[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <3ce4fc14-bd43-47e9-aa37-59bb3cd5d051@sirena.org.uk>
Date: Thu, 20 Jul 2023 13:27:10 +0100
From: Mark Brown <broonie@...nel.org>
To: Will Deacon <will@...nel.org>
Cc: Catalin Marinas <catalin.marinas@....com>,
Shuah Khan <shuah@...nel.org>,
David Spickett <David.Spickett@....com>,
linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
linux-kselftest@...r.kernel.org, stable@...r.kernel.org
Subject: Re: [PATCH 1/3] arm64/fpsimd: Ensure SME storage is allocated after
SVE VL changes
On Thu, Jul 20, 2023 at 11:52:36AM +0100, Will Deacon wrote:
> On Thu, Jul 13, 2023 at 09:06:04PM +0100, Mark Brown wrote:
> > Since the ABI does not specify that changing the SVE vector length disturbs
> > SME state, and since SVE code may not be aware of SME code in the process,
> > we shouldn't simply discard any ZA state. Instead immediately reallocate
> > the storage for SVE if SME is active, and disable SME if we change the SVE
> > vector length while there is no SME state active.
> What is the advantage of keep the old behaviour in this case? In other
> words, if it's acceptable to reallocate the state when SME is active, why
> not just reallocate in all cases?
It was minimising the changes to the status quo given how attached
people often are to these things.
Download attachment "signature.asc" of type "application/pgp-signature" (489 bytes)
Powered by blists - more mailing lists