[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20230601112027.GI38236@hirez.programming.kicks-ass.net>
Date: Thu, 1 Jun 2023 13:20:27 +0200
From: Peter Zijlstra <peterz@...radead.org>
To: Thomas Richter <tmricht@...ux.ibm.com>
Cc: Ian Rogers <irogers@...gle.com>, Marc Zyngier <maz@...nel.org>,
Oliver Upton <oliver.upton@...ux.dev>,
Ravi Bangoria <ravi.bangoria@....com>,
Nathan Chancellor <nathan@...nel.org>, namhyung@...nel.org,
eranian@...gle.com, acme@...nel.org, mark.rutland@....com,
jolsa@...nel.org, bp@...en8.de, kan.liang@...ux.intel.com,
adrian.hunter@...el.com, maddy@...ux.ibm.com, x86@...nel.org,
linux-perf-users@...r.kernel.org, linux-kernel@...r.kernel.org,
sandipan.das@....com, ananth.narayan@....com,
santosh.shukla@....com, kvmarm@...ts.linux.dev
Subject: Re: [PATCH v4 3/4] perf/core: Remove pmu linear searching code
On Thu, Jun 01, 2023 at 01:18:56PM +0200, Peter Zijlstra wrote:
> On Thu, Jun 01, 2023 at 01:02:30PM +0200, Thomas Richter wrote:
>
> > PS: I have the feeling, it gets complicated to have multiple hardware PMUs
> > per platform.
>
> Recently someone was poking around giving the pmu device a parent, this
> would, I think, result in sysfs links, which could be used to decide
> what's what.
>
> Core pmus would have the CPU device as their parent, while memory
> controller thingies would link to the relevant node or something.
>
> Ofc. all that's future-work/pending.
https://lkml.kernel.org/r/20230404134225.13408-1-Jonathan.Cameron%40huawei.com
https://lkml.kernel.org/r/20230531110011.13963-2-Jonathan.Cameron%40huawei.com
Powered by blists - more mailing lists