[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <05bedc84-3672-975c-87ee-f094ca766ee8@suse.cz>
Date: Tue, 21 Apr 2020 16:15:33 +0200
From: Vlastimil Babka <vbabka@...e.cz>
To: Bernard Zhao <bernard@...o.com>, Christoph Lameter <cl@...ux.com>,
Pekka Enberg <penberg@...nel.org>,
David Rientjes <rientjes@...gle.com>,
Joonsoo Kim <iamjoonsoo.kim@....com>,
Andrew Morton <akpm@...ux-foundation.org>, linux-mm@...ck.org,
linux-kernel@...r.kernel.org
Cc: kernel@...o.com
Subject: Re: [PATCH] kmalloc_index optimization(add kmalloc max size check)
On 4/17/20 9:09 AM, Bernard Zhao wrote:
> kmalloc size should never exceed KMALLOC_MAX_SIZE.
> kmalloc_index realise if size is exceed KMALLOC_MAX_SIZE, e.g 64M,
> kmalloc_index just return index 26, but never check with OS`s max
> kmalloc config KMALLOC_MAX_SIZE. This index`s kmalloc caches maybe
> not create in function create_kmalloc_caches.
> We can throw an warninginfo in kmalloc at the beginning, instead of
> being guaranteed by the buddy alloc behind.
>
> Signed-off-by: Bernard Zhao <bernard@...o.com>
kmalloc_index() is only called from kmalloc() and kmalloc_node() for
compile-time constant sizes up to KMALLOC_MAX_CACHE_SIZE, which is smaller
(SLUB, SLOB) or equal (SLAB) than KMALLOC_MAX_SIZE. So your patch is effectively
a no-op and we better shouldn't tease the compiler too much, so that
kmalloc_index() stays fully compile-time evaluated.
> ---
> include/linux/slab.h | 4 ++++
> 1 file changed, 4 insertions(+)
>
> diff --git a/include/linux/slab.h b/include/linux/slab.h
> index 6d45488..59b60d2 100644
> --- a/include/linux/slab.h
> +++ b/include/linux/slab.h
> @@ -351,6 +351,10 @@ static __always_inline unsigned int kmalloc_index(size_t size)
> if (!size)
> return 0;
>
> + /* size should never exceed KMALLOC_MAX_SIZE. */
> + if (size > KMALLOC_MAX_SIZE)
> + WARN(1, "size exceed max kmalloc size!\n");
> +
> if (size <= KMALLOC_MIN_SIZE)
> return KMALLOC_SHIFT_LOW;
>
>
Powered by blists - more mailing lists