lkml.org 
[lkml]   [2018]   [Feb]   [6]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH 01/17] perf report: Ask ordered events for --tasks option
Em Tue, Feb 06, 2018 at 07:59:51PM +0100, Jiri Olsa escreveu:
> On Tue, Feb 06, 2018 at 03:48:20PM -0300, Arnaldo Carvalho de Melo wrote:
> > Em Tue, Feb 06, 2018 at 07:17:57PM +0100, Jiri Olsa escreveu:
> > > If we have the time in, keep the events in time order.
> >
> > Try to be more verbose, what actual effect this will have in this particular
> > case?
> >
> > So, I had to try it to see the effects and explain them:
> >
> > --- /tmp/before 2018-02-06 15:40:29.536411625 -0300
> > +++ /tmp/after 2018-02-06 15:40:51.963403599 -0300
> > @@ -5,34 +5,34 @@
> > 2540 2540 1818 | gnome-terminal-
> > 3489 3489 2540 | bash
> > 32433 32433 3489 | perf
> > - 32434 32434 32433 | perf
> > + 32434 32434 32433 | make
> > 32441 32441 32434 | make
> > 32514 32514 32441 | make
> > 511 511 32514 | sh
> > - 512 512 511 | sh
> > + 512 512 511 | install
> >
> > We don't have perf calling perf calling make, etc, the second perf actually is
> > 'make', i.e. there was reordering of PERF_RECORD_COMM/PERF_RECORD_FORK:
> >
> > Look for FORK and COMM meta events, for those tids:
> >
> > [root@jouet acme]# perf report -D | egrep 'PERF_RECORD_(FORK|COMM)' | egrep '3243[34]'
> > 0 14774650990679 0x1a3cd8 [0x38]: PERF_RECORD_FORK(32433:32433):(3489:3489)
> > 1 14774652080381 0x1d6568 [0x30]: PERF_RECORD_COMM exec: perf:32433/32433
> > 1 14774742473340 0x1dbb48 [0x38]: PERF_RECORD_FORK(32434:32434):(32433:32433)
> > 0 14774752005779 0x1a4af8 [0x30]: PERF_RECORD_COMM exec: make:32434/32434
> > 0 14774753997960 0x1a5578 [0x38]: PERF_RECORD_FORK(32435:32435):(32434:32434)
> > 0 14774756070782 0x1a5618 [0x38]: PERF_RECORD_FORK(32438:32438):(32434:32434)
> > 0 14774757772939 0x1a5680 [0x38]: PERF_RECORD_FORK(32440:32440):(32434:32434)
> > 0 14774758230600 0x1a56e8 [0x38]: PERF_RECORD_FORK(32441:32441):(32434:32434)
> > [root@jouet acme]#
> >
> > So they are on different CPUs, thus ring buffers, and when we don't use
> > ordered_events, we end up mixing that up, right?
>
> right ;-) time sorted is always better..

Sure ;-) Adding the comments and applying...

- Arnaldo

\
 
 \ /
  Last update: 2018-02-06 20:21    [W:0.082 / U:0.760 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site