> > --- a/drivers/acpi/resource.c
>> > +++ b/drivers/acpi/resource.c
>> > @@ -200,7 +200,7 @@ bool acpi_dev_resource_address_space(struct acpi_resource *ares,
>> >
>> >         status = acpi_resource_to_address64(ares, &addr);
>> >         if (ACPI_FAILURE(status))
>> > -               return true;
>> > +               return false;
>> >
>> >         res->start = addr.minimum;
>> >         res->end = addr.maximum;
>> > --
>> > 1.7.10.4
>> >
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-pci" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>
> --
> I speak only for myself.
> Rafael J. Wysocki, Intel Open Source Technology Center.

Re: [Patch v7 01/18] ACPI, irq: fix regression casued by 6b9fb7082409

[Pavel Machek]

Hi!

> When IOAPIC is disabled, acpi_gsi_to_irq() should return gsi directly
> instead of calling mp_map_gsi_to_irq() to translate gsi to IRQ by IOAPIC.
> It fixes https://bugzilla.kernel.org/show_bug.cgi?id=84381.
> 
> Reported-by: Thomas Richter <thor@math.tu-berlin.de>
> Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
> Cc: rui.zhang@intel.com
> Cc: <stable@vger.kernel.org> # 3.17
> ---
>  arch/x86/kernel/acpi/boot.c |   13 +++++++++----
>  1 file changed, 9 insertions(+), 4 deletions(-)
> 
> diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c
> index b436fc735aa4..eceba9d9e116 100644
> --- a/arch/x86/kernel/acpi/boot.c
> +++ b/arch/x86/kernel/acpi/boot.c
> @@ -604,14 +604,19 @@ void __init acpi_pic_sci_set_trigger(unsigned int irq, u16 trigger)
>  
>  int acpi_gsi_to_irq(u32 gsi, unsigned int *irqp)
>  {
> -	int irq = mp_map_gsi_to_irq(gsi, IOAPIC_MAP_ALLOC | IOAPIC_MAP_CHECK);
> +	int irq;
>  
> -	if (irq >= 0) {
> +	if (acpi_irq_model == ACPI_IRQ_MODEL_PIC) {
> +		*irqp = gsi;

As you have multiple return points, anyway, I'd do return 0; here

> +	} else {
> +		irq = mp_map_gsi_to_irq(gsi,
> +					IOAPIC_MAP_ALLOC | IOAPIC_MAP_CHECK);
> +		if (irq < 0)
> +			return -1;
>  		*irqp = irq;
> -		return 0;
>  	}

...so that one level of nesting is avoided, and there are no problems
with fiting to 80 colums.

With that,

Acked-by: Pavel Machek <pavel@ucw.cz>

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [Patch v7 17/18] x86, irq: Introduce helper to check whether an IOAPIC has been registered

[Pavel Machek]

On Mon 2014-10-27 13:21:47, Jiang Liu wrote:
> Introduce acpi_ioapic_registered() to check whether an IOAPIC has already
> been registered, it will be used when enabling IOAPIC hotplug.
> 
> Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>

Nothing obviously wrong.

Acked-by: Pavel Machek <pavel@ucw.cz>

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: [Patch v7 01/18] ACPI, irq: fix regression casued by 6b9fb7082409

[Bjorn Helgaas]

The subject should describe the change you're making.  "Fix regression
caused by xxx" doesn't say anything about what the patch does.

On Sun, Oct 26, 2014 at 11:21 PM, Jiang Liu <jiang.liu@linux.intel.com> wrote:
> When IOAPIC is disabled, acpi_gsi_to_irq() should return gsi directly
> instead of calling mp_map_gsi_to_irq() to translate gsi to IRQ by IOAPIC.

I don't know what the convention is or if there even is one, but "GSI"
is even more of an acronym than "IRQ", so I'd capitalize it.

> It fixes https://bugzilla.kernel.org/show_bug.cgi?id=84381.

A line like the following may be useful to those who have to backport this.

Fixes: 6b9fb7082409 ("x86, ACPI, irq: Consolidate algorithm of mapping
(ioapic, pin) to IRQ number")

> Reported-by: Thomas Richter <thor@math.tu-berlin.de>
> Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com>
> Cc: rui.zhang@intel.com
> Cc: <stable@vger.kernel.org> # 3.17
> ---
>  arch/x86/kernel/acpi/boot.c |   13 +++++++++----
>  1 file changed, 9 insertions(+), 4 deletions(-)
>
> diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c
> index b436fc735aa4..eceba9d9e116 100644
> --- a/arch/x86/kernel/acpi/boot.c
> +++ b/arch/x86/kernel/acpi/boot.c
> @@ -604,14 +604,19 @@ void __init acpi_pic_sci_set_trigger(unsigned int irq, u16 trigger)
>
>  int acpi_gsi_to_irq(u32 gsi, unsigned int *irqp)
>  {
> -       int irq = mp_map_gsi_to_irq(gsi, IOAPIC_MAP_ALLOC | IOAPIC_MAP_CHECK);
> +       int irq;
>
> -       if (irq >= 0) {
> +       if (acpi_irq_model == ACPI_IRQ_MODEL_PIC) {
> +               *irqp = gsi;
> +       } else {
> +               irq = mp_map_gsi_to_irq(gsi,
> +                                       IOAPIC_MAP_ALLOC | IOAPIC_MAP_CHECK);
> +               if (irq < 0)
> +                       return -1;
>                 *irqp = irq;
> -               return 0;
>         }
>
> -       return -1;
> +       return 0;
>  }
>  EXPORT_SYMBOL_GPL(acpi_gsi_to_irq);
>
> --
> 1.7.10.4
>

Re: [Patch v7 01/18] ACPI, irq: fix regression casued by 6b9fb7082409

[Thomas Gleixner]

On Tue, 28 Oct 2014, Bjorn Helgaas wrote:
> The subject should describe the change you're making.  "Fix regression
> caused by xxx" doesn't say anything about what the patch does.
> 
> On Sun, Oct 26, 2014 at 11:21 PM, Jiang Liu <jiang.liu@linux.intel.com> wrote:
> > When IOAPIC is disabled, acpi_gsi_to_irq() should return gsi directly
> > instead of calling mp_map_gsi_to_irq() to translate gsi to IRQ by IOAPIC.
> 
> I don't know what the convention is or if there even is one, but "GSI"
> is even more of an acronym than "IRQ", so I'd capitalize it.
> 
> > It fixes https://bugzilla.kernel.org/show_bug.cgi?id=84381.
> 
> A line like the following may be useful to those who have to backport this.
> 
> Fixes: 6b9fb7082409 ("x86, ACPI, irq: Consolidate algorithm of mapping
> (ioapic, pin) to IRQ number")

The patch is already queued in tip x86/urgent with a proper changelog.

http://git.kernel.org/cgit/linux/kernel/git/tip/tip.git/commit/?id=961b6a7003acec4f9d70dabc1a253b783cb74272

Thanks,

	tglx

Re: [Patch v7 14/18] x86, irq, ACPI: Introduce a rwsem to protect IOAPIC operations from hotplug

[Thomas Gleixner]

On Mon, 27 Oct 2014, Jiang Liu wrote:
> We are going to support ACPI based IOAPIC hotplug, so introduce a rwsem
> to protect IOAPIC data structures from IOAPIC hotplug. We choose to
> serialize in ACPI instead of in the IOAPIC core because:
> 1) currently we are only plan to support ACPI based IOAPIC hotplug
> 2) it's much more cleaner and easier
> 3) It does't affect IOAPIC discovered by devicetree, SFI and mpparse.

I had a last intensive look at this series as I was about to merge
it. So I looked at the locking rules here again
 
> +/*
> + * Locks related to IOAPIC hotplug
> + * Hotplug side:
> + * 	->lock_device_hotplug()	//device_hotplug_lock
> + *		->acpi_ioapic_rwsem
> + *			->ioapic_lock
> + * Interrupt mapping side:
> + *	->acpi_ioapic_rwsem
> + *		->ioapic_mutex
> + *			->ioapic_lock
> + */

This looks sane, but I cannot figure out at all why this needs to be a
rwsem.

> +static DECLARE_RWSEM(acpi_ioapic_rwsem);

I think it should be a simple mutex because the rwsem does not protect
against concurrent execution what taken for read.

And the site which takes it for write is in the early boot process
where nothing runs in parallel AFAICT.

Thanks,

	tglx

Re: [Patch v7 14/18] x86, irq, ACPI: Introduce a rwsem to protect IOAPIC operations from hotplug

[Jiang Liu]

On 2014/11/2 2:59, Thomas Gleixner wrote:
> On Mon, 27 Oct 2014, Jiang Liu wrote:
>> We are going to support ACPI based IOAPIC hotplug, so introduce a rwsem
>> to protect IOAPIC data structures from IOAPIC hotplug. We choose to
>> serialize in ACPI instead of in the IOAPIC core because:
>> 1) currently we are only plan to support ACPI based IOAPIC hotplug
>> 2) it's much more cleaner and easier
>> 3) It does't affect IOAPIC discovered by devicetree, SFI and mpparse.
> 
> I had a last intensive look at this series as I was about to merge
> it. So I looked at the locking rules here again
>  
>> +/*
>> + * Locks related to IOAPIC hotplug
>> + * Hotplug side:
>> + * 	->lock_device_hotplug()	//device_hotplug_lock
>> + *		->acpi_ioapic_rwsem
>> + *			->ioapic_lock
>> + * Interrupt mapping side:
>> + *	->acpi_ioapic_rwsem
>> + *		->ioapic_mutex
>> + *			->ioapic_lock
>> + */
> 
> This looks sane, but I cannot figure out at all why this needs to be a
> rwsem.
> 
>> +static DECLARE_RWSEM(acpi_ioapic_rwsem);
> 
> I think it should be a simple mutex because the rwsem does not protect
> against concurrent execution what taken for read.
> 
> And the site which takes it for write is in the early boot process
> where nothing runs in parallel AFAICT.
Hi Thomas,
	You are right. It's not on hot path, so a mutex is better than
a rwsem here. I will send out an updated version soon.
Regards!
Gerry
> 
> Thanks,
> 
> 	tglx
>