[1/2] kasan: prevent cpu_quarantine corruption when CPU offline and cache shrink occur at same time

Commit Message

Andrew Morton April 27, 2022, 7:41 p.m. UTC

From: Zqiang <[e-mail]>
Subject: kasan: prevent cpu_quarantine corruption when CPU offline and cache shrink occur at same time

kasan_quarantine_remove_cache() is called in kmem_cache_shrink()/
destroy().  The kasan_quarantine_remove_cache() call is protected by
cpuslock in kmem_cache_destroy() to ensure serialization with
kasan_cpu_offline().

However the kasan_quarantine_remove_cache() call is not protected by
cpuslock in kmem_cache_shrink().  When a CPU is going offline and cache
shrink occurs at same time, the cpu_quarantine may be corrupted by
interrupt (per_cpu_remove_cache operation).

So add a cpu_quarantine offline flags check in per_cpu_remove_cache().

[[e-mail]: add comment, per Zqiang]
Link: https://lkml.kernel.org/r/[e-mail]
Signed-off-by: Zqiang <[e-mail]>
Reviewed-by: Dmitry Vyukov <[e-mail]>
Cc: Andrey Ryabinin <[e-mail]>
Cc: Alexander Potapenko <[e-mail]>
Cc: Andrey Konovalov <[e-mail]>
Cc: <[e-mail]>
Signed-off-by: Andrew Morton <[e-mail]>
---

 mm/kasan/quarantine.c |    7 +++++++
 1 file changed, 7 insertions(+)

Patch

--- a/mm/kasan/quarantine.c~kasan-prevent-cpu_quarantine-corruption-when-cpu-offline-and-cache-shrink-occur-at-same-time
+++ a/mm/kasan/quarantine.c
@@ -315,6 +315,13 @@  static void per_cpu_remove_cache(void *a
 	struct qlist_head *q;
 
 	q = this_cpu_ptr(&cpu_quarantine);
+	/*
+	 * Ensure the ordering between the writing to q->offline and
+	 * per_cpu_remove_cache.  Prevent cpu_quarantine from being corrupted
+	 * by interrupt.
+	 */
+	if (READ_ONCE(q->offline))
+		return;
 	qlist_move_cache(q, &to_free, cache);
 	qlist_free_all(&to_free, cache);
 }