Re: [PATCH v4] skbuff: Proactively round up to kmalloc bucket size

[Vlastimil Babka]

On 10/22/22 01:49, Kees Cook wrote:
> Instead of discovering the kmalloc bucket size _after_ allocation, round
> up proactively so the allocation is explicitly made for the full size,
> allowing the compiler to correctly reason about the resulting size of
> the buffer through the existing __alloc_size() hint.
> 
> This will allow for kernels built with CONFIG_UBSAN_BOUNDS or the
> coming dynamic bounds checking under CONFIG_FORTIFY_SOURCE to gain
> back the __alloc_size() hints that were temporarily reverted in commit
> 93dd04ab0b2b ("slab: remove __alloc_size attribute from __kmalloc_track_caller")
> 
> Cc: "David S. Miller" <davem@xxxxxxxxxxxxx>
> Cc: Eric Dumazet <edumazet@xxxxxxxxxx>
> Cc: Jakub Kicinski <kuba@xxxxxxxxxx>
> Cc: Paolo Abeni <pabeni@xxxxxxxxxx>
> Cc: netdev@xxxxxxxxxxxxxxx
> Cc: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> Cc: Nick Desaulniers <ndesaulniers@xxxxxxxxxx>
> Cc: David Rientjes <rientjes@xxxxxxxxxx>
> Cc: Vlastimil Babka <vbabka@xxxxxxx>
> Signed-off-by: Kees Cook <keescook@xxxxxxxxxxxx>

Acked-by: Vlastimil Babka <vbabka@xxxxxxx>

Nit below:

> ---
> v4: use kmalloc_size_roundup() in callers, not kmalloc_reserve()
> v3: https://lore.kernel.org/lkml/20221018093005.give.246-kees@xxxxxxxxxx
> v2: https://lore.kernel.org/lkml/20220923202822.2667581-4-keescook@xxxxxxxxxxxx
> ---
>  net/core/skbuff.c | 50 +++++++++++++++++++++++------------------------
>  1 file changed, 25 insertions(+), 25 deletions(-)
> 
> diff --git a/net/core/skbuff.c b/net/core/skbuff.c
> index 651a82d30b09..77af430296e2 100644
> --- a/net/core/skbuff.c
> +++ b/net/core/skbuff.c
> @@ -508,14 +508,14 @@ struct sk_buff *__alloc_skb(unsigned int size, gfp_t gfp_mask,
>  	 */
>  	size = SKB_DATA_ALIGN(size);
>  	size += SKB_DATA_ALIGN(sizeof(struct skb_shared_info));
> -	data = kmalloc_reserve(size, gfp_mask, node, &pfmemalloc);
> +	osize = kmalloc_size_roundup(size);
> +	data = kmalloc_reserve(osize, gfp_mask, node, &pfmemalloc);
>  	if (unlikely(!data))
>  		goto nodata;
>  	/* kmalloc(size) might give us more room than requested.

The line above should now say kmalloc_size_roundup(size), or maybe could be
deleted completely now?

>  	 * Put skb_shared_info exactly at the end of allocated zone,
>  	 * to allow max possible filling before reallocation.
>  	 */
> -	osize = ksize(data);
>  	size = SKB_WITH_OVERHEAD(osize);
>  	prefetchw(data + size);
>