linux-kernel - Re: [PATCH v2 5/6] slab: validate slab before using it in alloc_single_from

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <aMghx7WTJP9NHoNe@hyeyoo>
Date: Mon, 15 Sep 2025 23:25:11 +0900
From: Harry Yoo <harry.yoo@...cle.com>
To: Vlastimil Babka <vbabka@...e.cz>
Cc: "Matthew Wilcox (Oracle)" <willy@...radead.org>,
        Christoph Lameter <cl@...two.org>,
        David Rientjes <rientjes@...gle.com>,
        Roman Gushchin <roman.gushchin@...ux.dev>,
        Andrew Morton <akpm@...ux-foundation.org>, linux-mm@...ck.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 5/6] slab: validate slab before using it in
 alloc_single_from_partial()

On Mon, Sep 15, 2025 at 03:55:12PM +0200, Vlastimil Babka wrote:
> We touch slab->freelist and slab->inuse before checking the slab pointer
> is actually sane. Do that validation first, which will be safer. We can
> thus also remove the check from alloc_debug_processing().
> 
> This adds a new "s->flags & SLAB_CONSISTENCY_CHECKS" test but
> alloc_single_from_partial() is only called for caches with debugging
> enabled so it's acceptable.
> 
> In alloc_single_from_new_slab() we just created the struct slab and call
> alloc_debug_processing() to mainly set up redzones, tracking etc, while
> not really expecting the consistency checks to fail. Thus don't validate
> it there.
> 
> Signed-off-by: Vlastimil Babka <vbabka@...e.cz>
> ---
>  mm/slub.c | 20 ++++++++++++--------
>  1 file changed, 12 insertions(+), 8 deletions(-)
> 
> diff --git a/mm/slub.c b/mm/slub.c
> index 12ad42f3d2e066b02340f2c30a85422583af3c5d..e5b53d1debddd3fe0f941f579a1043a5b976e50b 100644
> --- a/mm/slub.c
> +++ b/mm/slub.c
> @@ -821,6 +821,8 @@ static inline unsigned int get_orig_size(struct kmem_cache *s, void *object)
>  	return *(unsigned int *)p;
>  }
>  
> +#ifdef CONFIG_SLUB_DEBUG
> +
>  /*
>   * For debugging context when we want to check if the struct slab pointer
>   * appears to be valid.
> @@ -830,7 +832,6 @@ static inline bool validate_slab_ptr(struct slab *slab)
>  	return PageSlab(slab_page(slab));
>  }
>  
> -#ifdef CONFIG_SLUB_DEBUG
>  static unsigned long object_map[BITS_TO_LONGS(MAX_OBJS_PER_PAGE)];
>  static DEFINE_SPINLOCK(object_map_lock);
>  
> @@ -1651,11 +1652,6 @@ static noinline bool alloc_debug_processing(struct kmem_cache *s,
>  			struct slab *slab, void *object, int orig_size)
>  {
>  	if (s->flags & SLAB_CONSISTENCY_CHECKS) {
> -		if (!validate_slab_ptr(slab)) {
> -			slab_err(s, slab, "Not a valid slab page");
> -			return false;
> -		}
> -
>  		if (!alloc_consistency_checks(s, slab, object))
>  			goto bad;
>  	}
> @@ -2825,13 +2821,21 @@ static void *alloc_single_from_partial(struct kmem_cache *s,
>  
>  	lockdep_assert_held(&n->list_lock);
>  
> +#ifdef SLUB_DEBUG

I'm sure you meant CONFIG_SLUB_DEBUG ;)

With that adjusted, looks good to me,
Reviewed-by: Harry Yoo <harry.yoo@...cle.com>

> +	if (s->flags & SLAB_CONSISTENCY_CHECKS) {
> +		if (!validate_slab_ptr(slab)) {
> +			slab_err(s, slab, "Not a valid slab page");
> +			return NULL;
> +		}
> +	}
> +#endif
> +
>  	object = slab->freelist;
>  	slab->freelist = get_freepointer(s, object);
>  	slab->inuse++;
>  
>  	if (!alloc_debug_processing(s, slab, object, orig_size)) {
> -		if (validate_slab_ptr(slab))
> -			remove_partial(n, slab);
> +		remove_partial(n, slab);
>  		return NULL;
>  	}

-- 
Cheers,
Harry / Hyeonggon