at 9:29 PM, Stefan Hajnoczi <stefa...@gmail.com> wrote: > > Nice, perf(1) is good for that. You can enable trace events and add > kprobes/uprobes to record timestamps when specific functions are entered. >
Thanks Stefan, One last question: Currently we can achieve hundreds of KIOPS. That means perf can easily capture millions of trace events per second. I found perf has quite high overhead at such a rate of trace events. Do you have any advices on tracing high IOPS tasks?
