From 9ec4fa271faf2db3b8e1419c998da1ca6b094eb6 Mon Sep 17 00:00:00 2001
From: Yinghai Lu <yinghai@kernel.org>
Date: Mon, 27 Apr 2009 17:57:18 -0700
Subject: irq, cpumask: correct CPUMASKS_OFFSTACK typo and fix fallout

CPUMASKS_OFFSTACK is not defined anywhere (it is CPUMASK_OFFSTACK).
It is a typo and init_allocate_desc_masks() is called before it set
affinity to all cpus...

Split init_alloc_desc_masks() into all_desc_masks() and init_desc_masks().

Also use CPUMASK_OFFSTACK in alloc_desc_masks().

[ Impact: fix smp_affinity copying/setup when moving irq_desc between CPUs ]

Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Acked-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Suresh Siddha <suresh.b.siddha@intel.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
LKML-Reference: <49F6546E.3040406@kernel.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 include/linux/irq.h | 27 ++++++++++++++++++---------
 1 file changed, 18 insertions(+), 9 deletions(-)

(limited to 'include/linux/irq.h')

diff --git a/include/linux/irq.h b/include/linux/irq.h
index b7cbeed972e..c4953cf27e5 100644
--- a/include/linux/irq.h
+++ b/include/linux/irq.h
@@ -424,27 +424,25 @@ extern int set_irq_msi(unsigned int irq, struct msi_desc *entry);
 
 #ifdef CONFIG_SMP
 /**
- * init_alloc_desc_masks - allocate cpumasks for irq_desc
+ * alloc_desc_masks - allocate cpumasks for irq_desc
  * @desc:	pointer to irq_desc struct
  * @cpu:	cpu which will be handling the cpumasks
  * @boot:	true if need bootmem
  *
  * Allocates affinity and pending_mask cpumask if required.
  * Returns true if successful (or not required).
- * Side effect: affinity has all bits set, pending_mask has all bits clear.
  */
-static inline bool init_alloc_desc_masks(struct irq_desc *desc, int cpu,
+static inline bool alloc_desc_masks(struct irq_desc *desc, int cpu,
 								bool boot)
 {
+#ifdef CONFIG_CPUMASK_OFFSTACK
 	int node;
 
 	if (boot) {
 		alloc_bootmem_cpumask_var(&desc->affinity);
-		cpumask_setall(desc->affinity);
 
 #ifdef CONFIG_GENERIC_PENDING_IRQ
 		alloc_bootmem_cpumask_var(&desc->pending_mask);
-		cpumask_clear(desc->pending_mask);
 #endif
 		return true;
 	}
@@ -453,18 +451,25 @@ static inline bool init_alloc_desc_masks(struct irq_desc *desc, int cpu,
 
 	if (!alloc_cpumask_var_node(&desc->affinity, GFP_ATOMIC, node))
 		return false;
-	cpumask_setall(desc->affinity);
 
 #ifdef CONFIG_GENERIC_PENDING_IRQ
 	if (!alloc_cpumask_var_node(&desc->pending_mask, GFP_ATOMIC, node)) {
 		free_cpumask_var(desc->affinity);
 		return false;
 	}
-	cpumask_clear(desc->pending_mask);
+#endif
 #endif
 	return true;
 }
 
+static inline void init_desc_masks(struct irq_desc *desc)
+{
+	cpumask_setall(desc->affinity);
+#ifdef CONFIG_GENERIC_PENDING_IRQ
+	cpumask_clear(desc->pending_mask);
+#endif
+}
+
 /**
  * init_copy_desc_masks - copy cpumasks for irq_desc
  * @old_desc:	pointer to old irq_desc struct
@@ -478,7 +483,7 @@ static inline bool init_alloc_desc_masks(struct irq_desc *desc, int cpu,
 static inline void init_copy_desc_masks(struct irq_desc *old_desc,
 					struct irq_desc *new_desc)
 {
-#ifdef CONFIG_CPUMASKS_OFFSTACK
+#ifdef CONFIG_CPUMASK_OFFSTACK
 	cpumask_copy(new_desc->affinity, old_desc->affinity);
 
 #ifdef CONFIG_GENERIC_PENDING_IRQ
@@ -499,12 +504,16 @@ static inline void free_desc_masks(struct irq_desc *old_desc,
 
 #else /* !CONFIG_SMP */
 
-static inline bool init_alloc_desc_masks(struct irq_desc *desc, int cpu,
+static inline bool alloc_desc_masks(struct irq_desc *desc, int cpu,
 								bool boot)
 {
 	return true;
 }
 
+static inline void init_desc_masks(struct irq_desc *desc)
+{
+}
+
 static inline void init_copy_desc_masks(struct irq_desc *old_desc,
 					struct irq_desc *new_desc)
 {
-- 
cgit v1.2.3-70-g09d2


From fcef5911c7ea89b80d5bfc727f402f37c9eefd57 Mon Sep 17 00:00:00 2001
From: Yinghai Lu <yinghai@kernel.org>
Date: Mon, 27 Apr 2009 17:58:23 -0700
Subject: x86/irq: remove leftover code from NUMA_MIGRATE_IRQ_DESC

The original feature of migrating irq_desc dynamic was too fragile
and was causing problems: it caused crashes on systems with lots of
cards with MSI-X when user-space irq-balancer was enabled.

We now have new patches that create irq_desc according to device
numa node. This patch removes the leftover bits of the dynamic balancer.

[ Impact: remove dead code ]

Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Suresh Siddha <suresh.b.siddha@intel.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Rusty Russell <rusty@rustcorp.com.au>
LKML-Reference: <49F654AF.8000808@kernel.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 arch/x86/Kconfig                  | 10 -------
 arch/x86/configs/x86_64_defconfig |  1 -
 arch/x86/kernel/apic/io_apic.c    | 56 +++------------------------------------
 include/linux/irq.h               | 10 -------
 kernel/irq/Makefile               |  2 +-
 kernel/irq/chip.c                 | 12 ++-------
 kernel/irq/handle.c               |  9 ++-----
 kernel/irq/numa_migrate.c         |  2 --
 8 files changed, 9 insertions(+), 93 deletions(-)

(limited to 'include/linux/irq.h')

diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index df9e885eee1..e1b2543f8ed 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -274,16 +274,6 @@ config SPARSE_IRQ
 
 	  If you don't know what to do here, say N.
 
-config NUMA_MIGRATE_IRQ_DESC
-	bool "Move irq desc when changing irq smp_affinity"
-	depends on SPARSE_IRQ && NUMA
-	depends on BROKEN
-	default n
-	---help---
-	  This enables moving irq_desc to cpu/node that irq will use handled.
-
-	  If you don't know what to do here, say N.
-
 config X86_MPPARSE
 	bool "Enable MPS table" if ACPI
 	default y
diff --git a/arch/x86/configs/x86_64_defconfig b/arch/x86/configs/x86_64_defconfig
index 9fe5d212ab4..27b8ce0f590 100644
--- a/arch/x86/configs/x86_64_defconfig
+++ b/arch/x86/configs/x86_64_defconfig
@@ -195,7 +195,6 @@ CONFIG_HIGH_RES_TIMERS=y
 CONFIG_GENERIC_CLOCKEVENTS_BUILD=y
 CONFIG_SMP=y
 CONFIG_SPARSE_IRQ=y
-# CONFIG_NUMA_MIGRATE_IRQ_DESC is not set
 CONFIG_X86_FIND_SMP_CONFIG=y
 CONFIG_X86_MPPARSE=y
 # CONFIG_X86_ELAN is not set
diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
index 30da617d18e..9fbf0f7ec7e 100644
--- a/arch/x86/kernel/apic/io_apic.c
+++ b/arch/x86/kernel/apic/io_apic.c
@@ -148,9 +148,6 @@ struct irq_cfg {
 	unsigned move_cleanup_count;
 	u8 vector;
 	u8 move_in_progress : 1;
-#ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC
-	u8 move_desc_pending : 1;
-#endif
 };
 
 /* irq_cfg is indexed by the sum of all RTEs in all I/O APICs. */
@@ -254,8 +251,7 @@ int arch_init_chip_data(struct irq_desc *desc, int cpu)
 	return 0;
 }
 
-#ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC
-
+/* for move_irq_desc */
 static void
 init_copy_irq_2_pin(struct irq_cfg *old_cfg, struct irq_cfg *cfg, int cpu)
 {
@@ -356,19 +352,7 @@ void arch_free_chip_data(struct irq_desc *old_desc, struct irq_desc *desc)
 		old_desc->chip_data = NULL;
 	}
 }
-
-static void
-set_extra_move_desc(struct irq_desc *desc, const struct cpumask *mask)
-{
-	struct irq_cfg *cfg = desc->chip_data;
-
-	if (!cfg->move_in_progress) {
-		/* it means that domain is not changed */
-		if (!cpumask_intersects(desc->affinity, mask))
-			cfg->move_desc_pending = 1;
-	}
-}
-#endif
+/* end for move_irq_desc */
 
 #else
 static struct irq_cfg *irq_cfg(unsigned int irq)
@@ -378,13 +362,6 @@ static struct irq_cfg *irq_cfg(unsigned int irq)
 
 #endif
 
-#ifndef CONFIG_NUMA_MIGRATE_IRQ_DESC
-static inline void
-set_extra_move_desc(struct irq_desc *desc, const struct cpumask *mask)
-{
-}
-#endif
-
 struct io_apic {
 	unsigned int index;
 	unsigned int unused[3];
@@ -592,9 +569,6 @@ set_desc_affinity(struct irq_desc *desc, const struct cpumask *mask)
 	if (assign_irq_vector(irq, cfg, mask))
 		return BAD_APICID;
 
-	/* check that before desc->addinity get updated */
-	set_extra_move_desc(desc, mask);
-
 	cpumask_copy(desc->affinity, mask);
 
 	return apic->cpu_mask_to_apicid_and(desc->affinity, cfg->domain);
@@ -2393,8 +2367,6 @@ migrate_ioapic_irq_desc(struct irq_desc *desc, const struct cpumask *mask)
 	if (assign_irq_vector(irq, cfg, mask))
 		return;
 
-	set_extra_move_desc(desc, mask);
-
 	dest = apic->cpu_mask_to_apicid_and(cfg->domain, mask);
 
 	irte.vector = cfg->vector;
@@ -2491,34 +2463,14 @@ static void irq_complete_move(struct irq_desc **descp)
 	struct irq_cfg *cfg = desc->chip_data;
 	unsigned vector, me;
 
-	if (likely(!cfg->move_in_progress)) {
-#ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC
-		if (likely(!cfg->move_desc_pending))
-			return;
-
-		/* domain has not changed, but affinity did */
-		me = smp_processor_id();
-		if (cpumask_test_cpu(me, desc->affinity)) {
-			*descp = desc = move_irq_desc(desc, me);
-			/* get the new one */
-			cfg = desc->chip_data;
-			cfg->move_desc_pending = 0;
-		}
-#endif
+	if (likely(!cfg->move_in_progress))
 		return;
-	}
 
 	vector = ~get_irq_regs()->orig_ax;
 	me = smp_processor_id();
 
-	if (vector == cfg->vector && cpumask_test_cpu(me, cfg->domain)) {
-#ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC
-		*descp = desc = move_irq_desc(desc, me);
-		/* get the new one */
-		cfg = desc->chip_data;
-#endif
+	if (vector == cfg->vector && cpumask_test_cpu(me, cfg->domain))
 		send_cleanup_vector(cfg);
-	}
 }
 #else
 static inline void irq_complete_move(struct irq_desc **descp) {}
diff --git a/include/linux/irq.h b/include/linux/irq.h
index c4953cf27e5..2a34cd6281d 100644
--- a/include/linux/irq.h
+++ b/include/linux/irq.h
@@ -212,16 +212,6 @@ extern struct irq_desc *move_irq_desc(struct irq_desc *old_desc, int cpu);
 
 extern struct irq_desc *irq_to_desc_alloc_cpu(unsigned int irq, int cpu);
 
-static inline struct irq_desc *
-irq_remap_to_desc(unsigned int irq, struct irq_desc *desc)
-{
-#ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC
-	return irq_to_desc(irq);
-#else
-	return desc;
-#endif
-}
-
 /*
  * Migration helpers for obsolete names, they will go away:
  */
diff --git a/kernel/irq/Makefile b/kernel/irq/Makefile
index 3394f8f5296..2f065277f8e 100644
--- a/kernel/irq/Makefile
+++ b/kernel/irq/Makefile
@@ -3,5 +3,5 @@ obj-y := handle.o manage.o spurious.o resend.o chip.o devres.o
 obj-$(CONFIG_GENERIC_IRQ_PROBE) += autoprobe.o
 obj-$(CONFIG_PROC_FS) += proc.o
 obj-$(CONFIG_GENERIC_PENDING_IRQ) += migration.o
-obj-$(CONFIG_NUMA_MIGRATE_IRQ_DESC) += numa_migrate.o
+obj-$(CONFIG_SPARSE_IRQ) += numa_migrate.o
 obj-$(CONFIG_PM_SLEEP) += pm.o
diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c
index c687ba4363f..13c68e71b72 100644
--- a/kernel/irq/chip.c
+++ b/kernel/irq/chip.c
@@ -359,7 +359,6 @@ handle_level_irq(unsigned int irq, struct irq_desc *desc)
 
 	spin_lock(&desc->lock);
 	mask_ack_irq(desc, irq);
-	desc = irq_remap_to_desc(irq, desc);
 
 	if (unlikely(desc->status & IRQ_INPROGRESS))
 		goto out_unlock;
@@ -438,7 +437,6 @@ handle_fasteoi_irq(unsigned int irq, struct irq_desc *desc)
 	desc->status &= ~IRQ_INPROGRESS;
 out:
 	desc->chip->eoi(irq);
-	desc = irq_remap_to_desc(irq, desc);
 
 	spin_unlock(&desc->lock);
 }
@@ -475,7 +473,6 @@ handle_edge_irq(unsigned int irq, struct irq_desc *desc)
 		    !desc->action)) {
 		desc->status |= (IRQ_PENDING | IRQ_MASKED);
 		mask_ack_irq(desc, irq);
-		desc = irq_remap_to_desc(irq, desc);
 		goto out_unlock;
 	}
 	kstat_incr_irqs_this_cpu(irq, desc);
@@ -483,7 +480,6 @@ handle_edge_irq(unsigned int irq, struct irq_desc *desc)
 	/* Start handling the irq */
 	if (desc->chip->ack)
 		desc->chip->ack(irq);
-	desc = irq_remap_to_desc(irq, desc);
 
 	/* Mark the IRQ currently in progress.*/
 	desc->status |= IRQ_INPROGRESS;
@@ -544,10 +540,8 @@ handle_percpu_irq(unsigned int irq, struct irq_desc *desc)
 	if (!noirqdebug)
 		note_interrupt(irq, desc, action_ret);
 
-	if (desc->chip->eoi) {
+	if (desc->chip->eoi)
 		desc->chip->eoi(irq);
-		desc = irq_remap_to_desc(irq, desc);
-	}
 }
 
 void
@@ -582,10 +576,8 @@ __set_irq_handler(unsigned int irq, irq_flow_handler_t handle, int is_chained,
 
 	/* Uninstall? */
 	if (handle == handle_bad_irq) {
-		if (desc->chip != &no_irq_chip) {
+		if (desc->chip != &no_irq_chip)
 			mask_ack_irq(desc, irq);
-			desc = irq_remap_to_desc(irq, desc);
-		}
 		desc->status |= IRQ_DISABLED;
 		desc->depth = 1;
 	}
diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c
index 882c7980010..3e0cbc44bd7 100644
--- a/kernel/irq/handle.c
+++ b/kernel/irq/handle.c
@@ -458,11 +458,8 @@ unsigned int __do_IRQ(unsigned int irq)
 		/*
 		 * No locking required for CPU-local interrupts:
 		 */
-		if (desc->chip->ack) {
+		if (desc->chip->ack)
 			desc->chip->ack(irq);
-			/* get new one */
-			desc = irq_remap_to_desc(irq, desc);
-		}
 		if (likely(!(desc->status & IRQ_DISABLED))) {
 			action_ret = handle_IRQ_event(irq, desc->action);
 			if (!noirqdebug)
@@ -473,10 +470,8 @@ unsigned int __do_IRQ(unsigned int irq)
 	}
 
 	spin_lock(&desc->lock);
-	if (desc->chip->ack) {
+	if (desc->chip->ack)
 		desc->chip->ack(irq);
-		desc = irq_remap_to_desc(irq, desc);
-	}
 	/*
 	 * REPLAY is when Linux resends an IRQ that was dropped earlier
 	 * WAITING is used by probe to mark irqs that are being tested
diff --git a/kernel/irq/numa_migrate.c b/kernel/irq/numa_migrate.c
index 5760d725162..ce72bc3f4ce 100644
--- a/kernel/irq/numa_migrate.c
+++ b/kernel/irq/numa_migrate.c
@@ -97,9 +97,7 @@ static struct irq_desc *__real_move_irq_desc(struct irq_desc *old_desc,
 
 	/* free the old one */
 	free_one_irq_desc(old_desc, desc);
-	spin_unlock(&old_desc->lock);
 	kfree(old_desc);
-	spin_lock(&desc->lock);
 
 	return desc;
 
-- 
cgit v1.2.3-70-g09d2


From d5dedd4507d307eb3f35f21b6e16f336fdc0d82a Mon Sep 17 00:00:00 2001
From: Yinghai Lu <yinghai@kernel.org>
Date: Mon, 27 Apr 2009 17:59:21 -0700
Subject: irq: change ->set_affinity() to return status

according to Ingo, change set_affinity() in irq_chip should return int,
because that way we can handle failure cases in a much cleaner way, in
the genirq layer.

v2: fix two typos

[ Impact: extend API ]

Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Suresh Siddha <suresh.b.siddha@intel.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: linux-arch@vger.kernel.org
LKML-Reference: <49F654E9.4070809@kernel.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 arch/alpha/kernel/sys_dp264.c         |  8 +++--
 arch/alpha/kernel/sys_titan.c         |  4 ++-
 arch/arm/common/gic.c                 |  4 ++-
 arch/cris/arch-v32/kernel/irq.c       |  4 ++-
 arch/ia64/hp/sim/hpsim_irq.c          |  3 +-
 arch/ia64/kernel/iosapic.c            | 10 +++---
 arch/ia64/kernel/msi_ia64.c           | 16 +++++----
 arch/ia64/sn/kernel/irq.c             |  4 ++-
 arch/ia64/sn/kernel/msi_sn.c          |  8 +++--
 arch/mips/cavium-octeon/octeon-irq.c  |  8 +++--
 arch/mips/include/asm/irq.h           |  2 +-
 arch/mips/kernel/irq-gic.c            |  5 +--
 arch/mips/mti-malta/malta-smtc.c      |  4 ++-
 arch/mips/sibyte/bcm1480/irq.c        |  8 +++--
 arch/mips/sibyte/sb1250/irq.c         |  8 +++--
 arch/parisc/kernel/irq.c              |  6 ++--
 arch/powerpc/platforms/pseries/xics.c | 12 ++++---
 arch/powerpc/sysdev/mpic.c            |  4 ++-
 arch/sparc/kernel/irq_64.c            | 12 +++++--
 arch/x86/kernel/apic/io_apic.c        | 64 ++++++++++++++++++++++-------------
 drivers/parisc/iosapic.c              |  6 ++--
 drivers/xen/events.c                  | 12 ++++---
 include/linux/irq.h                   |  2 +-
 23 files changed, 140 insertions(+), 74 deletions(-)

(limited to 'include/linux/irq.h')

diff --git a/arch/alpha/kernel/sys_dp264.c b/arch/alpha/kernel/sys_dp264.c
index 9c9d1fd4155..5bd5259324b 100644
--- a/arch/alpha/kernel/sys_dp264.c
+++ b/arch/alpha/kernel/sys_dp264.c
@@ -176,22 +176,26 @@ cpu_set_irq_affinity(unsigned int irq, cpumask_t affinity)
 	}
 }
 
-static void
+static int
 dp264_set_affinity(unsigned int irq, const struct cpumask *affinity)
 { 
 	spin_lock(&dp264_irq_lock);
 	cpu_set_irq_affinity(irq, *affinity);
 	tsunami_update_irq_hw(cached_irq_mask);
 	spin_unlock(&dp264_irq_lock);
+
+	return 0;
 }
 
-static void
+static int
 clipper_set_affinity(unsigned int irq, const struct cpumask *affinity)
 { 
 	spin_lock(&dp264_irq_lock);
 	cpu_set_irq_affinity(irq - 16, *affinity);
 	tsunami_update_irq_hw(cached_irq_mask);
 	spin_unlock(&dp264_irq_lock);
+
+	return 0;
 }
 
 static struct hw_interrupt_type dp264_irq_type = {
diff --git a/arch/alpha/kernel/sys_titan.c b/arch/alpha/kernel/sys_titan.c
index 27f840a4ad3..8dd239ebdb9 100644
--- a/arch/alpha/kernel/sys_titan.c
+++ b/arch/alpha/kernel/sys_titan.c
@@ -157,13 +157,15 @@ titan_cpu_set_irq_affinity(unsigned int irq, cpumask_t affinity)
 
 }
 
-static void
+static int
 titan_set_irq_affinity(unsigned int irq, const struct cpumask *affinity)
 { 
 	spin_lock(&titan_irq_lock);
 	titan_cpu_set_irq_affinity(irq - 16, *affinity);
 	titan_update_irq_hw(titan_cached_irq_mask);
 	spin_unlock(&titan_irq_lock);
+
+	return 0;
 }
 
 static void
diff --git a/arch/arm/common/gic.c b/arch/arm/common/gic.c
index c6884ba1d5e..90f6b7f52d4 100644
--- a/arch/arm/common/gic.c
+++ b/arch/arm/common/gic.c
@@ -109,7 +109,7 @@ static void gic_unmask_irq(unsigned int irq)
 }
 
 #ifdef CONFIG_SMP
-static void gic_set_cpu(unsigned int irq, const struct cpumask *mask_val)
+static int gic_set_cpu(unsigned int irq, const struct cpumask *mask_val)
 {
 	void __iomem *reg = gic_dist_base(irq) + GIC_DIST_TARGET + (gic_irq(irq) & ~3);
 	unsigned int shift = (irq % 4) * 8;
@@ -122,6 +122,8 @@ static void gic_set_cpu(unsigned int irq, const struct cpumask *mask_val)
 	val |= 1 << (cpu + shift);
 	writel(val, reg);
 	spin_unlock(&irq_controller_lock);
+
+	return 0;
 }
 #endif
 
diff --git a/arch/cris/arch-v32/kernel/irq.c b/arch/cris/arch-v32/kernel/irq.c
index df3925cb1c7..d70b445f4a8 100644
--- a/arch/cris/arch-v32/kernel/irq.c
+++ b/arch/cris/arch-v32/kernel/irq.c
@@ -325,12 +325,14 @@ static void end_crisv32_irq(unsigned int irq)
 {
 }
 
-void set_affinity_crisv32_irq(unsigned int irq, const struct cpumask *dest)
+int set_affinity_crisv32_irq(unsigned int irq, const struct cpumask *dest)
 {
 	unsigned long flags;
 	spin_lock_irqsave(&irq_lock, flags);
 	irq_allocations[irq - FIRST_IRQ].mask = *dest;
 	spin_unlock_irqrestore(&irq_lock, flags);
+
+	return 0;
 }
 
 static struct irq_chip crisv32_irq_type = {
diff --git a/arch/ia64/hp/sim/hpsim_irq.c b/arch/ia64/hp/sim/hpsim_irq.c
index cc0a3182db3..acb5047ab57 100644
--- a/arch/ia64/hp/sim/hpsim_irq.c
+++ b/arch/ia64/hp/sim/hpsim_irq.c
@@ -21,9 +21,10 @@ hpsim_irq_noop (unsigned int irq)
 {
 }
 
-static void
+static int
 hpsim_set_affinity_noop(unsigned int a, const struct cpumask *b)
 {
+	return 0;
 }
 
 static struct hw_interrupt_type irq_type_hp_sim = {
diff --git a/arch/ia64/kernel/iosapic.c b/arch/ia64/kernel/iosapic.c
index 166e0d839fa..f92cef47bf8 100644
--- a/arch/ia64/kernel/iosapic.c
+++ b/arch/ia64/kernel/iosapic.c
@@ -329,7 +329,7 @@ unmask_irq (unsigned int irq)
 }
 
 
-static void
+static int
 iosapic_set_affinity(unsigned int irq, const struct cpumask *mask)
 {
 #ifdef CONFIG_SMP
@@ -343,15 +343,15 @@ iosapic_set_affinity(unsigned int irq, const struct cpumask *mask)
 
 	cpu = cpumask_first_and(cpu_online_mask, mask);
 	if (cpu >= nr_cpu_ids)
-		return;
+		return -1;
 
 	if (irq_prepare_move(irq, cpu))
-		return;
+		return -1;
 
 	dest = cpu_physical_id(cpu);
 
 	if (!iosapic_intr_info[irq].count)
-		return;			/* not an IOSAPIC interrupt */
+		return -1;			/* not an IOSAPIC interrupt */
 
 	set_irq_affinity_info(irq, dest, redir);
 
@@ -376,7 +376,9 @@ iosapic_set_affinity(unsigned int irq, const struct cpumask *mask)
 		iosapic_write(iosapic, IOSAPIC_RTE_HIGH(rte_index), high32);
 		iosapic_write(iosapic, IOSAPIC_RTE_LOW(rte_index), low32);
 	}
+
 #endif
+	return 0;
 }
 
 /*
diff --git a/arch/ia64/kernel/msi_ia64.c b/arch/ia64/kernel/msi_ia64.c
index 2b15e233f7f..0f8ade9331b 100644
--- a/arch/ia64/kernel/msi_ia64.c
+++ b/arch/ia64/kernel/msi_ia64.c
@@ -12,7 +12,7 @@
 static struct irq_chip	ia64_msi_chip;
 
 #ifdef CONFIG_SMP
-static void ia64_set_msi_irq_affinity(unsigned int irq,
+static int ia64_set_msi_irq_affinity(unsigned int irq,
 				      const cpumask_t *cpu_mask)
 {
 	struct msi_msg msg;
@@ -20,10 +20,10 @@ static void ia64_set_msi_irq_affinity(unsigned int irq,
 	int cpu = first_cpu(*cpu_mask);
 
 	if (!cpu_online(cpu))
-		return;
+		return -1;
 
 	if (irq_prepare_move(irq, cpu))
-		return;
+		return -1;
 
 	read_msi_msg(irq, &msg);
 
@@ -39,6 +39,8 @@ static void ia64_set_msi_irq_affinity(unsigned int irq,
 
 	write_msi_msg(irq, &msg);
 	cpumask_copy(irq_desc[irq].affinity, cpumask_of(cpu));
+
+	return 0;
 }
 #endif /* CONFIG_SMP */
 
@@ -130,17 +132,17 @@ void arch_teardown_msi_irq(unsigned int irq)
 
 #ifdef CONFIG_DMAR
 #ifdef CONFIG_SMP
-static void dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask)
+static int dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask)
 {
 	struct irq_cfg *cfg = irq_cfg + irq;
 	struct msi_msg msg;
 	int cpu = cpumask_first(mask);
 
 	if (!cpu_online(cpu))
-		return;
+		return -1;
 
 	if (irq_prepare_move(irq, cpu))
-		return;
+		return -1;
 
 	dmar_msi_read(irq, &msg);
 
@@ -151,6 +153,8 @@ static void dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask)
 
 	dmar_msi_write(irq, &msg);
 	cpumask_copy(irq_desc[irq].affinity, mask);
+
+	return 0;
 }
 #endif /* CONFIG_SMP */
 
diff --git a/arch/ia64/sn/kernel/irq.c b/arch/ia64/sn/kernel/irq.c
index 66fd705e82c..764f26abac0 100644
--- a/arch/ia64/sn/kernel/irq.c
+++ b/arch/ia64/sn/kernel/irq.c
@@ -227,7 +227,7 @@ finish_up:
 	return new_irq_info;
 }
 
-static void sn_set_affinity_irq(unsigned int irq, const struct cpumask *mask)
+static int sn_set_affinity_irq(unsigned int irq, const struct cpumask *mask)
 {
 	struct sn_irq_info *sn_irq_info, *sn_irq_info_safe;
 	nasid_t nasid;
@@ -239,6 +239,8 @@ static void sn_set_affinity_irq(unsigned int irq, const struct cpumask *mask)
 	list_for_each_entry_safe(sn_irq_info, sn_irq_info_safe,
 				 sn_irq_lh[irq], list)
 		(void)sn_retarget_vector(sn_irq_info, nasid, slice);
+
+	return 0;
 }
 
 #ifdef CONFIG_SMP
diff --git a/arch/ia64/sn/kernel/msi_sn.c b/arch/ia64/sn/kernel/msi_sn.c
index 81e428943d7..fbbfb970120 100644
--- a/arch/ia64/sn/kernel/msi_sn.c
+++ b/arch/ia64/sn/kernel/msi_sn.c
@@ -151,7 +151,7 @@ int sn_setup_msi_irq(struct pci_dev *pdev, struct msi_desc *entry)
 }
 
 #ifdef CONFIG_SMP
-static void sn_set_msi_irq_affinity(unsigned int irq,
+static int sn_set_msi_irq_affinity(unsigned int irq,
 				    const struct cpumask *cpu_mask)
 {
 	struct msi_msg msg;
@@ -168,7 +168,7 @@ static void sn_set_msi_irq_affinity(unsigned int irq,
 	cpu = cpumask_first(cpu_mask);
 	sn_irq_info = sn_msi_info[irq].sn_irq_info;
 	if (sn_irq_info == NULL || sn_irq_info->irq_int_bit >= 0)
-		return;
+		return -1;
 
 	/*
 	 * Release XIO resources for the old MSI PCI address
@@ -189,7 +189,7 @@ static void sn_set_msi_irq_affinity(unsigned int irq,
 	new_irq_info = sn_retarget_vector(sn_irq_info, nasid, slice);
 	sn_msi_info[irq].sn_irq_info = new_irq_info;
 	if (new_irq_info == NULL)
-		return;
+		return -1;
 
 	/*
 	 * Map the xio address into bus space
@@ -206,6 +206,8 @@ static void sn_set_msi_irq_affinity(unsigned int irq,
 
 	write_msi_msg(irq, &msg);
 	cpumask_copy(irq_desc[irq].affinity, cpu_mask);
+
+	return 0;
 }
 #endif /* CONFIG_SMP */
 
diff --git a/arch/mips/cavium-octeon/octeon-irq.c b/arch/mips/cavium-octeon/octeon-irq.c
index 1c19af8daa6..d3a0c8154be 100644
--- a/arch/mips/cavium-octeon/octeon-irq.c
+++ b/arch/mips/cavium-octeon/octeon-irq.c
@@ -177,7 +177,7 @@ static void octeon_irq_ciu0_disable(unsigned int irq)
 }
 
 #ifdef CONFIG_SMP
-static void octeon_irq_ciu0_set_affinity(unsigned int irq, const struct cpumask *dest)
+static int octeon_irq_ciu0_set_affinity(unsigned int irq, const struct cpumask *dest)
 {
 	int cpu;
 	int bit = irq - OCTEON_IRQ_WORKQ0;	/* Bit 0-63 of EN0 */
@@ -199,6 +199,8 @@ static void octeon_irq_ciu0_set_affinity(unsigned int irq, const struct cpumask
 	 */
 	cvmx_read_csr(CVMX_CIU_INTX_EN0(cvmx_get_core_num() * 2));
 	write_unlock(&octeon_irq_ciu0_rwlock);
+
+	return 0;
 }
 #endif
 
@@ -292,7 +294,7 @@ static void octeon_irq_ciu1_disable(unsigned int irq)
 }
 
 #ifdef CONFIG_SMP
-static void octeon_irq_ciu1_set_affinity(unsigned int irq, const struct cpumask *dest)
+static int octeon_irq_ciu1_set_affinity(unsigned int irq, const struct cpumask *dest)
 {
 	int cpu;
 	int bit = irq - OCTEON_IRQ_WDOG0;	/* Bit 0-63 of EN1 */
@@ -315,6 +317,8 @@ static void octeon_irq_ciu1_set_affinity(unsigned int irq, const struct cpumask
 	 */
 	cvmx_read_csr(CVMX_CIU_INTX_EN1(cvmx_get_core_num() * 2 + 1));
 	write_unlock(&octeon_irq_ciu1_rwlock);
+
+	return 0;
 }
 #endif
 
diff --git a/arch/mips/include/asm/irq.h b/arch/mips/include/asm/irq.h
index 3214ade02d1..4f1eed107b0 100644
--- a/arch/mips/include/asm/irq.h
+++ b/arch/mips/include/asm/irq.h
@@ -49,7 +49,7 @@ static inline void smtc_im_ack_irq(unsigned int irq)
 #ifdef CONFIG_MIPS_MT_SMTC_IRQAFF
 #include <linux/cpumask.h>
 
-extern void plat_set_irq_affinity(unsigned int irq,
+extern int plat_set_irq_affinity(unsigned int irq,
 				  const struct cpumask *affinity);
 extern void smtc_forward_irq(unsigned int irq);
 
diff --git a/arch/mips/kernel/irq-gic.c b/arch/mips/kernel/irq-gic.c
index 87deb8f6c45..3f43c2e3aa5 100644
--- a/arch/mips/kernel/irq-gic.c
+++ b/arch/mips/kernel/irq-gic.c
@@ -155,7 +155,7 @@ static void gic_unmask_irq(unsigned int irq)
 
 static DEFINE_SPINLOCK(gic_lock);
 
-static void gic_set_affinity(unsigned int irq, const struct cpumask *cpumask)
+static int gic_set_affinity(unsigned int irq, const struct cpumask *cpumask)
 {
 	cpumask_t	tmp = CPU_MASK_NONE;
 	unsigned long	flags;
@@ -166,7 +166,7 @@ static void gic_set_affinity(unsigned int irq, const struct cpumask *cpumask)
 
 	cpumask_and(&tmp, cpumask, cpu_online_mask);
 	if (cpus_empty(tmp))
-		return;
+		return -1;
 
 	/* Assumption : cpumask refers to a single CPU */
 	spin_lock_irqsave(&gic_lock, flags);
@@ -190,6 +190,7 @@ static void gic_set_affinity(unsigned int irq, const struct cpumask *cpumask)
 	cpumask_copy(irq_desc[irq].affinity, cpumask);
 	spin_unlock_irqrestore(&gic_lock, flags);
 
+	return 0;
 }
 #endif
 
diff --git a/arch/mips/mti-malta/malta-smtc.c b/arch/mips/mti-malta/malta-smtc.c
index 5ba31888fef..499ffe5475d 100644
--- a/arch/mips/mti-malta/malta-smtc.c
+++ b/arch/mips/mti-malta/malta-smtc.c
@@ -114,7 +114,7 @@ struct plat_smp_ops msmtc_smp_ops = {
  */
 
 
-void plat_set_irq_affinity(unsigned int irq, const struct cpumask *affinity)
+int plat_set_irq_affinity(unsigned int irq, const struct cpumask *affinity)
 {
 	cpumask_t tmask;
 	int cpu = 0;
@@ -156,5 +156,7 @@ void plat_set_irq_affinity(unsigned int irq, const struct cpumask *affinity)
 
 	/* Do any generic SMTC IRQ affinity setup */
 	smtc_set_irq_affinity(irq, tmask);
+
+	return 0;
 }
 #endif /* CONFIG_MIPS_MT_SMTC_IRQAFF */
diff --git a/arch/mips/sibyte/bcm1480/irq.c b/arch/mips/sibyte/bcm1480/irq.c
index 352352b3cb2..4f256a131bf 100644
--- a/arch/mips/sibyte/bcm1480/irq.c
+++ b/arch/mips/sibyte/bcm1480/irq.c
@@ -50,7 +50,7 @@ static void enable_bcm1480_irq(unsigned int irq);
 static void disable_bcm1480_irq(unsigned int irq);
 static void ack_bcm1480_irq(unsigned int irq);
 #ifdef CONFIG_SMP
-static void bcm1480_set_affinity(unsigned int irq, const struct cpumask *mask);
+static int bcm1480_set_affinity(unsigned int irq, const struct cpumask *mask);
 #endif
 
 #ifdef CONFIG_PCI
@@ -109,7 +109,7 @@ void bcm1480_unmask_irq(int cpu, int irq)
 }
 
 #ifdef CONFIG_SMP
-static void bcm1480_set_affinity(unsigned int irq, const struct cpumask *mask)
+static int bcm1480_set_affinity(unsigned int irq, const struct cpumask *mask)
 {
 	int i = 0, old_cpu, cpu, int_on, k;
 	u64 cur_ints;
@@ -119,7 +119,7 @@ static void bcm1480_set_affinity(unsigned int irq, const struct cpumask *mask)
 
 	if (cpumask_weight(mask) != 1) {
 		printk("attempted to set irq affinity for irq %d to multiple CPUs\n", irq);
-		return;
+		return -1;
 	}
 	i = cpumask_first(mask);
 
@@ -155,6 +155,8 @@ static void bcm1480_set_affinity(unsigned int irq, const struct cpumask *mask)
 	}
 	spin_unlock(&bcm1480_imr_lock);
 	spin_unlock_irqrestore(&desc->lock, flags);
+
+	return 0;
 }
 #endif
 
diff --git a/arch/mips/sibyte/sb1250/irq.c b/arch/mips/sibyte/sb1250/irq.c
index c08ff582da6..e389507f1f9 100644
--- a/arch/mips/sibyte/sb1250/irq.c
+++ b/arch/mips/sibyte/sb1250/irq.c
@@ -50,7 +50,7 @@ static void enable_sb1250_irq(unsigned int irq);
 static void disable_sb1250_irq(unsigned int irq);
 static void ack_sb1250_irq(unsigned int irq);
 #ifdef CONFIG_SMP
-static void sb1250_set_affinity(unsigned int irq, const struct cpumask *mask);
+static int sb1250_set_affinity(unsigned int irq, const struct cpumask *mask);
 #endif
 
 #ifdef CONFIG_SIBYTE_HAS_LDT
@@ -103,7 +103,7 @@ void sb1250_unmask_irq(int cpu, int irq)
 }
 
 #ifdef CONFIG_SMP
-static void sb1250_set_affinity(unsigned int irq, const struct cpumask *mask)
+static int sb1250_set_affinity(unsigned int irq, const struct cpumask *mask)
 {
 	int i = 0, old_cpu, cpu, int_on;
 	u64 cur_ints;
@@ -114,7 +114,7 @@ static void sb1250_set_affinity(unsigned int irq, const struct cpumask *mask)
 
 	if (cpumask_weight(mask) > 1) {
 		printk("attempted to set irq affinity for irq %d to multiple CPUs\n", irq);
-		return;
+		return -1;
 	}
 
 	/* Convert logical CPU to physical CPU */
@@ -146,6 +146,8 @@ static void sb1250_set_affinity(unsigned int irq, const struct cpumask *mask)
 	}
 	spin_unlock(&sb1250_imr_lock);
 	spin_unlock_irqrestore(&desc->lock, flags);
+
+	return 0;
 }
 #endif
 
diff --git a/arch/parisc/kernel/irq.c b/arch/parisc/kernel/irq.c
index 4ea4229d765..8007f1e6572 100644
--- a/arch/parisc/kernel/irq.c
+++ b/arch/parisc/kernel/irq.c
@@ -130,15 +130,17 @@ int cpu_check_affinity(unsigned int irq, const struct cpumask *dest)
 	return cpu_dest;
 }
 
-static void cpu_set_affinity_irq(unsigned int irq, const struct cpumask *dest)
+static int cpu_set_affinity_irq(unsigned int irq, const struct cpumask *dest)
 {
 	int cpu_dest;
 
 	cpu_dest = cpu_check_affinity(irq, dest);
 	if (cpu_dest < 0)
-		return;
+		return -1;
 
 	cpumask_copy(&irq_desc[irq].affinity, dest);
+
+	return 0;
 }
 #endif
 
diff --git a/arch/powerpc/platforms/pseries/xics.c b/arch/powerpc/platforms/pseries/xics.c
index 80b513449f4..be3581a8c29 100644
--- a/arch/powerpc/platforms/pseries/xics.c
+++ b/arch/powerpc/platforms/pseries/xics.c
@@ -333,7 +333,7 @@ static void xics_eoi_lpar(unsigned int virq)
 	lpar_xirr_info_set((0xff << 24) | irq);
 }
 
-static void xics_set_affinity(unsigned int virq, const struct cpumask *cpumask)
+static int xics_set_affinity(unsigned int virq, const struct cpumask *cpumask)
 {
 	unsigned int irq;
 	int status;
@@ -342,14 +342,14 @@ static void xics_set_affinity(unsigned int virq, const struct cpumask *cpumask)
 
 	irq = (unsigned int)irq_map[virq].hwirq;
 	if (irq == XICS_IPI || irq == XICS_IRQ_SPURIOUS)
-		return;
+		return -1;
 
 	status = rtas_call(ibm_get_xive, 1, 3, xics_status, irq);
 
 	if (status) {
 		printk(KERN_ERR "%s: ibm,get-xive irq=%u returns %d\n",
 			__func__, irq, status);
-		return;
+		return -1;
 	}
 
 	/*
@@ -363,7 +363,7 @@ static void xics_set_affinity(unsigned int virq, const struct cpumask *cpumask)
 		printk(KERN_WARNING
 			"%s: No online cpus in the mask %s for irq %d\n",
 			__func__, cpulist, virq);
-		return;
+		return -1;
 	}
 
 	status = rtas_call(ibm_set_xive, 3, 1, NULL,
@@ -372,8 +372,10 @@ static void xics_set_affinity(unsigned int virq, const struct cpumask *cpumask)
 	if (status) {
 		printk(KERN_ERR "%s: ibm,set-xive irq=%u returns %d\n",
 			__func__, irq, status);
-		return;
+		return -1;
 	}
+
+	return 0;
 }
 
 static struct irq_chip xics_pic_direct = {
diff --git a/arch/powerpc/sysdev/mpic.c b/arch/powerpc/sysdev/mpic.c
index 21b95670159..f4cbd15cf22 100644
--- a/arch/powerpc/sysdev/mpic.c
+++ b/arch/powerpc/sysdev/mpic.c
@@ -807,7 +807,7 @@ static void mpic_end_ipi(unsigned int irq)
 
 #endif /* CONFIG_SMP */
 
-void mpic_set_affinity(unsigned int irq, const struct cpumask *cpumask)
+int mpic_set_affinity(unsigned int irq, const struct cpumask *cpumask)
 {
 	struct mpic *mpic = mpic_from_irq(irq);
 	unsigned int src = mpic_irq_to_hw(irq);
@@ -824,6 +824,8 @@ void mpic_set_affinity(unsigned int irq, const struct cpumask *cpumask)
 		mpic_irq_write(src, MPIC_INFO(IRQ_DESTINATION),
 			       mpic_physmask(cpus_addr(tmp)[0]));
 	}
+
+	return 0;
 }
 
 static unsigned int mpic_type_to_vecpri(struct mpic *mpic, unsigned int type)
diff --git a/arch/sparc/kernel/irq_64.c b/arch/sparc/kernel/irq_64.c
index 5deabe921a4..e5e78f9cfc9 100644
--- a/arch/sparc/kernel/irq_64.c
+++ b/arch/sparc/kernel/irq_64.c
@@ -318,10 +318,12 @@ static void sun4u_irq_enable(unsigned int virt_irq)
 	}
 }
 
-static void sun4u_set_affinity(unsigned int virt_irq,
+static int sun4u_set_affinity(unsigned int virt_irq,
 			       const struct cpumask *mask)
 {
 	sun4u_irq_enable(virt_irq);
+
+	return 0;
 }
 
 /* Don't do anything.  The desc->status check for IRQ_DISABLED in
@@ -377,7 +379,7 @@ static void sun4v_irq_enable(unsigned int virt_irq)
 		       ino, err);
 }
 
-static void sun4v_set_affinity(unsigned int virt_irq,
+static int sun4v_set_affinity(unsigned int virt_irq,
 			       const struct cpumask *mask)
 {
 	unsigned int ino = virt_irq_table[virt_irq].dev_ino;
@@ -388,6 +390,8 @@ static void sun4v_set_affinity(unsigned int virt_irq,
 	if (err != HV_EOK)
 		printk(KERN_ERR "sun4v_intr_settarget(%x,%lu): "
 		       "err(%d)\n", ino, cpuid, err);
+
+	return 0;
 }
 
 static void sun4v_irq_disable(unsigned int virt_irq)
@@ -445,7 +449,7 @@ static void sun4v_virq_enable(unsigned int virt_irq)
 		       dev_handle, dev_ino, err);
 }
 
-static void sun4v_virt_set_affinity(unsigned int virt_irq,
+static int sun4v_virt_set_affinity(unsigned int virt_irq,
 				    const struct cpumask *mask)
 {
 	unsigned long cpuid, dev_handle, dev_ino;
@@ -461,6 +465,8 @@ static void sun4v_virt_set_affinity(unsigned int virt_irq,
 		printk(KERN_ERR "sun4v_vintr_set_target(%lx,%lx,%lu): "
 		       "err(%d)\n",
 		       dev_handle, dev_ino, cpuid, err);
+
+	return 0;
 }
 
 static void sun4v_virq_disable(unsigned int virt_irq)
diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
index 9fbf0f7ec7e..5c7630b40a5 100644
--- a/arch/x86/kernel/apic/io_apic.c
+++ b/arch/x86/kernel/apic/io_apic.c
@@ -574,13 +574,14 @@ set_desc_affinity(struct irq_desc *desc, const struct cpumask *mask)
 	return apic->cpu_mask_to_apicid_and(desc->affinity, cfg->domain);
 }
 
-static void
+static int
 set_ioapic_affinity_irq_desc(struct irq_desc *desc, const struct cpumask *mask)
 {
 	struct irq_cfg *cfg;
 	unsigned long flags;
 	unsigned int dest;
 	unsigned int irq;
+	int ret = -1;
 
 	irq = desc->irq;
 	cfg = desc->chip_data;
@@ -591,18 +592,21 @@ set_ioapic_affinity_irq_desc(struct irq_desc *desc, const struct cpumask *mask)
 		/* Only the high 8 bits are valid. */
 		dest = SET_APIC_LOGICAL_ID(dest);
 		__target_IO_APIC_irq(irq, dest, cfg);
+		ret = 0;
 	}
 	spin_unlock_irqrestore(&ioapic_lock, flags);
+
+	return ret;
 }
 
-static void
+static int
 set_ioapic_affinity_irq(unsigned int irq, const struct cpumask *mask)
 {
 	struct irq_desc *desc;
 
 	desc = irq_to_desc(irq);
 
-	set_ioapic_affinity_irq_desc(desc, mask);
+	return set_ioapic_affinity_irq_desc(desc, mask);
 }
 #endif /* CONFIG_SMP */
 
@@ -2348,24 +2352,25 @@ static int ioapic_retrigger_irq(unsigned int irq)
  * Real vector that is used for interrupting cpu will be coming from
  * the interrupt-remapping table entry.
  */
-static void
+static int
 migrate_ioapic_irq_desc(struct irq_desc *desc, const struct cpumask *mask)
 {
 	struct irq_cfg *cfg;
 	struct irte irte;
 	unsigned int dest;
 	unsigned int irq;
+	int ret = -1;
 
 	if (!cpumask_intersects(mask, cpu_online_mask))
-		return;
+		return ret;
 
 	irq = desc->irq;
 	if (get_irte(irq, &irte))
-		return;
+		return ret;
 
 	cfg = desc->chip_data;
 	if (assign_irq_vector(irq, cfg, mask))
-		return;
+		return ret;
 
 	dest = apic->cpu_mask_to_apicid_and(cfg->domain, mask);
 
@@ -2381,27 +2386,30 @@ migrate_ioapic_irq_desc(struct irq_desc *desc, const struct cpumask *mask)
 		send_cleanup_vector(cfg);
 
 	cpumask_copy(desc->affinity, mask);
+
+	return 0;
 }
 
 /*
  * Migrates the IRQ destination in the process context.
  */
-static void set_ir_ioapic_affinity_irq_desc(struct irq_desc *desc,
+static int set_ir_ioapic_affinity_irq_desc(struct irq_desc *desc,
 					    const struct cpumask *mask)
 {
-	migrate_ioapic_irq_desc(desc, mask);
+	return migrate_ioapic_irq_desc(desc, mask);
 }
-static void set_ir_ioapic_affinity_irq(unsigned int irq,
+static int set_ir_ioapic_affinity_irq(unsigned int irq,
 				       const struct cpumask *mask)
 {
 	struct irq_desc *desc = irq_to_desc(irq);
 
-	set_ir_ioapic_affinity_irq_desc(desc, mask);
+	return set_ir_ioapic_affinity_irq_desc(desc, mask);
 }
 #else
-static inline void set_ir_ioapic_affinity_irq_desc(struct irq_desc *desc,
+static inline int set_ir_ioapic_affinity_irq_desc(struct irq_desc *desc,
 						   const struct cpumask *mask)
 {
+	return 0;
 }
 #endif
 
@@ -3318,7 +3326,7 @@ static int msi_compose_msg(struct pci_dev *pdev, unsigned int irq, struct msi_ms
 }
 
 #ifdef CONFIG_SMP
-static void set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask)
+static int set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask)
 {
 	struct irq_desc *desc = irq_to_desc(irq);
 	struct irq_cfg *cfg;
@@ -3327,7 +3335,7 @@ static void set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask)
 
 	dest = set_desc_affinity(desc, mask);
 	if (dest == BAD_APICID)
-		return;
+		return -1;
 
 	cfg = desc->chip_data;
 
@@ -3339,13 +3347,15 @@ static void set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask)
 	msg.address_lo |= MSI_ADDR_DEST_ID(dest);
 
 	write_msi_msg_desc(desc, &msg);
+
+	return 0;
 }
 #ifdef CONFIG_INTR_REMAP
 /*
  * Migrate the MSI irq to another cpumask. This migration is
  * done in the process context using interrupt-remapping hardware.
  */
-static void
+static int
 ir_set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask)
 {
 	struct irq_desc *desc = irq_to_desc(irq);
@@ -3354,11 +3364,11 @@ ir_set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask)
 	struct irte irte;
 
 	if (get_irte(irq, &irte))
-		return;
+		return -1;
 
 	dest = set_desc_affinity(desc, mask);
 	if (dest == BAD_APICID)
-		return;
+		return -1;
 
 	irte.vector = cfg->vector;
 	irte.dest_id = IRTE_DEST(dest);
@@ -3375,6 +3385,8 @@ ir_set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask)
 	 */
 	if (cfg->move_in_progress)
 		send_cleanup_vector(cfg);
+
+	return 0;
 }
 
 #endif
@@ -3528,7 +3540,7 @@ void arch_teardown_msi_irq(unsigned int irq)
 
 #if defined (CONFIG_DMAR) || defined (CONFIG_INTR_REMAP)
 #ifdef CONFIG_SMP
-static void dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask)
+static int dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask)
 {
 	struct irq_desc *desc = irq_to_desc(irq);
 	struct irq_cfg *cfg;
@@ -3537,7 +3549,7 @@ static void dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask)
 
 	dest = set_desc_affinity(desc, mask);
 	if (dest == BAD_APICID)
-		return;
+		return -1;
 
 	cfg = desc->chip_data;
 
@@ -3549,6 +3561,8 @@ static void dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask)
 	msg.address_lo |= MSI_ADDR_DEST_ID(dest);
 
 	dmar_msi_write(irq, &msg);
+
+	return 0;
 }
 
 #endif /* CONFIG_SMP */
@@ -3582,7 +3596,7 @@ int arch_setup_dmar_msi(unsigned int irq)
 #ifdef CONFIG_HPET_TIMER
 
 #ifdef CONFIG_SMP
-static void hpet_msi_set_affinity(unsigned int irq, const struct cpumask *mask)
+static int hpet_msi_set_affinity(unsigned int irq, const struct cpumask *mask)
 {
 	struct irq_desc *desc = irq_to_desc(irq);
 	struct irq_cfg *cfg;
@@ -3591,7 +3605,7 @@ static void hpet_msi_set_affinity(unsigned int irq, const struct cpumask *mask)
 
 	dest = set_desc_affinity(desc, mask);
 	if (dest == BAD_APICID)
-		return;
+		return -1;
 
 	cfg = desc->chip_data;
 
@@ -3603,6 +3617,8 @@ static void hpet_msi_set_affinity(unsigned int irq, const struct cpumask *mask)
 	msg.address_lo |= MSI_ADDR_DEST_ID(dest);
 
 	hpet_msi_write(irq, &msg);
+
+	return 0;
 }
 
 #endif /* CONFIG_SMP */
@@ -3659,7 +3675,7 @@ static void target_ht_irq(unsigned int irq, unsigned int dest, u8 vector)
 	write_ht_irq_msg(irq, &msg);
 }
 
-static void set_ht_irq_affinity(unsigned int irq, const struct cpumask *mask)
+static int set_ht_irq_affinity(unsigned int irq, const struct cpumask *mask)
 {
 	struct irq_desc *desc = irq_to_desc(irq);
 	struct irq_cfg *cfg;
@@ -3667,11 +3683,13 @@ static void set_ht_irq_affinity(unsigned int irq, const struct cpumask *mask)
 
 	dest = set_desc_affinity(desc, mask);
 	if (dest == BAD_APICID)
-		return;
+		return -1;
 
 	cfg = desc->chip_data;
 
 	target_ht_irq(irq, dest, cfg->vector);
+
+	return 0;
 }
 
 #endif
diff --git a/drivers/parisc/iosapic.c b/drivers/parisc/iosapic.c
index 73348c4047e..4a9cc92d4d1 100644
--- a/drivers/parisc/iosapic.c
+++ b/drivers/parisc/iosapic.c
@@ -702,7 +702,7 @@ static unsigned int iosapic_startup_irq(unsigned int irq)
 }
 
 #ifdef CONFIG_SMP
-static void iosapic_set_affinity_irq(unsigned int irq,
+static int iosapic_set_affinity_irq(unsigned int irq,
 				     const struct cpumask *dest)
 {
 	struct vector_info *vi = iosapic_get_vector(irq);
@@ -712,7 +712,7 @@ static void iosapic_set_affinity_irq(unsigned int irq,
 
 	dest_cpu = cpu_check_affinity(irq, dest);
 	if (dest_cpu < 0)
-		return;
+		return -1;
 
 	cpumask_copy(irq_desc[irq].affinity, cpumask_of(dest_cpu));
 	vi->txn_addr = txn_affinity_addr(irq, dest_cpu);
@@ -724,6 +724,8 @@ static void iosapic_set_affinity_irq(unsigned int irq,
 	iosapic_set_irt_data(vi, &dummy_d0, &d1);
 	iosapic_wr_irt_entry(vi, d0, d1);
 	spin_unlock_irqrestore(&iosapic_lock, flags);
+
+	return 0;
 }
 #endif
 
diff --git a/drivers/xen/events.c b/drivers/xen/events.c
index 30963af5dba..33389880279 100644
--- a/drivers/xen/events.c
+++ b/drivers/xen/events.c
@@ -688,13 +688,13 @@ void rebind_evtchn_irq(int evtchn, int irq)
 }
 
 /* Rebind an evtchn so that it gets delivered to a specific cpu */
-static void rebind_irq_to_cpu(unsigned irq, unsigned tcpu)
+static int rebind_irq_to_cpu(unsigned irq, unsigned tcpu)
 {
 	struct evtchn_bind_vcpu bind_vcpu;
 	int evtchn = evtchn_from_irq(irq);
 
 	if (!VALID_EVTCHN(evtchn))
-		return;
+		return -1;
 
 	/* Send future instances of this interrupt to other vcpu. */
 	bind_vcpu.port = evtchn;
@@ -707,13 +707,15 @@ static void rebind_irq_to_cpu(unsigned irq, unsigned tcpu)
 	 */
 	if (HYPERVISOR_event_channel_op(EVTCHNOP_bind_vcpu, &bind_vcpu) >= 0)
 		bind_evtchn_to_cpu(evtchn, tcpu);
-}
 
+	return 0;
+}
 
-static void set_affinity_irq(unsigned irq, const struct cpumask *dest)
+static int set_affinity_irq(unsigned irq, const struct cpumask *dest)
 {
 	unsigned tcpu = cpumask_first(dest);
-	rebind_irq_to_cpu(irq, tcpu);
+
+	return rebind_irq_to_cpu(irq, tcpu);
 }
 
 int resend_irq_on_evtchn(unsigned int irq)
diff --git a/include/linux/irq.h b/include/linux/irq.h
index 2a34cd6281d..8e4c18b2915 100644
--- a/include/linux/irq.h
+++ b/include/linux/irq.h
@@ -117,7 +117,7 @@ struct irq_chip {
 	void		(*eoi)(unsigned int irq);
 
 	void		(*end)(unsigned int irq);
-	void		(*set_affinity)(unsigned int irq,
+	int		(*set_affinity)(unsigned int irq,
 					const struct cpumask *dest);
 	int		(*retrigger)(unsigned int irq);
 	int		(*set_type)(unsigned int irq, unsigned int flow_type);
-- 
cgit v1.2.3-70-g09d2


From 85ac16d033370caf6f48d743c8dc8103700f5cc5 Mon Sep 17 00:00:00 2001
From: Yinghai Lu <yinghai@kernel.org>
Date: Mon, 27 Apr 2009 18:00:38 -0700
Subject: x86/irq: change irq_desc_alloc() to take node instead of cpu

This simplifies the node awareness of the code. All our allocators
only deal with a NUMA node ID locality not with CPU ids anyway - so
there's no need to maintain (and transform) a CPU id all across the
IRq layer.

v2: keep move_irq_desc related

[ Impact: cleanup, prepare IRQ code to be NUMA-aware ]

Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Suresh Siddha <suresh.b.siddha@intel.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
LKML-Reference: <49F65536.2020300@kernel.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 arch/x86/kernel/apic/io_apic.c | 58 +++++++++++++++++++-----------------------
 arch/x86/lguest/boot.c         |  2 +-
 drivers/pci/intr_remapping.c   | 15 +++++------
 drivers/xen/events.c           |  2 +-
 include/linux/interrupt.h      |  2 +-
 include/linux/irq.h            | 16 +++++-------
 kernel/irq/handle.c            | 28 ++++++++------------
 kernel/irq/internals.h         |  2 +-
 kernel/irq/numa_migrate.c      | 36 +++++++++-----------------
 kernel/softirq.c               |  2 +-
 10 files changed, 66 insertions(+), 97 deletions(-)

(limited to 'include/linux/irq.h')

diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
index 5c7630b40a5..560b887ba27 100644
--- a/arch/x86/kernel/apic/io_apic.c
+++ b/arch/x86/kernel/apic/io_apic.c
@@ -129,12 +129,9 @@ struct irq_pin_list {
 	struct irq_pin_list *next;
 };
 
-static struct irq_pin_list *get_one_free_irq_2_pin(int cpu)
+static struct irq_pin_list *get_one_free_irq_2_pin(int node)
 {
 	struct irq_pin_list *pin;
-	int node;
-
-	node = cpu_to_node(cpu);
 
 	pin = kzalloc_node(sizeof(*pin), GFP_ATOMIC, node);
 
@@ -209,12 +206,9 @@ static struct irq_cfg *irq_cfg(unsigned int irq)
 	return cfg;
 }
 
-static struct irq_cfg *get_one_free_irq_cfg(int cpu)
+static struct irq_cfg *get_one_free_irq_cfg(int node)
 {
 	struct irq_cfg *cfg;
-	int node;
-
-	node = cpu_to_node(cpu);
 
 	cfg = kzalloc_node(sizeof(*cfg), GFP_ATOMIC, node);
 	if (cfg) {
@@ -235,13 +229,13 @@ static struct irq_cfg *get_one_free_irq_cfg(int cpu)
 	return cfg;
 }
 
-int arch_init_chip_data(struct irq_desc *desc, int cpu)
+int arch_init_chip_data(struct irq_desc *desc, int node)
 {
 	struct irq_cfg *cfg;
 
 	cfg = desc->chip_data;
 	if (!cfg) {
-		desc->chip_data = get_one_free_irq_cfg(cpu);
+		desc->chip_data = get_one_free_irq_cfg(node);
 		if (!desc->chip_data) {
 			printk(KERN_ERR "can not alloc irq_cfg\n");
 			BUG_ON(1);
@@ -253,7 +247,7 @@ int arch_init_chip_data(struct irq_desc *desc, int cpu)
 
 /* for move_irq_desc */
 static void
-init_copy_irq_2_pin(struct irq_cfg *old_cfg, struct irq_cfg *cfg, int cpu)
+init_copy_irq_2_pin(struct irq_cfg *old_cfg, struct irq_cfg *cfg, int node)
 {
 	struct irq_pin_list *old_entry, *head, *tail, *entry;
 
@@ -262,7 +256,7 @@ init_copy_irq_2_pin(struct irq_cfg *old_cfg, struct irq_cfg *cfg, int cpu)
 	if (!old_entry)
 		return;
 
-	entry = get_one_free_irq_2_pin(cpu);
+	entry = get_one_free_irq_2_pin(node);
 	if (!entry)
 		return;
 
@@ -272,7 +266,7 @@ init_copy_irq_2_pin(struct irq_cfg *old_cfg, struct irq_cfg *cfg, int cpu)
 	tail		= entry;
 	old_entry	= old_entry->next;
 	while (old_entry) {
-		entry = get_one_free_irq_2_pin(cpu);
+		entry = get_one_free_irq_2_pin(node);
 		if (!entry) {
 			entry = head;
 			while (entry) {
@@ -312,12 +306,12 @@ static void free_irq_2_pin(struct irq_cfg *old_cfg, struct irq_cfg *cfg)
 }
 
 void arch_init_copy_chip_data(struct irq_desc *old_desc,
-				 struct irq_desc *desc, int cpu)
+				 struct irq_desc *desc, int node)
 {
 	struct irq_cfg *cfg;
 	struct irq_cfg *old_cfg;
 
-	cfg = get_one_free_irq_cfg(cpu);
+	cfg = get_one_free_irq_cfg(node);
 
 	if (!cfg)
 		return;
@@ -328,7 +322,7 @@ void arch_init_copy_chip_data(struct irq_desc *old_desc,
 
 	memcpy(cfg, old_cfg, sizeof(struct irq_cfg));
 
-	init_copy_irq_2_pin(old_cfg, cfg, cpu);
+	init_copy_irq_2_pin(old_cfg, cfg, node);
 }
 
 static void free_irq_cfg(struct irq_cfg *old_cfg)
@@ -615,13 +609,13 @@ set_ioapic_affinity_irq(unsigned int irq, const struct cpumask *mask)
  * shared ISA-space IRQs, so we have to support them. We are super
  * fast in the common case, and fast for shared ISA-space IRQs.
  */
-static void add_pin_to_irq_cpu(struct irq_cfg *cfg, int cpu, int apic, int pin)
+static void add_pin_to_irq_node(struct irq_cfg *cfg, int node, int apic, int pin)
 {
 	struct irq_pin_list *entry;
 
 	entry = cfg->irq_2_pin;
 	if (!entry) {
-		entry = get_one_free_irq_2_pin(cpu);
+		entry = get_one_free_irq_2_pin(node);
 		if (!entry) {
 			printk(KERN_ERR "can not alloc irq_2_pin to add %d - %d\n",
 					apic, pin);
@@ -641,7 +635,7 @@ static void add_pin_to_irq_cpu(struct irq_cfg *cfg, int cpu, int apic, int pin)
 		entry = entry->next;
 	}
 
-	entry->next = get_one_free_irq_2_pin(cpu);
+	entry->next = get_one_free_irq_2_pin(node);
 	entry = entry->next;
 	entry->apic = apic;
 	entry->pin = pin;
@@ -650,7 +644,7 @@ static void add_pin_to_irq_cpu(struct irq_cfg *cfg, int cpu, int apic, int pin)
 /*
  * Reroute an IRQ to a different pin.
  */
-static void __init replace_pin_at_irq_cpu(struct irq_cfg *cfg, int cpu,
+static void __init replace_pin_at_irq_node(struct irq_cfg *cfg, int node,
 				      int oldapic, int oldpin,
 				      int newapic, int newpin)
 {
@@ -670,7 +664,7 @@ static void __init replace_pin_at_irq_cpu(struct irq_cfg *cfg, int cpu,
 
 	/* why? call replace before add? */
 	if (!replaced)
-		add_pin_to_irq_cpu(cfg, cpu, newapic, newpin);
+		add_pin_to_irq_node(cfg, node, newapic, newpin);
 }
 
 static inline void io_apic_modify_irq(struct irq_cfg *cfg,
@@ -1612,7 +1606,7 @@ static void __init setup_IO_APIC_irqs(void)
 	int notcon = 0;
 	struct irq_desc *desc;
 	struct irq_cfg *cfg;
-	int cpu = boot_cpu_id;
+	int node = cpu_to_node(boot_cpu_id);
 
 	apic_printk(APIC_VERBOSE, KERN_DEBUG "init IO_APIC IRQs\n");
 
@@ -1647,13 +1641,13 @@ static void __init setup_IO_APIC_irqs(void)
 					apic->multi_timer_check(apic_id, irq))
 				continue;
 
-			desc = irq_to_desc_alloc_cpu(irq, cpu);
+			desc = irq_to_desc_alloc_node(irq, node);
 			if (!desc) {
 				printk(KERN_INFO "can not get irq_desc for %d\n", irq);
 				continue;
 			}
 			cfg = desc->chip_data;
-			add_pin_to_irq_cpu(cfg, cpu, apic_id, pin);
+			add_pin_to_irq_node(cfg, node, apic_id, pin);
 
 			setup_IO_APIC_irq(apic_id, pin, irq, desc,
 					irq_trigger(idx), irq_polarity(idx));
@@ -2863,7 +2857,7 @@ static inline void __init check_timer(void)
 {
 	struct irq_desc *desc = irq_to_desc(0);
 	struct irq_cfg *cfg = desc->chip_data;
-	int cpu = boot_cpu_id;
+	int node = cpu_to_node(boot_cpu_id);
 	int apic1, pin1, apic2, pin2;
 	unsigned long flags;
 	int no_pin1 = 0;
@@ -2929,7 +2923,7 @@ static inline void __init check_timer(void)
 		 * Ok, does IRQ0 through the IOAPIC work?
 		 */
 		if (no_pin1) {
-			add_pin_to_irq_cpu(cfg, cpu, apic1, pin1);
+			add_pin_to_irq_node(cfg, node, apic1, pin1);
 			setup_timer_IRQ0_pin(apic1, pin1, cfg->vector);
 		} else {
 			/* for edge trigger, setup_IO_APIC_irq already
@@ -2966,7 +2960,7 @@ static inline void __init check_timer(void)
 		/*
 		 * legacy devices should be connected to IO APIC #0
 		 */
-		replace_pin_at_irq_cpu(cfg, cpu, apic1, pin1, apic2, pin2);
+		replace_pin_at_irq_node(cfg, node, apic1, pin1, apic2, pin2);
 		setup_timer_IRQ0_pin(apic2, pin2, cfg->vector);
 		enable_8259A_irq(0);
 		if (timer_irq_works()) {
@@ -3185,7 +3179,7 @@ unsigned int create_irq_nr(unsigned int irq_want)
 	unsigned int new;
 	unsigned long flags;
 	struct irq_cfg *cfg_new = NULL;
-	int cpu = boot_cpu_id;
+	int node = cpu_to_node(boot_cpu_id);
 	struct irq_desc *desc_new = NULL;
 
 	irq = 0;
@@ -3194,7 +3188,7 @@ unsigned int create_irq_nr(unsigned int irq_want)
 
 	spin_lock_irqsave(&vector_lock, flags);
 	for (new = irq_want; new < nr_irqs; new++) {
-		desc_new = irq_to_desc_alloc_cpu(new, cpu);
+		desc_new = irq_to_desc_alloc_node(new, node);
 		if (!desc_new) {
 			printk(KERN_INFO "can not get irq_desc for %d\n", new);
 			continue;
@@ -3968,7 +3962,7 @@ int io_apic_set_pci_routing (int ioapic, int pin, int irq, int triggering, int p
 {
 	struct irq_desc *desc;
 	struct irq_cfg *cfg;
-	int cpu = boot_cpu_id;
+	int node = cpu_to_node(boot_cpu_id);
 
 	if (!IO_APIC_IRQ(irq)) {
 		apic_printk(APIC_QUIET,KERN_ERR "IOAPIC[%d]: Invalid reference to IRQ 0\n",
@@ -3976,7 +3970,7 @@ int io_apic_set_pci_routing (int ioapic, int pin, int irq, int triggering, int p
 		return -EINVAL;
 	}
 
-	desc = irq_to_desc_alloc_cpu(irq, cpu);
+	desc = irq_to_desc_alloc_node(irq, node);
 	if (!desc) {
 		printk(KERN_INFO "can not get irq_desc %d\n", irq);
 		return 0;
@@ -3987,7 +3981,7 @@ int io_apic_set_pci_routing (int ioapic, int pin, int irq, int triggering, int p
 	 */
 	if (irq >= NR_IRQS_LEGACY) {
 		cfg = desc->chip_data;
-		add_pin_to_irq_cpu(cfg, cpu, ioapic, pin);
+		add_pin_to_irq_node(cfg, node, ioapic, pin);
 	}
 
 	setup_IO_APIC_irq(ioapic, pin, irq, desc, triggering, polarity);
diff --git a/arch/x86/lguest/boot.c b/arch/x86/lguest/boot.c
index ca7ec44bafc..45acbcf2568 100644
--- a/arch/x86/lguest/boot.c
+++ b/arch/x86/lguest/boot.c
@@ -636,7 +636,7 @@ static void __init lguest_init_IRQ(void)
 
 void lguest_setup_irq(unsigned int irq)
 {
-	irq_to_desc_alloc_cpu(irq, 0);
+	irq_to_desc_alloc_node(irq, 0);
 	set_irq_chip_and_handler_name(irq, &lguest_irq_controller,
 				      handle_level_irq, "level");
 }
diff --git a/drivers/pci/intr_remapping.c b/drivers/pci/intr_remapping.c
index f5e0ea724a6..9eff36a293e 100644
--- a/drivers/pci/intr_remapping.c
+++ b/drivers/pci/intr_remapping.c
@@ -23,15 +23,12 @@ struct irq_2_iommu {
 };
 
 #ifdef CONFIG_GENERIC_HARDIRQS
-static struct irq_2_iommu *get_one_free_irq_2_iommu(int cpu)
+static struct irq_2_iommu *get_one_free_irq_2_iommu(int node)
 {
 	struct irq_2_iommu *iommu;
-	int node;
-
-	node = cpu_to_node(cpu);
 
 	iommu = kzalloc_node(sizeof(*iommu), GFP_ATOMIC, node);
-	printk(KERN_DEBUG "alloc irq_2_iommu on cpu %d node %d\n", cpu, node);
+	printk(KERN_DEBUG "alloc irq_2_iommu on node %d\n", node);
 
 	return iommu;
 }
@@ -48,7 +45,7 @@ static struct irq_2_iommu *irq_2_iommu(unsigned int irq)
 	return desc->irq_2_iommu;
 }
 
-static struct irq_2_iommu *irq_2_iommu_alloc_cpu(unsigned int irq, int cpu)
+static struct irq_2_iommu *irq_2_iommu_alloc_node(unsigned int irq, int node)
 {
 	struct irq_desc *desc;
 	struct irq_2_iommu *irq_iommu;
@@ -56,7 +53,7 @@ static struct irq_2_iommu *irq_2_iommu_alloc_cpu(unsigned int irq, int cpu)
 	/*
 	 * alloc irq desc if not allocated already.
 	 */
-	desc = irq_to_desc_alloc_cpu(irq, cpu);
+	desc = irq_to_desc_alloc_node(irq, node);
 	if (!desc) {
 		printk(KERN_INFO "can not get irq_desc for %d\n", irq);
 		return NULL;
@@ -65,14 +62,14 @@ static struct irq_2_iommu *irq_2_iommu_alloc_cpu(unsigned int irq, int cpu)
 	irq_iommu = desc->irq_2_iommu;
 
 	if (!irq_iommu)
-		desc->irq_2_iommu = get_one_free_irq_2_iommu(cpu);
+		desc->irq_2_iommu = get_one_free_irq_2_iommu(node);
 
 	return desc->irq_2_iommu;
 }
 
 static struct irq_2_iommu *irq_2_iommu_alloc(unsigned int irq)
 {
-	return irq_2_iommu_alloc_cpu(irq, boot_cpu_id);
+	return irq_2_iommu_alloc_node(irq, cpu_to_node(boot_cpu_id));
 }
 
 #else /* !CONFIG_SPARSE_IRQ */
diff --git a/drivers/xen/events.c b/drivers/xen/events.c
index 33389880279..be437c2bc94 100644
--- a/drivers/xen/events.c
+++ b/drivers/xen/events.c
@@ -335,7 +335,7 @@ static int find_unbound_irq(void)
 	if (irq == nr_irqs)
 		panic("No available IRQ to bind to: increase nr_irqs!\n");
 
-	desc = irq_to_desc_alloc_cpu(irq, 0);
+	desc = irq_to_desc_alloc_node(irq, 0);
 	if (WARN_ON(desc == NULL))
 		return -1;
 
diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h
index 91bb76f44f1..ff374ceface 100644
--- a/include/linux/interrupt.h
+++ b/include/linux/interrupt.h
@@ -566,6 +566,6 @@ struct irq_desc;
 extern int early_irq_init(void);
 extern int arch_probe_nr_irqs(void);
 extern int arch_early_irq_init(void);
-extern int arch_init_chip_data(struct irq_desc *desc, int cpu);
+extern int arch_init_chip_data(struct irq_desc *desc, int node);
 
 #endif
diff --git a/include/linux/irq.h b/include/linux/irq.h
index 8e4c18b2915..a09baf8f9d9 100644
--- a/include/linux/irq.h
+++ b/include/linux/irq.h
@@ -187,7 +187,7 @@ struct irq_desc {
 	spinlock_t		lock;
 #ifdef CONFIG_SMP
 	cpumask_var_t		affinity;
-	unsigned int		cpu;
+	unsigned int		node;
 #ifdef CONFIG_GENERIC_PENDING_IRQ
 	cpumask_var_t		pending_mask;
 #endif
@@ -201,16 +201,16 @@ struct irq_desc {
 } ____cacheline_internodealigned_in_smp;
 
 extern void arch_init_copy_chip_data(struct irq_desc *old_desc,
-					struct irq_desc *desc, int cpu);
+					struct irq_desc *desc, int node);
 extern void arch_free_chip_data(struct irq_desc *old_desc, struct irq_desc *desc);
 
 #ifndef CONFIG_SPARSE_IRQ
 extern struct irq_desc irq_desc[NR_IRQS];
 #else /* CONFIG_SPARSE_IRQ */
-extern struct irq_desc *move_irq_desc(struct irq_desc *old_desc, int cpu);
+extern struct irq_desc *move_irq_desc(struct irq_desc *old_desc, int node);
 #endif /* CONFIG_SPARSE_IRQ */
 
-extern struct irq_desc *irq_to_desc_alloc_cpu(unsigned int irq, int cpu);
+extern struct irq_desc *irq_to_desc_alloc_node(unsigned int irq, int node);
 
 /*
  * Migration helpers for obsolete names, they will go away:
@@ -422,12 +422,10 @@ extern int set_irq_msi(unsigned int irq, struct msi_desc *entry);
  * Allocates affinity and pending_mask cpumask if required.
  * Returns true if successful (or not required).
  */
-static inline bool alloc_desc_masks(struct irq_desc *desc, int cpu,
+static inline bool alloc_desc_masks(struct irq_desc *desc, int node,
 								bool boot)
 {
 #ifdef CONFIG_CPUMASK_OFFSTACK
-	int node;
-
 	if (boot) {
 		alloc_bootmem_cpumask_var(&desc->affinity);
 
@@ -437,8 +435,6 @@ static inline bool alloc_desc_masks(struct irq_desc *desc, int cpu,
 		return true;
 	}
 
-	node = cpu_to_node(cpu);
-
 	if (!alloc_cpumask_var_node(&desc->affinity, GFP_ATOMIC, node))
 		return false;
 
@@ -494,7 +490,7 @@ static inline void free_desc_masks(struct irq_desc *old_desc,
 
 #else /* !CONFIG_SMP */
 
-static inline bool alloc_desc_masks(struct irq_desc *desc, int cpu,
+static inline bool alloc_desc_masks(struct irq_desc *desc, int node,
 								bool boot)
 {
 	return true;
diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c
index 3e0cbc44bd7..a6368db2618 100644
--- a/kernel/irq/handle.c
+++ b/kernel/irq/handle.c
@@ -81,12 +81,10 @@ static struct irq_desc irq_desc_init = {
 	.lock       = __SPIN_LOCK_UNLOCKED(irq_desc_init.lock),
 };
 
-void init_kstat_irqs(struct irq_desc *desc, int cpu, int nr)
+void init_kstat_irqs(struct irq_desc *desc, int node, int nr)
 {
-	int node;
 	void *ptr;
 
-	node = cpu_to_node(cpu);
 	ptr = kzalloc_node(nr * sizeof(*desc->kstat_irqs), GFP_ATOMIC, node);
 
 	/*
@@ -94,33 +92,32 @@ void init_kstat_irqs(struct irq_desc *desc, int cpu, int nr)
 	 * init_copy_kstat_irqs() could still use old one
 	 */
 	if (ptr) {
-		printk(KERN_DEBUG "  alloc kstat_irqs on cpu %d node %d\n",
-			 cpu, node);
+		printk(KERN_DEBUG "  alloc kstat_irqs on node %d\n", node);
 		desc->kstat_irqs = ptr;
 	}
 }
 
-static void init_one_irq_desc(int irq, struct irq_desc *desc, int cpu)
+static void init_one_irq_desc(int irq, struct irq_desc *desc, int node)
 {
 	memcpy(desc, &irq_desc_init, sizeof(struct irq_desc));
 
 	spin_lock_init(&desc->lock);
 	desc->irq = irq;
 #ifdef CONFIG_SMP
-	desc->cpu = cpu;
+	desc->node = node;
 #endif
 	lockdep_set_class(&desc->lock, &irq_desc_lock_class);
-	init_kstat_irqs(desc, cpu, nr_cpu_ids);
+	init_kstat_irqs(desc, node, nr_cpu_ids);
 	if (!desc->kstat_irqs) {
 		printk(KERN_ERR "can not alloc kstat_irqs\n");
 		BUG_ON(1);
 	}
-	if (!alloc_desc_masks(desc, cpu, false)) {
+	if (!alloc_desc_masks(desc, node, false)) {
 		printk(KERN_ERR "can not alloc irq_desc cpumasks\n");
 		BUG_ON(1);
 	}
 	init_desc_masks(desc);
-	arch_init_chip_data(desc, cpu);
+	arch_init_chip_data(desc, node);
 }
 
 /*
@@ -189,11 +186,10 @@ struct irq_desc *irq_to_desc(unsigned int irq)
 	return NULL;
 }
 
-struct irq_desc *irq_to_desc_alloc_cpu(unsigned int irq, int cpu)
+struct irq_desc *irq_to_desc_alloc_node(unsigned int irq, int node)
 {
 	struct irq_desc *desc;
 	unsigned long flags;
-	int node;
 
 	if (irq >= nr_irqs) {
 		WARN(1, "irq (%d) >= nr_irqs (%d) in irq_to_desc_alloc\n",
@@ -212,15 +208,13 @@ struct irq_desc *irq_to_desc_alloc_cpu(unsigned int irq, int cpu)
 	if (desc)
 		goto out_unlock;
 
-	node = cpu_to_node(cpu);
 	desc = kzalloc_node(sizeof(*desc), GFP_ATOMIC, node);
-	printk(KERN_DEBUG "  alloc irq_desc for %d on cpu %d node %d\n",
-		 irq, cpu, node);
+	printk(KERN_DEBUG "  alloc irq_desc for %d on node %d\n", irq, node);
 	if (!desc) {
 		printk(KERN_ERR "can not alloc irq_desc\n");
 		BUG_ON(1);
 	}
-	init_one_irq_desc(irq, desc, cpu);
+	init_one_irq_desc(irq, desc, node);
 
 	irq_desc_ptrs[irq] = desc;
 
@@ -270,7 +264,7 @@ struct irq_desc *irq_to_desc(unsigned int irq)
 	return (irq < NR_IRQS) ? irq_desc + irq : NULL;
 }
 
-struct irq_desc *irq_to_desc_alloc_cpu(unsigned int irq, int cpu)
+struct irq_desc *irq_to_desc_alloc_node(unsigned int irq, int node)
 {
 	return irq_to_desc(irq);
 }
diff --git a/kernel/irq/internals.h b/kernel/irq/internals.h
index de5f412f6a9..73468253143 100644
--- a/kernel/irq/internals.h
+++ b/kernel/irq/internals.h
@@ -16,7 +16,7 @@ extern void __disable_irq(struct irq_desc *desc, unsigned int irq, bool susp);
 extern void __enable_irq(struct irq_desc *desc, unsigned int irq, bool resume);
 
 extern struct lock_class_key irq_desc_lock_class;
-extern void init_kstat_irqs(struct irq_desc *desc, int cpu, int nr);
+extern void init_kstat_irqs(struct irq_desc *desc, int node, int nr);
 extern void clear_kstat_irqs(struct irq_desc *desc);
 extern spinlock_t sparse_irq_lock;
 
diff --git a/kernel/irq/numa_migrate.c b/kernel/irq/numa_migrate.c
index ce72bc3f4ce..2f69bee57bf 100644
--- a/kernel/irq/numa_migrate.c
+++ b/kernel/irq/numa_migrate.c
@@ -15,9 +15,9 @@
 
 static void init_copy_kstat_irqs(struct irq_desc *old_desc,
 				 struct irq_desc *desc,
-				 int cpu, int nr)
+				 int node, int nr)
 {
-	init_kstat_irqs(desc, cpu, nr);
+	init_kstat_irqs(desc, node, nr);
 
 	if (desc->kstat_irqs != old_desc->kstat_irqs)
 		memcpy(desc->kstat_irqs, old_desc->kstat_irqs,
@@ -34,20 +34,20 @@ static void free_kstat_irqs(struct irq_desc *old_desc, struct irq_desc *desc)
 }
 
 static bool init_copy_one_irq_desc(int irq, struct irq_desc *old_desc,
-		 struct irq_desc *desc, int cpu)
+		 struct irq_desc *desc, int node)
 {
 	memcpy(desc, old_desc, sizeof(struct irq_desc));
-	if (!alloc_desc_masks(desc, cpu, false)) {
+	if (!alloc_desc_masks(desc, node, false)) {
 		printk(KERN_ERR "irq %d: can not get new irq_desc cpumask "
 				"for migration.\n", irq);
 		return false;
 	}
 	spin_lock_init(&desc->lock);
-	desc->cpu = cpu;
+	desc->node = node;
 	lockdep_set_class(&desc->lock, &irq_desc_lock_class);
-	init_copy_kstat_irqs(old_desc, desc, cpu, nr_cpu_ids);
+	init_copy_kstat_irqs(old_desc, desc, node, nr_cpu_ids);
 	init_copy_desc_masks(old_desc, desc);
-	arch_init_copy_chip_data(old_desc, desc, cpu);
+	arch_init_copy_chip_data(old_desc, desc, node);
 	return true;
 }
 
@@ -59,12 +59,11 @@ static void free_one_irq_desc(struct irq_desc *old_desc, struct irq_desc *desc)
 }
 
 static struct irq_desc *__real_move_irq_desc(struct irq_desc *old_desc,
-						int cpu)
+						int node)
 {
 	struct irq_desc *desc;
 	unsigned int irq;
 	unsigned long flags;
-	int node;
 
 	irq = old_desc->irq;
 
@@ -76,7 +75,6 @@ static struct irq_desc *__real_move_irq_desc(struct irq_desc *old_desc,
 	if (desc && old_desc != desc)
 		goto out_unlock;
 
-	node = cpu_to_node(cpu);
 	desc = kzalloc_node(sizeof(*desc), GFP_ATOMIC, node);
 	if (!desc) {
 		printk(KERN_ERR "irq %d: can not get new irq_desc "
@@ -85,7 +83,7 @@ static struct irq_desc *__real_move_irq_desc(struct irq_desc *old_desc,
 		desc = old_desc;
 		goto out_unlock;
 	}
-	if (!init_copy_one_irq_desc(irq, old_desc, desc, cpu)) {
+	if (!init_copy_one_irq_desc(irq, old_desc, desc, node)) {
 		/* still use old one */
 		kfree(desc);
 		desc = old_desc;
@@ -107,24 +105,14 @@ out_unlock:
 	return desc;
 }
 
-struct irq_desc *move_irq_desc(struct irq_desc *desc, int cpu)
+struct irq_desc *move_irq_desc(struct irq_desc *desc, int node)
 {
-	int old_cpu;
-	int node, old_node;
-
 	/* those all static, do move them */
 	if (desc->irq < NR_IRQS_LEGACY)
 		return desc;
 
-	old_cpu = desc->cpu;
-	if (old_cpu != cpu) {
-		node = cpu_to_node(cpu);
-		old_node = cpu_to_node(old_cpu);
-		if (old_node != node)
-			desc = __real_move_irq_desc(desc, cpu);
-		else
-			desc->cpu = cpu;
-	}
+	if (desc->node != node)
+		desc = __real_move_irq_desc(desc, node);
 
 	return desc;
 }
diff --git a/kernel/softirq.c b/kernel/softirq.c
index b525dd34851..f674f332a02 100644
--- a/kernel/softirq.c
+++ b/kernel/softirq.c
@@ -828,7 +828,7 @@ int __init __weak arch_early_irq_init(void)
 	return 0;
 }
 
-int __weak arch_init_chip_data(struct irq_desc *desc, int cpu)
+int __weak arch_init_chip_data(struct irq_desc *desc, int node)
 {
 	return 0;
 }
-- 
cgit v1.2.3-70-g09d2


From d047f53a2ecce37e3bdf79eac5a326fbaadb3628 Mon Sep 17 00:00:00 2001
From: Yinghai Lu <yinghai@kernel.org>
Date: Mon, 27 Apr 2009 18:02:23 -0700
Subject: x86/irq: change MSI irq_desc to be more numa aware

Try to get irq_desc on the home node in create_irq_nr().

v2: don't check if we can move it when sparse_irq is not used
v3: use move_irq_des, if that node is not what we want

[ Impact: optimization, make MSI IRQ descriptors more NUMA aware ]

Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Suresh Siddha <suresh.b.siddha@intel.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Rusty Russell <rusty@rustcorp.com.au>
LKML-Reference: <49F6559F.7070005@kernel.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 arch/x86/kernel/apic/io_apic.c | 17 +++++++++++++----
 include/linux/irq.h            |  2 +-
 2 files changed, 14 insertions(+), 5 deletions(-)

(limited to 'include/linux/irq.h')

diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
index 82376e021b5..9cd4806cdf5 100644
--- a/arch/x86/kernel/apic/io_apic.c
+++ b/arch/x86/kernel/apic/io_apic.c
@@ -3172,14 +3172,13 @@ static int nr_irqs_gsi = NR_IRQS_LEGACY;
 /*
  * Dynamic irq allocate and deallocation
  */
-unsigned int create_irq_nr(unsigned int irq_want)
+unsigned int create_irq_nr(unsigned int irq_want, int node)
 {
 	/* Allocate an unused irq */
 	unsigned int irq;
 	unsigned int new;
 	unsigned long flags;
 	struct irq_cfg *cfg_new = NULL;
-	int node = cpu_to_node(boot_cpu_id);
 	struct irq_desc *desc_new = NULL;
 
 	irq = 0;
@@ -3197,6 +3196,13 @@ unsigned int create_irq_nr(unsigned int irq_want)
 
 		if (cfg_new->vector != 0)
 			continue;
+
+#ifdef CONFIG_NUMA_IRQ_DESC
+		/* different node ?*/
+		if (desc_new->node != node)
+			desc = move_irq_desc(desc, node);
+#endif
+
 		if (__assign_irq_vector(new, cfg_new, apic->target_cpus()) == 0)
 			irq = new;
 		break;
@@ -3214,11 +3220,12 @@ unsigned int create_irq_nr(unsigned int irq_want)
 
 int create_irq(void)
 {
+	int node = cpu_to_node(boot_cpu_id);
 	unsigned int irq_want;
 	int irq;
 
 	irq_want = nr_irqs_gsi;
-	irq = create_irq_nr(irq_want);
+	irq = create_irq_nr(irq_want, node);
 
 	if (irq == 0)
 		irq = -1;
@@ -3476,15 +3483,17 @@ int arch_setup_msi_irqs(struct pci_dev *dev, int nvec, int type)
 	unsigned int irq_want;
 	struct intel_iommu *iommu = NULL;
 	int index = 0;
+	int node;
 
 	/* x86 doesn't support multiple MSI yet */
 	if (type == PCI_CAP_ID_MSI && nvec > 1)
 		return 1;
 
+	node = dev_to_node(&dev->dev);
 	irq_want = nr_irqs_gsi;
 	sub_handle = 0;
 	list_for_each_entry(msidesc, &dev->msi_list, list) {
-		irq = create_irq_nr(irq_want);
+		irq = create_irq_nr(irq_want, node);
 		if (irq == 0)
 			return -1;
 		irq_want = irq + 1;
diff --git a/include/linux/irq.h b/include/linux/irq.h
index a09baf8f9d9..4b95ddb5304 100644
--- a/include/linux/irq.h
+++ b/include/linux/irq.h
@@ -376,7 +376,7 @@ extern void set_irq_noprobe(unsigned int irq);
 extern void set_irq_probe(unsigned int irq);
 
 /* Handle dynamic irq creation and destruction */
-extern unsigned int create_irq_nr(unsigned int irq_want);
+extern unsigned int create_irq_nr(unsigned int irq_want, int node);
 extern int create_irq(void);
 extern void destroy_irq(unsigned int irq);
 
-- 
cgit v1.2.3-70-g09d2


From 15e957d08dd4a841359cfec59ecb74041e0097aa Mon Sep 17 00:00:00 2001
From: Yinghai Lu <yinghai@kernel.org>
Date: Thu, 30 Apr 2009 01:17:50 -0700
Subject: x86/irq: use move_irq_desc() in create_irq_nr()

move_irq_desc() will try to move irq_desc to the home node if
the allocated one is not correct, in create_irq_nr().

( This can happen on devices that are on different nodes that
  are using MSI, when drivers are loaded and unloaded randomly. )

v2: fix non-smp build
v3: add NUMA_IRQ_DESC to eliminate #ifdefs

[ Impact: improve irq descriptor locality on NUMA systems ]

Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Suresh Siddha <suresh.b.siddha@intel.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Rusty Russell <rusty@rustcorp.com.au>
LKML-Reference: <49F95EAE.2050903@kernel.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 arch/x86/Kconfig               |  4 ++++
 arch/x86/kernel/apic/io_apic.c |  6 +-----
 include/linux/irq.h            | 11 +++++++++--
 kernel/irq/Makefile            |  2 +-
 4 files changed, 15 insertions(+), 8 deletions(-)

(limited to 'include/linux/irq.h')

diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index e1b2543f8ed..674e21e9f0a 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -274,6 +274,10 @@ config SPARSE_IRQ
 
 	  If you don't know what to do here, say N.
 
+config NUMA_IRQ_DESC
+	def_bool y
+	depends on SPARSE_IRQ && NUMA
+
 config X86_MPPARSE
 	bool "Enable MPS table" if ACPI
 	default y
diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
index 9cd4806cdf5..e583291fe6c 100644
--- a/arch/x86/kernel/apic/io_apic.c
+++ b/arch/x86/kernel/apic/io_apic.c
@@ -3197,11 +3197,7 @@ unsigned int create_irq_nr(unsigned int irq_want, int node)
 		if (cfg_new->vector != 0)
 			continue;
 
-#ifdef CONFIG_NUMA_IRQ_DESC
-		/* different node ?*/
-		if (desc_new->node != node)
-			desc = move_irq_desc(desc, node);
-#endif
+		desc_new = move_irq_desc(desc_new, node);
 
 		if (__assign_irq_vector(new, cfg_new, apic->target_cpus()) == 0)
 			irq = new;
diff --git a/include/linux/irq.h b/include/linux/irq.h
index 4b95ddb5304..eedbb8e5e0c 100644
--- a/include/linux/irq.h
+++ b/include/linux/irq.h
@@ -206,9 +206,16 @@ extern void arch_free_chip_data(struct irq_desc *old_desc, struct irq_desc *desc
 
 #ifndef CONFIG_SPARSE_IRQ
 extern struct irq_desc irq_desc[NR_IRQS];
-#else /* CONFIG_SPARSE_IRQ */
+#endif
+
+#ifdef CONFIG_NUMA_IRQ_DESC
 extern struct irq_desc *move_irq_desc(struct irq_desc *old_desc, int node);
-#endif /* CONFIG_SPARSE_IRQ */
+#else
+static inline struct irq_desc *move_irq_desc(struct irq_desc *desc, int node)
+{
+	return desc;
+}
+#endif
 
 extern struct irq_desc *irq_to_desc_alloc_node(unsigned int irq, int node);
 
diff --git a/kernel/irq/Makefile b/kernel/irq/Makefile
index 2f065277f8e..7d047808419 100644
--- a/kernel/irq/Makefile
+++ b/kernel/irq/Makefile
@@ -3,5 +3,5 @@ obj-y := handle.o manage.o spurious.o resend.o chip.o devres.o
 obj-$(CONFIG_GENERIC_IRQ_PROBE) += autoprobe.o
 obj-$(CONFIG_PROC_FS) += proc.o
 obj-$(CONFIG_GENERIC_PENDING_IRQ) += migration.o
-obj-$(CONFIG_SPARSE_IRQ) += numa_migrate.o
+obj-$(CONFIG_NUMA_IRQ_DESC) += numa_migrate.o
 obj-$(CONFIG_PM_SLEEP) += pm.o
-- 
cgit v1.2.3-70-g09d2


From 38c7fed2f5ffee17e1fa3e0f78b0e1bf43d52d13 Mon Sep 17 00:00:00 2001
From: Yinghai Lu <yinghai@kernel.org>
Date: Mon, 25 May 2009 15:10:58 +0300
Subject: x86: remove some alloc_bootmem_cpumask_var calling

Now that we set up the slab allocator earlier, we can get rid of some
alloc_bootmem_cpumask_var() calls in boot code.

Cc: Ingo Molnar <mingo@elte.hu>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Pekka Enberg <penberg@cs.helsinki.fi>
---
 arch/x86/kernel/apic/io_apic.c |  4 ++--
 include/linux/irq.h            | 18 +++++++-----------
 kernel/cpuset.c                |  2 +-
 kernel/profile.c               |  6 ------
 lib/cpumask.c                  | 11 ++---------
 5 files changed, 12 insertions(+), 29 deletions(-)

(limited to 'include/linux/irq.h')

diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
index 1946fac42ab..139201a562b 100644
--- a/arch/x86/kernel/apic/io_apic.c
+++ b/arch/x86/kernel/apic/io_apic.c
@@ -185,8 +185,8 @@ int __init arch_early_irq_init(void)
 	for (i = 0; i < count; i++) {
 		desc = irq_to_desc(i);
 		desc->chip_data = &cfg[i];
-		alloc_bootmem_cpumask_var(&cfg[i].domain);
-		alloc_bootmem_cpumask_var(&cfg[i].old_domain);
+		alloc_cpumask_var(&cfg[i].domain, GFP_NOWAIT);
+		alloc_cpumask_var(&cfg[i].old_domain, GFP_NOWAIT);
 		if (i < NR_IRQS_LEGACY)
 			cpumask_setall(cfg[i].domain);
 	}
diff --git a/include/linux/irq.h b/include/linux/irq.h
index eedbb8e5e0c..1e50c34f006 100644
--- a/include/linux/irq.h
+++ b/include/linux/irq.h
@@ -430,23 +430,19 @@ extern int set_irq_msi(unsigned int irq, struct msi_desc *entry);
  * Returns true if successful (or not required).
  */
 static inline bool alloc_desc_masks(struct irq_desc *desc, int node,
-								bool boot)
+							bool boot)
 {
-#ifdef CONFIG_CPUMASK_OFFSTACK
-	if (boot) {
-		alloc_bootmem_cpumask_var(&desc->affinity);
+	gfp_t gfp = GFP_ATOMIC;
 
-#ifdef CONFIG_GENERIC_PENDING_IRQ
-		alloc_bootmem_cpumask_var(&desc->pending_mask);
-#endif
-		return true;
-	}
+	if (boot)
+		gfp = GFP_NOWAIT;
 
-	if (!alloc_cpumask_var_node(&desc->affinity, GFP_ATOMIC, node))
+#ifdef CONFIG_CPUMASK_OFFSTACK
+	if (!alloc_cpumask_var_node(&desc->affinity, gfp, node))
 		return false;
 
 #ifdef CONFIG_GENERIC_PENDING_IRQ
-	if (!alloc_cpumask_var_node(&desc->pending_mask, GFP_ATOMIC, node)) {
+	if (!alloc_cpumask_var_node(&desc->pending_mask, gfp, node)) {
 		free_cpumask_var(desc->affinity);
 		return false;
 	}
diff --git a/kernel/cpuset.c b/kernel/cpuset.c
index 026faccca86..d5a7e17474e 100644
--- a/kernel/cpuset.c
+++ b/kernel/cpuset.c
@@ -1857,7 +1857,7 @@ struct cgroup_subsys cpuset_subsys = {
 
 int __init cpuset_init_early(void)
 {
-	alloc_bootmem_cpumask_var(&top_cpuset.cpus_allowed);
+	alloc_cpumask_var(&top_cpuset.cpus_allowed, GFP_NOWAIT);
 
 	top_cpuset.mems_generation = cpuset_mems_generation++;
 	return 0;
diff --git a/kernel/profile.c b/kernel/profile.c
index 7724e0409ba..28cf26ad2d2 100644
--- a/kernel/profile.c
+++ b/kernel/profile.c
@@ -111,12 +111,6 @@ int __ref profile_init(void)
 	/* only text is profiled */
 	prof_len = (_etext - _stext) >> prof_shift;
 	buffer_bytes = prof_len*sizeof(atomic_t);
-	if (!slab_is_available()) {
-		prof_buffer = alloc_bootmem(buffer_bytes);
-		alloc_bootmem_cpumask_var(&prof_cpu_mask);
-		cpumask_copy(prof_cpu_mask, cpu_possible_mask);
-		return 0;
-	}
 
 	if (!alloc_cpumask_var(&prof_cpu_mask, GFP_KERNEL))
 		return -ENOMEM;
diff --git a/lib/cpumask.c b/lib/cpumask.c
index eb23aaa0c7b..7bb4142a502 100644
--- a/lib/cpumask.c
+++ b/lib/cpumask.c
@@ -92,15 +92,8 @@ int cpumask_any_but(const struct cpumask *mask, unsigned int cpu)
  */
 bool alloc_cpumask_var_node(cpumask_var_t *mask, gfp_t flags, int node)
 {
-	if (likely(slab_is_available()))
-		*mask = kmalloc_node(cpumask_size(), flags, node);
-	else {
-#ifdef CONFIG_DEBUG_PER_CPU_MAPS
-		printk(KERN_ERR
-			"=> alloc_cpumask_var: kmalloc not available!\n");
-#endif
-		*mask = NULL;
-	}
+	*mask = kmalloc_node(cpumask_size(), flags, node);
+
 #ifdef CONFIG_DEBUG_PER_CPU_MAPS
 	if (!*mask) {
 		printk(KERN_ERR "=> alloc_cpumask_var: failed!\n");
-- 
cgit v1.2.3-70-g09d2


From ab33dcff40d7a9a28587e4425621e4cbc4089e03 Mon Sep 17 00:00:00 2001
From: Randy Dunlap <randy.dunlap@oracle.com>
Date: Sat, 13 Jun 2009 20:01:00 -0700
Subject: genirq, irq.h: Fix kernel-doc warnings

Fix kernel-doc warnings in linux/irq.h:

  Warning(include/linux/irq.h:201): No description found for parameter 'node'
  Warning(include/linux/irq.h:201): Excess struct/union/enum/typedef member 'cpu' description in 'irq_desc'
  Warning(include/linux/irq.h:434): No description found for parameter 'node'
  Warning(include/linux/irq.h:434): Excess function parameter 'cpu' description in 'alloc_desc_masks'

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
LKML-Reference: <4A3467EC.50006@oracle.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 include/linux/irq.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'include/linux/irq.h')

diff --git a/include/linux/irq.h b/include/linux/irq.h
index b7cbeed972e..0c001c15752 100644
--- a/include/linux/irq.h
+++ b/include/linux/irq.h
@@ -157,7 +157,7 @@ struct irq_2_iommu;
  * @irqs_unhandled:	stats field for spurious unhandled interrupts
  * @lock:		locking for SMP
  * @affinity:		IRQ affinity on SMP
- * @cpu:		cpu index useful for balancing
+ * @node:		node index useful for balancing
  * @pending_mask:	pending rebalanced interrupts
  * @threads_active:	number of irqaction threads currently running
  * @wait_for_threads:	wait queue for sync_irq to wait for threaded handlers
@@ -426,7 +426,7 @@ extern int set_irq_msi(unsigned int irq, struct msi_desc *entry);
 /**
  * init_alloc_desc_masks - allocate cpumasks for irq_desc
  * @desc:	pointer to irq_desc struct
- * @cpu:	cpu which will be handling the cpumasks
+ * @node:	node which will be handling the cpumasks
  * @boot:	true if need bootmem
  *
  * Allocates affinity and pending_mask cpumask if required.
-- 
cgit v1.2.3-70-g09d2