[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <172433479466.350842.6920589600831615538.b4-ty@arm.com>
Date: Thu, 22 Aug 2024 15:05:53 +0100
From: Suzuki K Poulose <suzuki.poulose@....com>
To: anshuman.khandual@....com,
James Clark <james.clark@...aro.org>,
gankulkarni@...amperecomputing.com,
mike.leach@...aro.org,
leo.yan@...ux.dev,
coresight@...ts.linaro.org
Cc: Suzuki K Poulose <suzuki.poulose@....com>,
Arnaldo Carvalho de Melo <acme@...nel.org>,
John Garry <john.g.garry@...cle.com>,
Maxime Coquelin <mcoquelin.stm32@...il.com>,
Peter Zijlstra <peterz@...radead.org>,
Ian Rogers <irogers@...gle.com>,
Alexandre Torgue <alexandre.torgue@...s.st.com>,
Mark Rutland <mark.rutland@....com>,
linux-perf-users@...r.kernel.org,
Ingo Molnar <mingo@...hat.com>,
linux-stm32@...md-mailman.stormreply.com,
"Liang, Kan" <kan.liang@...ux.intel.com>,
Will Deacon <will@...nel.org>,
Namhyung Kim <namhyung@...nel.org>,
linux-kernel@...r.kernel.org,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
linux-arm-kernel@...ts.infradead.org,
Adrian Hunter <adrian.hunter@...el.com>,
Jiri Olsa <jolsa@...nel.org>,
James Clark <james.clark@....com>
Subject: Re: (subset) [PATCH v6 00/17] coresight: Use per-sink trace ID maps for Perf sessions
On Mon, 22 Jul 2024 11:11:42 +0100, James Clark wrote:
> This will allow sessions with more than CORESIGHT_TRACE_IDS_MAX ETMs
> as long as there are fewer than that many ETMs connected to each sink.
>
> Each sink owns its own trace ID map, and any Perf session connecting to
> that sink will allocate from it, even if the sink is currently in use by
> other users. This is similar to the existing behavior where the dynamic
> trace IDs are constant as long as there is any concurrent Perf session
> active. It's not completely optimal because slightly more IDs will be
> used than necessary, but the optimal solution involves tracking the PIDs
> of each session and allocating ID maps based on the session owner. This
> is difficult to do with the combination of per-thread and per-cpu modes
> and some scheduling issues. The complexity of this isn't likely to worth
> it because even with multiple users they'd just see a difference in the
> ordering of ID allocations rather than hitting any limits (unless the
> hardware does have too many ETMs connected to one sink).
>
> [...]
Applied, the kernel driver changes to coresight/next. Thanks!
[09/17] coresight: Remove unused ETM Perf stubs
https://git.kernel.org/coresight/c/34172002bdac
[10/17] coresight: Clarify comments around the PID of the sink owner
https://git.kernel.org/coresight/c/eda1d11979c0
[11/17] coresight: Move struct coresight_trace_id_map to common header
https://git.kernel.org/coresight/c/acb0184fe9bc
[12/17] coresight: Expose map arguments in trace ID API
https://git.kernel.org/coresight/c/7e52877868ae
[13/17] coresight: Make CPU id map a property of a trace ID map
https://git.kernel.org/coresight/c/d53c8253c782
[14/17] coresight: Use per-sink trace ID maps for Perf sessions
https://git.kernel.org/coresight/c/5ad628a76176
[15/17] coresight: Remove pending trace ID release mechanism
https://git.kernel.org/coresight/c/de0029fdde86
[16/17] coresight: Emit sink ID in the HW_ID packets
https://git.kernel.org/coresight/c/487eec8da80a
[17/17] coresight: Make trace ID map spinlock local to the map
https://git.kernel.org/coresight/c/988d40a4d4e7
Best regards,
--
Suzuki K Poulose <suzuki.poulose@....com>
Powered by blists - more mailing lists