Re: [PATCH] seqlock: Use WRITE_ONCE() when updating sequence

[Peter Zijlstra <peterz@infradead.org>]

On Tue, Dec 17, 2024 at 03:17:36PM -0800, Daniel Xu wrote:
> `sequence` is a concurrently accessed shared variable on the reader
> side. Therefore, it needs to be wrapped in WRITE_ONCE() in order to
> prevent unwanted compiler optimizations like store tearing.
> 
> Signed-off-by: Daniel Xu <dxu@dxuuu.xyz>
> ---
>  include/linux/seqlock.h | 14 +++++++-------
>  1 file changed, 7 insertions(+), 7 deletions(-)
> 
> diff --git a/include/linux/seqlock.h b/include/linux/seqlock.h
> index 5298765d6ca4..f4c6f2507742 100644
> --- a/include/linux/seqlock.h
> +++ b/include/linux/seqlock.h
> @@ -45,7 +45,7 @@ static inline void __seqcount_init(seqcount_t *s, const char *name,
>  	 * Make sure we are not reinitializing a held lock:
>  	 */
>  	lockdep_init_map(&s->dep_map, name, key, 0);
> -	s->sequence = 0;
> +	WRITE_ONCE(s->sequence, 0);
>  }
>  
>  #ifdef CONFIG_DEBUG_LOCK_ALLOC
> @@ -405,7 +405,7 @@ do {									\
>  static inline void do_raw_write_seqcount_begin(seqcount_t *s)
>  {
>  	kcsan_nestable_atomic_begin();
> -	s->sequence++;
> +	WRITE_ONCE(s->sequence, READ_ONCE(s->sequence) + 1);
>  	smp_wmb();
>  }

This results in significantly worse code-gen, it will change an inc to
memory with a load,inc,store.