Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:3 as permitted sender) client-ip=2620:137:e000::3:3;
Date:   Wed, 22 Nov 2023 06:43:24 +0700
From:   Bagas Sanjaya <bagasdotme@gmail.com>
To:     Hector Martin <marcan@marcan.st>,
        Linux perf Profiling <linux-perf-users@vger.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Cc:     Marc Zyngier <maz@kernel.org>,
        Asahi Linux Mailing List <asahi@lists.linux.dev>,
        Ian Rogers <irogers@google.com>,
        Kan Liang <kan.liang@linux.intel.com>,
        Arnaldo Carvalho de Melo <acme@redhat.com>
Subject: Re: [REGRESSION] Perf (userspace) broken on big.LITTLE systems since
 v6.5
Message-ID: <ZV1AnNB2CSbAUFVg@archie.me>
References: <08f1f185-e259-4014-9ca4-6411d5c1bc65@marcan.st>
MIME-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha512;
        protocol="application/pgp-signature"; boundary="fu/8SvnJ7Xc8DEer"
Content-Disposition: inline
In-Reply-To: <08f1f185-e259-4014-9ca4-6411d5c1bc65@marcan.st>
Precedence: bulk


--fu/8SvnJ7Xc8DEer
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Tue, Nov 21, 2023 at 09:08:48PM +0900, Hector Martin wrote:
> Perf broke on all Apple ARM64 systems (tested almost everything), and
> according to maz also on Juno (so, probably all big.LITTLE) since v6.5.
>=20
> Test command:
>=20
> sudo taskset -c 0 ./perf stat -e apple_icestorm_pmu/cycles/ -e
> apple_firestorm_pmu/cycles/ -e cycles ls
>=20
> Since this is taskset to CPU #0 (LITTLE core, icestorm), only events for
> icestorm are expected.
>=20
> I bisected the breakage to two distinct points:
>=20
> 5ea8f2ccffb is the first bad commit. With its parent, the output is as
> expected (same as v6.4):
>=20
>          3,297,462      apple_icestorm_pmu/cycles/
>=20
>      <not counted>      apple_firestorm_pmu/cycles/
>                        (0.00%)
>      <not counted>      cycles
>                        (0.00%)
>=20
> With 5ea8f2ccffb everything breaks:
>=20
>    <not supported>      apple_icestorm_pmu/cycles/
>=20
>    <not supported>      apple_firestorm_pmu/cycles/
>=20
>      <not counted>      cycles
>                        (0.00%)
>=20
> Somewhere along the way to 82fe2e45cdb00 things get even worse (didn't
> bother bisecting this range). With its parent:
>=20
>    <not supported>      apple_icestorm_pmu/cycles/
>=20
>    <not supported>      apple_firestorm_pmu/cycles/
>=20
>    <not supported>      apple_icestorm_pmu/cycles/
>=20
>    <not supported>      apple_firestorm_pmu/cycles/
>=20
> Then 82fe2e45cdb00 leads to the current v6.5 behavior:
>=20
>      <not counted>      apple_icestorm_pmu/cycles/
>                        (0.00%)
>      <not counted>      apple_firestorm_pmu/cycles/
>                        (0.00%)
>      <not counted>      cycles
>                        (0.00%)
>=20
> If I taskset the task to CPU#2 (big core, firestorm), I get events:
>=20
>          1,454,858      apple_icestorm_pmu/cycles/
>=20
>          1,454,760      apple_firestorm_pmu/cycles/
>=20
>          1,454,384      cycles
>=20
>=20
> So the current behavior is that all output seems to come from the
> firestorm PMU event counter, regardless of requested event.
>=20
> This is all unchanged and still broken in v6.7-rc2.
>=20

Thanks for the regression report (and it has been handled well already).
I'm adding it to regzbot for tracking:

#regzbot ^introduced: 5ea8f2ccffb239

--=20
An old man doll... just what I always wanted! - Clara

--fu/8SvnJ7Xc8DEer
Content-Type: application/pgp-signature; name="signature.asc"

-----BEGIN PGP SIGNATURE-----

iHUEABYKAB0WIQSSYQ6Cy7oyFNCHrUH2uYlJVVFOowUCZV1AlwAKCRD2uYlJVVFO
o8aCAP9VLRaP2SCBxEoyHuPgsn0Q5WQW3SmEqOCX6a8pq9iEYAD9FltPgjVA6zjk
h9C6hLjE6H67lgw50GnDhziUTCtluAY=
=9D7c
-----END PGP SIGNATURE-----

--fu/8SvnJ7Xc8DEer--