lkml.org 
[lkml]   [2017]   [Nov]   [2]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH 1/2] perf mmap: Fix perf backward recording
On Wed, Nov 01, 2017 at 08:56:32PM +0800, Wangnan (F) wrote:
>
>
> On 2017/11/1 20:39, Jiri Olsa wrote:
> > On Wed, Nov 01, 2017 at 08:10:49PM +0800, Wangnan (F) wrote:
> >
> > SNIP
> >
> > > > > > diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
> > > > > > index c6c891e154a6..27ebe355e794 100644
> > > > > > --- a/tools/perf/util/evlist.c
> > > > > > +++ b/tools/perf/util/evlist.c
> > > > > > @@ -838,6 +838,11 @@ static int perf_evlist__mmap_per_evsel(struct perf_evlist *evlist, int idx,
> > > > > > if (*output == -1) {
> > > > > > *output = fd;
> > > > > > + if (evsel->attr.write_backward)
> > > > > > + mp->prot = PROT_READ;
> > > > > > + else
> > > > > > + mp->prot = PROT_READ | PROT_WRITE;
> > > > > > +
> > > > > If evlist->overwrite is true, PROT_WRITE should be unset even if
> > > > > write_backward is
> > > > > not set. If you want to delay the setting of mp->prot, you need to consider
> > > > > both evlist->overwrite and evsel->attr.write_backward.
> > > > I thought evsel->attr.write_backward should be set when
> > > > evlist->overwrite is set. Do you mean following case?
> > > >
> > > > perf record --overwrite -e 'cycles/no-overwrite/'
> > > >
> > > No. evlist->overwrite is unrelated to '--overwrite'. This is why I
> > > said the concept of 'overwrite' and 'backward' is ambiguous.
> > >
> > > perf record never sets 'evlist->overwrite'. What '--overwrite' actually
> > > does is setting write_backward. Some testcases needs overwrite evlist.
> > did not check deeply, but so why can't we do the below?
> >
> > jirka
> >
> >
> > ---
> > diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
> > index f4d9fc54b382..4643c487bd29 100644
> > --- a/tools/perf/builtin-record.c
> > +++ b/tools/perf/builtin-record.c
> > @@ -300,7 +300,7 @@ static int record__mmap_evlist(struct record *rec,
> > struct record_opts *opts = &rec->opts;
> > char msg[512];
> > - if (perf_evlist__mmap_ex(evlist, opts->mmap_pages, false,
> > + if (perf_evlist__mmap_ex(evlist, opts->mmap_pages, opts->overwrite,
> > opts->auxtrace_mmap_pages,
> > opts->auxtrace_snapshot_mode) < 0) {
>
> perf_evlist__mmap_ex's overwrite argument is used to control evlist->mmap.
>
> Consider Namhyung's example:
>
> perf record --overwrite -e 'cycles/no-overwrite/'
>
> In this case both evlist->mmap and evlist->backward_mmap would be set to overwrite.
> 'cycles' will be put into evlist->mmap, so it will be set to overwrite incorrectly.

right, missed the separate mmaps..

so we have some code that uses evlist->overwrite, which is always
set to 'false' in perf record.. but in the crucial checks like
for perf_mmap__consume we use the 'backward' bool to save the day

that might need some consolidation as well.. we could keep the
overwrite flag in the struct perf_mmap.. that could simplify the code

jirka

\
 
 \ /
  Last update: 2017-11-02 16:13    [W:0.313 / U:0.392 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site