Re: [PATCH v2 2/2] mm/vmalloc: Use __this_cpu_try_cmpxchg() in preload_this_cpu_lock()

[Uladzislau Rezki]

On Tue, May 28, 2024 at 04:43:14PM +0200, Uros Bizjak wrote:
> Use __this_cpu_try_cmpxchg() instead of
> __this_cpu_cmpxchg (*ptr, old, new) == old in
> preload_this_cpu_lock().  x86 CMPXCHG instruction returns
> success in ZF flag, so this change saves a compare after cmpxchg.
> 
> The generated code improves from:
> 
>     4bb6:	48 85 f6             	test   %rsi,%rsi
>     4bb9:	0f 84 10 fa ff ff    	je     45cf <...>
>     4bbf:	4c 89 e8             	mov    %r13,%rax
>     4bc2:	65 48 0f b1 35 00 00 	cmpxchg %rsi,%gs:0x0(%rip)
>     4bc9:	00 00
>     4bcb:	48 85 c0             	test   %rax,%rax
>     4bce:	0f 84 fb f9 ff ff    	je     45cf <...>
> 
> to:
> 
>     4bb6:	48 85 f6             	test   %rsi,%rsi
>     4bb9:	0f 84 10 fa ff ff    	je     45cf <...>
>     4bbf:	4c 89 e8             	mov    %r13,%rax
>     4bc2:	65 48 0f b1 35 00 00 	cmpxchg %rsi,%gs:0x0(%rip)
>     4bc9:	00 00
>     4bcb:	0f 84 fe f9 ff ff    	je     45cf <...>
> 
> No functional change intended.
> 
> Signed-off-by: Uros Bizjak <ubizjak@xxxxxxxxx>
> Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
> Cc: Uladzislau Rezki <urezki@xxxxxxxxx>
> Cc: Christoph Hellwig <hch@xxxxxxxxxxxxx>
> Cc: Lorenzo Stoakes <lstoakes@xxxxxxxxx>
> Cc: Dennis Zhou <dennis@xxxxxxxxxx>
> Cc: Tejun Heo <tj@xxxxxxxxxx>
> Cc: Christoph Lameter <cl@xxxxxxxxx>
> ---
> v2: Show generated code improvement in the commit message.
> ---
>  mm/vmalloc.c | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/mm/vmalloc.c b/mm/vmalloc.c
> index 5d3aa2dc88a8..4f34d935d648 100644
> --- a/mm/vmalloc.c
> +++ b/mm/vmalloc.c
> @@ -1816,7 +1816,7 @@ static void free_vmap_area(struct vmap_area *va)
>  static inline void
>  preload_this_cpu_lock(spinlock_t *lock, gfp_t gfp_mask, int node)
>  {
> -	struct vmap_area *va = NULL;
> +	struct vmap_area *va = NULL, *tmp;
>  
>  	/*
>  	 * Preload this CPU with one extra vmap_area object. It is used
> @@ -1832,7 +1832,8 @@ preload_this_cpu_lock(spinlock_t *lock, gfp_t gfp_mask, int node)
>  
>  	spin_lock(lock);
>  
> -	if (va && __this_cpu_cmpxchg(ne_fit_preload_node, NULL, va))
> +	tmp = NULL;
> +	if (va && !__this_cpu_try_cmpxchg(ne_fit_preload_node, &tmp, va))
>  		kmem_cache_free(vmap_area_cachep, va);
>  }
>  
> -- 
> 2.42.0
> 
Reviewed-by: Uladzislau Rezki (Sony) <urezki@xxxxxxxxx>

Thanks!

--
Uladzislau Rezki