[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20180711154758.4k2phfrzl2r5m34o@lakrids.cambridge.arm.com>
Date: Wed, 11 Jul 2018 16:47:58 +0100
From: Mark Rutland <mark.rutland@....com>
To: Ard Biesheuvel <ard.biesheuvel@...aro.org>,
Dave Martin <Dave.Martin@....com>
Cc: "Yandong.Zhao" <yandong77520@...il.com>,
linux-arm-kernel <linux-arm-kernel@...ts.infradead.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Will Deacon <will.deacon@....com>,
Catalin Marinas <catalin.marinas@....com>,
zhaoyd@...ndersoft.com, zhaoxb@...ndersoft.com,
fanlc0801@...ndersoft.com
Subject: Re: [PATCH] arm64: neon: Fix function may_use_simd() return error
status
On Wed, Jul 11, 2018 at 09:20:03AM +0200, Ard Biesheuvel wrote:
> On 11 July 2018 at 03:09, Yandong.Zhao <yandong77520@...il.com> wrote:
> > From: Yandong Zhao <yandong77520@...il.com>
> >
> > It does not matter if the caller of may_use_simd() migrates to
> > another cpu after the call, but it is still important that the
> > kernel_neon_busy percpu instance that is read matches the cpu the
> > task is running on at the time of the read.
> >
> > This means that raw_cpu_read() is not sufficient. kernel_neon_busy
> > may appear true if the caller migrates during the execution of
> > raw_cpu_read() and the next task to be scheduled in on the initial
> > cpu calls kernel_neon_begin().
> >
> > This patch replaces raw_cpu_read() with this_cpu_read() to protect
> > against this race.
> >
> > Signed-off-by: Yandong Zhao <yandong77520@...il.com>
>
> I had a bit of trouble disentangling the per-cpu spaghetti to decide
> whether this may trigger warnings when CONFIG_DEBUG_PREEMPT=y, but I
> don't think so. So assuming this is *not* the case:
It shouldn't, since:
* this_cpu_*() are prempt-safe
* __this_cpu_*() are not preempt-safe (and warn when preemptible)
* raw_cpu_*() are not preempt safe (but don't warn when preemptible)
> Acked-by: Ard Biesheuvel <ard.biesheuvel@...aro.org>
>
>
> > ---
> > arch/arm64/include/asm/simd.h | 5 +++--
> > 1 file changed, 3 insertions(+), 2 deletions(-)
> >
> > diff --git a/arch/arm64/include/asm/simd.h b/arch/arm64/include/asm/simd.h
> > index fa8b3fe..784a8c2 100644
> > --- a/arch/arm64/include/asm/simd.h
> > +++ b/arch/arm64/include/asm/simd.h
> > @@ -29,7 +29,8 @@
> > static __must_check inline bool may_use_simd(void)
> > {
> > /*
> > - * The raw_cpu_read() is racy if called with preemption enabled.
> > + * The this_cpu_read() is racy if called with preemption enabled,
> > + * since the task may subsequently migrate to another CPU.
> > * This is not a bug: kernel_neon_busy is only set when
> > * preemption is disabled, so we cannot migrate to another CPU
> > * while it is set, nor can we migrate to a CPU where it is set.
It would be nice if we could clarify the "is racy" part here.
How about:
/*
* kernel_neon_busy is only set while preemption is disabled,
* and is clear whenever preemption is enabled. Since
* this_cpu_read() is atomic w.r.t. preemption, kernel_neon_busy
* cannot change under our feet -- if it's set we cannot be
* migrated, and if it's clear we cannot be migrated to a CPU
* where it is set.
*/
With that:
Reviewed-by: Mark Rutland <mark.rutland@.....com>
Thanks,
Mark.
> > @@ -42,7 +43,7 @@ static __must_check inline bool may_use_simd(void)
> > * false.
> > */
> > return !in_irq() && !irqs_disabled() && !in_nmi() &&
> > - !raw_cpu_read(kernel_neon_busy);
> > + !this_cpu_read(kernel_neon_busy);
> > }
> >
> > #else /* ! CONFIG_KERNEL_MODE_NEON */
> > --
> > 1.9.1
> >
Powered by blists - more mailing lists