Hello Dongli,
On Thu, Aug 13, 2020 at 11:57:51PM -0700, Dongli Zhang wrote:
On 8/11/20 5:46 AM, Eugeniu Rosca wrote:
Commit 52f23478081ae0 ("mm/slub.c: fix corrupted freechain in deactivate_slab()") suffered an update when picked up from LKML [1].
Specifically, relocating 'freelist = NULL' into 'freelist_corrupted()' created a no-op statement. Fix it by sticking to the behavior intended in the original patch [1]. Prefer the lowest-line-count solution.
[1] https://lore.kernel.org/linux-mm/20200331031450.12182-1-dongli.zhang@oracle....
Fixes: 52f23478081ae0 ("mm/slub.c: fix corrupted freechain in deactivate_slab()") Cc: Andrew Morton akpm@linux-foundation.org Cc: Dongli Zhang dongli.zhang@oracle.com Cc: stable@vger.kernel.org Signed-off-by: Eugeniu Rosca erosca@de.adit-jv.com
mm/slub.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/mm/slub.c b/mm/slub.c index 68c02b2eecd9..9a3e963b02a3 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -677,7 +677,6 @@ static bool freelist_corrupted(struct kmem_cache *s, struct page *page, if ((s->flags & SLAB_CONSISTENCY_CHECKS) && !check_valid_pointer(s, page, nextfree)) { object_err(s, page, freelist, "Freechain corrupt");
freelist = NULL;@@ -2184,8 +2183,10 @@ static void deactivate_slab(struct kmem_cache *s, struct page *page, * 'freelist' is already corrupted. So isolate all objects * starting at 'freelist'. */
if (freelist_corrupted(s, page, freelist, nextfree))
if (freelist_corrupted(s, page, freelist, nextfree)) {freelist = NULL;This is good to me.
However, this would confuse people when CONFIG_SLUB_DEBUG is not defined.
While reading the source code, people may be curious why to reset freelist when CONFIG_SLUB_DEBUG is even not defined.
This is a fair point. To address it, the `freelist = NULL` assignment should be then moved into the body of freelist_corrupted(). If no concerns on that, I will soon push a v2 implementing this proposal.