[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20200110115132.GA8786@arrakis.emea.arm.com>
Date: Fri, 10 Jan 2020 11:51:32 +0000
From: Catalin Marinas <catalin.marinas@....com>
To: Suzuki K Poulose <suzuki.poulose@....com>
Cc: linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
will@...nel.org, maz@...nel.org, mark.rutland@....com,
dave.martin@....com, ard.biesheuvel@...aro.org,
christoffer.dall@....com, Will Deacon <will.deacon@....com>
Subject: Re: [PATCH v2 2/7] arm64: fpsimd: Make sure SVE setup is complete
before SIMD is used
On Tue, Dec 17, 2019 at 06:33:57PM +0000, Suzuki K Poulose wrote:
> In-kernel users of NEON rely on may_use_simd() to check if the SIMD
> can be used. However, we must initialize the SVE before SIMD can
> be used. Add a sanity check to make sure that we have completed the
> SVE setup before anyone uses the SIMD.
>
> Cc: Ard Biesheuvel <ard.biesheuvel@...aro.org>
> Cc: Mark Rutland <mark.rutland@....com>
> Cc: Will Deacon <will.deacon@....com>
> Cc: Catalin Marinas <catalin.marinas@....com>
> Signed-off-by: Suzuki K Poulose <suzuki.poulose@....com>
> ---
> Discussion here : https://lkml.kernel.org/r/20191014145204.GS27757@arm.com
Re-reading this thread, I think the conclusion was more towards having a
WARN_ON in system_supports_fpsimd() (or may_use_simd()). We don't expect
code to start using neon before the SMP is initialised (other than
early_initcall(), the rest run after the secondary CPUs are brought up).
--
Catalin
Powered by blists - more mailing lists