]> git.baikalelectronics.ru Git - kernel.git/commitdiff
kasan: add memzero init for unaligned size at DEBUG
authorYee Lee <yee.lee@mediatek.com>
Thu, 15 Jul 2021 04:26:37 +0000 (21:26 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Thu, 15 Jul 2021 17:13:49 +0000 (10:13 -0700)
Issue: when SLUB debug is on, hwtag kasan_unpoison() would overwrite the
redzone of object with unaligned size.

An additional memzero_explicit() path is added to replacing init by hwtag
instruction for those unaligned size at SLUB debug mode.

The penalty is acceptable since they are only enabled in debug mode, not
production builds.  A block of comment is added for explanation.

Link: https://lkml.kernel.org/r/20210705103229.8505-3-yee.lee@mediatek.com
Signed-off-by: Yee Lee <yee.lee@mediatek.com>
Suggested-by: Andrey Konovalov <andreyknvl@gmail.com>
Suggested-by: Marco Elver <elver@google.com>
Reviewed-by: Marco Elver <elver@google.com>
Reviewed-by: Andrey Konovalov <andreyknvl@gmail.com>
Cc: Andrey Ryabinin <ryabinin.a.a@gmail.com>
Cc: Alexander Potapenko <glider@google.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Nicholas Tang <nicholas.tang@mediatek.com>
Cc: Kuan-Ying Lee <Kuan-Ying.Lee@mediatek.com>
Cc: Chinwen Chang <chinwen.chang@mediatek.com>
Cc: Matthew Wilcox <willy@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/kasan/kasan.h

index 98e3059bfea455685d252da6c516de27e3271269..d739cdd1621ab47e56cc599706f6f16fb413e8d6 100644 (file)
@@ -9,6 +9,7 @@
 #ifdef CONFIG_KASAN_HW_TAGS
 
 #include <linux/static_key.h>
+#include "../slab.h"
 
 DECLARE_STATIC_KEY_FALSE(kasan_flag_stacktrace);
 extern bool kasan_flag_async __ro_after_init;
@@ -387,6 +388,17 @@ static inline void kasan_unpoison(const void *addr, size_t size, bool init)
 
        if (WARN_ON((unsigned long)addr & KASAN_GRANULE_MASK))
                return;
+       /*
+        * Explicitly initialize the memory with the precise object size to
+        * avoid overwriting the SLAB redzone. This disables initialization in
+        * the arch code and may thus lead to performance penalty. The penalty
+        * is accepted since SLAB redzones aren't enabled in production builds.
+        */
+       if (__slub_debug_enabled() &&
+           init && ((unsigned long)size & KASAN_GRANULE_MASK)) {
+               init = false;
+               memzero_explicit((void *)addr, size);
+       }
        size = round_up(size, KASAN_GRANULE_SIZE);
 
        hw_set_mem_tag_range((void *)addr, size, tag, init);