]> git.ipfire.org Git - thirdparty/kernel/stable.git/commit
mm: kmsan: fix poisoning of high-order non-compound pages
authorRyan Roberts <ryan.roberts@arm.com>
Sun, 4 Jan 2026 13:43:47 +0000 (13:43 +0000)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Fri, 23 Jan 2026 10:21:29 +0000 (11:21 +0100)
commitb5cb343f0c3d071848452ca856a3bacef7943933
tree5be7948faab4baf8320d22b9e616144c3e178012
parent17f95d348589b2d65b0799f35a3c617e61cc4a28
mm: kmsan: fix poisoning of high-order non-compound pages

commit 4795d205d78690a46b60164f44b8bb7b3e800865 upstream.

kmsan_free_page() is called by the page allocator's free_pages_prepare()
during page freeing.  Its job is to poison all the memory covered by the
page.  It can be called with an order-0 page, a compound high-order page
or a non-compound high-order page.  But page_size() only works for order-0
and compound pages.  For a non-compound high-order page it will
incorrectly return PAGE_SIZE.

The implication is that the tail pages of a high-order non-compound page
do not get poisoned at free, so any invalid access while they are free
could go unnoticed.  It looks like the pages will be poisoned again at
allocation time, so that would bookend the window.

Fix this by using the order parameter to calculate the size.

Link: https://lkml.kernel.org/r/20260104134348.3544298-1-ryan.roberts@arm.com
Fixes: b073d7f8aee4 ("mm: kmsan: maintain KMSAN metadata for page operations")
Signed-off-by: Ryan Roberts <ryan.roberts@arm.com>
Reviewed-by: Alexander Potapenko <glider@google.com>
Tested-by: Alexander Potapenko <glider@google.com>
Cc: Dmitriy Vyukov <dvyukov@google.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Marco Elver <elver@google.com>
Cc: Ryan Roberts <ryan.roberts@arm.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
mm/kmsan/shadow.c