This is the mail archive of the systemtap@sourceware.org mailing list for the systemtap project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH -tip v6 00/22] kprobes: introduce NOKPROBE_SYMBOL(), cleanup and fixes crash bugs


(2013/12/23 22:04), Ingo Molnar wrote:
> 
> * Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> wrote:
> 
>>> I can reproduce one weirdness, with just 13 probes added, 'perf probe 
>>> -l' [which should really be 'perf probe list'!] executes very slowly:
>>>
>>>  # perf stat --null --repeat 3 perf probe -l
>>>
>>>  Performance counter stats for 'perf probe -l' (3 runs):
>>>
>>>        0.763640098 seconds time elapsed                                          ( +-  1.61% )
>>>
>>> 0.7 seconds is ridiculously long.
>>
>> This comes from the perf-probe's internal issue. It tries to get 
>> actual lines and files by using debuginfo. I guess most of the time 
>> consumed when processing huge kernel debuginfo.
> 
> No, I don't have any debuginfo, I never use it.

Then it should be a performance bug. I need to take a look.

BTW, this is just a bug of perf probe, for example,

 # wc -l /sys/kernel/debug/tracing/kprobe_events
 3204 /sys/kernel/debug/tracing/kprobe_events

 # ./perf stat --null --repeat 3 cat /sys/kernel/debug/tracing/kprobe_events

  Performance counter stats for 'cat /sys/kernel/debug/tracing/kprobe_events' (3 runs):

       0.032075943 seconds time elapsed                                          ( +- 38.72% )

So even with 3204 probes, listing the kprobe events only takes 0.03 sec.

Thank you,

-- 
Masami HIRAMATSU
IT Management Research Dept. Linux Technology Center
Hitachi, Ltd., Yokohama Research Laboratory
E-mail: masami.hiramatsu.pt@hitachi.com



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]