From: Namhyung Kim <namhyung@kernel.org>
To: Sandipan Das <sandipan.das@amd.com>
Cc: Ian Rogers <irogers@google.com>,
linux-perf-users@vger.kernel.org, linux-kernel@vger.kernel.org,
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>,
James Clark <james.clark@linaro.org>,
x86@kernel.org, "H . Peter Anvin" <hpa@zytor.com>,
Ravi Bangoria <ravi.bangoria@amd.com>,
Ananth Narayan <ananth.narayan@amd.com>
Subject: Re: [PATCH v2 1/2] perf vendor events amd: Update Zen 5 core events
Date: Mon, 6 Jul 2026 10:29:39 -0700 [thread overview]
Message-ID: <akvmA_ondqn-lj9W@google.com> (raw)
In-Reply-To: <bf0b2ab8-c2b1-4020-8506-6e2188e78250@amd.com>
Hello,
On Mon, Jul 06, 2026 at 10:51:20PM +0530, Sandipan Das wrote:
> On 06-07-2026 22:44, Ian Rogers wrote:
> > On Mon, Jun 22, 2026 at 10:28 PM Sandipan Das <sandipan.das@amd.com> wrote:
> >>
> >> Update definitions for the following events.
> >> * PMCx00A - Add missing unit masks
> >> * PMCx00B - Add missing unit masks and fix descriptions
> >> * PMCx00C - Add missing unit masks
> >> * PMCx00D - Add missing unit masks
> >> * PMCx025 - Add missing unit masks and fix descriptions
> >>
> >> Fixes: 45c072f2537a ("perf vendor events amd: Add Zen 5 core events")
> >> Signed-off-by: Sandipan Das <sandipan.das@amd.com>
> >> ---
> >>
> >> Previous versions can be found at:
> >> v1: https://lore.kernel.org/all/f16d1391366762dee0d49f00697f5e6ea913e2c1.1780330378.git.sandipan.das@amd.com/
> >>
> >> Changes in v2:
> >> * Fix event description pointed out by sashiko.
> >>
> >> ---
> >> .../arch/x86/amdzen5/floating-point.json | 130 +++++++++++++++---
> >> .../arch/x86/amdzen5/load-store.json | 8 +-
> >> 2 files changed, 120 insertions(+), 18 deletions(-)
> >>
> >> diff --git a/tools/perf/pmu-events/arch/x86/amdzen5/floating-point.json b/tools/perf/pmu-events/arch/x86/amdzen5/floating-point.json
> >> index 9204bfb1d69e..569975b53cc3 100644
> >> --- a/tools/perf/pmu-events/arch/x86/amdzen5/floating-point.json
> >> +++ b/tools/perf/pmu-events/arch/x86/amdzen5/floating-point.json
> >> @@ -179,6 +179,30 @@
> >> "BriefDescription": "Retired scalar floating-point blend ops.",
> >> "UMask": "0x09"
> >> },
> >> + {
> >> + "EventName": "fp_ops_retired_by_type.scalar_mov",
> >> + "EventCode": "0x0a",
> >> + "BriefDescription": "Retired scalar floating-point MOV ops.",
> >> + "UMask": "0x0a"
> >> + },
> >> + {
> >> + "EventName": "fp_ops_retired_by_type.scalar_shuffle",
> >> + "EventCode": "0x0a",
> >> + "BriefDescription": "Retired scalar floating-point shuffle ops (may include instructions not necessarily thought of as including shuffles e.g. horizontal add, dot product, and certain MOV instructions).",
> >> + "UMask": "0x0b"
> >> + },
> >> + {
> >> + "EventName": "fp_ops_retired_by_type.scalar_bfloat",
> >> + "EventCode": "0x0a",
> >> + "BriefDescription": "Retired scalar floating-point bfloat ops.",
> >> + "UMask": "0x0c"
> >> + },
> >> + {
> >> + "EventName": "fp_ops_retired_by_type.scalar_logical",
> >> + "EventCode": "0x0a",
> >> + "BriefDescription": "Retired scalar floating-point logical ops.",
> >> + "UMask": "0x0d"
> >> + },
> >> {
> >> "EventName": "fp_ops_retired_by_type.scalar_other",
> >> "EventCode": "0x0a",
> >> @@ -245,12 +269,24 @@
> >> "BriefDescription": "Retired vector floating-point blend ops.",
> >> "UMask": "0x90"
> >> },
> >> + {
> >> + "EventName": "fp_ops_retired_by_type.vector_mov",
> >> + "EventCode": "0x0a",
> >> + "BriefDescription": "Retired vector floating-point MOV ops.",
> >> + "UMask": "0xa0"
> >> + },
> >> {
> >> "EventName": "fp_ops_retired_by_type.vector_shuffle",
> >> "EventCode": "0x0a",
> >> "BriefDescription": "Retired vector floating-point shuffle ops (may include instructions not necessarily thought of as including shuffles e.g. horizontal add, dot product, and certain MOV instructions).",
> >> "UMask": "0xb0"
> >> },
> >> + {
> >> + "EventName": "fp_ops_retired_by_type.vector_bfloat",
> >> + "EventCode": "0x0a",
> >> + "BriefDescription": "Retired vector floating-point bfloat ops.",
> >> + "UMask": "0xc0"
> >> + },
> >> {
> >> "EventName": "fp_ops_retired_by_type.vector_logical",
> >> "EventCode": "0x0a",
> >> @@ -278,7 +314,7 @@
> >> {
> >> "EventName": "sse_avx_ops_retired.mmx_add",
> >> "EventCode": "0x0b",
> >> - "BriefDescription": "Retired MMX integer add.",
> >> + "BriefDescription": "Retired MMX integer add ops.",
> >> "UMask": "0x01"
> >> },
> >> {
> >> @@ -299,16 +335,34 @@
> >> "BriefDescription": "Retired MMX integer multiply-accumulate ops.",
> >> "UMask": "0x04"
> >> },
> >> + {
> >> + "EventName": "sse_avx_ops_retired.mmx_aes",
> >> + "EventCode": "0x0b",
> >> + "BriefDescription": "Retired MMX integer AES ops.",
> >> + "UMask": "0x05"
> >> + },
> >> + {
> >> + "EventName": "sse_avx_ops_retired.mmx_sha",
> >> + "EventCode": "0x0b",
> >> + "BriefDescription": "Retired MMX integer SHA ops.",
> >> + "UMask": "0x06"
> >> + },
> >> {
> >> "EventName": "sse_avx_ops_retired.mmx_cmp",
> >> "EventCode": "0x0b",
> >> "BriefDescription": "Retired MMX integer compare ops.",
> >> "UMask": "0x07"
> >> },
> >> + {
> >> + "EventName": "sse_avx_ops_retired.mmx_cvt",
> >> + "EventCode": "0x0b",
> >> + "BriefDescription": "Retired MMX integer convert or pack ops.",
> >> + "UMask": "0x08"
> >> + },
> >> {
> >> "EventName": "sse_avx_ops_retired.mmx_shift",
> >> "EventCode": "0x0b",
> >> - "BriefDescription": "Retired MMX integer shift ops.",
> >> + "BriefDescription": "Retired MMX integer shift or rotate ops.",
> >> "UMask": "0x09"
> >> },
> >> {
> >> @@ -324,9 +378,9 @@
> >> "UMask": "0x0b"
> >> },
> >> {
> >> - "EventName": "sse_avx_ops_retired.mmx_pack",
> >> + "EventName": "sse_avx_ops_retired.mmx_vnni",
> >
> > Just a heads up, when renaming events it may be a good idea to keep
> > the old name and mark the event as deprecated, then add the event
> > again with the new name. This avoids events breaking when the perf
> > tool is updated. Personally I feel vendors should get to decide about
> > this kind of thing, but there have been some strong sensitivities
> > around event names in the past.
> >
>
> Thanks for the suggestion. I agree that renames can be problematic
> for existing scripted workflows.
>
> Namhyung, please let me know if you want to revert these changes.
> I can send out v3.
It'd be nice if you can send an incremental fix on top of the current
perf-tools-next.
Thanks,
Namhyung
prev parent reply other threads:[~2026-07-06 17:29 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-06-23 5:28 [PATCH v2 1/2] perf vendor events amd: Update Zen 5 core events Sandipan Das
2026-06-23 5:28 ` [PATCH v2 2/2] perf vendor events amd: Update Zen 6 " Sandipan Das
2026-07-06 16:58 ` [PATCH v2 1/2] perf vendor events amd: Update Zen 5 " Namhyung Kim
2026-07-06 17:14 ` Ian Rogers
2026-07-06 17:21 ` Sandipan Das
2026-07-06 17:29 ` Namhyung Kim [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=akvmA_ondqn-lj9W@google.com \
--to=namhyung@kernel.org \
--cc=acme@kernel.org \
--cc=adrian.hunter@intel.com \
--cc=alexander.shishkin@linux.intel.com \
--cc=ananth.narayan@amd.com \
--cc=hpa@zytor.com \
--cc=irogers@google.com \
--cc=james.clark@linaro.org \
--cc=jolsa@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-perf-users@vger.kernel.org \
--cc=mark.rutland@arm.com \
--cc=mingo@redhat.com \
--cc=peterz@infradead.org \
--cc=ravi.bangoria@amd.com \
--cc=sandipan.das@amd.com \
--cc=x86@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox