[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAMj1kXE8Yyn0orHe3DjFvCQ3k5mUKJF10BcJ3fJTessRoMShBA@mail.gmail.com>
Date: Fri, 23 Jul 2021 16:12:49 +0200
From: Ard Biesheuvel <ardb@...nel.org>
To: Valentin Schneider <valentin.schneider@....com>
Cc: Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Linux ARM <linux-arm-kernel@...ts.infradead.org>,
linux-rt-users@...r.kernel.org, X86 ML <x86@...nel.org>,
Catalin Marinas <catalin.marinas@....com>,
Will Deacon <will@...nel.org>,
Thomas Gleixner <tglx@...utronix.de>,
Steven Rostedt <rostedt@...dmis.org>,
Daniel Bristot de Oliveira <bristot@...hat.com>,
"Peter Zijlstra (Intel)" <peterz@...radead.org>,
Ingo Molnar <mingo@...nel.org>, Borislav Petkov <bp@...en8.de>,
"H. Peter Anvin" <hpa@...or.com>, Mark Brown <broonie@...nel.org>,
Dave Martin <Dave.Martin@....com>
Subject: Re: [PATCH 0/3] sched, x86, arm64: PREEMPT_RT, FPU and preemption
On Thu, 22 Jul 2021 at 19:53, Valentin Schneider
<valentin.schneider@....com> wrote:
>
> Hi folks,
>
> This stems from some more v5.13-rt1 breakage on arm64. This time per-CPU access
> safety isn't sufficient, we really need to keep preemption disabled.
>
> In a happy accident I stumbled on
>
> cba08c5dc6dc ("x86/fpu: Make kernel FPU protection RT friendly")
>
> so I packaged what was done there into some common helpers and plastered them
> over the problematic areas.
>
> Cheers,
> Valentin
>
> Valentin Schneider (3):
> sched/preempt: Introduce preempt_{enable, disable}_bh()
> x86/fpu: Make FPU protection reuse common helper
> arm64/fpsimd: Fix FPSIMD context handling vs PREEMPT_RT
>
Thanks for fixing this.
Reviewed-by: Ard Biesheuvel <ardb@...nel.org>
> arch/arm64/kernel/fpsimd.c | 6 ++++--
> arch/x86/include/asm/fpu/api.h | 19 ++-----------------
> include/linux/bottom_half.h | 26 ++++++++++++++++++++++++++
> 3 files changed, 32 insertions(+), 19 deletions(-)
>
> --
> 2.25.1
>
Powered by blists - more mailing lists