On 02/06/15 01:09, Leonid Yegoshin wrote:
> Repleace smp_mb() in arch_write_unlock() and __clear_bit_unlock() to

Replace.

> smp_mb__before_llsc() call which does "release" barrier functionality.
> 
> It seems like it was missed in commit f252ffd50c97dae87b45f1dbad24f71358ccfbd6
> during introduction of "acquire" and "release" semantics.
> 
> Signed-off-by: Leonid Yegoshin <Leonid.Yegoshin@imgtec.com>
> ---
>  arch/mips/include/asm/bitops.h   |    2 +-
>  arch/mips/include/asm/spinlock.h |    2 +-
>  2 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/mips/include/asm/bitops.h b/arch/mips/include/asm/bitops.h
> index 0cf29bd5dc5c..ce9666cf1499 100644
> --- a/arch/mips/include/asm/bitops.h
> +++ b/arch/mips/include/asm/bitops.h
> @@ -469,7 +469,7 @@ static inline int test_and_change_bit(unsigned long nr,
>   */
>  static inline void __clear_bit_unlock(unsigned long nr, volatile unsigned long *addr)
>  {
> -	smp_mb();
> +	smp_mb__before_llsc();
>  	__clear_bit(nr, addr);
>  }
>  
> diff --git a/arch/mips/include/asm/spinlock.h b/arch/mips/include/asm/spinlock.h
> index 1fca2e0793dc..7c7f3b2bd3de 100644
> --- a/arch/mips/include/asm/spinlock.h
> +++ b/arch/mips/include/asm/spinlock.h
> @@ -317,7 +317,7 @@ static inline void arch_write_lock(arch_rwlock_t *rw)
>  
>  static inline void arch_write_unlock(arch_rwlock_t *rw)
>  {
> -	smp_mb();
> +	smp_mb__before_llsc();

arch_write_unlock appears to just use sw, not sc, and __clear_bit
appears to be implemented in plain C, so is smp_mb__before_llsc() really
appropriate? Would smp_release() be more accurate/correct in both cases?

Cheers
James

>  
>  	__asm__ __volatile__(
>  	"				# arch_write_unlock	\n"
> 
>