On Sat, 19 Feb 2022 at 02:26, Peter Collingbourne pcc@google.com wrote:
This is a followup to commit f649dc0e0d7b ("kasan: fix unit tests with CONFIG_UBSAN_LOCAL_BOUNDS enabled") that fixes tests that fail as a result of __alloc_size annotations being added to the kernel allocator functions.
Link: https://linux-review.googlesource.com/id/I4334cafc5db600fda5cebb851b2ee9fd09... Signed-off-by: Peter Collingbourne pcc@google.com Cc: stable@vger.kernel.org # 5.16.x Fixes: c37495d6254c ("slab: add __alloc_size attributes for better bounds checking")
lib/test_kasan.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/lib/test_kasan.c b/lib/test_kasan.c index 26a5c9007653..3bf8801d0e66 100644 --- a/lib/test_kasan.c +++ b/lib/test_kasan.c @@ -177,7 +177,8 @@ static void kmalloc_node_oob_right(struct kunit *test) */ static void kmalloc_pagealloc_oob_right(struct kunit *test) {
char *ptr;
/* See comment in kasan_global_oob_right. */char *volatile ptr; size_t size = KMALLOC_MAX_CACHE_SIZE + 10;
I think more recently we've been using OPTIMIZER_HIDE_VAR() to hide things from the compiler. Does OPTIMIZER_HIDE_VAR(ptr) right before the access also work in this case?
I leave it to you which you think is cleaner - I'm guessing that we might want to avoid volatile if we can.
KASAN_TEST_NEEDS_CONFIG_ON(test, CONFIG_SLUB);@@ -272,7 +273,8 @@ static void kmalloc_large_oob_right(struct kunit *test) static void krealloc_more_oob_helper(struct kunit *test, size_t size1, size_t size2) {
char *ptr1, *ptr2;
/* See comment in kasan_global_oob_right. */char *ptr1, *volatile ptr2; size_t middle; KUNIT_ASSERT_LT(test, size1, size2);@@ -304,7 +306,8 @@ static void krealloc_more_oob_helper(struct kunit *test, static void krealloc_less_oob_helper(struct kunit *test, size_t size1, size_t size2) {
char *ptr1, *ptr2;
/* See comment in kasan_global_oob_right. */char *ptr1, *volatile ptr2; size_t middle; KUNIT_ASSERT_LT(test, size2, size1);-- 2.35.1.473.g83b2b277ed-goog