Linux CPU之mpstat

大家好，欢迎来到IT知识分享网。

前言

NAME mpstat - Report processors related statistics. 

vmstat用来观测系统整体的性能情况，pidstat来观测单个进程的性能情况，那么mpstat是用来观测单个CPU的性能情况。

一、mpstat 简介

mpstat命令输出每个可用处理器的标准输出活动，输出显示中cpu0是第一个处理器。还报告了所有处理器的全局的平均活动。mpstat命令可以在SMP和UP机器上使用，但在后者（UP机器上）中，只打印全局平均活动。如果未指定参数，则默认报告为CPU整体利用率报告。

[root@localhost ~]# mpstat Linux 3.10.0-957.el7.x86_64 (localhost.localdomain) 11/28/2022 _x86_64_ (4 CPU) 02:48:09 PM CPU %usr %nice %sys %iowait %irq %soft %steal %guest %gnice %idle 02:48:09 PM all 0.31 0.00 0.28 0.01 0.00 0.00 0.00 0.00 0.00 99.40 

 mpstat ...... [ interval [ count ] ] 

interval参数指定每个报告之间的时间量（以秒为单位）。值为0（或根本没有参数）表示自系统启动（引导）以来将报告处理器统计信息。如果没有将interval参数设置为零，则可以将count参数与interval参数一起指定。count的值决定了间隔几秒生成的报告的数量。如果指定interval参数而不指定count参数，则mpstat命令将连续生成报表。

[root@localhost ~]# mpstat 2 5 Linux 3.10.0-957.el7.x86_64 (localhost.localdomain) 11/28/2022 _x86_64_ (4 CPU) 02:55:10 PM CPU %usr %nice %sys %iowait %irq %soft %steal %guest %gnice %idle 02:55:12 PM all 0.50 0.00 0.50 0.00 0.00 0.00 0.00 0.00 0.00 99.00 02:55:14 PM all 0.25 0.00 0.63 0.00 0.00 0.00 0.00 0.00 0.00 99.12 02:55:16 PM all 0.38 0.00 0.50 0.00 0.00 0.00 0.00 0.00 0.00 99.12 02:55:18 PM all 0.50 0.00 0.62 0.00 0.00 0.00 0.00 0.00 0.00 98.88 02:55:20 PM all 0.38 0.00 0.50 0.00 0.00 0.00 0.00 0.00 0.00 99.12 Average: all 0.40 0.00 0.55 0.00 0.00 0.00 0.00 0.00 0.00 99.05 

以两秒间隔显示所有处理器之间的五个全局统计数据报告。

就是读取 /proc/stat 文件中的数据：

open("/proc/stat", O_RDONLY) = 3 

二、mpstat -P

 -P { 
    cpu [,...] | ON | ALL } 

指示要报告统计信息的处理器编号，cpu是处理器编号，处理器0是第一个处理器。ON关键字表示要为每个在线处理器报告统计信息，而ALL关键字表示要报告所有处理器的统计信息。

输出处理器1（第二个处理器）的报告统计信息：

[root@localhost ~]# mpstat -P 1 Linux 3.10.0-957.el7.x86_64 (localhost.localdomain) 11/28/2022 _x86_64_ (4 CPU) 03:08:34 PM CPU %usr %nice %sys %iowait %irq %soft %steal %guest %gnice %idle 03:08:34 PM 1 0.28 0.00 0.29 0.00 0.00 0.00 0.00 0.00 0.00 99.43 

输出每个在线处理器报告统计信息：

[root@localhost ~]# mpstat -P ON Linux 3.10.0-957.el7.x86_64 (localhost.localdomain) 11/28/2022 _x86_64_ (4 CPU) 03:09:59 PM CPU %usr %nice %sys %iowait %irq %soft %steal %guest %gnice %idle 03:09:59 PM all 0.31 0.00 0.28 0.01 0.00 0.00 0.00 0.00 0.00 99.40 03:09:59 PM 0 0.31 0.00 0.27 0.00 0.00 0.00 0.00 0.00 0.00 99.41 03:09:59 PM 1 0.28 0.00 0.29 0.00 0.00 0.00 0.00 0.00 0.00 99.43 03:09:59 PM 2 0.33 0.00 0.29 0.00 0.00 0.00 0.00 0.00 0.00 99.38 03:09:59 PM 3 0.34 0.00 0.28 0.01 0.00 0.00 0.00 0.00 0.00 99.37 

输出所有处理器的统计信息：

[root@localhost ~]# mpstat -P ALL Linux 3.10.0-957.el7.x86_64 (localhost.localdomain) 11/28/2022 _x86_64_ (4 CPU) 03:11:24 PM CPU %usr %nice %sys %iowait %irq %soft %steal %guest %gnice %idle 03:11:24 PM all 0.31 0.00 0.28 0.01 0.00 0.00 0.00 0.00 0.00 99.40 03:11:24 PM 0 0.31 0.00 0.27 0.00 0.00 0.00 0.00 0.00 0.00 99.41 03:11:24 PM 1 0.28 0.00 0.29 0.00 0.00 0.00 0.00 0.00 0.00 99.43 03:11:24 PM 2 0.33 0.00 0.29 0.00 0.00 0.00 0.00 0.00 0.00 99.38 03:11:24 PM 3 0.34 0.00 0.28 0.01 0.00 0.00 0.00 0.00 0.00 99.37 

每个字段的含义：

CPU Processor number. The keyword all indicates that statistics are calculated as averages among all processors. %usr Show the percentage of CPU utilization that occurred while executing at the user level (application). %nice Show the percentage of CPU utilization that occurred while executing at the user level with nice priority. %sys Show the percentage of CPU utilization that occurred while executing at the system level (kernel). Note that this does not include time spent servicing hardware and soft‐ ware interrupts. %iowait Show the percentage of time that the CPU or CPUs were idle during which the system had an outstanding disk I/O request. %irq Show the percentage of time spent by the CPU or CPUs to service hardware interrupts. %soft Show the percentage of time spent by the CPU or CPUs to service software interrupts. %steal Show the percentage of time spent in involuntary wait by the virtual CPU or CPUs while the hypervisor was servicing another virtual processor. %guest Show the percentage of time spent by the CPU or CPUs to run a virtual processor. %gnice Show the percentage of time spent by the CPU or CPUs to run a niced guest. %idle Show the percentage of time that the CPU or CPUs were idle and the system did not have an outstanding disk I/O request. 

具体请参考：Linux top命令的cpu使用率和内存使用率 这篇文章中关于各个字段的释义。

三、mpstat -I

-I { 
    SUM | CPU | SCPU | ALL } 

报告中断统计信息

3.1 mpstat -I SUM

[root@localhost ~]# mpstat -I SUM Linux 3.10.0-957.el7.x86_64 (localhost.localdomain) 11/28/2022 _x86_64_ (4 CPU) 03:21:23 PM CPU intr/s 03:21:23 PM all 93.37 

3.2 mpstat -I CPU

3.2.1 数据来源

[root@localhost ~]# mpstat -I CPU Linux 3.10.0-957.el7.x86_64 (localhost.localdomain) 11/28/2022 _x86_64_ (4 CPU) 03:23:41 PM CPU 0/s 1/s 8/s 9/s 12/s 16/s 20/s 120/s 121/s 122/s 123/s 124/s 125/s 126/s 127/s NMI/s LOC/s SPU/s PMI/s IWI/s RTR/s RES/s CAL/s TLB/s TRM/s THR/s DFR/s MCE/s MCP/s ERR/s MIS/s PIN/s NPI/s PIW/s 03:23:41 PM 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.02 4.03 0.00 0.00 0.00 20.01 0.00 0.00 0.19 0.00 0.21 0.00 0.11 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 03:23:41 PM 1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 21.62 0.00 0.00 0.21 0.00 0.15 0.00 0.14 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 03:23:41 PM 2 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 22.75 0.00 0.00 0.17 0.00 0.15 0.00 0.07 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 03:23:41 PM 3 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.20 0.00 0.00 0.00 0.00 22.86 0.00 0.00 0.25 0.00 0.15 0.00 0.08 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 

open("/proc/interrupts", O_RDONLY) = 3 

[root@localhost ~]# cat /proc/interrupts CPU0 CPU1 CPU2 CPU3 0: 55 0 0 0 IR-IO-APIC-edge timer 1: 4 0 0 0 IR-IO-APIC-edge i8042 8: 1 0 0 0 IR-IO-APIC-edge rtc0 9: 4 0 0 0 IR-IO-APIC-fasteoi acpi 12: 3 3 0 0 IR-IO-APIC-edge i8042 16: 0 0 0 0 IR-IO-APIC-fasteoi i801_smbus 20: 0 0 0 0 IR-IO-APIC-fasteoi idma64.0 120: 0 0 0 0 DMAR_MSI-edge dmar0 121: 0 0 0 0 DMAR_MSI-edge dmar1 122: 0 0 0 0 IR-PCI-MSI-edge aerdrv, PCIe PME 123: 148 16 10 2 IR-PCI-MSI-edge xhci_hcd 124: 4738 519 421 54943 IR-PCI-MSI-edge 0000:00:17.0 125:  0 0 0 IR-PCI-MSI-edge enp1s0 126: 38 1 109 9 IR-PCI-MSI-edge i915 127: 541 136 191 85 IR-PCI-MSI-edge snd_hda_intel:card0 NMI: 56 52 55 56 Non-maskable interrupts LOC:     Local timer interrupts SPU: 0 0 0 0 Spurious interrupts PMI: 56 52 55 56 Performance monitoring interrupts IWI: 52427 58892 47990 68017 IRQ work interrupts RTR: 0 0 0 0 APIC ICR read retries RES: 56937 40801 42634 41527 Rescheduling interrupts CAL: 1150 1147 1194 1149 Function call interrupts TLB: 28982 39043 19609 20986 TLB shootdowns TRM: 0 0 0 0 Thermal event interrupts THR: 0 0 0 0 Threshold APIC interrupts DFR: 0 0 0 0 Deferred Error APIC interrupts MCE: 0 0 0 0 Machine check exceptions MCP: 918 918 918 918 Machine check polls ERR: 0 MIS: 0 PIN: 0 0 0 0 Posted-interrupt notification event NPI: 0 0 0 0 Nested posted-interrupt event PIW: 0 0 0 0 Posted-interrupt wakeup event 

3.2.2 内核源码解析

// linux-3.10/fs/proc/interrupts.c /* * /proc/interrupts */ static void *int_seq_start(struct seq_file *f, loff_t *pos) { 
    return (*pos <= nr_irqs) ? pos : NULL; } static void *int_seq_next(struct seq_file *f, void *v, loff_t *pos) { 
    (*pos)++; if (*pos > nr_irqs) return NULL; return pos; } static void int_seq_stop(struct seq_file *f, void *v) { 
    /* Nothing to do */ } static const struct seq_operations int_seq_ops = { 
    .start = int_seq_start, .next = int_seq_next, .stop = int_seq_stop, .show = show_interrupts }; static int interrupts_open(struct inode *inode, struct file *filp) { 
    return seq_open(filp, &int_seq_ops); } static const struct file_operations proc_interrupts_operations = { 
    .open = interrupts_open, .read = seq_read, .llseek = seq_lseek, .release = seq_release, }; static int __init proc_interrupts_init(void) { 
    proc_create("interrupts", 0, NULL, &proc_interrupts_operations); return 0; } module_init(proc_interrupts_init); 

其中show_interrupts函数：

// linux-3.10/kernel/irq/proc.c int show_interrupts(struct seq_file *p, void *v) { 
    ...... arch_show_interrupts(p, prec); ...... } 

arch_show_interrupts是一个与架构有关的函数，对于x86架构：

// linux-3.10/arch/x86/kernel/irq.c #define irq_stats(x) (&per_cpu(irq_stat, x)) /* * /proc/interrupts printing for arch specific interrupts */ int arch_show_interrupts(struct seq_file *p, int prec) { 
    int j; seq_printf(p, "%*s: ", prec, "NMI"); for_each_online_cpu(j) seq_printf(p, "%10u ", irq_stats(j)->__nmi_count); seq_printf(p, " Non-maskable interrupts\n"); #ifdef CONFIG_X86_LOCAL_APIC seq_printf(p, "%*s: ", prec, "LOC"); for_each_online_cpu(j) seq_printf(p, "%10u ", irq_stats(j)->apic_timer_irqs); seq_printf(p, " Local timer interrupts\n"); seq_printf(p, "%*s: ", prec, "SPU"); for_each_online_cpu(j) seq_printf(p, "%10u ", irq_stats(j)->irq_spurious_count); seq_printf(p, " Spurious interrupts\n"); seq_printf(p, "%*s: ", prec, "PMI"); for_each_online_cpu(j) seq_printf(p, "%10u ", irq_stats(j)->apic_perf_irqs); seq_printf(p, " Performance monitoring interrupts\n"); seq_printf(p, "%*s: ", prec, "IWI"); for_each_online_cpu(j) seq_printf(p, "%10u ", irq_stats(j)->apic_irq_work_irqs); seq_printf(p, " IRQ work interrupts\n"); seq_printf(p, "%*s: ", prec, "RTR"); for_each_online_cpu(j) seq_printf(p, "%10u ", irq_stats(j)->icr_read_retry_count); seq_printf(p, " APIC ICR read retries\n"); #endif if (x86_platform_ipi_callback) { 
    seq_printf(p, "%*s: ", prec, "PLT"); for_each_online_cpu(j) seq_printf(p, "%10u ", irq_stats(j)->x86_platform_ipis); seq_printf(p, " Platform interrupts\n"); } #ifdef CONFIG_SMP seq_printf(p, "%*s: ", prec, "RES"); for_each_online_cpu(j) seq_printf(p, "%10u ", irq_stats(j)->irq_resched_count); seq_printf(p, " Rescheduling interrupts\n"); seq_printf(p, "%*s: ", prec, "CAL"); for_each_online_cpu(j) seq_printf(p, "%10u ", irq_stats(j)->irq_call_count - irq_stats(j)->irq_tlb_count); seq_printf(p, " Function call interrupts\n"); seq_printf(p, "%*s: ", prec, "TLB"); for_each_online_cpu(j) seq_printf(p, "%10u ", irq_stats(j)->irq_tlb_count); seq_printf(p, " TLB shootdowns\n"); #endif #ifdef CONFIG_X86_THERMAL_VECTOR seq_printf(p, "%*s: ", prec, "TRM"); for_each_online_cpu(j) seq_printf(p, "%10u ", irq_stats(j)->irq_thermal_count); seq_printf(p, " Thermal event interrupts\n"); #endif #ifdef CONFIG_X86_MCE_THRESHOLD seq_printf(p, "%*s: ", prec, "THR"); for_each_online_cpu(j) seq_printf(p, "%10u ", irq_stats(j)->irq_threshold_count); seq_printf(p, " Threshold APIC interrupts\n"); #endif #ifdef CONFIG_X86_MCE seq_printf(p, "%*s: ", prec, "MCE"); for_each_online_cpu(j) seq_printf(p, "%10u ", per_cpu(mce_exception_count, j)); seq_printf(p, " Machine check exceptions\n"); seq_printf(p, "%*s: ", prec, "MCP"); for_each_online_cpu(j) seq_printf(p, "%10u ", per_cpu(mce_poll_count, j)); seq_printf(p, " Machine check polls\n"); #endif seq_printf(p, "%*s: %10u\n", prec, "ERR", atomic_read(&irq_err_count)); #if defined(CONFIG_X86_IO_APIC) seq_printf(p, "%*s: %10u\n", prec, "MIS", atomic_read(&irq_mis_count)); #endif return 0; } 

可以看到主要是从 per-cpu内存区读取相应的数据，关于x86_64 per-cpu相关知识请参考：Linux per-cpu

// linux-3.10/arch/x86/include/asm/hardirq.h typedef struct { 
    unsigned int __softirq_pending; unsigned int __nmi_count; /* arch dependent */ #ifdef CONFIG_X86_LOCAL_APIC unsigned int apic_timer_irqs; /* arch dependent */ unsigned int irq_spurious_count; unsigned int icr_read_retry_count; #endif #ifdef CONFIG_HAVE_KVM unsigned int kvm_posted_intr_ipis; #endif unsigned int x86_platform_ipis; /* arch dependent */ unsigned int apic_perf_irqs; unsigned int apic_irq_work_irqs; #ifdef CONFIG_SMP unsigned int irq_resched_count; unsigned int irq_call_count; /* * irq_tlb_count is double-counted in irq_call_count, so it must be * subtracted from irq_call_count when displaying irq_call_count */ unsigned int irq_tlb_count; #endif #ifdef CONFIG_X86_THERMAL_VECTOR unsigned int irq_thermal_count; #endif #ifdef CONFIG_X86_MCE_THRESHOLD unsigned int irq_threshold_count; #endif } ____cacheline_aligned irq_cpustat_t; DECLARE_PER_CPU_SHARED_ALIGNED(irq_cpustat_t, irq_stat); 

// linux-3.10/include/linux/irq_cpustat.h /* * Simple wrappers reducing source bloat. Define all irq_stat fields * here, even ones that are arch dependent. That way we get common * definitions instead of differing sets for each arch. */ #ifndef __ARCH_IRQ_STAT extern irq_cpustat_t irq_stat[]; /* defined in asm/hardirq.h */ #define __IRQ_STAT(cpu, member) (irq_stat[cpu].member) #endif 

3.3 mpstat -I SCPU

3.3.1 数据来源

[root@localhost ~]# mpstat -I SCPU Linux 3.10.0-957.el7.x86_64 (localhost.localdomain) 11/28/2022 _x86_64_ (4 CPU) 04:48:54 PM CPU HI/s TIMER/s NET_TX/s NET_RX/s BLOCK/s BLOCK_IOPOLL/s TASKLET/s SCHED/s HRTIMER/s RCU/s 04:48:54 PM 0 0.00 10.62 0.17 4.26 0.02 0.00 0.02 6.63 0.00 3.92 04:48:54 PM 1 0.00 12.78 0.00 0.03 0.00 0.00 0.00 7.19 0.00 4.89 04:48:54 PM 2 0.00 12.41 0.00 0.03 0.00 0.00 0.00 7.28 0.00 4.48 04:48:54 PM 3 0.00 12.97 0.00 0.02 0.19 0.00 0.00 7.13 0.00 4.90 

数据来源读取 /proc/softirqs 文件，/proc/softirqs 提供了软中断的运行情况：

open("/proc/softirqs", O_RDONLY) = 3 

[root@localhost ~]# cat /proc/softirqs CPU0 CPU1 CPU2 CPU3 HI: 29 12 81 4 TIMER:     NET_TX: 46707 2 2 1 NET_RX:  8070 7563 6755 BLOCK: 5432 776 578 53783 BLOCK_IOPOLL: 0 0 0 0 TASKLET: 5769 0 0 0 SCHED:     HRTIMER: 0 0 0 0 RCU:     

软中断包括了 10 个类别，分别对应不同的工作类型。比如 NET_RX 表示网络接收中断，而 NET_TX 表示网络发送中断。

参数详解：

优先级高的软中断（比如：0）在优先级低的软中断（比如：9）优先执行。

当内核中出现大量的软中断时（当软中断比较多时，普通进程优先级低于软中断，那么普通进程无法获得足够多的处理器时间），软中断会以内核线程的方式运行的，每个 CPU 都对应一个软中断内核线程，这个软中断内核线程就叫做 ksoftirqd/CPU 编号。

[root@localhost ~]# top -n 1 | grep ksoftirqd 3 root 20 0 0 0 0 S 0.0 0.0 0:00.89 ksoftirqd/0 14 root 20 0 0 0 0 S 0.0 0.0 0:00.10 ksoftirqd/1 19 root 20 0 0 0 0 S 0.0 0.0 0:00.13 ksoftirqd/2 24 root 20 0 0 0 0 S 0.0 0.0 0:00.18 ksoftirqd/3 

或者：

[root@localhost ~]# ps aux | grep ksoftirq root 3 0.0 0.0 0 0 ? S Nov25 0:00 [ksoftirqd/0] root 14 0.0 0.0 0 0 ? S Nov25 0:00 [ksoftirqd/1] root 19 0.0 0.0 0 0 ? S Nov25 0:00 [ksoftirqd/2] root 24 0.0 0.0 0 0 ? S Nov25 0:00 [ksoftirqd/3] 

这些线程的名字外面都有中括号，这说明 ps 无法获取它们的命令行参数（cmline）。一般来说，ps 的输出中，名字括在中括号里的，一般都是内核线程。

3.3.2 内核源码解析

（1）softirqs

// linux-3.10/fs/proc/softirqs.c /* * /proc/softirqs ... display the number of softirqs */ static int show_softirqs(struct seq_file *p, void *v) { 
    int i, j; seq_puts(p, " "); for_each_possible_cpu(i) seq_printf(p, "CPU%-8d", i); seq_putc(p, '\n'); for (i = 0; i < NR_SOFTIRQS; i++) { 
    seq_printf(p, "%12s:", softirq_to_name[i]); for_each_possible_cpu(j) seq_printf(p, " %10u", kstat_softirqs_cpu(i, j)); seq_putc(p, '\n'); } return 0; } static int softirqs_open(struct inode *inode, struct file *file) { 
    return single_open(file, show_softirqs, NULL); } static const struct file_operations proc_softirqs_operations = { 
    .open = softirqs_open, .read = seq_read, .llseek = seq_lseek, .release = single_release, }; static int __init proc_softirqs_init(void) { 
    proc_create("softirqs", 0, NULL, &proc_softirqs_operations); return 0; } module_init(proc_softirqs_init); 

// linux-3.10/include/linux/interrupt.h /* PLEASE, avoid to allocate new softirqs, if you need not _really_ high frequency threaded job scheduling. For almost all the purposes tasklets are more than enough. F.e. all serial device BHs et al. should be converted to tasklets, not to softirqs. */ enum { 
    HI_SOFTIRQ=0, TIMER_SOFTIRQ, NET_TX_SOFTIRQ, NET_RX_SOFTIRQ, BLOCK_SOFTIRQ, BLOCK_IOPOLL_SOFTIRQ, TASKLET_SOFTIRQ, SCHED_SOFTIRQ, HRTIMER_SOFTIRQ, RCU_SOFTIRQ, /* Preferable RCU should always be the last softirq */ NR_SOFTIRQS }; /* map softirq index to softirq name. update 'softirq_to_name' in * kernel/softirq.c when adding a new softirq. */ extern char *softirq_to_name[NR_SOFTIRQS]; 

// linux-3.10/kernel/softirq.c static struct softirq_action softirq_vec[NR_SOFTIRQS] __cacheline_aligned_in_smp; char *softirq_to_name[NR_SOFTIRQS] = { 
    "HI", "TIMER", "NET_TX", "NET_RX", "BLOCK", "BLOCK_IOPOLL", "TASKLET", "SCHED", "HRTIMER", "RCU" }; 

定义per-cpu变量 ：struct kernel_stat kstat，并且将 kstat 符号导出

// linux-3.10/kernel/sched/core.c DEFINE_PER_CPU(struct kernel_stat, kstat); EXPORT_PER_CPU_SYMBOL(kstat); 

[root@localhost ~]# cat /proc/kallsyms | grep '\<kstat\>' 0000000000015b60 A kstat 

[root@localhost ~]# cat /proc/kallsyms | grep '\<__per_cpu_start\>' 0000000000000000 A __per_cpu_start [root@localhost ~]# cat /proc/kallsyms | grep '\<__per_cpu_end\>' 000000000001d000 A __per_cpu_end 

kstat 在 _per_cpu_start 和 __per_cpu_end 范围内，是内核中的per-cpu变量。

读取 softirqs 数据：

// linux-3.10/include/linux/kernel_stat.h struct kernel_stat { 
    #ifndef CONFIG_GENERIC_HARDIRQS unsigned int irqs[NR_IRQS]; #endif unsigned long irqs_sum; unsigned int softirqs[NR_SOFTIRQS]; }; DECLARE_PER_CPU(struct kernel_stat, kstat); #define kstat_cpu(cpu) per_cpu(kstat, cpu) static inline unsigned int kstat_softirqs_cpu(unsigned int irq, int cpu) { 
    return kstat_cpu(cpu).softirqs[irq]; } 

（2）ksoftirqd

定义ksoftirqd，用一个struct task_struct表示，存放在per-cpu内存中：

// linux-3.10/kernel/softirq.c DEFINE_PER_CPU(struct task_struct *, ksoftirqd); /* * we cannot loop indefinitely here to avoid userspace starvation, * but we also don't want to introduce a worst case 1/HZ latency * to the pending events, so lets the scheduler to balance * the softirq load for us. */ /* 不能在这里无限循环以避免用户空间不足，但我们也不想为 the pending events 引入最坏的1/HZ延迟 因此让调度器为我们平衡 the softirq 负载。 */ static void wakeup_softirqd(void) { 
    /* Interrupts are disabled: no need to stop preemption */ struct task_struct *tsk = __this_cpu_read(ksoftirqd); if (tsk && tsk->state != TASK_RUNNING) wake_up_process(tsk); } 

声明 ksoftirqd：

// linux-3.10/include/linux/interrupt.h DECLARE_PER_CPU(struct task_struct *, ksoftirqd); static inline struct task_struct *this_cpu_ksoftirqd(void) { 
    return this_cpu_read(ksoftirqd); } 

参考资料

Linux内核 3.10.0