[PATCH] x86/preempt: do not clear PREEMPT_NEED_RESCHED on preempt count reset

[PATCH] x86/preempt: do not clear PREEMPT_NEED_RESCHED on preempt count reset

[Martin Schwidefsky]

The per-cpu preempt count of x86 contains two values, the actual preempt
count and the inverted PREEMPT_NEED_RESCHED bit. If a corrupted preempt
count is detected the preempt_count_set function is used to reset the
preempt count.

In case the inverted PREEMPT_NEED_RESCHED bit is zero at the time of the
reset, the preemption indication is lost. Use raw_cpu_cmpxchg_4 to reset
only the count part and leave the PREEMPT_NEED_RESCHED bit as it is.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
---
 arch/x86/include/asm/preempt.h | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/arch/x86/include/asm/preempt.h b/arch/x86/include/asm/preempt.h
index 17f2186..ec1f3c6 100644
--- a/arch/x86/include/asm/preempt.h
+++ b/arch/x86/include/asm/preempt.h
@@ -24,7 +24,13 @@ static __always_inline int preempt_count(void)
 
 static __always_inline void preempt_count_set(int pc)
 {
-	raw_cpu_write_4(__preempt_count, pc);
+	int old, new;
+
+	do {
+		old = raw_cpu_read_4(__preempt_count);
+		new = (old & PREEMPT_NEED_RESCHED) |
+			(pc & ~PREEMPT_NEED_RESCHED);
+	} while (raw_cpu_cmpxchg_4(__preempt_count, old, new) != old);
 }
 
 /*
-- 
1.9.1

Re: [PATCH] x86/preempt: do not clear PREEMPT_NEED_RESCHED on preempt count reset

[Peter Zijlstra]

On Mon, Nov 07, 2016 at 02:01:00PM +0100, Martin Schwidefsky wrote:
> The per-cpu preempt count of x86 contains two values, the actual preempt
> count and the inverted PREEMPT_NEED_RESCHED bit. If a corrupted preempt
> count is detected the preempt_count_set function is used to reset the
> preempt count.
> 
> In case the inverted PREEMPT_NEED_RESCHED bit is zero at the time of the
> reset, the preemption indication is lost. Use raw_cpu_cmpxchg_4 to reset
> only the count part and leave the PREEMPT_NEED_RESCHED bit as it is.
> 
> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>

Thanks Martin!

I don't suppose this really hurts too much; if we hit a warn where this
restore is needed there are bigger problems, but given you've done the
patch already and it does improve consistency, I've taken it.

> ---
>  arch/x86/include/asm/preempt.h | 8 +++++++-
>  1 file changed, 7 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/x86/include/asm/preempt.h b/arch/x86/include/asm/preempt.h
> index 17f2186..ec1f3c6 100644
> --- a/arch/x86/include/asm/preempt.h
> +++ b/arch/x86/include/asm/preempt.h
> @@ -24,7 +24,13 @@ static __always_inline int preempt_count(void)
>  
>  static __always_inline void preempt_count_set(int pc)
>  {
> -	raw_cpu_write_4(__preempt_count, pc);
> +	int old, new;
> +
> +	do {
> +		old = raw_cpu_read_4(__preempt_count);
> +		new = (old & PREEMPT_NEED_RESCHED) |
> +			(pc & ~PREEMPT_NEED_RESCHED);
> +	} while (raw_cpu_cmpxchg_4(__preempt_count, old, new) != old);
>  }
>  
>  /*
> -- 
> 1.9.1
> 

[tip:sched/core] sched/x86: Do not clear PREEMPT_NEED_RESCHED on preempt count reset

[tip-bot for Martin Schwidefsky]

Commit-ID:  f285144f81e814f39342dbf5321d6ba939890b1b
Gitweb:     http://git.kernel.org/tip/f285144f81e814f39342dbf5321d6ba939890b1b
Author:     Martin Schwidefsky <schwidefsky@de.ibm.com>
AuthorDate: Mon, 7 Nov 2016 14:01:00 +0100
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Wed, 16 Nov 2016 10:29:04 +0100

sched/x86: Do not clear PREEMPT_NEED_RESCHED on preempt count reset

The per-cpu preempt count of x86 contains two values, the actual preempt
count and the inverted PREEMPT_NEED_RESCHED bit. If a corrupted preempt
count is detected the preempt_count_set() function is used to reset the
preempt count.

In case the inverted PREEMPT_NEED_RESCHED bit is zero at the time of the
reset, the preemption indication is lost. Use raw_cpu_cmpxchg_4() to reset
only the count part and leave the PREEMPT_NEED_RESCHED bit as it is.

This improves the kernel's behavior when it runs into preempt count leaks
and tries to fix them up.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1478523660-733-1-git-send-email-schwidefsky@de.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/include/asm/preempt.h | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/arch/x86/include/asm/preempt.h b/arch/x86/include/asm/preempt.h
index 17f2186..ec1f3c6 100644
--- a/arch/x86/include/asm/preempt.h
+++ b/arch/x86/include/asm/preempt.h
@@ -24,7 +24,13 @@ static __always_inline int preempt_count(void)
 
 static __always_inline void preempt_count_set(int pc)
 {
-	raw_cpu_write_4(__preempt_count, pc);
+	int old, new;
+
+	do {
+		old = raw_cpu_read_4(__preempt_count);
+		new = (old & PREEMPT_NEED_RESCHED) |
+			(pc & ~PREEMPT_NEED_RESCHED);
+	} while (raw_cpu_cmpxchg_4(__preempt_count, old, new) != old);
 }
 
 /*