linux-kernel - Re: [PATCH v1 2/2] perf script: Fix for `perf script +F metric` with leader sampling

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <Zp_Lrlz1S5uVlovH@tassilo>
Date: Tue, 23 Jul 2024 08:26:38 -0700
From: Andi Kleen <ak@...ux.intel.com>
To: Ian Rogers <irogers@...gle.com>
Cc: James Clark <james.clark@...aro.org>,
	Peter Zijlstra <peterz@...radead.org>,
	Ingo Molnar <mingo@...hat.com>,
	Arnaldo Carvalho de Melo <acme@...nel.org>,
	Namhyung Kim <namhyung@...nel.org>,
	Mark Rutland <mark.rutland@....com>,
	Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
	Jiri Olsa <jolsa@...nel.org>,
	Adrian Hunter <adrian.hunter@...el.com>,
	Kan Liang <kan.liang@...ux.intel.com>,
	linux-perf-users@...r.kernel.org, linux-kernel@...r.kernel.org,
	Athira Rajeev <atrajeev@...ux.vnet.ibm.com>
Subject: Re: [PATCH v1 2/2] perf script: Fix for `perf script +F metric` with
 leader sampling

> > For some reason I only get the metric: lines when I record with -a. I
> > noticed this because Andi's test doesn't use -a so it fails.
> >
> > I'm not sure if that's expected or it's related to your disclaimer below?
> 
> It is. When you don't do -a the cpu map just contains -1 and for some

The intention was to make it work without -a

> reason it is busted. The whole indirections to arrays of arrays,

That's what I tried to fix earlier with my patch that remapped to index 0,
but it ran into the problem that it conflicted with the perf stat report setup.

The metrics code doesn't really need a map, all it wants to do is to
process the last contiguous group. 

That is at some point perhaps it would be nice to extend it to multiple
groups, but I wasn't trying to support that here.

> counts, stats, aggregations, with indices into various other arrays
> and a lack of helpers. The code works for perf stat, but there is a
> lot of complexity that I don't fully grok in that. Here I've tried to

I don't really know how we ended up with such convoluted code, but that's
a different issue.

-Andi