From 8c007bfdf1bab536f824d91fccc76596c18aba78 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Sun, 31 Jan 2010 20:30:23 +0000 Subject: powerpc: Reduce footprint of irq_stat PowerPC is currently using asm-generic/hardirq.h which statically allocates an NR_CPUS irq_stat array. Switch to an arch specific implementation which uses per cpu data: On a kernel with NR_CPUS=1024, this saves quite a lot of memory: text data bss dec hex filename 8767938 2944132 1636796 13348866 cbb002 vmlinux.baseline 8767779 2944260 1505724 13217763 c9afe3 vmlinux.irq_cpustat A saving of around 128kB. Signed-off-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt --- arch/powerpc/include/asm/hardirq.h | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) (limited to 'arch/powerpc/include/asm/hardirq.h') diff --git a/arch/powerpc/include/asm/hardirq.h b/arch/powerpc/include/asm/hardirq.h index fb3c05a0cbbf..9bf3467581b1 100644 --- a/arch/powerpc/include/asm/hardirq.h +++ b/arch/powerpc/include/asm/hardirq.h @@ -1 +1,22 @@ -#include +#ifndef _ASM_POWERPC_HARDIRQ_H +#define _ASM_POWERPC_HARDIRQ_H + +#include +#include + +typedef struct { + unsigned int __softirq_pending; +} ____cacheline_aligned irq_cpustat_t; + +DECLARE_PER_CPU_SHARED_ALIGNED(irq_cpustat_t, irq_stat); + +#define __ARCH_IRQ_STAT + +#define local_softirq_pending() __get_cpu_var(irq_stat).__softirq_pending + +static inline void ack_bad_irq(unsigned int irq) +{ + printk(KERN_CRIT "unexpected IRQ trap at vector %02x\n", irq); +} + +#endif /* _ASM_POWERPC_HARDIRQ_H */ -- cgit v1.2.3 From 89713ed10815401a1bfe12e3a076b64048381b56 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Sun, 31 Jan 2010 20:34:06 +0000 Subject: powerpc: Add timer, performance monitor and machine check counts to /proc/interrupts With NO_HZ it is useful to know how often the decrementer is going off. The patch below adds an entry for it and also adds it into the /proc/stat summaries. While here, I added performance monitoring and machine check exceptions. I found it useful to keep an eye on the PMU exception rate when using the perf tool. Since it's possible to take a completely handled machine check on a System p box it also sounds like a good idea to keep a machine check summary. The event naming matches x86 to keep gratuitous differences to a minimum. Signed-off-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt --- arch/powerpc/include/asm/hardirq.h | 9 +++++++++ arch/powerpc/kernel/irq.c | 35 +++++++++++++++++++++++++++++++++++ arch/powerpc/kernel/time.c | 2 ++ arch/powerpc/kernel/traps.c | 4 ++++ 4 files changed, 50 insertions(+) (limited to 'arch/powerpc/include/asm/hardirq.h') diff --git a/arch/powerpc/include/asm/hardirq.h b/arch/powerpc/include/asm/hardirq.h index 9bf3467581b1..cd2d4be882aa 100644 --- a/arch/powerpc/include/asm/hardirq.h +++ b/arch/powerpc/include/asm/hardirq.h @@ -6,6 +6,9 @@ typedef struct { unsigned int __softirq_pending; + unsigned int timer_irqs; + unsigned int pmu_irqs; + unsigned int mce_exceptions; } ____cacheline_aligned irq_cpustat_t; DECLARE_PER_CPU_SHARED_ALIGNED(irq_cpustat_t, irq_stat); @@ -19,4 +22,10 @@ static inline void ack_bad_irq(unsigned int irq) printk(KERN_CRIT "unexpected IRQ trap at vector %02x\n", irq); } +extern u64 arch_irq_stat_cpu(unsigned int cpu); +#define arch_irq_stat_cpu arch_irq_stat_cpu + +extern u64 arch_irq_stat(void); +#define arch_irq_stat arch_irq_stat + #endif /* _ASM_POWERPC_HARDIRQ_H */ diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c index b9cbb4570048..710505240f2f 100644 --- a/arch/powerpc/kernel/irq.c +++ b/arch/powerpc/kernel/irq.c @@ -196,6 +196,21 @@ static int show_other_interrupts(struct seq_file *p, int prec) } #endif /* CONFIG_PPC32 && CONFIG_TAU_INT */ + seq_printf(p, "%*s: ", prec, "LOC"); + for_each_online_cpu(j) + seq_printf(p, "%10u ", per_cpu(irq_stat, j).timer_irqs); + seq_printf(p, " Local timer interrupts\n"); + + seq_printf(p, "%*s: ", prec, "CNT"); + for_each_online_cpu(j) + seq_printf(p, "%10u ", per_cpu(irq_stat, j).pmu_irqs); + seq_printf(p, " Performance monitoring interrupts\n"); + + seq_printf(p, "%*s: ", prec, "MCE"); + for_each_online_cpu(j) + seq_printf(p, "%10u ", per_cpu(irq_stat, j).mce_exceptions); + seq_printf(p, " Machine check exceptions\n"); + seq_printf(p, "%*s: %10u\n", prec, "BAD", ppc_spurious_interrupts); return 0; @@ -258,6 +273,26 @@ out: return 0; } +/* + * /proc/stat helpers + */ +u64 arch_irq_stat_cpu(unsigned int cpu) +{ + u64 sum = per_cpu(irq_stat, cpu).timer_irqs; + + sum += per_cpu(irq_stat, cpu).pmu_irqs; + sum += per_cpu(irq_stat, cpu).mce_exceptions; + + return sum; +} + +u64 arch_irq_stat(void) +{ + u64 sum = ppc_spurious_interrupts; + + return sum; +} + #ifdef CONFIG_HOTPLUG_CPU void fixup_irqs(cpumask_t map) { diff --git a/arch/powerpc/kernel/time.c b/arch/powerpc/kernel/time.c index ed1c0f58344a..1b16b9a3e49a 100644 --- a/arch/powerpc/kernel/time.c +++ b/arch/powerpc/kernel/time.c @@ -575,6 +575,8 @@ void timer_interrupt(struct pt_regs * regs) trace_timer_interrupt_entry(regs); + __get_cpu_var(irq_stat).timer_irqs++; + /* Ensure a positive value is written to the decrementer, or else * some CPUs will continuue to take decrementer exceptions */ set_dec(DECREMENTER_MAX); diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c index 0a320dbd950a..895da29e7db8 100644 --- a/arch/powerpc/kernel/traps.c +++ b/arch/powerpc/kernel/traps.c @@ -483,6 +483,8 @@ void machine_check_exception(struct pt_regs *regs) { int recover = 0; + __get_cpu_var(irq_stat).mce_exceptions++; + /* See if any machine dependent calls. In theory, we would want * to call the CPU first, and call the ppc_md. one if the CPU * one returns a positive number. However there is existing code @@ -965,6 +967,8 @@ void vsx_unavailable_exception(struct pt_regs *regs) void performance_monitor_exception(struct pt_regs *regs) { + __get_cpu_var(irq_stat).pmu_irqs++; + perf_irq(regs); } -- cgit v1.2.3 From 17081102a6e0fe32cf47cdbdf8f2e9ab55273b08 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Sun, 31 Jan 2010 20:34:36 +0000 Subject: powerpc: Convert global "BAD" interrupt to per cpu spurious I often get asked if BAD interrupts are really bad. On some boxes (eg IBM machines running a hypervisor) there are valid cases where are presented with an interrupt that is not for us. These cases are common enough to show up as thousands of BAD interrupts a day. Tone them down by calling them spurious. Since they can be a significant cause of OS jitter, we may as well log them per cpu so we know where they are occurring. Signed-off-by: Anton Blanchard Signed-off-by: Benjamin Herrenschmidt --- arch/powerpc/include/asm/hardirq.h | 4 +--- arch/powerpc/kernel/irq.c | 19 +++++++------------ 2 files changed, 8 insertions(+), 15 deletions(-) (limited to 'arch/powerpc/include/asm/hardirq.h') diff --git a/arch/powerpc/include/asm/hardirq.h b/arch/powerpc/include/asm/hardirq.h index cd2d4be882aa..3147a2970125 100644 --- a/arch/powerpc/include/asm/hardirq.h +++ b/arch/powerpc/include/asm/hardirq.h @@ -9,6 +9,7 @@ typedef struct { unsigned int timer_irqs; unsigned int pmu_irqs; unsigned int mce_exceptions; + unsigned int spurious_irqs; } ____cacheline_aligned irq_cpustat_t; DECLARE_PER_CPU_SHARED_ALIGNED(irq_cpustat_t, irq_stat); @@ -25,7 +26,4 @@ static inline void ack_bad_irq(unsigned int irq) extern u64 arch_irq_stat_cpu(unsigned int cpu); #define arch_irq_stat_cpu arch_irq_stat_cpu -extern u64 arch_irq_stat(void); -#define arch_irq_stat arch_irq_stat - #endif /* _ASM_POWERPC_HARDIRQ_H */ diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c index 710505240f2f..9ae77e52f9d3 100644 --- a/arch/powerpc/kernel/irq.c +++ b/arch/powerpc/kernel/irq.c @@ -77,7 +77,6 @@ DEFINE_PER_CPU_SHARED_ALIGNED(irq_cpustat_t, irq_stat); EXPORT_PER_CPU_SYMBOL(irq_stat); int __irq_offset_value; -static int ppc_spurious_interrupts; #ifdef CONFIG_PPC32 EXPORT_SYMBOL(__irq_offset_value); @@ -201,6 +200,11 @@ static int show_other_interrupts(struct seq_file *p, int prec) seq_printf(p, "%10u ", per_cpu(irq_stat, j).timer_irqs); seq_printf(p, " Local timer interrupts\n"); + seq_printf(p, "%*s: ", prec, "SPU"); + for_each_online_cpu(j) + seq_printf(p, "%10u ", per_cpu(irq_stat, j).spurious_irqs); + seq_printf(p, " Spurious interrupts\n"); + seq_printf(p, "%*s: ", prec, "CNT"); for_each_online_cpu(j) seq_printf(p, "%10u ", per_cpu(irq_stat, j).pmu_irqs); @@ -211,8 +215,6 @@ static int show_other_interrupts(struct seq_file *p, int prec) seq_printf(p, "%10u ", per_cpu(irq_stat, j).mce_exceptions); seq_printf(p, " Machine check exceptions\n"); - seq_printf(p, "%*s: %10u\n", prec, "BAD", ppc_spurious_interrupts); - return 0; } @@ -282,13 +284,7 @@ u64 arch_irq_stat_cpu(unsigned int cpu) sum += per_cpu(irq_stat, cpu).pmu_irqs; sum += per_cpu(irq_stat, cpu).mce_exceptions; - - return sum; -} - -u64 arch_irq_stat(void) -{ - u64 sum = ppc_spurious_interrupts; + sum += per_cpu(irq_stat, cpu).spurious_irqs; return sum; } @@ -404,8 +400,7 @@ void do_IRQ(struct pt_regs *regs) if (irq != NO_IRQ && irq != NO_IRQ_IGNORE) handle_one_irq(irq); else if (irq != NO_IRQ_IGNORE) - /* That's not SMP safe ... but who cares ? */ - ppc_spurious_interrupts++; + __get_cpu_var(irq_stat).spurious_irqs++; irq_exit(); set_irq_regs(old_regs); -- cgit v1.2.3