[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Yb8t7R577H3DiJGz@krava>
Date: Sun, 19 Dec 2021 14:04:45 +0100
From: Jiri Olsa <jolsa@...hat.com>
To: "Frank Ch. Eigler" <fche@...hat.com>
Cc: Namhyung Kim <namhyung@...nel.org>,
Arnaldo Carvalho de Melo <acme@...nel.org>,
lkml <linux-kernel@...r.kernel.org>,
Peter Zijlstra <a.p.zijlstra@...llo.nl>,
Ingo Molnar <mingo@...nel.org>,
Mark Rutland <mark.rutland@....com>,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
Michael Petlan <mpetlan@...hat.com>,
Ian Rogers <irogers@...gle.com>,
linux-perf-users <linux-perf-users@...r.kernel.org>
Subject: Re: [RFC] perf record: Disable debuginfod by default
On Fri, Dec 10, 2021 at 11:50:10AM -0500, Frank Ch. Eigler wrote:
> Hi -
>
> On Fri, Dec 10, 2021 at 01:23:32PM +0100, Jiri Olsa wrote:
> > [...]
> > at the end of the perf record we populate buildid cache
> > with profiled binaries for the current perf.data
> >
> > **IF** there's DEBUGINFOD_URLS defined, that code will
> > also ask debuginfod for binaries it could not find on
> > the system
>
> Consider doing this only at the end of the run, and in the background,
> just as a prefetch for the perf report step? The main downside there
> could be if one runs many perf record jobs in close proximity,
> overlapping larger prefetch download tasks. That might waste some
> network traffic.
right, IMO it's too much for default behaviour but I think could
do this under new (config) option, if there's a need for that
thanks,
jirka
Powered by blists - more mailing lists