From: Abel Wu <wuyun.wu@xxxxxxxxxx>
Subject: mm/slub: make add_full() condition more explicit
The commit below is incomplete, as it didn't handle the add_full() part.
commit a4d3f8916c65 ("slub: remove useless kmem_cache_debug() before
remove_full()")
This patch checks for SLAB_STORE_USER instead of kmem_cache_debug(), since
that should be the only context in which we need the list_lock for
add_full().
Link: https://lkml.kernel.org/r/20200811020240.1231-1-wuyun.wu@xxxxxxxxxx
Signed-off-by: Abel Wu <wuyun.wu@xxxxxxxxxx>
Cc: Christoph Lameter <cl@xxxxxxxxx>
Cc: Pekka Enberg <penberg@xxxxxxxxxx>
Cc: David Rientjes <rientjes@xxxxxxxxxx>
Cc: Joonsoo Kim <iamjoonsoo.kim@xxxxxxx>
Cc: Liu Xiang <liu.xiang6@xxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
mm/slub.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
--- a/mm/slub.c~mm-slub-make-add_full-condition-more-explicit
+++ a/mm/slub.c
@@ -2245,7 +2245,8 @@ redo:
}
} else {
m = M_FULL;
- if (kmem_cache_debug(s) && !lock) {
+#ifdef CONFIG_SLUB_DEBUG
+ if ((s->flags & SLAB_STORE_USER) && !lock) {
lock = 1;
/*
* This also ensures that the scanning of full
@@ -2254,6 +2255,7 @@ redo:
*/
spin_lock(&n->list_lock);
}
+#endif
}
if (l != m) {
_
