> On 三月 17, 2016, 1:02 a.m., Ben Mahler wrote: > > src/linux/perf.cpp, lines 435-437 > > <https://reviews.apache.org/r/44379/diff/2/?file=1293838#file1293838line435> > > > > Hm.. this comment is really hard for me to understand, if OS vendors > > enhance the format, how did you know that they only use this particular > > format and not others?
Thanks for the review! The format itself doesn't matter, the order of the format does, that's why we use size of tokens to parse the format. Moreover, this format is full fledged which should be supported by all means. btw, it's unlikely for my custom to depoly upstream kernel in production, that's why I stay on Centos7 stable kernel in most cases. - fan ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/44379/#review123962 ----------------------------------------------------------- On 三月 10, 2016, 3:26 a.m., fan du wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/44379/ > ----------------------------------------------------------- > > (Updated 三月 10, 2016, 3:26 a.m.) > > > Review request for mesos and Ben Mahler. > > > Bugs: MESOS-4705 > https://issues.apache.org/jira/browse/MESOS-4705 > > > Repository: mesos > > > Description > ------- > > For vanilla kernel <= 3.12, 'perf stat' format is: > value,event,cgroup > Tested with 3.12, 3.11, 3.10 vanilla kernel. > > OS vendors may enhance perf tools with additional format: > value,unit,event,cgroup,running,ratio > Tested on Centos 7 with kernel 3.10.0-{123 OR 229}.el7.x86_64 > > > Diffs > ----- > > src/linux/perf.cpp 1c113a2b3f57877e132bbd65e01fb2f045132128 > > Diff: https://reviews.apache.org/r/44379/diff/ > > > Testing > ------- > > 1. {Found and Test} with Serenity, ema filter could get perf event statistics > correctly as expected. > 2. ./bin/mesos-tests.sh --gtest_filter=PerfEventIsolatorTest* > --log_dir=/tmp/mesos/ > > > Thanks, > > fan du > >
