Re: [PATCH 1/2] slab: __GFP_ZERO is incompatible with a constructor

[Michal Hocko]

On Tue 10-04-18 05:53:50, Matthew Wilcox wrote:
> From: Matthew Wilcox <mawilcox@xxxxxxxxxxxxx>
> 
> __GFP_ZERO requests that the object be initialised to all-zeroes,
> while the purpose of a constructor is to initialise an object to a
> particular pattern.  We cannot do both.  Add a warning to catch any
> users who mistakenly pass a __GFP_ZERO flag when allocating a slab with
> a constructor.
> 
> Fixes: d07dbea46405 ("Slab allocators: support __GFP_ZERO in all allocators")
> Signed-off-by: Matthew Wilcox <mawilcox@xxxxxxxxxxxxx>
> Cc: stable@xxxxxxxxxxxxxxx
> ---
>  mm/slab.c | 6 ++++--
>  mm/slob.c | 4 +++-
>  mm/slub.c | 6 ++++--
>  3 files changed, 11 insertions(+), 5 deletions(-)
> 
> diff --git a/mm/slab.c b/mm/slab.c
> index 38d3f4fd17d7..8b2cb7db85db 100644
> --- a/mm/slab.c
> +++ b/mm/slab.c
> @@ -3313,8 +3313,10 @@ slab_alloc_node(struct kmem_cache *cachep, gfp_t flags, int nodeid,
>  	local_irq_restore(save_flags);
>  	ptr = cache_alloc_debugcheck_after(cachep, flags, ptr, caller);
>  
> -	if (unlikely(flags & __GFP_ZERO) && ptr)
> -		memset(ptr, 0, cachep->object_size);
> +	if (unlikely(flags & __GFP_ZERO) && ptr) {
> +		if (!WARN_ON_ONCE(cachep->ctor))
> +			memset(ptr, 0, cachep->object_size);
> +	}
>  
>  	slab_post_alloc_hook(cachep, flags, 1, &ptr);
>  	return ptr;

Why don't we need to cover this in slab_alloc and kmem_cache_alloc_bulk as well?

Other than that this patch makes sense to me.
-- 
Michal Hocko
SUSE Labs