Re: Summary of perf-tools-next changes and patches sent to linux-perf-users

[Arnaldo Carvalho de Melo <acme@kernel.org>]

On Mon, May 25, 2026 at 09:21:44PM -0300, Arnaldo Carvalho de Melo wrote:
> On Mon, May 25, 2026 at 09:15:35PM -0300, Arnaldo Carvalho de Melo wrote:
> > On Mon, May 25, 2026 at 04:35:51PM -0700, Ian Rogers wrote:
> > > I've been working on a script [1] to summarize what's been sent to the
> > > linux-perf-users mailing list (using nntp) and compare it to what's
> > > merged into the perf-tools-next branch (using git). The idea is to
> > > make it slightly easier to track down patches that might have been
> > > missed. The script shows what has been merged and what's pending, what
> > > Sashiko thinks, and whose turn it is to work on the patch (author,
> > > reviewer or maintainer) using some naive tag analysis. The output is
> > > quite large, so rather than adding it below, I will send the last 4
> > > weeks in separate e-mail replies to this message.
> > 
> > I'm doing this locally as well, helps.
> 
> For instance, from earlier today: 
> 
> $ cat /tmp/outstanding.txt
> 
> Outstanding patches for perf-tools-next — 2026-05-25
> (checked against next2/perf-tools-next)
> =====================================================
> 
> Tier 2 — Likely ready, needs closer look
> -----------------------------------------

I had alredy processed Tier 1 at this point so deleted it from the file
:-)

- Arnaldo
 
> 1. perf stat: Fix uncore metric scaling across aggregation modes
>    Author: Chun-Tse Shao <ctshao@google.com>
>    Version: v3, 2 patches
>    Reviews: Namhyung reviewed v1; no reviews on v3 yet
>    Message-ID: <20260521201505.124690-1-ctshao@google.com>
> 
> 2. perf python: Improvements and callchain support
>    Author: Ian Rogers <irogers@google.com>
>    Version: v9, 23 patches
>    Message-ID: <20260522220435.2378363-1-irogers@google.com>
> 
> 3. perf data ctf: Replace libbabeltrace with babeltrace2-ctf-writer
>    Author: Michael Jeanson <mjeanson@efficios.com>
>    Version: v1, 1 patch + Ian's 2-patch feature test
>    Message-ID: <20260512194710.162215-1-mjeanson@efficios.com>
>    Ian's follow-up: <20260512215621.468685-1-irogers@google.com>
> 
> 4. perf test: Harness improvements
>    Author: Ian Rogers <irogers@google.com>
>    Version: v1, 14 patches
>    Message-ID: <20260513230450.529380-1-irogers@google.com>
> 
> Tier 3 — Single patches, needs review
> --------------------------------------
> 
> 5. perf bench: Add --write-size option to sched pipe
>    Author: Breno Leitao <leitao@debian.org>
>    Version: v1, 1 patch
>    Message-ID: <20260515-perf_bench_pipe-v1-1-3c5b805ba178@debian.org>
> 
> 6. perf evsel: Fix error handling in tp_format lookup
>    Author: Hongling Zeng <zenghongling@kylinos.cn>
>    Version: v1, 1 patch
>    Message-ID: <20260521025738.17867-1-zenghongling@kylinos.cn>
> 
> 7. perf kvm stat: Update the exit reason mappings
>    Author: Gautam Menghani <gautam@linux.ibm.com>
>    Version: v1, 1 patch
>    Reviews: Ian reviewed (needs v2 per feedback)
>    Message-ID: <20260512120408.59332-1-gautam@linux.ibm.com>
> 
> 8. perf build: Fix Python extension build with GCC 16 hardening
>    Author: Jan Polensky <japo@linux.ibm.com>
>    Version: v1, 1 patch
>    Reviews: Thomas Richter Reviewed-by
>    Message-ID: <20260520125637.105681-1-japo@linux.ibm.com>
> 
> Already merged in next2/perf-tools-next
> -----------------------------------------
> 
> - perf tests: uncore/topdown event sorting (Ian, v8) — 9f116f4b811b
> - perf build: parallel build cleanup (Ian, v7) — 85cc481af5ff + 5 more
> - perf build: Update kernel headers (Namhyung, v3) — aa13e4b120f9..5a433107fab6
> - Fix SIGCHLD vs pause() race (Swapnil, v4) — 35c9fb22000f
> - Fix kwork ASAN + libunwind test (Ian, v1) — e898c505b0ee, fdec0a81cad5
> - Fix cgroup metric with BPF counters (Ian, v4) — 65bd18199ad5, bd2a5be1fe73
> - tool_pmu enable/disable (Ian, v2) — 0b18bced5444
> - libunwind multiple remote (Ian, v5) — fdec0a81cad5
> - tools/build: LLVM versioning (James, v2) — 98e68cb77823
> - perf inject intel-PT LBR/brstack (Ian, v3) — daac18e7c42c
> - perf trace --show-cpu (Aaron, v4) — e445b78ffb8d
> - stat delay duration_time test (Ian) — f83deb058025