[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <69afb239-da54-452d-8ab4-2d80dbdf8dce@sirena.org.uk>
Date: Wed, 2 Jul 2025 12:16:16 +0100
From: Mark Brown <broonie@...nel.org>
To: kan.liang@...ux.intel.com
Cc: peterz@...radead.org, mingo@...hat.com, acme@...nel.org,
namhyung@...nel.org, tglx@...utronix.de,
dave.hansen@...ux.intel.com, irogers@...gle.com,
adrian.hunter@...el.com, jolsa@...nel.org,
alexander.shishkin@...ux.intel.com, linux-kernel@...r.kernel.org,
dapeng1.mi@...ux.intel.com, ak@...ux.intel.com, zide.chen@...el.com,
mark.rutland@....com, ravi.bangoria@....com
Subject: Re: [RFC PATCH V2 06/13] perf: Support SIMD registers
On Thu, Jun 26, 2025 at 12:56:03PM -0700, kan.liang@...ux.intel.com wrote:
> * { u64 abi; # enum perf_sample_regs_abi
> - * u64 regs[weight(mask)]; } && PERF_SAMPLE_REGS_USER
> + * u64 regs[weight(mask)];
> + * struct {
> + * u16 nr_vectors;
> + * u16 vector_qwords;
> + * u16 nr_pred;
> + * u16 pred_qwords;
> + * u64 data[nr_vectors * vector_qwords + nr_pred * pred_qwords];
> + * } && (abi & PERF_SAMPLE_REGS_ABI_SIMD)
> + * } && PERF_SAMPLE_REGS_USER
I'm not super familiar with perf but I think this should work for arm64,
it supplies the vector length through the _qwords and we can handle FFR
being optional by varying the number of predicate registers.
Download attachment "signature.asc" of type "application/pgp-signature" (489 bytes)
Powered by blists - more mailing lists