From ca109491f612aab5c8152207631c0444f63da97f Mon Sep 17 00:00:00 2001
From: Peter Zijlstra <peterz@infradead.org>
Date: Tue, 25 Nov 2008 12:43:51 +0100
Subject: hrtimer: removing all ur callback modes

Impact: cleanup, move all hrtimer processing into hardirq context

This is an attempt at removing some of the hrtimer complexity by
reducing the number of callback modes to 1.

This means that all hrtimer callback functions will be ran from HARD-irq
context.

I went through all the 30 odd hrtimer callback functions in the kernel
and saw only one that I'm not quite sure of, which is the one in
net/can/bcm.c - hence I'm CC-ing the folks responsible for that code.

Furthermore, the hrtimer core now calls callbacks directly with IRQs
disabled in case you try to enqueue an expired timer. If this timer is a
periodic timer (which should use hrtimer_forward() to advance its time)
then it might be possible to end up in an inf. recursive loop due to the
fact that hrtimer_forward() doesn't round up to the next timer
granularity, and therefore keeps on calling the callback - obviously
this needs a fix.

Aside from that, this seems to compile and actually boot on my dual core
test box - although I'm sure there are some bugs in, me not hitting any
makes me certain :-)

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 include/linux/interrupt.h | 3 ---
 1 file changed, 3 deletions(-)

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

diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h
index f58a0cf8929a..d6210a97a8ca 100644
--- a/include/linux/interrupt.h
+++ b/include/linux/interrupt.h
@@ -251,9 +251,6 @@ enum
 	BLOCK_SOFTIRQ,
 	TASKLET_SOFTIRQ,
 	SCHED_SOFTIRQ,
-#ifdef CONFIG_HIGH_RES_TIMERS
-	HRTIMER_SOFTIRQ,
-#endif
 	RCU_SOFTIRQ, 	/* Preferable RCU should always be the last softirq */
 
 	NR_SOFTIRQS
-- 
cgit v1.3-7-g2ca7


From 0de26520c7cabf36e1de090ea8092f011a6106ce Mon Sep 17 00:00:00 2001
From: Rusty Russell <rusty@rustcorp.com.au>
Date: Sat, 13 Dec 2008 21:20:26 +1030
Subject: cpumask: make irq_set_affinity() take a const struct cpumask

Impact: change existing irq_chip API

Not much point with gentle transition here: the struct irq_chip's
setaffinity method signature needs to change.

Fortunately, not widely used code, but hits a few architectures.

Note: In irq_select_affinity() I save a temporary in by mangling
irq_desc[irq].affinity directly.  Ingo, does this break anything?

(Folded in fix from KOSAKI Motohiro)

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Mike Travis <travis@sgi.com>
Reviewed-by: Grant Grundler <grundler@parisc-linux.org>
Acked-by: Ingo Molnar <mingo@redhat.com>
Cc: ralf@linux-mips.org
Cc: grundler@parisc-linux.org
Cc: jeremy@xensource.com
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
---
 arch/alpha/kernel/irq.c               |  2 +-
 arch/alpha/kernel/sys_dp264.c         |  8 ++--
 arch/alpha/kernel/sys_titan.c         |  4 +-
 arch/arm/common/gic.c                 |  4 +-
 arch/arm/kernel/irq.c                 |  2 +-
 arch/arm/oprofile/op_model_mpcore.c   |  4 +-
 arch/cris/arch-v32/kernel/irq.c       |  4 +-
 arch/ia64/hp/sim/hpsim_irq.c          |  2 +-
 arch/ia64/kernel/iosapic.c            | 12 +++---
 arch/ia64/kernel/irq.c                |  9 ++--
 arch/ia64/kernel/msi_ia64.c           | 12 +++---
 arch/ia64/kernel/smpboot.c            |  4 +-
 arch/ia64/sn/kernel/irq.c             |  6 +--
 arch/ia64/sn/kernel/msi_sn.c          |  7 +--
 arch/mips/include/asm/irq.h           |  3 +-
 arch/mips/kernel/cevt-bcm1480.c       |  2 +-
 arch/mips/kernel/cevt-sb1250.c        |  2 +-
 arch/mips/kernel/irq-gic.c            |  6 +--
 arch/mips/mti-malta/malta-smtc.c      |  6 +--
 arch/mips/sibyte/bcm1480/irq.c        |  8 ++--
 arch/mips/sibyte/sb1250/irq.c         |  8 ++--
 arch/parisc/kernel/irq.c              |  6 +--
 arch/powerpc/kernel/irq.c             |  2 +-
 arch/powerpc/platforms/pseries/xics.c |  6 +--
 arch/powerpc/sysdev/mpic.c            |  4 +-
 arch/powerpc/sysdev/mpic.h            |  2 +-
 arch/sparc64/kernel/irq.c             | 11 +++--
 arch/sparc64/kernel/of_device.c       |  2 +-
 arch/sparc64/kernel/pci_msi.c         |  2 +-
 arch/x86/kernel/hpet.c                |  4 +-
 arch/x86/kernel/io_apic.c             | 81 +++++++++++++++++------------------
 arch/x86/kernel/irq_32.c              |  2 +-
 arch/x86/kernel/irq_64.c              |  2 +-
 drivers/parisc/iosapic.c              |  7 +--
 drivers/xen/events.c                  |  6 +--
 include/linux/interrupt.h             |  4 +-
 include/linux/irq.h                   |  3 +-
 kernel/irq/chip.c                     |  2 +-
 kernel/irq/manage.c                   | 22 +++++-----
 kernel/irq/migration.c                | 14 +++---
 kernel/irq/proc.c                     | 29 ++++++++-----
 kernel/time/tick-common.c             |  6 +--
 42 files changed, 171 insertions(+), 161 deletions(-)

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

diff --git a/arch/alpha/kernel/irq.c b/arch/alpha/kernel/irq.c
index c626a821cdcb..d0f1620007f7 100644
--- a/arch/alpha/kernel/irq.c
+++ b/arch/alpha/kernel/irq.c
@@ -55,7 +55,7 @@ int irq_select_affinity(unsigned int irq)
 	last_cpu = cpu;
 
 	irq_desc[irq].affinity = cpumask_of_cpu(cpu);
-	irq_desc[irq].chip->set_affinity(irq, cpumask_of_cpu(cpu));
+	irq_desc[irq].chip->set_affinity(irq, cpumask_of(cpu));
 	return 0;
 }
 #endif /* CONFIG_SMP */
diff --git a/arch/alpha/kernel/sys_dp264.c b/arch/alpha/kernel/sys_dp264.c
index c71b0fd7a61f..ab44c164d9d4 100644
--- a/arch/alpha/kernel/sys_dp264.c
+++ b/arch/alpha/kernel/sys_dp264.c
@@ -177,19 +177,19 @@ cpu_set_irq_affinity(unsigned int irq, cpumask_t affinity)
 }
 
 static void
-dp264_set_affinity(unsigned int irq, cpumask_t affinity)
+dp264_set_affinity(unsigned int irq, const struct cpumask *affinity)
 { 
 	spin_lock(&dp264_irq_lock);
-	cpu_set_irq_affinity(irq, affinity);
+	cpu_set_irq_affinity(irq, *affinity);
 	tsunami_update_irq_hw(cached_irq_mask);
 	spin_unlock(&dp264_irq_lock);
 }
 
 static void
-clipper_set_affinity(unsigned int irq, cpumask_t affinity)
+clipper_set_affinity(unsigned int irq, const struct cpumask *affinity)
 { 
 	spin_lock(&dp264_irq_lock);
-	cpu_set_irq_affinity(irq - 16, affinity);
+	cpu_set_irq_affinity(irq - 16, *affinity);
 	tsunami_update_irq_hw(cached_irq_mask);
 	spin_unlock(&dp264_irq_lock);
 }
diff --git a/arch/alpha/kernel/sys_titan.c b/arch/alpha/kernel/sys_titan.c
index 52c91ccc1648..27f840a4ad3d 100644
--- a/arch/alpha/kernel/sys_titan.c
+++ b/arch/alpha/kernel/sys_titan.c
@@ -158,10 +158,10 @@ titan_cpu_set_irq_affinity(unsigned int irq, cpumask_t affinity)
 }
 
 static void
-titan_set_irq_affinity(unsigned int irq, cpumask_t affinity)
+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_cpu_set_irq_affinity(irq - 16, *affinity);
 	titan_update_irq_hw(titan_cached_irq_mask);
 	spin_unlock(&titan_irq_lock);
 }
diff --git a/arch/arm/common/gic.c b/arch/arm/common/gic.c
index 7fc9860a97d7..c6884ba1d5ed 100644
--- a/arch/arm/common/gic.c
+++ b/arch/arm/common/gic.c
@@ -109,11 +109,11 @@ static void gic_unmask_irq(unsigned int irq)
 }
 
 #ifdef CONFIG_SMP
-static void gic_set_cpu(unsigned int irq, cpumask_t mask_val)
+static void 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;
-	unsigned int cpu = first_cpu(mask_val);
+	unsigned int cpu = cpumask_first(mask_val);
 	u32 val;
 
 	spin_lock(&irq_controller_lock);
diff --git a/arch/arm/kernel/irq.c b/arch/arm/kernel/irq.c
index 2f3eb795fa6e..7141cee1fab7 100644
--- a/arch/arm/kernel/irq.c
+++ b/arch/arm/kernel/irq.c
@@ -174,7 +174,7 @@ static void route_irq(struct irq_desc *desc, unsigned int irq, unsigned int cpu)
 	pr_debug("IRQ%u: moving from cpu%u to cpu%u\n", irq, desc->cpu, cpu);
 
 	spin_lock_irq(&desc->lock);
-	desc->chip->set_affinity(irq, cpumask_of_cpu(cpu));
+	desc->chip->set_affinity(irq, cpumask_of(cpu));
 	spin_unlock_irq(&desc->lock);
 }
 
diff --git a/arch/arm/oprofile/op_model_mpcore.c b/arch/arm/oprofile/op_model_mpcore.c
index 4de366e8b4c5..6d6bd5899240 100644
--- a/arch/arm/oprofile/op_model_mpcore.c
+++ b/arch/arm/oprofile/op_model_mpcore.c
@@ -260,10 +260,10 @@ static void em_stop(void)
 static void em_route_irq(int irq, unsigned int cpu)
 {
 	struct irq_desc *desc = irq_desc + irq;
-	cpumask_t mask = cpumask_of_cpu(cpu);
+	const struct cpumask *mask = cpumask_of(cpu);
 
 	spin_lock_irq(&desc->lock);
-	desc->affinity = mask;
+	desc->affinity = *mask;
 	desc->chip->set_affinity(irq, mask);
 	spin_unlock_irq(&desc->lock);
 }
diff --git a/arch/cris/arch-v32/kernel/irq.c b/arch/cris/arch-v32/kernel/irq.c
index 173c141ac9ba..295131fee710 100644
--- a/arch/cris/arch-v32/kernel/irq.c
+++ b/arch/cris/arch-v32/kernel/irq.c
@@ -325,11 +325,11 @@ static void end_crisv32_irq(unsigned int irq)
 {
 }
 
-void set_affinity_crisv32_irq(unsigned int irq, cpumask_t dest)
+void 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;
+	irq_allocations[irq - FIRST_IRQ].mask = *dest;
 	spin_unlock_irqrestore(&irq_lock, flags);
 }
 
diff --git a/arch/ia64/hp/sim/hpsim_irq.c b/arch/ia64/hp/sim/hpsim_irq.c
index c2f58ff364e7..cc0a3182db3c 100644
--- a/arch/ia64/hp/sim/hpsim_irq.c
+++ b/arch/ia64/hp/sim/hpsim_irq.c
@@ -22,7 +22,7 @@ hpsim_irq_noop (unsigned int irq)
 }
 
 static void
-hpsim_set_affinity_noop (unsigned int a, cpumask_t b)
+hpsim_set_affinity_noop(unsigned int a, const struct cpumask *b)
 {
 }
 
diff --git a/arch/ia64/kernel/iosapic.c b/arch/ia64/kernel/iosapic.c
index 5c4674ae8aea..c8adecd5b416 100644
--- a/arch/ia64/kernel/iosapic.c
+++ b/arch/ia64/kernel/iosapic.c
@@ -330,25 +330,25 @@ unmask_irq (unsigned int irq)
 
 
 static void
-iosapic_set_affinity (unsigned int irq, cpumask_t mask)
+iosapic_set_affinity(unsigned int irq, const struct cpumask *mask)
 {
 #ifdef CONFIG_SMP
 	u32 high32, low32;
-	int dest, rte_index;
+	int cpu, dest, rte_index;
 	int redir = (irq & IA64_IRQ_REDIRECTED) ? 1 : 0;
 	struct iosapic_rte_info *rte;
 	struct iosapic *iosapic;
 
 	irq &= (~IA64_IRQ_REDIRECTED);
 
-	cpus_and(mask, mask, cpu_online_map);
-	if (cpus_empty(mask))
+	cpu = cpumask_first_and(cpu_online_mask, mask);
+	if (cpu >= nr_cpu_ids)
 		return;
 
-	if (irq_prepare_move(irq, first_cpu(mask)))
+	if (irq_prepare_move(irq, cpu))
 		return;
 
-	dest = cpu_physical_id(first_cpu(mask));
+	dest = cpu_physical_id(cpu);
 
 	if (!iosapic_intr_info[irq].count)
 		return;			/* not an IOSAPIC interrupt */
diff --git a/arch/ia64/kernel/irq.c b/arch/ia64/kernel/irq.c
index 7fd18f54c056..0b6db53fedcf 100644
--- a/arch/ia64/kernel/irq.c
+++ b/arch/ia64/kernel/irq.c
@@ -133,7 +133,6 @@ unsigned int vectors_in_migration[NR_IRQS];
  */
 static void migrate_irqs(void)
 {
-	cpumask_t	mask;
 	irq_desc_t *desc;
 	int 		irq, new_cpu;
 
@@ -152,15 +151,14 @@ static void migrate_irqs(void)
 		if (desc->status == IRQ_PER_CPU)
 			continue;
 
-		cpus_and(mask, irq_desc[irq].affinity, cpu_online_map);
-		if (any_online_cpu(mask) == NR_CPUS) {
+		if (cpumask_any_and(&irq_desc[irq].affinity, cpu_online_mask)
+		    >= nr_cpu_ids) {
 			/*
 			 * Save it for phase 2 processing
 			 */
 			vectors_in_migration[irq] = irq;
 
 			new_cpu = any_online_cpu(cpu_online_map);
-			mask = cpumask_of_cpu(new_cpu);
 
 			/*
 			 * Al three are essential, currently WARN_ON.. maybe panic?
@@ -168,7 +166,8 @@ static void migrate_irqs(void)
 			if (desc->chip && desc->chip->disable &&
 				desc->chip->enable && desc->chip->set_affinity) {
 				desc->chip->disable(irq);
-				desc->chip->set_affinity(irq, mask);
+				desc->chip->set_affinity(irq,
+							 cpumask_of(new_cpu));
 				desc->chip->enable(irq);
 			} else {
 				WARN_ON((!(desc->chip) || !(desc->chip->disable) ||
diff --git a/arch/ia64/kernel/msi_ia64.c b/arch/ia64/kernel/msi_ia64.c
index 702a09c13238..890339339035 100644
--- a/arch/ia64/kernel/msi_ia64.c
+++ b/arch/ia64/kernel/msi_ia64.c
@@ -49,11 +49,12 @@
 static struct irq_chip	ia64_msi_chip;
 
 #ifdef CONFIG_SMP
-static void ia64_set_msi_irq_affinity(unsigned int irq, cpumask_t cpu_mask)
+static void ia64_set_msi_irq_affinity(unsigned int irq,
+				      const cpumask_t *cpu_mask)
 {
 	struct msi_msg msg;
 	u32 addr, data;
-	int cpu = first_cpu(cpu_mask);
+	int cpu = first_cpu(*cpu_mask);
 
 	if (!cpu_online(cpu))
 		return;
@@ -166,12 +167,11 @@ void arch_teardown_msi_irq(unsigned int irq)
 
 #ifdef CONFIG_DMAR
 #ifdef CONFIG_SMP
-static void dmar_msi_set_affinity(unsigned int irq, cpumask_t mask)
+static void dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask)
 {
 	struct irq_cfg *cfg = irq_cfg + irq;
 	struct msi_msg msg;
-	int cpu = first_cpu(mask);
-
+	int cpu = cpumask_first(mask);
 
 	if (!cpu_online(cpu))
 		return;
@@ -187,7 +187,7 @@ static void dmar_msi_set_affinity(unsigned int irq, cpumask_t mask)
 	msg.address_lo |= MSI_ADDR_DESTID_CPU(cpu_physical_id(cpu));
 
 	dmar_msi_write(irq, &msg);
-	irq_desc[irq].affinity = mask;
+	irq_desc[irq].affinity = *mask;
 }
 #endif /* CONFIG_SMP */
 
diff --git a/arch/ia64/kernel/smpboot.c b/arch/ia64/kernel/smpboot.c
index 4ede6e571c38..11463994a7d5 100644
--- a/arch/ia64/kernel/smpboot.c
+++ b/arch/ia64/kernel/smpboot.c
@@ -682,7 +682,7 @@ int migrate_platform_irqs(unsigned int cpu)
 {
 	int new_cpei_cpu;
 	irq_desc_t *desc = NULL;
-	cpumask_t 	mask;
+	const struct cpumask *mask;
 	int 		retval = 0;
 
 	/*
@@ -695,7 +695,7 @@ int migrate_platform_irqs(unsigned int cpu)
 			 * Now re-target the CPEI to a different processor
 			 */
 			new_cpei_cpu = any_online_cpu(cpu_online_map);
-			mask = cpumask_of_cpu(new_cpei_cpu);
+			mask = cpumask_of(new_cpei_cpu);
 			set_cpei_target_cpu(new_cpei_cpu);
 			desc = irq_desc + ia64_cpe_irq;
 			/*
diff --git a/arch/ia64/sn/kernel/irq.c b/arch/ia64/sn/kernel/irq.c
index 0c66dbdd1d72..66fd705e82c0 100644
--- a/arch/ia64/sn/kernel/irq.c
+++ b/arch/ia64/sn/kernel/irq.c
@@ -227,14 +227,14 @@ finish_up:
 	return new_irq_info;
 }
 
-static void sn_set_affinity_irq(unsigned int irq, cpumask_t mask)
+static void 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;
 	int slice;
 
-	nasid = cpuid_to_nasid(first_cpu(mask));
-	slice = cpuid_to_slice(first_cpu(mask));
+	nasid = cpuid_to_nasid(cpumask_first(mask));
+	slice = cpuid_to_slice(cpumask_first(mask));
 
 	list_for_each_entry_safe(sn_irq_info, sn_irq_info_safe,
 				 sn_irq_lh[irq], list)
diff --git a/arch/ia64/sn/kernel/msi_sn.c b/arch/ia64/sn/kernel/msi_sn.c
index 83f190ffe350..ca553b0429ce 100644
--- a/arch/ia64/sn/kernel/msi_sn.c
+++ b/arch/ia64/sn/kernel/msi_sn.c
@@ -151,7 +151,8 @@ 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, cpumask_t cpu_mask)
+static void sn_set_msi_irq_affinity(unsigned int irq,
+				    const struct cpumask *cpu_mask)
 {
 	struct msi_msg msg;
 	int slice;
@@ -164,7 +165,7 @@ static void sn_set_msi_irq_affinity(unsigned int irq, cpumask_t cpu_mask)
 	struct sn_pcibus_provider *provider;
 	unsigned int cpu;
 
-	cpu = first_cpu(cpu_mask);
+	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;
@@ -204,7 +205,7 @@ static void sn_set_msi_irq_affinity(unsigned int irq, cpumask_t cpu_mask)
 	msg.address_lo = (u32)(bus_addr & 0x00000000ffffffff);
 
 	write_msi_msg(irq, &msg);
-	irq_desc[irq].affinity = cpu_mask;
+	irq_desc[irq].affinity = *cpu_mask;
 }
 #endif /* CONFIG_SMP */
 
diff --git a/arch/mips/include/asm/irq.h b/arch/mips/include/asm/irq.h
index a58f0eecc68f..abc62aa744ac 100644
--- a/arch/mips/include/asm/irq.h
+++ b/arch/mips/include/asm/irq.h
@@ -49,7 +49,8 @@ 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, cpumask_t affinity);
+extern void 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/cevt-bcm1480.c b/arch/mips/kernel/cevt-bcm1480.c
index 0a57f86945f1..d7e21bc8cd21 100644
--- a/arch/mips/kernel/cevt-bcm1480.c
+++ b/arch/mips/kernel/cevt-bcm1480.c
@@ -148,6 +148,6 @@ void __cpuinit sb1480_clockevent_init(void)
 	action->name	= name;
 	action->dev_id	= cd;
 
-	irq_set_affinity(irq, cpumask_of_cpu(cpu));
+	irq_set_affinity(irq, cpumask_of(cpu));
 	setup_irq(irq, action);
 }
diff --git a/arch/mips/kernel/cevt-sb1250.c b/arch/mips/kernel/cevt-sb1250.c
index 63ac3ad462bc..0f188cd46e03 100644
--- a/arch/mips/kernel/cevt-sb1250.c
+++ b/arch/mips/kernel/cevt-sb1250.c
@@ -147,6 +147,6 @@ void __cpuinit sb1250_clockevent_init(void)
 	action->name	= name;
 	action->dev_id	= cd;
 
-	irq_set_affinity(irq, cpumask_of_cpu(cpu));
+	irq_set_affinity(irq, cpumask_of(cpu));
 	setup_irq(irq, action);
 }
diff --git a/arch/mips/kernel/irq-gic.c b/arch/mips/kernel/irq-gic.c
index f0a4bb19e096..494a49a317e9 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, cpumask_t cpumask)
+static void gic_set_affinity(unsigned int irq, const struct cpumask *cpumask)
 {
 	cpumask_t	tmp = CPU_MASK_NONE;
 	unsigned long	flags;
@@ -164,7 +164,7 @@ static void gic_set_affinity(unsigned int irq, cpumask_t cpumask)
 	pr_debug(KERN_DEBUG "%s called\n", __func__);
 	irq -= _irqbase;
 
-	cpus_and(tmp, cpumask, cpu_online_map);
+	cpumask_and(&tmp, cpumask, cpu_online_mask);
 	if (cpus_empty(tmp))
 		return;
 
@@ -187,7 +187,7 @@ static void gic_set_affinity(unsigned int irq, cpumask_t cpumask)
 		set_bit(irq, pcpu_masks[first_cpu(tmp)].pcpu_mask);
 
 	}
-	irq_desc[irq].affinity = cpumask;
+	irq_desc[irq].affinity = *cpumask;
 	spin_unlock_irqrestore(&gic_lock, flags);
 
 }
diff --git a/arch/mips/mti-malta/malta-smtc.c b/arch/mips/mti-malta/malta-smtc.c
index f84a46a8ae6e..aabd7274507b 100644
--- a/arch/mips/mti-malta/malta-smtc.c
+++ b/arch/mips/mti-malta/malta-smtc.c
@@ -114,9 +114,9 @@ struct plat_smp_ops msmtc_smp_ops = {
  */
 
 
-void plat_set_irq_affinity(unsigned int irq, cpumask_t affinity)
+void plat_set_irq_affinity(unsigned int irq, const struct cpumask *affinity)
 {
-	cpumask_t tmask = affinity;
+	cpumask_t tmask = *affinity;
 	int cpu = 0;
 	void smtc_set_irq_affinity(unsigned int irq, cpumask_t aff);
 
@@ -139,7 +139,7 @@ void plat_set_irq_affinity(unsigned int irq, cpumask_t affinity)
 	 * be made to forward to an offline "CPU".
 	 */
 
-	for_each_cpu_mask(cpu, affinity) {
+	for_each_cpu(cpu, affinity) {
 		if ((cpu_data[cpu].vpe_id != 0) || !cpu_online(cpu))
 			cpu_clear(cpu, tmask);
 	}
diff --git a/arch/mips/sibyte/bcm1480/irq.c b/arch/mips/sibyte/bcm1480/irq.c
index a35818ed4263..12b465d404df 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, cpumask_t mask);
+static void 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, cpumask_t mask)
+static void bcm1480_set_affinity(unsigned int irq, const struct cpumask *mask)
 {
 	int i = 0, old_cpu, cpu, int_on, k;
 	u64 cur_ints;
@@ -117,11 +117,11 @@ static void bcm1480_set_affinity(unsigned int irq, cpumask_t mask)
 	unsigned long flags;
 	unsigned int irq_dirty;
 
-	if (cpus_weight(mask) != 1) {
+	if (cpumask_weight(mask) != 1) {
 		printk("attempted to set irq affinity for irq %d to multiple CPUs\n", irq);
 		return;
 	}
-	i = first_cpu(mask);
+	i = cpumask_first(mask);
 
 	/* Convert logical CPU to physical CPU */
 	cpu = cpu_logical_map(i);
diff --git a/arch/mips/sibyte/sb1250/irq.c b/arch/mips/sibyte/sb1250/irq.c
index a5158483986e..808ac2959b8c 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, cpumask_t mask);
+static void sb1250_set_affinity(unsigned int irq, const struct cpumask *mask);
 #endif
 
 #ifdef CONFIG_SIBYTE_HAS_LDT
@@ -103,16 +103,16 @@ void sb1250_unmask_irq(int cpu, int irq)
 }
 
 #ifdef CONFIG_SMP
-static void sb1250_set_affinity(unsigned int irq, cpumask_t mask)
+static void sb1250_set_affinity(unsigned int irq, const struct cpumask *mask)
 {
 	int i = 0, old_cpu, cpu, int_on;
 	u64 cur_ints;
 	struct irq_desc *desc = irq_desc + irq;
 	unsigned long flags;
 
-	i = first_cpu(mask);
+	i = cpumask_first(mask);
 
-	if (cpus_weight(mask) > 1) {
+	if (cpumask_weight(mask) > 1) {
 		printk("attempted to set irq affinity for irq %d to multiple CPUs\n", irq);
 		return;
 	}
diff --git a/arch/parisc/kernel/irq.c b/arch/parisc/kernel/irq.c
index 23ef950df008..4cea935e2f99 100644
--- a/arch/parisc/kernel/irq.c
+++ b/arch/parisc/kernel/irq.c
@@ -131,12 +131,12 @@ int cpu_check_affinity(unsigned int irq, cpumask_t *dest)
 	return 0;
 }
 
-static void cpu_set_affinity_irq(unsigned int irq, cpumask_t dest)
+static void cpu_set_affinity_irq(unsigned int irq, const struct cpumask *dest)
 {
-	if (cpu_check_affinity(irq, &dest))
+	if (cpu_check_affinity(irq, dest))
 		return;
 
-	irq_desc[irq].affinity = dest;
+	irq_desc[irq].affinity = *dest;
 }
 #endif
 
diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c
index ac222d0ab12e..23b8b5e36f98 100644
--- a/arch/powerpc/kernel/irq.c
+++ b/arch/powerpc/kernel/irq.c
@@ -237,7 +237,7 @@ void fixup_irqs(cpumask_t map)
 			mask = map;
 		}
 		if (irq_desc[irq].chip->set_affinity)
-			irq_desc[irq].chip->set_affinity(irq, mask);
+			irq_desc[irq].chip->set_affinity(irq, &mask);
 		else if (irq_desc[irq].action && !(warned++))
 			printk("Cannot set affinity for irq %i\n", irq);
 	}
diff --git a/arch/powerpc/platforms/pseries/xics.c b/arch/powerpc/platforms/pseries/xics.c
index 64d24310ce7e..424b335a71c8 100644
--- a/arch/powerpc/platforms/pseries/xics.c
+++ b/arch/powerpc/platforms/pseries/xics.c
@@ -332,7 +332,7 @@ static void xics_eoi_lpar(unsigned int virq)
 	lpar_xirr_info_set((0xff << 24) | irq);
 }
 
-static void xics_set_affinity(unsigned int virq, cpumask_t cpumask)
+static void xics_set_affinity(unsigned int virq, const struct cpumask *cpumask)
 {
 	unsigned int irq;
 	int status;
@@ -358,7 +358,7 @@ static void xics_set_affinity(unsigned int virq, cpumask_t cpumask)
 	irq_server = get_irq_server(virq, 1);
 	if (irq_server == -1) {
 		char cpulist[128];
-		cpumask_scnprintf(cpulist, sizeof(cpulist), &cpumask);
+		cpumask_scnprintf(cpulist, sizeof(cpulist), cpumask);
 		printk(KERN_WARNING
 			"%s: No online cpus in the mask %s for irq %d\n",
 			__func__, cpulist, virq);
@@ -845,7 +845,7 @@ void xics_migrate_irqs_away(void)
 
 		/* Reset affinity to all cpus */
 		irq_desc[virq].affinity = CPU_MASK_ALL;
-		desc->chip->set_affinity(virq, CPU_MASK_ALL);
+		desc->chip->set_affinity(virq, cpu_all_mask);
 unlock:
 		spin_unlock_irqrestore(&desc->lock, flags);
 	}
diff --git a/arch/powerpc/sysdev/mpic.c b/arch/powerpc/sysdev/mpic.c
index 1890fb085cde..5d7f9f0c93c3 100644
--- a/arch/powerpc/sysdev/mpic.c
+++ b/arch/powerpc/sysdev/mpic.c
@@ -817,7 +817,7 @@ static void mpic_end_ipi(unsigned int irq)
 
 #endif /* CONFIG_SMP */
 
-void mpic_set_affinity(unsigned int irq, cpumask_t cpumask)
+void 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);
@@ -829,7 +829,7 @@ void mpic_set_affinity(unsigned int irq, cpumask_t cpumask)
 	} else {
 		cpumask_t tmp;
 
-		cpus_and(tmp, cpumask, cpu_online_map);
+		cpumask_and(&tmp, cpumask, cpu_online_mask);
 
 		mpic_irq_write(src, MPIC_INFO(IRQ_DESTINATION),
 			       mpic_physmask(cpus_addr(tmp)[0]));
diff --git a/arch/powerpc/sysdev/mpic.h b/arch/powerpc/sysdev/mpic.h
index 6209c62a426d..3cef2af10f42 100644
--- a/arch/powerpc/sysdev/mpic.h
+++ b/arch/powerpc/sysdev/mpic.h
@@ -36,6 +36,6 @@ static inline int mpic_pasemi_msi_init(struct mpic *mpic)
 
 extern int mpic_set_irq_type(unsigned int virq, unsigned int flow_type);
 extern void mpic_set_vector(unsigned int virq, unsigned int vector);
-extern void mpic_set_affinity(unsigned int irq, cpumask_t cpumask);
+extern void mpic_set_affinity(unsigned int irq, const struct cpumask *cpumask);
 
 #endif /* _POWERPC_SYSDEV_MPIC_H */
diff --git a/arch/sparc64/kernel/irq.c b/arch/sparc64/kernel/irq.c
index 52fc836f464d..4aaf18e83c8c 100644
--- a/arch/sparc64/kernel/irq.c
+++ b/arch/sparc64/kernel/irq.c
@@ -312,7 +312,8 @@ static void sun4u_irq_enable(unsigned int virt_irq)
 	}
 }
 
-static void sun4u_set_affinity(unsigned int virt_irq, cpumask_t mask)
+static void sun4u_set_affinity(unsigned int virt_irq,
+			       const struct cpumask *mask)
 {
 	sun4u_irq_enable(virt_irq);
 }
@@ -362,7 +363,8 @@ static void sun4v_irq_enable(unsigned int virt_irq)
 		       ino, err);
 }
 
-static void sun4v_set_affinity(unsigned int virt_irq, cpumask_t mask)
+static void sun4v_set_affinity(unsigned int virt_irq,
+			       const struct cpumask *mask)
 {
 	unsigned int ino = virt_irq_table[virt_irq].dev_ino;
 	unsigned long cpuid = irq_choose_cpu(virt_irq);
@@ -429,7 +431,8 @@ static void sun4v_virq_enable(unsigned int virt_irq)
 		       dev_handle, dev_ino, err);
 }
 
-static void sun4v_virt_set_affinity(unsigned int virt_irq, cpumask_t mask)
+static void sun4v_virt_set_affinity(unsigned int virt_irq,
+				    const struct cpumask *mask)
 {
 	unsigned long cpuid, dev_handle, dev_ino;
 	int err;
@@ -788,7 +791,7 @@ void fixup_irqs(void)
 		    !(irq_desc[irq].status & IRQ_PER_CPU)) {
 			if (irq_desc[irq].chip->set_affinity)
 				irq_desc[irq].chip->set_affinity(irq,
-					irq_desc[irq].affinity);
+					&irq_desc[irq].affinity);
 		}
 		spin_unlock_irqrestore(&irq_desc[irq].lock, flags);
 	}
diff --git a/arch/sparc64/kernel/of_device.c b/arch/sparc64/kernel/of_device.c
index 0f616ae3246c..df2efb7fc14c 100644
--- a/arch/sparc64/kernel/of_device.c
+++ b/arch/sparc64/kernel/of_device.c
@@ -780,7 +780,7 @@ out:
 	if (nid != -1) {
 		cpumask_t numa_mask = node_to_cpumask(nid);
 
-		irq_set_affinity(irq, numa_mask);
+		irq_set_affinity(irq, &numa_mask);
 	}
 
 	return irq;
diff --git a/arch/sparc64/kernel/pci_msi.c b/arch/sparc64/kernel/pci_msi.c
index 2e680f34f727..0d0cd815e83e 100644
--- a/arch/sparc64/kernel/pci_msi.c
+++ b/arch/sparc64/kernel/pci_msi.c
@@ -288,7 +288,7 @@ static int bringup_one_msi_queue(struct pci_pbm_info *pbm,
 	if (nid != -1) {
 		cpumask_t numa_mask = node_to_cpumask(nid);
 
-		irq_set_affinity(irq, numa_mask);
+		irq_set_affinity(irq, &numa_mask);
 	}
 	err = request_irq(irq, sparc64_msiq_interrupt, 0,
 			  "MSIQ",
diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c
index 067d8de913f6..940f25851e1e 100644
--- a/arch/x86/kernel/hpet.c
+++ b/arch/x86/kernel/hpet.c
@@ -301,7 +301,7 @@ static void hpet_set_mode(enum clock_event_mode mode,
 			struct hpet_dev *hdev = EVT_TO_HPET_DEV(evt);
 			hpet_setup_msi_irq(hdev->irq);
 			disable_irq(hdev->irq);
-			irq_set_affinity(hdev->irq, cpumask_of_cpu(hdev->cpu));
+			irq_set_affinity(hdev->irq, cpumask_of(hdev->cpu));
 			enable_irq(hdev->irq);
 		}
 		break;
@@ -449,7 +449,7 @@ static int hpet_setup_irq(struct hpet_dev *dev)
 		return -1;
 
 	disable_irq(dev->irq);
-	irq_set_affinity(dev->irq, cpumask_of_cpu(dev->cpu));
+	irq_set_affinity(dev->irq, cpumask_of(dev->cpu));
 	enable_irq(dev->irq);
 
 	printk(KERN_DEBUG "hpet: %s irq %d for MSI\n",
diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c
index 9043251210fb..1184210e6d0c 100644
--- a/arch/x86/kernel/io_apic.c
+++ b/arch/x86/kernel/io_apic.c
@@ -361,7 +361,8 @@ static void __target_IO_APIC_irq(unsigned int irq, unsigned int dest, u8 vector)
 
 static int assign_irq_vector(int irq, cpumask_t mask);
 
-static void set_ioapic_affinity_irq(unsigned int irq, cpumask_t mask)
+static void set_ioapic_affinity_irq(unsigned int irq,
+				    const struct cpumask *mask)
 {
 	struct irq_cfg *cfg;
 	unsigned long flags;
@@ -369,15 +370,14 @@ static void set_ioapic_affinity_irq(unsigned int irq, cpumask_t mask)
 	cpumask_t tmp;
 	struct irq_desc *desc;
 
-	cpus_and(tmp, mask, cpu_online_map);
-	if (cpus_empty(tmp))
+	if (!cpumask_intersects(mask, cpu_online_mask))
 		return;
 
 	cfg = irq_cfg(irq);
-	if (assign_irq_vector(irq, mask))
+	if (assign_irq_vector(irq, *mask))
 		return;
 
-	cpus_and(tmp, cfg->domain, mask);
+	cpumask_and(&tmp, &cfg->domain, mask);
 	dest = cpu_mask_to_apicid(tmp);
 	/*
 	 * Only the high 8 bits are valid.
@@ -387,7 +387,7 @@ static void set_ioapic_affinity_irq(unsigned int irq, cpumask_t mask)
 	desc = irq_to_desc(irq);
 	spin_lock_irqsave(&ioapic_lock, flags);
 	__target_IO_APIC_irq(irq, dest, cfg->vector);
-	desc->affinity = mask;
+	cpumask_copy(&desc->affinity, mask);
 	spin_unlock_irqrestore(&ioapic_lock, flags);
 }
 #endif /* CONFIG_SMP */
@@ -2189,7 +2189,7 @@ static void ir_irq_migration(struct work_struct *work)
 				continue;
 			}
 
-			desc->chip->set_affinity(irq, desc->pending_mask);
+			desc->chip->set_affinity(irq, &desc->pending_mask);
 			spin_unlock_irqrestore(&desc->lock, flags);
 		}
 	}
@@ -2198,18 +2198,19 @@ static void ir_irq_migration(struct work_struct *work)
 /*
  * Migrates the IRQ destination in the process context.
  */
-static void set_ir_ioapic_affinity_irq(unsigned int irq, cpumask_t mask)
+static void set_ir_ioapic_affinity_irq(unsigned int irq,
+				       const struct cpumask *mask)
 {
 	struct irq_desc *desc = irq_to_desc(irq);
 
 	if (desc->status & IRQ_LEVEL) {
 		desc->status |= IRQ_MOVE_PENDING;
-		desc->pending_mask = mask;
+		cpumask_copy(&desc->pending_mask, mask);
 		migrate_irq_remapped_level(irq);
 		return;
 	}
 
-	migrate_ioapic_irq(irq, mask);
+	migrate_ioapic_irq(irq, *mask);
 }
 #endif
 
@@ -3027,7 +3028,7 @@ 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, cpumask_t mask)
+static void set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask)
 {
 	struct irq_cfg *cfg;
 	struct msi_msg msg;
@@ -3035,15 +3036,14 @@ static void set_msi_irq_affinity(unsigned int irq, cpumask_t mask)
 	cpumask_t tmp;
 	struct irq_desc *desc;
 
-	cpus_and(tmp, mask, cpu_online_map);
-	if (cpus_empty(tmp))
+	if (!cpumask_intersects(mask, cpu_online_mask))
 		return;
 
-	if (assign_irq_vector(irq, mask))
+	if (assign_irq_vector(irq, *mask))
 		return;
 
 	cfg = irq_cfg(irq);
-	cpus_and(tmp, cfg->domain, mask);
+	cpumask_and(&tmp, &cfg->domain, mask);
 	dest = cpu_mask_to_apicid(tmp);
 
 	read_msi_msg(irq, &msg);
@@ -3055,7 +3055,7 @@ static void set_msi_irq_affinity(unsigned int irq, cpumask_t mask)
 
 	write_msi_msg(irq, &msg);
 	desc = irq_to_desc(irq);
-	desc->affinity = mask;
+	cpumask_copy(&desc->affinity, mask);
 }
 
 #ifdef CONFIG_INTR_REMAP
@@ -3063,7 +3063,8 @@ static void set_msi_irq_affinity(unsigned int irq, cpumask_t mask)
  * Migrate the MSI irq to another cpumask. This migration is
  * done in the process context using interrupt-remapping hardware.
  */
-static void ir_set_msi_irq_affinity(unsigned int irq, cpumask_t mask)
+static void ir_set_msi_irq_affinity(unsigned int irq,
+				    const struct cpumask *mask)
 {
 	struct irq_cfg *cfg;
 	unsigned int dest;
@@ -3071,18 +3072,17 @@ static void ir_set_msi_irq_affinity(unsigned int irq, cpumask_t mask)
 	struct irte irte;
 	struct irq_desc *desc;
 
-	cpus_and(tmp, mask, cpu_online_map);
-	if (cpus_empty(tmp))
+	if (!cpumask_intersects(mask, cpu_online_mask))
 		return;
 
 	if (get_irte(irq, &irte))
 		return;
 
-	if (assign_irq_vector(irq, mask))
+	if (assign_irq_vector(irq, *mask))
 		return;
 
 	cfg = irq_cfg(irq);
-	cpus_and(tmp, cfg->domain, mask);
+	cpumask_and(&tmp, &cfg->domain, mask);
 	dest = cpu_mask_to_apicid(tmp);
 
 	irte.vector = cfg->vector;
@@ -3106,7 +3106,7 @@ static void ir_set_msi_irq_affinity(unsigned int irq, cpumask_t mask)
 	}
 
 	desc = irq_to_desc(irq);
-	desc->affinity = mask;
+	cpumask_copy(&desc->affinity, mask);
 }
 #endif
 #endif /* CONFIG_SMP */
@@ -3308,7 +3308,7 @@ void arch_teardown_msi_irq(unsigned int irq)
 
 #ifdef CONFIG_DMAR
 #ifdef CONFIG_SMP
-static void dmar_msi_set_affinity(unsigned int irq, cpumask_t mask)
+static void dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask)
 {
 	struct irq_cfg *cfg;
 	struct msi_msg msg;
@@ -3316,15 +3316,14 @@ static void dmar_msi_set_affinity(unsigned int irq, cpumask_t mask)
 	cpumask_t tmp;
 	struct irq_desc *desc;
 
-	cpus_and(tmp, mask, cpu_online_map);
-	if (cpus_empty(tmp))
+	if (!cpumask_intersects(mask, cpu_online_mask))
 		return;
 
-	if (assign_irq_vector(irq, mask))
+	if (assign_irq_vector(irq, *mask))
 		return;
 
 	cfg = irq_cfg(irq);
-	cpus_and(tmp, cfg->domain, mask);
+	cpumask_and(&tmp, &cfg->domain, mask);
 	dest = cpu_mask_to_apicid(tmp);
 
 	dmar_msi_read(irq, &msg);
@@ -3336,7 +3335,7 @@ static void dmar_msi_set_affinity(unsigned int irq, cpumask_t mask)
 
 	dmar_msi_write(irq, &msg);
 	desc = irq_to_desc(irq);
-	desc->affinity = mask;
+	cpumask_copy(&desc->affinity, mask);
 }
 #endif /* CONFIG_SMP */
 
@@ -3369,7 +3368,7 @@ int arch_setup_dmar_msi(unsigned int irq)
 #ifdef CONFIG_HPET_TIMER
 
 #ifdef CONFIG_SMP
-static void hpet_msi_set_affinity(unsigned int irq, cpumask_t mask)
+static void hpet_msi_set_affinity(unsigned int irq, const struct cpumask *mask)
 {
 	struct irq_cfg *cfg;
 	struct irq_desc *desc;
@@ -3377,15 +3376,14 @@ static void hpet_msi_set_affinity(unsigned int irq, cpumask_t mask)
 	unsigned int dest;
 	cpumask_t tmp;
 
-	cpus_and(tmp, mask, cpu_online_map);
-	if (cpus_empty(tmp))
+	if (!cpumask_intersects(mask, cpu_online_mask))
 		return;
 
-	if (assign_irq_vector(irq, mask))
+	if (assign_irq_vector(irq, *mask))
 		return;
 
 	cfg = irq_cfg(irq);
-	cpus_and(tmp, cfg->domain, mask);
+	cpumask_and(&tmp, &cfg->domain, mask);
 	dest = cpu_mask_to_apicid(tmp);
 
 	hpet_msi_read(irq, &msg);
@@ -3397,7 +3395,7 @@ static void hpet_msi_set_affinity(unsigned int irq, cpumask_t mask)
 
 	hpet_msi_write(irq, &msg);
 	desc = irq_to_desc(irq);
-	desc->affinity = mask;
+	cpumask_copy(&desc->affinity, mask);
 }
 #endif /* CONFIG_SMP */
 
@@ -3451,27 +3449,26 @@ 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, cpumask_t mask)
+static void set_ht_irq_affinity(unsigned int irq, const struct cpumask *mask)
 {
 	struct irq_cfg *cfg;
 	unsigned int dest;
 	cpumask_t tmp;
 	struct irq_desc *desc;
 
-	cpus_and(tmp, mask, cpu_online_map);
-	if (cpus_empty(tmp))
+	if (!cpumask_intersects(mask, cpu_online_mask))
 		return;
 
-	if (assign_irq_vector(irq, mask))
+	if (assign_irq_vector(irq, *mask))
 		return;
 
 	cfg = irq_cfg(irq);
-	cpus_and(tmp, cfg->domain, mask);
+	cpumask_and(&tmp, &cfg->domain, mask);
 	dest = cpu_mask_to_apicid(tmp);
 
 	target_ht_irq(irq, dest, cfg->vector);
 	desc = irq_to_desc(irq);
-	desc->affinity = mask;
+	cpumask_copy(&desc->affinity, mask);
 }
 #endif
 
@@ -3794,10 +3791,10 @@ void __init setup_ioapic_dest(void)
 
 #ifdef CONFIG_INTR_REMAP
 			if (intr_remapping_enabled)
-				set_ir_ioapic_affinity_irq(irq, mask);
+				set_ir_ioapic_affinity_irq(irq, &mask);
 			else
 #endif
-				set_ioapic_affinity_irq(irq, mask);
+				set_ioapic_affinity_irq(irq, &mask);
 		}
 
 	}
diff --git a/arch/x86/kernel/irq_32.c b/arch/x86/kernel/irq_32.c
index a51382672de0..87870a49be4e 100644
--- a/arch/x86/kernel/irq_32.c
+++ b/arch/x86/kernel/irq_32.c
@@ -251,7 +251,7 @@ void fixup_irqs(cpumask_t map)
 			mask = map;
 		}
 		if (desc->chip->set_affinity)
-			desc->chip->set_affinity(irq, mask);
+			desc->chip->set_affinity(irq, &mask);
 		else if (desc->action && !(warned++))
 			printk("Cannot set affinity for irq %i\n", irq);
 	}
diff --git a/arch/x86/kernel/irq_64.c b/arch/x86/kernel/irq_64.c
index 60eb84eb77a0..7d37f847544d 100644
--- a/arch/x86/kernel/irq_64.c
+++ b/arch/x86/kernel/irq_64.c
@@ -116,7 +116,7 @@ void fixup_irqs(cpumask_t map)
 			desc->chip->mask(irq);
 
 		if (desc->chip->set_affinity)
-			desc->chip->set_affinity(irq, mask);
+			desc->chip->set_affinity(irq, &mask);
 		else if (!(warned++))
 			set_affinity = 0;
 
diff --git a/drivers/parisc/iosapic.c b/drivers/parisc/iosapic.c
index 7beffcab2745..9dedbbd218c3 100644
--- a/drivers/parisc/iosapic.c
+++ b/drivers/parisc/iosapic.c
@@ -704,16 +704,17 @@ static unsigned int iosapic_startup_irq(unsigned int irq)
 }
 
 #ifdef CONFIG_SMP
-static void iosapic_set_affinity_irq(unsigned int irq, cpumask_t dest)
+static void iosapic_set_affinity_irq(unsigned int irq,
+				     const struct cpumask *dest)
 {
 	struct vector_info *vi = iosapic_get_vector(irq);
 	u32 d0, d1, dummy_d0;
 	unsigned long flags;
 
-	if (cpu_check_affinity(irq, &dest))
+	if (cpu_check_affinity(irq, dest))
 		return;
 
-	vi->txn_addr = txn_affinity_addr(irq, first_cpu(dest));
+	vi->txn_addr = txn_affinity_addr(irq, cpumask_first(dest));
 
 	spin_lock_irqsave(&iosapic_lock, flags);
 	/* d1 contains the destination CPU, so only want to set that
diff --git a/drivers/xen/events.c b/drivers/xen/events.c
index 1e3b934a4cf7..eba5ec5b020e 100644
--- a/drivers/xen/events.c
+++ b/drivers/xen/events.c
@@ -579,7 +579,7 @@ void rebind_evtchn_irq(int evtchn, int irq)
 	spin_unlock(&irq_mapping_update_lock);
 
 	/* new event channels are always bound to cpu 0 */
-	irq_set_affinity(irq, cpumask_of_cpu(0));
+	irq_set_affinity(irq, cpumask_of(0));
 
 	/* Unmask the event channel. */
 	enable_irq(irq);
@@ -608,9 +608,9 @@ static void rebind_irq_to_cpu(unsigned irq, unsigned tcpu)
 }
 
 
-static void set_affinity_irq(unsigned irq, cpumask_t dest)
+static void set_affinity_irq(unsigned irq, const struct cpumask *dest)
 {
-	unsigned tcpu = first_cpu(dest);
+	unsigned tcpu = cpumask_first(dest);
 	rebind_irq_to_cpu(irq, tcpu);
 }
 
diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h
index f58a0cf8929a..48e63934fabe 100644
--- a/include/linux/interrupt.h
+++ b/include/linux/interrupt.h
@@ -109,13 +109,13 @@ extern void enable_irq(unsigned int irq);
 
 extern cpumask_t irq_default_affinity;
 
-extern int irq_set_affinity(unsigned int irq, cpumask_t cpumask);
+extern int irq_set_affinity(unsigned int irq, const struct cpumask *cpumask);
 extern int irq_can_set_affinity(unsigned int irq);
 extern int irq_select_affinity(unsigned int irq);
 
 #else /* CONFIG_SMP */
 
-static inline int irq_set_affinity(unsigned int irq, cpumask_t cpumask)
+static inline int irq_set_affinity(unsigned int irq, const struct cpumask *m)
 {
 	return -EINVAL;
 }
diff --git a/include/linux/irq.h b/include/linux/irq.h
index 3dddfa703ebd..ab70fd604d3a 100644
--- a/include/linux/irq.h
+++ b/include/linux/irq.h
@@ -113,7 +113,8 @@ struct irq_chip {
 	void		(*eoi)(unsigned int irq);
 
 	void		(*end)(unsigned int irq);
-	void		(*set_affinity)(unsigned int irq, cpumask_t dest);
+	void		(*set_affinity)(unsigned int irq,
+					const struct cpumask *dest);
 	int		(*retrigger)(unsigned int irq);
 	int		(*set_type)(unsigned int irq, unsigned int flow_type);
 	int		(*set_wake)(unsigned int irq, unsigned int on);
diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c
index 10b5092e9bfe..58d8e31daa49 100644
--- a/kernel/irq/chip.c
+++ b/kernel/irq/chip.c
@@ -45,7 +45,7 @@ void dynamic_irq_init(unsigned int irq)
 	desc->irq_count = 0;
 	desc->irqs_unhandled = 0;
 #ifdef CONFIG_SMP
-	cpus_setall(desc->affinity);
+	cpumask_setall(&desc->affinity);
 #endif
 	spin_unlock_irqrestore(&desc->lock, flags);
 }
diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c
index 801addda3c43..10ad2f87ed9a 100644
--- a/kernel/irq/manage.c
+++ b/kernel/irq/manage.c
@@ -79,7 +79,7 @@ int irq_can_set_affinity(unsigned int irq)
  *	@cpumask:	cpumask
  *
  */
-int irq_set_affinity(unsigned int irq, cpumask_t cpumask)
+int irq_set_affinity(unsigned int irq, const struct cpumask *cpumask)
 {
 	struct irq_desc *desc = irq_to_desc(irq);
 	unsigned long flags;
@@ -91,14 +91,14 @@ int irq_set_affinity(unsigned int irq, cpumask_t cpumask)
 
 #ifdef CONFIG_GENERIC_PENDING_IRQ
 	if (desc->status & IRQ_MOVE_PCNTXT || desc->status & IRQ_DISABLED) {
-		desc->affinity = cpumask;
+		cpumask_copy(&desc->affinity, cpumask);
 		desc->chip->set_affinity(irq, cpumask);
 	} else {
 		desc->status |= IRQ_MOVE_PENDING;
-		desc->pending_mask = cpumask;
+		cpumask_copy(&desc->pending_mask, cpumask);
 	}
 #else
-	desc->affinity = cpumask;
+	cpumask_copy(&desc->affinity, cpumask);
 	desc->chip->set_affinity(irq, cpumask);
 #endif
 	desc->status |= IRQ_AFFINITY_SET;
@@ -112,26 +112,24 @@ int irq_set_affinity(unsigned int irq, cpumask_t cpumask)
  */
 int do_irq_select_affinity(unsigned int irq, struct irq_desc *desc)
 {
-	cpumask_t mask;
-
 	if (!irq_can_set_affinity(irq))
 		return 0;
 
-	cpus_and(mask, cpu_online_map, irq_default_affinity);
-
 	/*
 	 * Preserve an userspace affinity setup, but make sure that
 	 * one of the targets is online.
 	 */
 	if (desc->status & (IRQ_AFFINITY_SET | IRQ_NO_BALANCING)) {
-		if (cpus_intersects(desc->affinity, cpu_online_map))
-			mask = desc->affinity;
+		if (cpumask_any_and(&desc->affinity, cpu_online_mask)
+		    < nr_cpu_ids)
+			goto set_affinity;
 		else
 			desc->status &= ~IRQ_AFFINITY_SET;
 	}
 
-	desc->affinity = mask;
-	desc->chip->set_affinity(irq, mask);
+	cpumask_and(&desc->affinity, cpu_online_mask, &irq_default_affinity);
+set_affinity:
+	desc->chip->set_affinity(irq, &desc->affinity);
 
 	return 0;
 }
diff --git a/kernel/irq/migration.c b/kernel/irq/migration.c
index 9db681d95814..bd72329e630c 100644
--- a/kernel/irq/migration.c
+++ b/kernel/irq/migration.c
@@ -4,7 +4,6 @@
 void move_masked_irq(int irq)
 {
 	struct irq_desc *desc = irq_to_desc(irq);
-	cpumask_t tmp;
 
 	if (likely(!(desc->status & IRQ_MOVE_PENDING)))
 		return;
@@ -19,7 +18,7 @@ void move_masked_irq(int irq)
 
 	desc->status &= ~IRQ_MOVE_PENDING;
 
-	if (unlikely(cpus_empty(desc->pending_mask)))
+	if (unlikely(cpumask_empty(&desc->pending_mask)))
 		return;
 
 	if (!desc->chip->set_affinity)
@@ -27,8 +26,6 @@ void move_masked_irq(int irq)
 
 	assert_spin_locked(&desc->lock);
 
-	cpus_and(tmp, desc->pending_mask, cpu_online_map);
-
 	/*
 	 * If there was a valid mask to work with, please
 	 * do the disable, re-program, enable sequence.
@@ -41,10 +38,13 @@ void move_masked_irq(int irq)
 	 * For correct operation this depends on the caller
 	 * masking the irqs.
 	 */
-	if (likely(!cpus_empty(tmp))) {
-		desc->chip->set_affinity(irq,tmp);
+	if (likely(cpumask_any_and(&desc->pending_mask, cpu_online_mask)
+		   < nr_cpu_ids)) {
+		cpumask_and(&desc->affinity,
+			    &desc->pending_mask, cpu_online_mask);
+		desc->chip->set_affinity(irq, &desc->affinity);
 	}
-	cpus_clear(desc->pending_mask);
+	cpumask_clear(&desc->pending_mask);
 }
 
 void move_native_irq(int irq)
diff --git a/kernel/irq/proc.c b/kernel/irq/proc.c
index f293349d49d0..8e91c9762520 100644
--- a/kernel/irq/proc.c
+++ b/kernel/irq/proc.c
@@ -40,33 +40,42 @@ static ssize_t irq_affinity_proc_write(struct file *file,
 		const char __user *buffer, size_t count, loff_t *pos)
 {
 	unsigned int irq = (int)(long)PDE(file->f_path.dentry->d_inode)->data;
-	cpumask_t new_value;
+	cpumask_var_t new_value;
 	int err;
 
 	if (!irq_to_desc(irq)->chip->set_affinity || no_irq_affinity ||
 	    irq_balancing_disabled(irq))
 		return -EIO;
 
-	err = cpumask_parse_user(buffer, count, &new_value);
+	if (!alloc_cpumask_var(&new_value, GFP_KERNEL))
+		return -ENOMEM;
+
+	err = cpumask_parse_user(buffer, count, new_value);
 	if (err)
-		return err;
+		goto free_cpumask;
 
-	if (!is_affinity_mask_valid(new_value))
-		return -EINVAL;
+	if (!is_affinity_mask_valid(*new_value)) {
+		err = -EINVAL;
+		goto free_cpumask;
+	}
 
 	/*
 	 * Do not allow disabling IRQs completely - it's a too easy
 	 * way to make the system unusable accidentally :-) At least
 	 * one online CPU still has to be targeted.
 	 */
-	if (!cpus_intersects(new_value, cpu_online_map))
+	if (!cpumask_intersects(new_value, cpu_online_mask)) {
 		/* Special case for empty set - allow the architecture
 		   code to set default SMP affinity. */
-		return irq_select_affinity_usr(irq) ? -EINVAL : count;
-
-	irq_set_affinity(irq, new_value);
+		err = irq_select_affinity_usr(irq) ? -EINVAL : count;
+	} else {
+		irq_set_affinity(irq, new_value);
+		err = count;
+	}
 
-	return count;
+free_cpumask:
+	free_cpumask_var(new_value);
+	return err;
 }
 
 static int irq_affinity_proc_open(struct inode *inode, struct file *file)
diff --git a/kernel/time/tick-common.c b/kernel/time/tick-common.c
index df12434b43ca..ab65d217583f 100644
--- a/kernel/time/tick-common.c
+++ b/kernel/time/tick-common.c
@@ -136,7 +136,7 @@ void tick_setup_periodic(struct clock_event_device *dev, int broadcast)
  */
 static void tick_setup_device(struct tick_device *td,
 			      struct clock_event_device *newdev, int cpu,
-			      const cpumask_t *cpumask)
+			      const struct cpumask *cpumask)
 {
 	ktime_t next_event;
 	void (*handler)(struct clock_event_device *) = NULL;
@@ -171,8 +171,8 @@ static void tick_setup_device(struct tick_device *td,
 	 * When the device is not per cpu, pin the interrupt to the
 	 * current cpu:
 	 */
-	if (!cpus_equal(newdev->cpumask, *cpumask))
-		irq_set_affinity(newdev->irq, *cpumask);
+	if (!cpumask_equal(&newdev->cpumask, cpumask))
+		irq_set_affinity(newdev->irq, cpumask);
 
 	/*
 	 * When global broadcasting is active, check if the current
-- 
cgit v1.3-7-g2ca7


From 43a256322ac1fc105c181b3cade3b9bfc0b63ca1 Mon Sep 17 00:00:00 2001
From: Yinghai Lu <yhlu.kernel@gmail.com>
Date: Sun, 28 Dec 2008 16:01:13 -0800
Subject: sparseirq: move __weak symbols into separate compilation unit

GCC has a bug with __weak alias functions: if the functions are in
the same compilation unit as their call site, GCC can decide to
inline them - and thus rob the linker of the opportunity to override
the weak alias with the real thing.

So move all the IRQ handling related __weak symbols to kernel/irq/chip.c.

Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 include/linux/interrupt.h |  6 ++++++
 include/linux/irq.h       |  3 ---
 init/main.c               | 10 ----------
 kernel/irq/manage.c       |  9 ---------
 kernel/softirq.c          | 20 ++++++++++++++++++++
 5 files changed, 26 insertions(+), 22 deletions(-)

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

diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h
index 777f89e00b4a..d9a370325ae2 100644
--- a/include/linux/interrupt.h
+++ b/include/linux/interrupt.h
@@ -467,4 +467,10 @@ static inline void init_irq_proc(void)
 
 int show_interrupts(struct seq_file *p, void *v);
 
+struct irq_desc;
+
+extern int early_irq_init(void);
+extern int arch_early_irq_init(void);
+extern int arch_init_chip_data(struct irq_desc *desc, int cpu);
+
 #endif
diff --git a/include/linux/irq.h b/include/linux/irq.h
index 0e40af4bac40..d64a6d49bdef 100644
--- a/include/linux/irq.h
+++ b/include/linux/irq.h
@@ -193,9 +193,6 @@ struct irq_desc {
 	const char		*name;
 } ____cacheline_internodealigned_in_smp;
 
-extern int early_irq_init(void);
-extern int arch_early_irq_init(void);
-extern int arch_init_chip_data(struct irq_desc *desc, int cpu);
 extern void arch_init_copy_chip_data(struct irq_desc *old_desc,
 					struct irq_desc *desc, int cpu);
 extern void arch_free_chip_data(struct irq_desc *old_desc, struct irq_desc *desc);
diff --git a/init/main.c b/init/main.c
index c314aa15370e..2c183abbf61c 100644
--- a/init/main.c
+++ b/init/main.c
@@ -539,16 +539,6 @@ void __init __weak thread_info_cache_init(void)
 {
 }
 
-int __init __weak arch_early_irq_init(void)
-{
-	return 0;
-}
-
-int __init __weak early_irq_init(void)
-{
-	return arch_early_irq_init();
-}
-
 asmlinkage void __init start_kernel(void)
 {
 	char * command_line;
diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c
index c2741b02ad38..46953a06f4a8 100644
--- a/kernel/irq/manage.c
+++ b/kernel/irq/manage.c
@@ -261,15 +261,6 @@ void enable_irq(unsigned int irq)
 }
 EXPORT_SYMBOL(enable_irq);
 
-/*
- * [ Not in kernel/irq/handle.c, so that GCC does not
- *   inline the __weak alias: ]
- */
-int __weak arch_init_chip_data(struct irq_desc *desc, int cpu)
-{
-	return 0;
-}
-
 static int set_irq_wake_real(unsigned int irq, unsigned int on)
 {
 	struct irq_desc *desc = irq_to_desc(irq);
diff --git a/kernel/softirq.c b/kernel/softirq.c
index e7c69a720d69..daf46358d2dd 100644
--- a/kernel/softirq.c
+++ b/kernel/softirq.c
@@ -797,3 +797,23 @@ int on_each_cpu(void (*func) (void *info), void *info, int wait)
 }
 EXPORT_SYMBOL(on_each_cpu);
 #endif
+
+/*
+ * [ These __weak aliases are kept in a separate compilation unit, so that
+ *   GCC does not inline them incorrectly. ]
+ */
+
+int __init __weak early_irq_init(void)
+{
+	return 0;
+}
+
+int __init __weak arch_early_irq_init(void)
+{
+	return 0;
+}
+
+int __weak arch_init_chip_data(struct irq_desc *desc, int cpu)
+{
+	return 0;
+}
-- 
cgit v1.3-7-g2ca7


From d036e67b40f52bdd95392390108defbac7e53837 Mon Sep 17 00:00:00 2001
From: Rusty Russell <rusty@rustcorp.com.au>
Date: Thu, 1 Jan 2009 10:12:26 +1030
Subject: cpumask: convert kernel/irq

Impact: Reduce stack usage, use new cpumask API.  ALPHA mod!

Main change is that irq_default_affinity becomes a cpumask_var_t, so
treat it as a pointer (this effects alpha).

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
---
 arch/alpha/kernel/irq.c   |  3 ++-
 include/linux/interrupt.h |  2 +-
 kernel/irq/manage.c       | 11 +++++++++--
 kernel/irq/proc.c         | 32 +++++++++++++++++++++-----------
 4 files changed, 33 insertions(+), 15 deletions(-)

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

diff --git a/arch/alpha/kernel/irq.c b/arch/alpha/kernel/irq.c
index d0f1620007f7..703731accda6 100644
--- a/arch/alpha/kernel/irq.c
+++ b/arch/alpha/kernel/irq.c
@@ -50,7 +50,8 @@ int irq_select_affinity(unsigned int irq)
 	if (!irq_desc[irq].chip->set_affinity || irq_user_affinity[irq])
 		return 1;
 
-	while (!cpu_possible(cpu) || !cpu_isset(cpu, irq_default_affinity))
+	while (!cpu_possible(cpu) ||
+	       !cpumask_test_cpu(cpu, irq_default_affinity))
 		cpu = (cpu < (NR_CPUS-1) ? cpu + 1 : 0);
 	last_cpu = cpu;
 
diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h
index dfaee6bd265b..91f1ef8e5810 100644
--- a/include/linux/interrupt.h
+++ b/include/linux/interrupt.h
@@ -109,7 +109,7 @@ extern void enable_irq(unsigned int irq);
 
 #if defined(CONFIG_SMP) && defined(CONFIG_GENERIC_HARDIRQS)
 
-extern cpumask_t irq_default_affinity;
+extern cpumask_var_t irq_default_affinity;
 
 extern int irq_set_affinity(unsigned int irq, const struct cpumask *cpumask);
 extern int irq_can_set_affinity(unsigned int irq);
diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c
index 61c4a9b62165..cd0cd8dcb345 100644
--- a/kernel/irq/manage.c
+++ b/kernel/irq/manage.c
@@ -16,8 +16,15 @@
 #include "internals.h"
 
 #ifdef CONFIG_SMP
+cpumask_var_t irq_default_affinity;
 
-cpumask_t irq_default_affinity = CPU_MASK_ALL;
+static int init_irq_default_affinity(void)
+{
+	alloc_cpumask_var(&irq_default_affinity, GFP_KERNEL);
+	cpumask_setall(irq_default_affinity);
+	return 0;
+}
+core_initcall(init_irq_default_affinity);
 
 /**
  *	synchronize_irq - wait for pending IRQ handlers (on other CPUs)
@@ -127,7 +134,7 @@ int do_irq_select_affinity(unsigned int irq, struct irq_desc *desc)
 			desc->status &= ~IRQ_AFFINITY_SET;
 	}
 
-	cpumask_and(&desc->affinity, cpu_online_mask, &irq_default_affinity);
+	cpumask_and(&desc->affinity, cpu_online_mask, irq_default_affinity);
 set_affinity:
 	desc->chip->set_affinity(irq, &desc->affinity);
 
diff --git a/kernel/irq/proc.c b/kernel/irq/proc.c
index d2c0e5ee53c5..2abd3a7716ed 100644
--- a/kernel/irq/proc.c
+++ b/kernel/irq/proc.c
@@ -20,7 +20,7 @@ static struct proc_dir_entry *root_irq_dir;
 static int irq_affinity_proc_show(struct seq_file *m, void *v)
 {
 	struct irq_desc *desc = irq_to_desc((long)m->private);
-	cpumask_t *mask = &desc->affinity;
+	const struct cpumask *mask = &desc->affinity;
 
 #ifdef CONFIG_GENERIC_PENDING_IRQ
 	if (desc->status & IRQ_MOVE_PENDING)
@@ -93,7 +93,7 @@ static const struct file_operations irq_affinity_proc_fops = {
 
 static int default_affinity_show(struct seq_file *m, void *v)
 {
-	seq_cpumask(m, &irq_default_affinity);
+	seq_cpumask(m, irq_default_affinity);
 	seq_putc(m, '\n');
 	return 0;
 }
@@ -101,27 +101,37 @@ static int default_affinity_show(struct seq_file *m, void *v)
 static ssize_t default_affinity_write(struct file *file,
 		const char __user *buffer, size_t count, loff_t *ppos)
 {
-	cpumask_t new_value;
+	cpumask_var_t new_value;
 	int err;
 
-	err = cpumask_parse_user(buffer, count, &new_value);
+	if (!alloc_cpumask_var(&new_value, GFP_KERNEL))
+		return -ENOMEM;
+
+	err = cpumask_parse_user(buffer, count, new_value);
 	if (err)
-		return err;
+		goto out;
 
-	if (!is_affinity_mask_valid(new_value))
-		return -EINVAL;
+	if (!is_affinity_mask_valid(new_value)) {
+		err = -EINVAL;
+		goto out;
+	}
 
 	/*
 	 * Do not allow disabling IRQs completely - it's a too easy
 	 * way to make the system unusable accidentally :-) At least
 	 * one online CPU still has to be targeted.
 	 */
-	if (!cpus_intersects(new_value, cpu_online_map))
-		return -EINVAL;
+	if (!cpumask_intersects(new_value, cpu_online_mask)) {
+		err = -EINVAL;
+		goto out;
+	}
 
-	irq_default_affinity = new_value;
+	cpumask_copy(irq_default_affinity, new_value);
+	err = count;
 
-	return count;
+out:
+	free_cpumask_var(new_value);
+	return err;
 }
 
 static int default_affinity_open(struct inode *inode, struct file *file)
-- 
cgit v1.3-7-g2ca7


From a6037b61c2f5fc99c57c15b26d7cfa58bbb34008 Mon Sep 17 00:00:00 2001
From: Peter Zijlstra <a.p.zijlstra@chello.nl>
Date: Mon, 5 Jan 2009 11:28:22 +0100
Subject: hrtimer: fix recursion deadlock by re-introducing the softirq

Impact: fix rare runtime deadlock

There are a few sites that do:

  spin_lock_irq(&foo)
  hrtimer_start(&bar)
    __run_hrtimer(&bar)
      func()
        spin_lock(&foo)

which obviously deadlocks. In order to avoid this, never call __run_hrtimer()
from hrtimer_start*() context, but instead defer this to softirq context.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 include/linux/interrupt.h |  3 ++-
 kernel/hrtimer.c          | 60 +++++++++++++++++++++--------------------------
 2 files changed, 29 insertions(+), 34 deletions(-)

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

diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h
index 0702c4d7bdf0..2062833f5f7a 100644
--- a/include/linux/interrupt.h
+++ b/include/linux/interrupt.h
@@ -253,7 +253,8 @@ enum
 	BLOCK_SOFTIRQ,
 	TASKLET_SOFTIRQ,
 	SCHED_SOFTIRQ,
-	RCU_SOFTIRQ, 	/* Preferable RCU should always be the last softirq */
+	HRTIMER_SOFTIRQ,
+	RCU_SOFTIRQ,	/* Preferable RCU should always be the last softirq */
 
 	NR_SOFTIRQS
 };
diff --git a/kernel/hrtimer.c b/kernel/hrtimer.c
index 8010a67cead0..b68e98f4e4c1 100644
--- a/kernel/hrtimer.c
+++ b/kernel/hrtimer.c
@@ -634,7 +634,6 @@ static inline void hrtimer_init_timer_hres(struct hrtimer *timer)
 {
 }
 
-static void __run_hrtimer(struct hrtimer *timer);
 
 /*
  * When High resolution timers are active, try to reprogram. Note, that in case
@@ -646,13 +645,9 @@ static inline int hrtimer_enqueue_reprogram(struct hrtimer *timer,
 					    struct hrtimer_clock_base *base)
 {
 	if (base->cpu_base->hres_active && hrtimer_reprogram(timer, base)) {
-		/*
-		 * XXX: recursion check?
-		 * hrtimer_forward() should round up with timer granularity
-		 * so that we never get into inf recursion here,
-		 * it doesn't do that though
-		 */
-		__run_hrtimer(timer);
+		spin_unlock(&base->cpu_base->lock);
+		raise_softirq_irqoff(HRTIMER_SOFTIRQ);
+		spin_lock(&base->cpu_base->lock);
 		return 1;
 	}
 	return 0;
@@ -705,11 +700,6 @@ static inline int hrtimer_enqueue_reprogram(struct hrtimer *timer,
 }
 static inline void hrtimer_init_hres(struct hrtimer_cpu_base *base) { }
 static inline void hrtimer_init_timer_hres(struct hrtimer *timer) { }
-static inline int hrtimer_reprogram(struct hrtimer *timer,
-				    struct hrtimer_clock_base *base)
-{
-	return 0;
-}
 
 #endif /* CONFIG_HIGH_RES_TIMERS */
 
@@ -780,9 +770,11 @@ EXPORT_SYMBOL_GPL(hrtimer_forward);
  *
  * The timer is inserted in expiry order. Insertion into the
  * red black tree is O(log(n)). Must hold the base lock.
+ *
+ * Returns 1 when the new timer is the leftmost timer in the tree.
  */
-static void enqueue_hrtimer(struct hrtimer *timer,
-			    struct hrtimer_clock_base *base, int reprogram)
+static int enqueue_hrtimer(struct hrtimer *timer,
+			   struct hrtimer_clock_base *base)
 {
 	struct rb_node **link = &base->active.rb_node;
 	struct rb_node *parent = NULL;
@@ -814,20 +806,8 @@ static void enqueue_hrtimer(struct hrtimer *timer,
 	 * Insert the timer to the rbtree and check whether it
 	 * replaces the first pending timer
 	 */
-	if (leftmost) {
-		/*
-		 * Reprogram the clock event device. When the timer is already
-		 * expired hrtimer_enqueue_reprogram has either called the
-		 * callback or added it to the pending list and raised the
-		 * softirq.
-		 *
-		 * This is a NOP for !HIGHRES
-		 */
-		if (reprogram && hrtimer_enqueue_reprogram(timer, base))
-			return;
-
+	if (leftmost)
 		base->first = &timer->node;
-	}
 
 	rb_link_node(&timer->node, parent, link);
 	rb_insert_color(&timer->node, &base->active);
@@ -836,6 +816,8 @@ static void enqueue_hrtimer(struct hrtimer *timer,
 	 * state of a possibly running callback.
 	 */
 	timer->state |= HRTIMER_STATE_ENQUEUED;
+
+	return leftmost;
 }
 
 /*
@@ -912,7 +894,7 @@ hrtimer_start_range_ns(struct hrtimer *timer, ktime_t tim, unsigned long delta_n
 {
 	struct hrtimer_clock_base *base, *new_base;
 	unsigned long flags;
-	int ret;
+	int ret, leftmost;
 
 	base = lock_hrtimer_base(timer, &flags);
 
@@ -940,12 +922,16 @@ hrtimer_start_range_ns(struct hrtimer *timer, ktime_t tim, unsigned long delta_n
 
 	timer_stats_hrtimer_set_start_info(timer);
 
+	leftmost = enqueue_hrtimer(timer, new_base);
+
 	/*
 	 * Only allow reprogramming if the new base is on this CPU.
 	 * (it might still be on another CPU if the timer was pending)
+	 *
+	 * XXX send_remote_softirq() ?
 	 */
-	enqueue_hrtimer(timer, new_base,
-			new_base->cpu_base == &__get_cpu_var(hrtimer_bases));
+	if (leftmost && new_base->cpu_base == &__get_cpu_var(hrtimer_bases))
+		hrtimer_enqueue_reprogram(timer, new_base);
 
 	unlock_hrtimer_base(timer, &flags);
 
@@ -1163,7 +1149,7 @@ static void __run_hrtimer(struct hrtimer *timer)
 	 */
 	if (restart != HRTIMER_NORESTART) {
 		BUG_ON(timer->state != HRTIMER_STATE_CALLBACK);
-		enqueue_hrtimer(timer, base, 0);
+		enqueue_hrtimer(timer, base);
 	}
 	timer->state &= ~HRTIMER_STATE_CALLBACK;
 }
@@ -1277,6 +1263,11 @@ void hrtimer_peek_ahead_timers(void)
 	local_irq_restore(flags);
 }
 
+static void run_hrtimer_softirq(struct softirq_action *h)
+{
+	hrtimer_peek_ahead_timers();
+}
+
 #endif	/* CONFIG_HIGH_RES_TIMERS */
 
 /*
@@ -1532,7 +1523,7 @@ static void migrate_hrtimer_list(struct hrtimer_clock_base *old_base,
 		 * is done, which will run all expired timers and re-programm
 		 * the timer device.
 		 */
-		enqueue_hrtimer(timer, new_base, 0);
+		enqueue_hrtimer(timer, new_base);
 
 		/* Clear the migration state bit */
 		timer->state &= ~HRTIMER_STATE_MIGRATE;
@@ -1610,6 +1601,9 @@ void __init hrtimers_init(void)
 	hrtimer_cpu_notify(&hrtimers_nb, (unsigned long)CPU_UP_PREPARE,
 			  (void *)(long)smp_processor_id());
 	register_cpu_notifier(&hrtimers_nb);
+#ifdef CONFIG_HIGH_RES_TIMERS
+	open_softirq(HRTIMER_SOFTIRQ, run_hrtimer_softirq);
+#endif
 }
 
 /**
-- 
cgit v1.3-7-g2ca7


From 8c3659347efb43857b2c2d7bc63a9c7d68d1a608 Mon Sep 17 00:00:00 2001
From: Jesper Juhl <jj@chaosbits.net>
Date: Tue, 6 Jan 2009 14:41:14 -0800
Subject: include/linux/interrupt.h: do not include linux/irqnr.h twice

Signed-off-by: Jesper Juhl <jj@chaosbits.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
---
 include/linux/interrupt.h | 1 -
 1 file changed, 1 deletion(-)

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

diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h
index 0702c4d7bdf0..af886b26c9d1 100644
--- a/include/linux/interrupt.h
+++ b/include/linux/interrupt.h
@@ -14,7 +14,6 @@
 #include <linux/irqflags.h>
 #include <linux/smp.h>
 #include <linux/percpu.h>
-#include <linux/irqnr.h>
 
 #include <asm/atomic.h>
 #include <asm/ptrace.h>
-- 
cgit v1.3-7-g2ca7