mm: zbud: fix condition check on allocation size
authorHeesub Shin <heesub.shin@samsung.com>
Wed, 31 Jul 2013 20:53:40 +0000 (13:53 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Wed, 31 Jul 2013 21:41:03 +0000 (14:41 -0700)
zbud_alloc() incorrectly verifies the size of allocation limit.  It
should deny the allocation request greater than (PAGE_SIZE -
ZHDR_SIZE_ALIGNED - CHUNK_SIZE), not (PAGE_SIZE - ZHDR_SIZE_ALIGNED)
which has no remaining spaces for its buddy.  There is no point in
spending the entire zbud page storing only a single page, since we don't
have any benefits.

Signed-off-by: Heesub Shin <heesub.shin@samsung.com>
Acked-by: Seth Jennings <sjenning@linux.vnet.ibm.com>
Cc: Bob Liu <bob.liu@oracle.com>
Cc: Dongjun Shin <d.j.shin@samsung.com>
Cc: Sunae Seo <sunae.seo@samsung.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/zbud.c

index 9bb4710..ad1e781 100644 (file)
--- a/mm/zbud.c
+++ b/mm/zbud.c
@@ -257,7 +257,7 @@ int zbud_alloc(struct zbud_pool *pool, int size, gfp_t gfp,
 
        if (size <= 0 || gfp & __GFP_HIGHMEM)
                return -EINVAL;
-       if (size > PAGE_SIZE - ZHDR_SIZE_ALIGNED)
+       if (size > PAGE_SIZE - ZHDR_SIZE_ALIGNED - CHUNK_SIZE)
                return -ENOSPC;
        chunks = size_to_chunks(size);
        spin_lock(&pool->lock);