From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail-pl1-f170.google.com (mail-pl1-f170.google.com [209.85.214.170]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 91D9E2C6AF for ; Thu, 22 Aug 2024 15:15:28 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.214.170 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1724339730; cv=none; b=GPvxYnM9dlOpgzs+Yc8doFzbR6aEen0GSW9PF3uCCAgv+YUk0AW2tp6NTsOmNvGQw6yKyn6bFgx3iFxnMvTUXfyQWhiqTstnTamr01tTXGVX7BOqyWLvVDPXzNdqf+c73dcdNC3w+PXVbE0BVvVfPXoWrhpvCfusiEB8I8vW/9U= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1724339730; c=relaxed/simple; bh=BUzcI7OukSS13o2EOotZtgvbOMnfxnRYGYS1AuA8RtU=; h=MIME-Version:References:In-Reply-To:From:Date:Message-ID:Subject: To:Cc:Content-Type; b=cZKaolTrwdO/GIEw+w4QOAmLSPQbVHU/B2TOV2vDbCWO3khQ4rMMT+Es/lKnvxJW61yQuKcMZeGWX2uagBa1p7qOCI4482P/WiVFhokoXZ5KL5HWM9xmtzDdHe8ye4vHACHpdMeJ5hIRvT18JWDLFXzIGh/t/0B83eLqLPhbPB4= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=mCiRi6V5; arc=none smtp.client-ip=209.85.214.170 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="mCiRi6V5" Received: by mail-pl1-f170.google.com with SMTP id d9443c01a7336-202018541afso165865ad.1 for ; Thu, 22 Aug 2024 08:15:28 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1724339728; x=1724944528; darn=vger.kernel.org; h=content-transfer-encoding:cc:to:subject:message-id:date:from :in-reply-to:references:mime-version:from:to:cc:subject:date :message-id:reply-to; bh=lck+dKvH8IO6d/hyhyQtHkkuQFQqgFpWof7AZCeNw9c=; b=mCiRi6V5OYD6/cIdWyohJ6kPWdgwZvAlWko3k2veUEhm5Un+tb2ilasYX3OZiOekDL Jmt0t3LvvdGvyMewB6QPtCyqSHxdAg7gW+Wojj0Wueo1UmD4mczMvIaDxhHfNhbiWGrU j5LVavgwHNWtfKy3FkJEqlEOQCrYccvHyXrIZJKcvtTEvLPZ1SzMjJA7WLeztJg67q8x UIAsDoQVvnLXVINfBI4l9GICF5+xINhwgf/FP0c+xw6wZnXCECFPso6RO8uX7rzKZKLC qYvLUXcuwBrnu5woJ37Cg0WSyhYXw17bajPya+xafziun1KIB90hRj4GBre/sXoNfkVr TXTw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1724339728; x=1724944528; h=content-transfer-encoding:cc:to:subject:message-id:date:from :in-reply-to:references:mime-version:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=lck+dKvH8IO6d/hyhyQtHkkuQFQqgFpWof7AZCeNw9c=; b=swFIIwl3gu7E5QrTiaR3kbF+6zg2gxl44RYAJ1JoBufxBSorEDN/vn05R0wF3tPMQs 6y0Vl095dA4niD0XwkMlPSHoybu+DzMKZNZ04w93OP9EDMpNy+H0EG/ELoTxbQJ6lgG4 wc8zhSVyh45S8vj7/Zsip+6aYSXwoMbVnPpgx509vMitCyIzJe/zb17VGcPoH94t/Y5I 5Chzx8OtX70gP/Me4nIxu7n87uxUkwH8zIfsHmin7YUyK7+yLGcIuauX3xbDGKuFBtU4 EUAv2wRD1PjqbObPLUHs1ncfig5Cm5EUaxwjfPuCm9CG6TgMRRPSVjprC7Sut9UytmPB K4+A== X-Forwarded-Encrypted: i=1; AJvYcCVua5r1bh4KE7Mc6q336arez73jX72SIM5f3bGpI+vfobtqKNUzSwzt8VRfpS6TFu559mE7MbPqpZnOZbCNmS+z@vger.kernel.org X-Gm-Message-State: AOJu0Yw29aeq34PxZDj08Mfnxv5UAYG6rokgFHhbtAco2JSF8ZCy7vyL Li7snAISt5B4Ypw6+wVnrnaV+UZOqVvDy9Y1Gp8YU4hqr5p+3Bg2HcM8Kq2m1wFEupw0jgVmL7m s4dIl9VhUGyye+0t2mUsOK9Pq5HzO8/+jkd/a X-Google-Smtp-Source: AGHT+IHNIxkXRfs6rii3n8WwQKyCifwdGwWDQT62P2QnRmbheBcHUcPmp8diG+t5fKpMkFMff43JVN8PjuaoB4dtrLg= X-Received: by 2002:a17:902:ec92:b0:1fa:fe30:8fce with SMTP id d9443c01a7336-203855b162bmr2936145ad.23.1724339727292; Thu, 22 Aug 2024 08:15:27 -0700 (PDT) Precedence: bulk X-Mailing-List: linux-perf-users@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 References: <05E55194-C265-4BDA-911D-B9E57EED3CBB@nutanix.com> <65A24EAA-A348-4E77-AF5B-933B41C9CAB6@nutanix.com> <067B5698-9C34-4283-888B-DC691F3462A8@nutanix.com> In-Reply-To: <067B5698-9C34-4283-888B-DC691F3462A8@nutanix.com> From: Ian Rogers Date: Thu, 22 Aug 2024 08:15:05 -0700 Message-ID: Subject: Re: Perf test failures for 10.2 PMU event map aliases To: Jon Kohler Cc: "adrian.hunter@intel.com" , "linux-perf-users@vger.kernel.org" , LKML , Kan Liang , "alexander.shishkin@linux.intel.com" Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable On Thu, Aug 22, 2024 at 8:01=E2=80=AFAM Jon Kohler wrote: > > > > > On Aug 20, 2024, at 11:17=E2=80=AFAM, Ian Rogers w= rote: > > > > !-------------------------------------------------------------------| > > CAUTION: External Email > > > > |-------------------------------------------------------------------! > > > > On Tue, Aug 20, 2024 at 6:54=E2=80=AFAM Jon Kohler wr= ote: > >> > >> > >> > >>> On Aug 20, 2024, at 1:41=E2=80=AFAM, Ian Rogers = wrote: > >>> > >>> !-------------------------------------------------------------------| > >>> CAUTION: External Email > >>> > >>> |-------------------------------------------------------------------! > >>> > >>> On Mon, Aug 19, 2024 at 7:06=E2=80=AFPM Jon Kohler = wrote: > >>>> > >>>> Reaching out to the perf community for feedback on the following > >>>> observed test failure. On 6.6.y, I see persistent failures with test > >>>> 10.2 PMU event map aliases, complaining about testing aliases uncore > >>>> PMU mismatches. I've included two outputs below, one with a bit of > >>>> hacky print debugging. > >>>> > >>>> Using Intel(R) Xeon(R) Gold 6154 CPU: > >>>> 10.2: PMU event map aliases = : > >>>> --- start --- > >>>> test child forked, pid 962901 > >>>> Using CPUID GenuineIntel-6-55-4 > >>> > >>> Hi Jon, > >>> > >>> Sorry for the brief reply but I thought some quick hints might unbloc= k > >>> you on this. The CPUID lines up with a SkylakeX: > >>> https://urldefense.proofpoint.com/v2/url?u=3Dhttps-3A__git.kernel.org= _pub_scm_linux_kernel_git_perf_perf-2Dtools-2Dnext.git_tree_tools_perf_pmu-= 2Devents_arch_x86_mapfile.csv-3Fh-3Dperf-2Dtools-2Dnext-23n33&d=3DDwIFaQ&c= =3Ds883GpUCOChKOHiocYtGcg&r=3DNGPRGGo37mQiSXgHKm5rCQ&m=3DRJx661xzakrB42hsUs= FD1HhJczkgpaYur9lHVtl7j36__CBOqYfKf4Dnq0xdpBZl&s=3DF-eXsmTASgRsptt5Gahro6fR= yMwEQdjZ6PtY7vhzIKM&e=3D > >>> > >>>> testing core PMU cpu aliases: pass > >>>> JKDBG: pmu nr total 3 pmu->sysfs_aliases 3 pmu->sys_json_alias= es 0 > >>>> JKDBG: pmu cpu_aliases_added nr total 4 pmu->cpu_json_aliases = 1 > >>>> testing aliases uncore PMU uncore_imc_0: mismatch expected ali= ases > >>>> (1) vs found (4) > >>>> test child finished with -1 > >>>> ---- end ---- > >>>> PMU events subtest 2: FAILED! > >>>> > >>>> Using Intel(R) Xeon(R) Platinum 8352Y: > >>>> 10.2: PMU event map aliases = : > >>>> --- start --- > >>>> test child forked, pid 1765070 > >>>> Using CPUID GenuineIntel-6-6A-6 > >>> > >>> This is an IcelakeX: > >>> https://urldefense.proofpoint.com/v2/url?u=3Dhttps-3A__git.kernel.org= _pub_scm_linux_kernel_git_perf_perf-2Dtools-2Dnext.git_tree_tools_perf_pmu-= 2Devents_arch_x86_mapfile.csv-3Fh-3Dperf-2Dtools-2Dnext-23n18&d=3DDwIFaQ&c= =3Ds883GpUCOChKOHiocYtGcg&r=3DNGPRGGo37mQiSXgHKm5rCQ&m=3DRJx661xzakrB42hsUs= FD1HhJczkgpaYur9lHVtl7j36__CBOqYfKf4Dnq0xdpBZl&s=3D6DwD4ZmywAtcwCnRjx7wRfmd= W_G65wHIuyZJIc__2yc&e=3D > >>> > >>>> testing core PMU cpu aliases: pass > >>>> testing aliases uncore PMU uncore_imc_free_running_0: mismatch > >>>> expected aliases (1) vs found (6) > >>>> test child finished with -1 > >>>> ---- end ---- > >>>> PMU events subtest 2: FAILED! > >>>> > >>>> Digging in more, looking at pmu_aliases_parse, I see that we'll disc= ard > >>>> scale and unit files in pmu_alias_info_file, which leaves us with 3x > >>>> aliases in the uncore_imc_0 in the first case and 5x aliases in the > >>>> uncore_imc_free_running_0 second case. > >>>> > >>>> # From 6154-based system: > >>>> ls -lhat /sys/devices/uncore_imc_0/events > >>> > >>> The "uncore_" prefix and the "_0" suffix are optional, the naming > >>> matching is case insensitive. In the event json the events are listed > >>> here: > >>> https://urldefense.proofpoint.com/v2/url?u=3Dhttps-3A__git.kernel.org= _pub_scm_linux_kernel_git_perf_perf-2Dtools-2Dnext.git_tree_tools_perf_pmu-= 2Devents_arch_x86_skylakex_uncore-2Dmemory.json-3Fh-3Dperf-2Dtools-2Dnext&d= =3DDwIFaQ&c=3Ds883GpUCOChKOHiocYtGcg&r=3DNGPRGGo37mQiSXgHKm5rCQ&m=3DRJx661x= zakrB42hsUsFD1HhJczkgpaYur9lHVtl7j36__CBOqYfKf4Dnq0xdpBZl&s=3DFpAgVwLmTyXUF= QIMZ_gbPlH9aXvrmcJ8CZaW3tKIaj4&e=3D > >>> > >>>> total 0 > >>>> -r--r--r--. 1 root root 4.0K Aug 19 18:50 cas_count_read.scale > >>>> -r--r--r--. 1 root root 4.0K Aug 19 18:50 cas_count_read.unit > >>>> -r--r--r--. 1 root root 4.0K Aug 19 18:50 cas_count_write.scale > >>>> -r--r--r--. 1 root root 4.0K Aug 19 18:50 cas_count_write.unit > >>>> -r--r--r--. 1 root root 4.0K Aug 9 15:30 cas_count_read > >>>> -r--r--r--. 1 root root 4.0K Aug 9 15:30 cas_count_write > >>>> -r--r--r--. 1 root root 4.0K Aug 9 15:30 clockticks > >>> > >>> This should be 3 sysfs events (I don't like the term alias), note tha= t > >>> we load the sysfs and json events lazily to avoid overhead. > >>> > >>>> drwxr-xr-x. 2 root root 0 Jul 17 03:40 . > >>>> drwxr-xr-x. 5 root root 0 Jul 17 02:52 .. > >>>> > >>>> # From the 8352Y-based system: > >>>> ls -lhat /sys/bus/event_source/devices/uncore_imc_free_running_0/eve= nts > >>>> total 0 > >>>> -r--r--r--. 1 root root 4.0K Aug 20 01:44 ddrt_read.scale > >>>> -r--r--r--. 1 root root 4.0K Aug 20 01:44 ddrt_read.unit > >>>> -r--r--r--. 1 root root 4.0K Aug 20 01:44 ddrt_write.scale > >>>> -r--r--r--. 1 root root 4.0K Aug 20 01:44 ddrt_write.unit > >>>> -r--r--r--. 1 root root 4.0K Aug 20 01:44 read.scale > >>>> -r--r--r--. 1 root root 4.0K Aug 20 01:44 read.unit > >>>> -r--r--r--. 1 root root 4.0K Aug 20 01:44 write.scale > >>>> -r--r--r--. 1 root root 4.0K Aug 20 01:44 write.unit > >>>> -r--r--r--. 1 root root 4.0K Aug 19 21:33 dclk > >>>> -r--r--r--. 1 root root 4.0K Aug 19 21:33 ddrt_read > >>>> -r--r--r--. 1 root root 4.0K Aug 19 21:33 ddrt_write > >>>> -r--r--r--. 1 root root 4.0K Aug 19 21:33 read > >>>> -r--r--r--. 1 root root 4.0K Aug 19 21:33 write > >>> > >>> This is 5 sysfs events, the json events are here: > >>> https://urldefense.proofpoint.com/v2/url?u=3Dhttps-3A__git.kernel.org= _pub_scm_linux_kernel_git_perf_perf-2Dtools-2Dnext.git_tree_tools_perf_pmu-= 2Devents_arch_x86_icelakex_uncore-2Dmemory.json-3Fh-3Dperf-2Dtools-2Dnext-2= 3n134&d=3DDwIFaQ&c=3Ds883GpUCOChKOHiocYtGcg&r=3DNGPRGGo37mQiSXgHKm5rCQ&m=3D= RJx661xzakrB42hsUsFD1HhJczkgpaYur9lHVtl7j36__CBOqYfKf4Dnq0xdpBZl&s=3DMrHuUC= ZFqrNrd05IPyq4fuZDH4_owkEw0xHcc7bvGvU&e=3D > >>> Note, the "Unit", meaning the PMU should be imc_free_running to match > >>> this device. > >>> > >>>> drwxr-xr-x. 2 root root 0 Aug 15 03:15 . > >>>> drwxr-xr-x. 5 root root 0 Aug 15 02:42 .. > >>>> > >>>> Looking at the structure of __test_uncore_pmu_event_aliases, however= , > >>>> I'm not quite sure how this is supposed to work. I've annotated a wa= lk > >>>> through below to highlight where things are going off the rails. > >>>> > >>>> static int __test_uncore_pmu_event_aliases(struct perf_pmu_test_pmu = *test_pmu) > >>>> { > >>>> ... > >>>> /* Count how many aliases we generated */ > >>>> alias_count =3D perf_pmu__num_events(pmu); > >>>> // alias_count =3D=3D 4 in the 6154-based system > >>>> // alias_count =3D=3D 6 in the 8352Y-based system > >>>> > >>>> /* Count how many aliases we expect from the known table */ > >>>> for (table =3D &test_pmu->aliases[0]; *table; table++) > >>>> to_match_count++; > >>>> // this is looking at aliases in struct perf_p= mu_test_pmu > >>>> // table, which for uncore_imc_0 is a single e= ntry for > >>>> // &uncore_imc_cache_hits. > >>>> // > >>>> // for the 8352Y case, likewise, we only have = a single alias > >>>> // in the table for &uncore_imc_free_running_c= ache_miss. > >>>> // > >>>> // in both cases, to_match_count =3D=3D 1 > >>>> > >>>> // Compare 4 vs 1 or 6 vs 1 > >>>> if (alias_count !=3D to_match_count) { > >>>> pr_debug("testing aliases uncore PMU %s: mismatch expe= cted aliases (%d) vs found (%d)\n", > >>>> pmu_name, to_match_count /* 1 */, alias_count= /* 4 */); > >>>> return -1; > >>>> // we seemed doomed to hit this conditional al= ways, no? > >>>> } > >>>> ... > >>>> } > >>>> > >>>> I did a walkthrough of the latest mainline code, and don't see a mar= ked > >>>> difference that jump off the page to me that'd correct this behavior= , > >>>> and would love a helping hand to point in the right direction on thi= s. > >>>> > >>>> What am I missing here? > >>> > >>> I'll need some more time to dig into this. Hopefully the pointers abo= ve help. > >> > >> Thanks for the quick reply and pointers, I appreciate it. The tricky b= it still > >> remains, as the code I posted to above seems to solely depend on the > >> info filled into struct perf_pmu_test_pmu, right? If so, I don=E2=80= =99t see how the > >> dots connect from this test to the other events in sysfs/json=E2=80=99= s. > > > > So looking at the test it is using the testcpu: > > https://urldefense.proofpoint.com/v2/url?u=3Dhttps-3A__git.kernel.org_p= ub_scm_linux_kernel_git_perf_perf-2Dtools-2Dnext.git_tree_tools_perf_tests_= pmu-2Devents.c-3Fh-3Dperf-2Dtools-2Dnext-23n602&d=3DDwIFaQ&c=3Ds883GpUCOChK= OHiocYtGcg&r=3DNGPRGGo37mQiSXgHKm5rCQ&m=3DupMgwNSdGAw5sdDUTdoyvXhLy4KhFUYqP= dxZKx8Ov-ZxDYERFVy8PU040wwDAYPp&s=3DerGg8kUByjl_j5R0D0PxRZjTZhvazxwC9KW8rOT= 9Pp4&e=3D > > the json for that is here: > > https://urldefense.proofpoint.com/v2/url?u=3Dhttps-3A__git.kernel.org_p= ub_scm_linux_kernel_git_perf_perf-2Dtools-2Dnext.git_tree_tools_perf_pmu-2D= events_arch_test_test-5Fsoc_cpu-3Fh-3Dperf-2Dtools-2Dnext&d=3DDwIFaQ&c=3Ds8= 83GpUCOChKOHiocYtGcg&r=3DNGPRGGo37mQiSXgHKm5rCQ&m=3DupMgwNSdGAw5sdDUTdoyvXh= Ly4KhFUYqPdxZKx8Ov-ZxDYERFVy8PU040wwDAYPp&s=3Dz535_TbF_oJLjEoRuhbbqzB9Xo5Mw= WWmOcP0pgMulWY&e=3D > > The names in the test are based on ones seen on real CPUs, so this may > > be leading to the confusion. > > Hey Ian, > I was able to debug this a bit more. The following diff fixes this test o= n my system. > > Even though we were supposed to be using the test data only, the sysfs en= tries > from my systems, which happened to have similar names, threw a wrench in > this test. > > With this diff, we just use the JSON aliases that were added. > > Happy to send this out as a formal patch, but wanted to get the list=E2= =80=99s 2cents > first, as I feel like I=E2=80=99m missing something :) > > Jon > > diff --git a/tools/perf/tests/pmu-events.c b/tools/perf/tests/pmu-events.= c > index f5321fbdee79..893dc7afee76 100644 > --- a/tools/perf/tests/pmu-events.c > +++ b/tools/perf/tests/pmu-events.c > @@ -584,6 +584,9 @@ static int __test_uncore_pmu_event_aliases(struct per= f_pmu_test_pmu *test_pmu) > const struct pmu_events_table *events_table; > int res =3D 0; > > + /* CPU events come from struct pmu_event pmu_events__test_soc_cpu > + * and sys events come from struct pmu_event pmu_events__test_soc= _sys > + */ > events_table =3D find_core_events_table("testarch", "testcpu"); > if (!events_table) > return -1; > @@ -593,10 +596,14 @@ static int __test_uncore_pmu_event_aliases(struct p= erf_pmu_test_pmu *test_pmu) > pmu->sysfs_aliases_loaded =3D true; > pmu_add_sys_aliases(pmu); > > - /* Count how many aliases we generated */ > - alias_count =3D perf_pmu__num_events(pmu); > + /* How many events we gathered for this PMU in test_soc. > + * Note: we specifically do not use perf_pmu__num_events as that = may > + * include spurious system events from the system under test, whi= ch > + * may have similarly named PMUs. Thanks Jon, should we just rename the PMUs in the test json files? For example, rather than "CBO" here: https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git/tr= ee/tools/perf/pmu-events/arch/test/test_soc/cpu/uncore.json?h=3Dperf-tools-= next#n10 we can have "test_pmu1". Thanks for investigating this! Ian > + */ > + alias_count =3D pmu->cpu_json_aliases + pmu->sys_json_aliases; > > - /* Count how many aliases we expect from the known table */ > + /* How many aliases we expect from struct perf_pmu_test_pmu test_= pmus */ > for (table =3D &test_pmu->aliases[0]; *table; table++) > to_match_count++; > > > > > Thanks, > > Ian >