Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.33 as permitted sender) client-ip=23.128.96.33;
MIME-Version: 1.0
References: <20230925062323.840799-1-irogers@google.com> <CAM9d7chaKGciJgDLyN1hhDXNuFuCHRFcYAjTHY0bEMKpfBDZBg@mail.gmail.com>
 <CAP-5=fUqVB7W+a3o72VrhccWbos_XDu=jc53TmEZC26h_hBacA@mail.gmail.com>
In-Reply-To: <CAP-5=fUqVB7W+a3o72VrhccWbos_XDu=jc53TmEZC26h_hBacA@mail.gmail.com>
From:   Namhyung Kim <namhyung@kernel.org>
Date:   Mon, 2 Oct 2023 21:10:43 -0700
Message-ID: <CAM9d7cieEnuVZ7S0Yab6c3N=qKqgNSP3DL2JNFje1HFA7dcTtA@mail.gmail.com>
Subject: Re: [PATCH v1] perf pmus: Make PMU alias name loading lazy
To:     Ian Rogers <irogers@google.com>
Cc:     Peter Zijlstra <peterz@infradead.org>,
        Ingo Molnar <mingo@redhat.com>,
        Arnaldo Carvalho de Melo <acme@kernel.org>,
        Mark Rutland <mark.rutland@arm.com>,
        Alexander Shishkin <alexander.shishkin@linux.intel.com>,
        Jiri Olsa <jolsa@kernel.org>,
        Adrian Hunter <adrian.hunter@intel.com>,
        Leo Yan <leo.yan@linaro.org>,
        Ravi Bangoria <ravi.bangoria@amd.com>,
        James Clark <james.clark@arm.com>,
        Kan Liang <kan.liang@linux.intel.com>,
        linux-perf-users@vger.kernel.org, linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 8BIT
Precedence: bulk

On Wed, Sep 27, 2023 at 10:19 PM Ian Rogers <irogers@google.com> wrote:
>
> On Wed, Sep 27, 2023 at 10:00 PM Namhyung Kim <namhyung@kernel.org> wrote:
> >
> > Hi Ian,
> >
> > On Sun, Sep 24, 2023 at 11:24 PM Ian Rogers <irogers@google.com> wrote:
> > >
> > > PMU alias names were computed when the first perf_pmu is created,
> > > scanning all PMUs in event sources for a file called alias that
> > > generally doesn't exist. Switch to trying to load the file when all
> > > PMU related files are loaded in lookup. This would cause a PMU name
> > > lookup of an alias name to fail if no PMUs were loaded, so in that
> > > case all PMUs are loaded and the find repeated. The overhead is
> > > similar but in the (very) general case not all PMUs are scanned for
> > > the alias file.
> > >
> > > As the overhead occurs once per invocation it doesn't show in perf
> > > bench internals pmu-scan. On a tigerlake machine, the number of openat
> > > system calls for an event of cpu/cycles/ with perf stat reduces from
> > > 94 to 69 (ie 25 fewer openat calls).
> >
> > I think the pmu-scan bench could show the difference as it
> > calls perf_pmu__destroy() in the loop body.  So every call to
> > perf_pmu__scan() should start from nothing, right?
>
> The PMU alias name list was funny. It is/was maintained in the x86
> specific PMU code and the destroy didn't clear the list. This change
> adds an openat to loading a PMU for the alias, so pmu-scan shows a
> very small slow down. However, in the more normal cases we're reducing
> the number of openats by 25%.

I think that's ok.  Applied to perf-tools-next, thanks!