[lkml]   [2009]   [Jun]   [10]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [PATCH 1/3] ftrace: add tracepoint for timer
Hi Thomas:

Sorry for the delayed reply.

> Please stop this handwaving about potential use cases. I still have
> not seen a single technical argument which explains the benefit of
> those extra trace points. If we follow your "maybe its needed" logic
> then we have to add yet another dozen of tracepoints into the same
> functions to gather all possible states which might be there.
> Tracing is important but it needs to be done unintrusive to the code
> and you have to apply common sense which information is really
> valuable and necessary. Gathering random crap just because it might be
> necessary at some point is useless.

I see your point, and I agree with you on this.

> Get your gear together, sit down and think hard about which
> information a tracer or a probe needs to have to give us a useful
> insight into a particular object or function.
> As long as there is no technical convincing argument that the gathered
> information is really useful I'm not going to apply any of those
> patches to the code I maintain.

Those tracepoints are wanted and useful:

1) We can detect a timer's delay if jiffies info is added, like this:

XXX: timer_start: timer=e0b374e0 func=test_timerfuc expires=4294941565
XXX: timer_expire: timer=e0b374e0 func=test_timerfuc jiffies=4294941567

We expect the timer expires at 4294941565, actually the timer expires at
4294941567, so it is delayed by 2 jiffies.

2) We can monitor the lifecycle and behaviors of a timer, for example, when
monitoring dirty writeback timer, I found reading /proc/sys/vm/dirty_writeback_centisecs
will reset the timer:

pdflush-244 [000] 0.554131: timer_start: timer=c077ef20 func=wb_timer_fn expires=2757949
<idle>-0 [000] 5.544769: timer_expire: timer=c077ef20 func=wb_timer_fn
<idle>-0 [000] 5.544770: timer_cancel: timer=c077ef20 func=wb_timer_fn
pdflush-244 [000] 5.544807: timer_start: timer=c077ef20 func=wb_timer_fn expires=2762949
cat-2326 [000] 5.618852: timer_cancel: timer=c077ef20 func=wb_timer_fn
cat-2326 [000] 6.618858: timer_start: timer=c077ef20 func=wb_timer_fn expires=2764025

3) We are developing "flight-record" using crash. It can read out the ring
buffer from the dump file, and let us know what was happening before
kernel panic, so we may find the cause of the panic. This requires
well-defined tracepoints in different kernel subsystems, especially some
core subsystems, timer is surely one of them.

I'll revise those patches according to your suggestions.

Xiao Guangrong

> Thanks,
> tglx

 \ /
  Last update: 2009-06-10 12:03    [W:0.049 / U:1.380 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site