* [PATCH] arm64: neon: Fix function may_use_simd() return error status
@ 2018-07-12 3:29 Yandong.Zhao
2018-07-12 9:56 ` Dave Martin
0 siblings, 1 reply; 13+ messages in thread
From: Yandong.Zhao @ 2018-07-12 3:29 UTC (permalink / raw)
To: Dave.Martin
Cc: linux-arm-kernel, linux-kernel, ard.biesheuvel, will.deacon,
catalin.marinas, zhaoyd, zhaoxb, fanlc0801, Yandong Zhao
From: Yandong Zhao <yandong77520@gmail.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.
Fixes: cb84d11e1625 ("arm64: neon: Remove support for nested or hardirq kernel-mode NEON")
Acked-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Dave Martin <Dave.Martin@arm.com>
Reviewed-by: Mark Rutland <mark.rutland@arm..com>
Signed-off-by: Yandong Zhao <yandong77520@gmail.com>
---
arch/arm64/include/asm/simd.h | 19 +++++++------------
1 file changed, 7 insertions(+), 12 deletions(-)
diff --git a/arch/arm64/include/asm/simd.h b/arch/arm64/include/asm/simd.h
index fa8b3fe..6495cc5 100644
--- a/arch/arm64/include/asm/simd.h
+++ b/arch/arm64/include/asm/simd.h
@@ -29,20 +29,15 @@
static __must_check inline bool may_use_simd(void)
{
/*
- * The raw_cpu_read() is racy if called with preemption enabled.
- * 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.
- * So, if we find it clear on some CPU then we're guaranteed to
- * find it clear on any CPU we could migrate to.
- *
- * If we are in between kernel_neon_begin()...kernel_neon_end(),
- * the flag will be set, but preemption is also disabled, so we
- * can't migrate to another CPU and spuriously see it become
- * false.
+ * 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.
*/
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
^ permalink raw reply related [flat|nested] 13+ messages in thread
* Re: [PATCH] arm64: neon: Fix function may_use_simd() return error status
2018-07-12 3:29 [PATCH] arm64: neon: Fix function may_use_simd() return error status Yandong.Zhao
@ 2018-07-12 9:56 ` Dave Martin
0 siblings, 0 replies; 13+ messages in thread
From: Dave Martin @ 2018-07-12 9:56 UTC (permalink / raw)
To: Yandong.Zhao
Cc: zhaoxb, ard.biesheuvel, catalin.marinas, will.deacon,
linux-kernel, zhaoyd, linux-arm-kernel, fanlc0801
On Thu, Jul 12, 2018 at 11:29:38AM +0800, Yandong.Zhao wrote:
> From: Yandong Zhao <yandong77520@gmail.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.
>
> Fixes: cb84d11e1625 ("arm64: neon: Remove support for nested or hardirq kernel-mode NEON")
> Acked-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
> Reviewed-by: Dave Martin <Dave.Martin@arm.com>
> Reviewed-by: Mark Rutland <mark.rutland@arm..com>
> Signed-off-by: Yandong Zhao <yandong77520@gmail.com>
> ---
> arch/arm64/include/asm/simd.h | 19 +++++++------------
> 1 file changed, 7 insertions(+), 12 deletions(-)
>
> diff --git a/arch/arm64/include/asm/simd.h b/arch/arm64/include/asm/simd.h
> index fa8b3fe..6495cc5 100644
> --- a/arch/arm64/include/asm/simd.h
> +++ b/arch/arm64/include/asm/simd.h
> @@ -29,20 +29,15 @@
> static __must_check inline bool may_use_simd(void)
> {
> /*
> - * The raw_cpu_read() is racy if called with preemption enabled.
> - * 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.
> - * So, if we find it clear on some CPU then we're guaranteed to
> - * find it clear on any CPU we could migrate to.
> - *
> - * If we are in between kernel_neon_begin()...kernel_neon_end(),
> - * the flag will be set, but preemption is also disabled, so we
> - * can't migrate to another CPU and spuriously see it become
> - * false.
> + * 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.
> */
This new explanation looks fine to me.
[...]
Cheers
---Dave
^ permalink raw reply [flat|nested] 13+ messages in thread
* [PATCH] arm64: neon: Fix function may_use_simd() return error status
@ 2018-07-11 11:06 Yandong.Zhao
2018-07-11 12:05 ` Will Deacon
0 siblings, 1 reply; 13+ messages in thread
From: Yandong.Zhao @ 2018-07-11 11:06 UTC (permalink / raw)
To: Dave.Martin
Cc: linux-arm-kernel, linux-kernel, ard.biesheuvel, will.deacon,
catalin.marinas, zhaoyd, zhaoxb, fanlc0801, Yandong Zhao
From: Yandong Zhao <yandong77520@gmail.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.
Fixes: cb84d11e1625 ("arm64: neon: Remove support for nested or hardirq kernel-mode NEON")
Reviewed-by: Dave Martin <Dave.Martin@arm.com>
Signed-off-by: Yandong Zhao <yandong77520@gmail.com>
---
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.
@@ -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
^ permalink raw reply related [flat|nested] 13+ messages in thread
* Re: [PATCH] arm64: neon: Fix function may_use_simd() return error status
2018-07-11 11:06 Yandong.Zhao
@ 2018-07-11 12:05 ` Will Deacon
2018-07-11 12:58 ` Dave Martin
0 siblings, 1 reply; 13+ messages in thread
From: Will Deacon @ 2018-07-11 12:05 UTC (permalink / raw)
To: Yandong.Zhao
Cc: Dave.Martin, linux-arm-kernel, linux-kernel, ard.biesheuvel,
catalin.marinas, zhaoyd, zhaoxb, fanlc0801
On Wed, Jul 11, 2018 at 07:06:28PM +0800, Yandong.Zhao wrote:
> From: Yandong Zhao <yandong77520@gmail.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.
>
> Fixes: cb84d11e1625 ("arm64: neon: Remove support for nested or hardirq kernel-mode NEON")
> Reviewed-by: Dave Martin <Dave.Martin@arm.com>
> Signed-off-by: Yandong Zhao <yandong77520@gmail.com>
> ---
> arch/arm64/include/asm/simd.h | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
Does this need to go to stable?
Will
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [PATCH] arm64: neon: Fix function may_use_simd() return error status
2018-07-11 12:05 ` Will Deacon
@ 2018-07-11 12:58 ` Dave Martin
0 siblings, 0 replies; 13+ messages in thread
From: Dave Martin @ 2018-07-11 12:58 UTC (permalink / raw)
To: Will Deacon
Cc: Yandong.Zhao, zhaoxb, ard.biesheuvel, catalin.marinas,
linux-kernel, zhaoyd, linux-arm-kernel, fanlc0801
On Wed, Jul 11, 2018 at 01:05:19PM +0100, Will Deacon wrote:
> On Wed, Jul 11, 2018 at 07:06:28PM +0800, Yandong.Zhao wrote:
> > From: Yandong Zhao <yandong77520@gmail.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.
> >
> > Fixes: cb84d11e1625 ("arm64: neon: Remove support for nested or hardirq kernel-mode NEON")
> > Reviewed-by: Dave Martin <Dave.Martin@arm.com>
> > Signed-off-by: Yandong Zhao <yandong77520@gmail.com>
> > ---
> > arch/arm64/include/asm/simd.h | 5 +++--
> > 1 file changed, 3 insertions(+), 2 deletions(-)
>
> Does this need to go to stable?
It should, yes, so we should probably add
Cc: stable@vger.kernel.org
Are you OK to pick that up?
Cheers
---Dave
^ permalink raw reply [flat|nested] 13+ messages in thread
* [PATCH] arm64: neon: Fix function may_use_simd() return error status
@ 2018-07-11 1:09 Yandong.Zhao
2018-07-11 7:20 ` Ard Biesheuvel
2018-07-11 10:55 ` Dave Martin
0 siblings, 2 replies; 13+ messages in thread
From: Yandong.Zhao @ 2018-07-11 1:09 UTC (permalink / raw)
To: Dave.Martin
Cc: linux-arm-kernel, linux-kernel, ard.biesheuvel, will.deacon,
catalin.marinas, zhaoyd, zhaoxb, fanlc0801, Yandong Zhao
From: Yandong Zhao <yandong77520@gmail.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@gmail.com>
---
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.
@@ -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
^ permalink raw reply related [flat|nested] 13+ messages in thread
* Re: [PATCH] arm64: neon: Fix function may_use_simd() return error status
2018-07-11 1:09 Yandong.Zhao
@ 2018-07-11 7:20 ` Ard Biesheuvel
2018-07-11 15:47 ` Mark Rutland
2018-07-11 10:55 ` Dave Martin
1 sibling, 1 reply; 13+ messages in thread
From: Ard Biesheuvel @ 2018-07-11 7:20 UTC (permalink / raw)
To: Yandong.Zhao, Dave Martin
Cc: linux-arm-kernel, Linux Kernel Mailing List, Will Deacon,
Catalin Marinas, zhaoyd, zhaoxb, fanlc0801
On 11 July 2018 at 03:09, Yandong.Zhao <yandong77520@gmail.com> wrote:
> From: Yandong Zhao <yandong77520@gmail.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@gmail.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:
Acked-by: Ard Biesheuvel <ard.biesheuvel@linaro.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.
> @@ -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
>
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [PATCH] arm64: neon: Fix function may_use_simd() return error status
2018-07-11 7:20 ` Ard Biesheuvel
@ 2018-07-11 15:47 ` Mark Rutland
2018-07-11 16:03 ` Will Deacon
0 siblings, 1 reply; 13+ messages in thread
From: Mark Rutland @ 2018-07-11 15:47 UTC (permalink / raw)
To: Ard Biesheuvel, Dave Martin
Cc: Yandong.Zhao, linux-arm-kernel, Linux Kernel Mailing List,
Will Deacon, Catalin Marinas, zhaoyd, zhaoxb, fanlc0801
On Wed, Jul 11, 2018 at 09:20:03AM +0200, Ard Biesheuvel wrote:
> On 11 July 2018 at 03:09, Yandong.Zhao <yandong77520@gmail.com> wrote:
> > From: Yandong Zhao <yandong77520@gmail.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@gmail.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@linaro.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@arm..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
> >
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [PATCH] arm64: neon: Fix function may_use_simd() return error status
2018-07-11 15:47 ` Mark Rutland
@ 2018-07-11 16:03 ` Will Deacon
2018-07-11 16:07 ` Mark Rutland
0 siblings, 1 reply; 13+ messages in thread
From: Will Deacon @ 2018-07-11 16:03 UTC (permalink / raw)
To: Mark Rutland
Cc: Ard Biesheuvel, Dave Martin, Yandong.Zhao, linux-arm-kernel,
Linux Kernel Mailing List, Catalin Marinas, zhaoyd, zhaoxb,
fanlc0801
On Wed, Jul 11, 2018 at 04:47:58PM +0100, Mark Rutland wrote:
> On Wed, Jul 11, 2018 at 09:20:03AM +0200, Ard Biesheuvel wrote:
> > On 11 July 2018 at 03:09, Yandong.Zhao <yandong77520@gmail.com> wrote:
> > > 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@arm..com>
Thanks. Applied with the updated comment and your tag..
Will
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [PATCH] arm64: neon: Fix function may_use_simd() return error status
2018-07-11 16:03 ` Will Deacon
@ 2018-07-11 16:07 ` Mark Rutland
0 siblings, 0 replies; 13+ messages in thread
From: Mark Rutland @ 2018-07-11 16:07 UTC (permalink / raw)
To: Will Deacon
Cc: Ard Biesheuvel, Dave Martin, Yandong.Zhao, linux-arm-kernel,
Linux Kernel Mailing List, Catalin Marinas, zhaoyd, zhaoxb,
fanlc0801
On Wed, Jul 11, 2018 at 05:03:15PM +0100, Will Deacon wrote:
> On Wed, Jul 11, 2018 at 04:47:58PM +0100, Mark Rutland wrote:
> > On Wed, Jul 11, 2018 at 09:20:03AM +0200, Ard Biesheuvel wrote:
> > > On 11 July 2018 at 03:09, Yandong.Zhao <yandong77520@gmail.com> wrote:
> > > > 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@arm..com>
>
> Thanks. Applied with the updated comment and your tag..
Cheer!
Mar.
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [PATCH] arm64: neon: Fix function may_use_simd() return error status
2018-07-11 1:09 Yandong.Zhao
2018-07-11 7:20 ` Ard Biesheuvel
@ 2018-07-11 10:55 ` Dave Martin
1 sibling, 0 replies; 13+ messages in thread
From: Dave Martin @ 2018-07-11 10:55 UTC (permalink / raw)
To: Yandong.Zhao
Cc: zhaoxb, ard.biesheuvel, catalin.marinas, will.deacon,
linux-kernel, zhaoyd, linux-arm-kernel, fanlc0801
On Wed, Jul 11, 2018 at 09:09:59AM +0800, Yandong.Zhao wrote:
> From: Yandong Zhao <yandong77520@gmail.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@gmail.com>
Looks ok to me. You can add the following tags:
Fixes: cb84d11e1625 ("arm64: neon: Remove support for nested or hardirq kernel-mode NEON")
Reviewed-by: Dave Martin <Dave.Martin@arm.com>
Cheers
---Dave
> ---
> 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.
> @@ -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
>
>
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply [flat|nested] 13+ messages in thread
* [PATCH] arm64: neon: Fix function may_use_simd() return error status
@ 2018-07-10 2:21 Yandong.Zhao
2018-07-10 13:11 ` Dave Martin
0 siblings, 1 reply; 13+ messages in thread
From: Yandong.Zhao @ 2018-07-10 2:21 UTC (permalink / raw)
To: Dave.Martin
Cc: linux-arm-kernel, catalin.marinas, will.deacon, ard.biesheuvel,
linux-kernel, zhaoyd, zhaoxb, fanlc0801, Yandong Zhao
From: Yandong Zhao <yandong77520@gmail.com>
Operations for contexts where we do not want to do any checks for
preemptions. Unless strictly necessary, always use this_cpu_read()
instead. Because of the kernel_neon_busy here we have to make sure
that it is the current cpu.
Signed-off-by: Yandong Zhao <yandong77520@gmail.com>
---
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..8b97f8b 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 my 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.
@@ -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
^ permalink raw reply related [flat|nested] 13+ messages in thread
* Re: [PATCH] arm64: neon: Fix function may_use_simd() return error status
2018-07-10 2:21 Yandong.Zhao
@ 2018-07-10 13:11 ` Dave Martin
0 siblings, 0 replies; 13+ messages in thread
From: Dave Martin @ 2018-07-10 13:11 UTC (permalink / raw)
To: Yandong.Zhao
Cc: zhaoxb, ard.biesheuvel, catalin.marinas, will.deacon,
linux-kernel, zhaoyd, linux-arm-kernel, fanlc0801
On Tue, Jul 10, 2018 at 10:21:40AM +0800, Yandong.Zhao wrote:
> From: Yandong Zhao <yandong77520@gmail.com>
>
> Operations for contexts where we do not want to do any checks for
> preemptions. Unless strictly necessary, always use this_cpu_read()
> instead. Because of the kernel_neon_busy here we have to make sure
> that it is the current cpu.
I find this wording a bit confusing.
Does the following make look OK to you?
--8<--
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.
-->8--
>
> Signed-off-by: Yandong Zhao <yandong77520@gmail.com>
> ---
> 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..8b97f8b 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 my subsequently migrate to another CPU.
"my" -> "may"
(apologies if I was responsible for that typo)
[...]
Cheers
---Dave
^ permalink raw reply [flat|nested] 13+ messages in thread
end of thread, other threads:[~2018-07-12 9:56 UTC | newest]
Thread overview: 13+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-07-12 3:29 [PATCH] arm64: neon: Fix function may_use_simd() return error status Yandong.Zhao
2018-07-12 9:56 ` Dave Martin
-- strict thread matches above, loose matches on Subject: below --
2018-07-11 11:06 Yandong.Zhao
2018-07-11 12:05 ` Will Deacon
2018-07-11 12:58 ` Dave Martin
2018-07-11 1:09 Yandong.Zhao
2018-07-11 7:20 ` Ard Biesheuvel
2018-07-11 15:47 ` Mark Rutland
2018-07-11 16:03 ` Will Deacon
2018-07-11 16:07 ` Mark Rutland
2018-07-11 10:55 ` Dave Martin
2018-07-10 2:21 Yandong.Zhao
2018-07-10 13:11 ` Dave Martin
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).