kmemleak: allow to coexist with fault injection

kmemleak does one slab allocation per user allocation.  So if slab fault
injection is enabled to any degree, kmemleak instantly fails to allocate
and turns itself off.  However, it's useful to use kmemleak with fault
injection to find leaks on error paths.  On the other hand, checking
kmemleak itself is not so useful because (1) it's a debugging tool and
(2) it has a very regular allocation pattern (basically a single
allocation site, so it either works or not).

Turn off fault injection for kmemleak allocations.

Link: http://lkml.kernel.org/r/20180109192243.19316-1-dvyukov@google.com
Signed-off-by: Dmitry Vyukov <dvyukov@google.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

diff --git a/mm/kmemleak.c b/mm/kmemleak.c
index d73c14294f3a61c2385741b447aa31d203a2bc72..f656ca27f6c20596322ce2aea248b4b04085bb29 100644 (file)
--- a/mm/kmemleak.c
+++ b/mm/kmemleak.c
@@ -127,7 +127,7 @@
 /* GFP bitmask for kmemleak internal allocations */
 #define gfp_kmemleak_mask(gfp) (((gfp) & (GFP_KERNEL | GFP_ATOMIC)) | \
                                 __GFP_NORETRY | __GFP_NOMEMALLOC | \
-                                __GFP_NOWARN)
+                                __GFP_NOWARN | __GFP_NOFAIL)
 
 /* scanning area inside a memory block */
 struct kmemleak_scan_area {