Doug / smarc-fsl-linux-kernel

Download as
Browse Code ยป

Commit f2c66cd8eeddedb440f33bc0f5cec1ed7ae376cb

Authored by KAMEZAWA Hiroyuki
Committed by Linus Torvalds
1 parent 19cd56c48d
Exists in master and in 7 other branches imx_3.0.35_4.1.0, imx_3.10.17_1.0.1_ga, imx_3.10.53_1.1.0_ga, imx_3.14.28_1.0.0_ga, smarc-imx_3.10.53_1.1.0_ga, smarc-imx_3.14.28_1.0.0_ga, smarc-l5.0.0_1.0.0-ga

/proc/stat: scalability of irq num per cpu 

/proc/stat shows the total number of all interrupts to each cpu.  But when
the number of IRQs are very large, it take very long time and 'cat
/proc/stat' takes more than 10 secs.  This is because sum of all irq
events are counted when /proc/stat is read.  This patch adds "sum of all
irq" counter percpu and reduce read costs.

The cost of reading /proc/stat is important because it's used by major
applications as 'top', 'ps', 'w', etc....

A test on a mechin (4096cpu, 256 nodes, 4592 irqs) shows

 %time cat /proc/stat > /dev/null
 Before Patch:  12.627 sec
 After  Patch:  2.459 sec

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Tested-by: Jack Steiner <steiner@sgi.com>
Acked-by: Jack Steiner <steiner@sgi.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Showing 2 changed files with 13 additions and 5 deletions Side-by-side Diff

... ... @@ -52,9 +52,7 @@
52 52 guest = cputime64_add(guest, kstat_cpu(i).cpustat.guest);
53 53 guest_nice = cputime64_add(guest_nice,
54 54 kstat_cpu(i).cpustat.guest_nice);
55   - for_each_irq_nr(j) {
56   - sum += kstat_irqs_cpu(j, i);
57   - }
  55 + sum += kstat_cpu_irqs_sum(i);
58 56 sum += arch_irq_stat_cpu(i);
59 57  
60 58 for (j = 0; j < NR_SOFTIRQS; j++) {
include/linux/kernel_stat.h
Diff comments   View file @ f2c66cd
... ... @@ -33,6 +33,7 @@
33 33 #ifndef CONFIG_GENERIC_HARDIRQS
34 34 unsigned int irqs[NR_IRQS];
35 35 #endif
  36 + unsigned long irqs_sum;
36 37 unsigned int softirqs[NR_SOFTIRQS];
37 38 };
38 39  
... ... @@ -54,6 +55,7 @@
54 55 struct irq_desc *desc)
55 56 {
56 57 kstat_this_cpu.irqs[irq]++;
  58 + kstat_this_cpu.irqs_sum++;
57 59 }
58 60  
59 61 static inline unsigned int kstat_irqs_cpu(unsigned int irq, int cpu)
... ... @@ -65,8 +67,9 @@
65 67 extern unsigned int kstat_irqs_cpu(unsigned int irq, int cpu);
66 68 #define kstat_irqs_this_cpu(DESC) \
67 69 ((DESC)->kstat_irqs[smp_processor_id()])
68   -#define kstat_incr_irqs_this_cpu(irqno, DESC) \
69   - ((DESC)->kstat_irqs[smp_processor_id()]++)
  70 +#define kstat_incr_irqs_this_cpu(irqno, DESC) do {\
  71 + ((DESC)->kstat_irqs[smp_processor_id()]++);\
  72 + kstat_this_cpu.irqs_sum++; } while (0)
70 73  
71 74 #endif
72 75  
... ... @@ -94,6 +97,13 @@
94 97 return sum;
95 98 }
96 99  
  100 +/*
  101 + * Number of interrupts per cpu, since bootup
  102 + */
  103 +static inline unsigned int kstat_cpu_irqs_sum(unsigned int cpu)
  104 +{
  105 + return kstat_cpu(cpu).irqs_sum;
  106 +}
97 107  
98 108 /*
99 109 * Lock/unlock the current runqueue - to extract task statistics: