跟踪内核函数

• CONFIG_FTRACE
• CONFIG_HAVE_FUNCTION_TRACER
• CONFIG_HAVE_FUNCTION_GRAPH_TRACER
• CONFIG_HAVE_DYNAMIC_FTRACE
• CONFIG_HAVE_FTRACE_MCOUNT_RECORD

​

将 ftrace 信息转储到 kmsg 缓冲区

• 要增大 ftrace 环形缓冲区的大小，请运行以下命令：
  echo 200 > /sys/kernel/debug/tracing/buffer_size_kb
  

​

启用工作队列跟踪

• 要启用工作队列跟踪，请运行以下命令：
  mount -t debugfs none /sys/kernel/debug
  
  echo 1 > /sys/kernel/debug/tracing/events/workqueue/enable
  
   echo workqueue:workqueue_queue_work > /sys/kernel/debug/tracing/set_event
  
   cat /sys/kernel/debug/tracing/trace_pipe
  
  cat /sys/kernel/debug/tracing/per_cpu/cpu1/trace
  
  示例输出：
   # tracer: nop
  
   #
  
   # entries-in-buffer/entries-written: 8682/8682 #P:1
  
   #
  
   # \_-----=\> irqs-off
  
   # / \_----=\> need-resched
  
   # | / \_---=\> hardirq/softirq
  
   # || / \_--=\> preempt-depth
  
   # ||| / delay
  
   # TASK-PID CPU# |||| TIMESTAMP FUNCTION
  
   # | | | |||| | |
  
   \<...\>-4783 \[001\] d.s4 7524.354249: workqueue\_queue\_work: work struct=f2d91ee4 function=free\_css\_set\_work workqueue=f6427d80 req\_cpu=1 cpu=1
  
   \<idle\>-0 \[001\] d.h4 7524.424196: workqueue\_queue\_work: work struct=c10c32a8 function=def\_work\_fn workqueue=f55c7880 req\_cpu=1 cpu=4
  
   e.process.gapps-4758 \[001\] dNs4 7524.454227: workqueue\_queue\_work: work struct=c4727fa4 function=free\_css\_set\_work workqueue=f6427d80 req\_cpu=1 cpu=1
  
   Binder\_D-1693 \[001\] d.s3 7524.504198: workqueue\_queue\_work: work struct=cec6275c function=do\_dbs\_timer workqueue=f5424680 req\_cpu=1 cpu=1
  
   \<...\>-4832 \[001\] d.h3 7524.574194: workqueue\_queue\_work: work struct=c10c32a8 function=def\_work\_fn workqueue=f55c7880 req\_cpu=1 cpu=4
  

​

启用中断请求（IRQ）跟踪

• 要启用中断跟踪，请运行以下命令：
  mount -t debugfs none /sys/kernel/debug
  
  echo 1 > /sys/kernel/debug/tracing/events/irq/irq_handler_entry/enable
  
  cat /sys/kernel/debug/tracing/trace
  
  cat /sys/kernel/debug/tracing/trace_pipe
  
  示例输出：
   TASK-PID CPU# TIMESTAMP FUNCTION
  
   | | | | |
  
   adbd-302 \[000\] 295.075568: irq\_handler\_entry: irq=132 name=msm\_otg
  
   adbd-302 \[000\] 295.075599: irq\_handler\_entry: irq=132 name=msm\_hsusb
  
   adbd-302 \[000\] 295.075782: irq\_handler\_entry: irq=132 name=msm\_otg
  
   adbd-302 \[000\] 295.075782: irq\_handler\_entry: irq=132 name=msm\_hsusb
  
   \<idle\>-0 \[000\] 295.076270: irq\_handler\_entry: irq=132 name=msm\_otg
  
   \<idle\>-0 \[000\] 295.076270: irq\_handler\_entry: irq=132 name=msm\_hsusb
  
   \<idle\>-0 \[000\] 295.077155: irq\_handler\_entry: irq=18 name=gp\_timer
  
   \<idle\>-0 \[000\] 295.087166: irq\_handler\_entry: irq=18 name=gp\_timer
  
   \<idle\>-0 \[000\] 295.097146: irq\_handler\_entry: irq=18 name=gp\_timer
  

​

启用调度器跟踪

• 要跟踪任务之间的上下文切换，请运行以下命令以启用 sched_switch 跟踪器：
  mount -t debugfs none /sys/kernel/debug
  
  echo 1 > /sys/kernel/debug/tracing/events/sched/sched_switch/enable
  
  cat /sys/kernel/debug/tracing/trace
  
  cat /sys/kernel/debug/tracing/trace_pipe
  
  示例输出：
   TASK-PID CPU# TIMESTAMP FUNCTION
  
   | | | | |
  
   WindowManagerPo-484 \[000\] 994.583135: sched\_switch: prev\_comm=WindowManagerPo
  
   SurfaceFlinger-334 \[000\] 994.583652: sched\_switch: prev\_comm=SurfaceFlinger
  
   WindowManagerPo-484 \[000\] 994.584320: sched\_switch: prev\_comm=WindowManagerPo
  
   GL updater-675 \[000\] 994.584370: sched\_switch: prev\_comm=GL updater
  
   WindowManagerPo-484 \[000\] 994.584424: sched\_switch: prev\_comm=WindowManagerPo
  

​

查找 IRQ 延迟

• 要查找系统中的最大 IRQ 延迟和抢占延迟，请按如下方式为 IRQOFF 和 PREEMPTIONOFF 启用 ftrace 配置：

  CONFIG_IRQSOFF_TRACER =Y
  CONFIG_PREEMPT_TRACER =Y
  

  要使此配置生效，请重新编译内核。此配置可以有效检测毫秒级延迟。

• 要启用跟踪，请运行以下命令：

  echo 0 > /sys/kernel/debug/tracing/tracing_enabled
  

  echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
  

  echo irqsoff > /sys/kernel/debug/tracing/current_tracer
  

  echo 1 > /sys/kernel/debug/tracing/tracing_enabled
  

  cat /sys/kernel/debug/tracing/trace
  

• 要查找系统中观测到的最大延迟，请配置 tracing_max_latency = 0。要检测任何超过指定限制的延迟，请以微秒为单位设置阈值级别。例如，

  echo 2000 > /sys/kernel/debug/tracing/tracing_thresh
  

cat /sys/kernel/debug/tracing/trace

# tracer: irqsoff

#

# WARNING: FUNCTION TRACING IS CORRUPTED

# MAY BE MISSING FUNCTION EVENTS

# irqsoff latency trace v1.1.5 on 3.4.0-perf-g7736d93-dirty

# --------------------------------------------------------------------

# latency: 16757 us, #4/4, CPU#0 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:1)

# -----------------

# | task: EventThread-534 (uid:1000 nice:-9 policy:0 rt\_prio:0)

# -----------------

# =\> started at: _raw_spin\_lock\_irqsave

# =\> ended at: _raw_spin\_unlock\_irqrestore

#

#

# \_------=\> CPU#

# / \_-----=\> irqs-off

# | / \_----=\> need-resched

# || / \_---=\> hardirq/softirq

# ||| / \_--=\> preempt-depth

# |||| / delay

# cmd pid ||||| time | caller

# / ||||| | /

\<...\>-534 0d... 0us!: _raw_spin\_lock\_irqsave

\<...\>-534 0d..1 16756us\+: _raw_spin\_unlock\_irqrestore

\<...\>-534 0d..1 16758us\+: trace\_hardirqs\_on \<-\_raw\_spin\_unlock\_irqrestore

\<...\>-534 0d..1 16780us : \<stack trace\>

=\> trace\_hardirqs\_on

=\> _raw_spin\_unlock\_irqrestore

=\> clk\_enable

=\> mdss\_dsi\_clk\_enable

=\> mdss\_dsi\_clk\_ctrl

=\> mdss\_dsi\_clk\_req

=\> mdss\_dsi\_event\_handler

=\> mdss\_mdp\_ctl\_intf\_event

=\> mdss\_mdp\_cmd\_add\_vsync\_handler

=\> mdss\_mdp\_overlay\_vsync\_ctrl

=\> mdss\_mdp\_overlay\_ioctl\_handler

=\> mdss\_fb\_ioctl

=\> do\_fb\_ioctl

=\> fb\_ioctl

=\> do\_vfs\_ioctl

=\> sys\_ioctl

=\> ret\_fast\_syscall

/sys/kernel/debug/tracing #