mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* [PATCH] block: Use WRITE_ONCE for cache->nr_irq update
@ 2024-11-04 11:37 Zilin Guan
  0 siblings, 0 replies; only message in thread
From: Zilin Guan @ 2024-11-04 11:37 UTC (permalink / raw)
  To: axboe; +Cc: linux-block, linux-kernel, Zilin Guan

In function bio_put_percpu_cache(), cache->nr_irq is read indirectly
in line 776 using READ_ONCE(). This ensures safe access in a
multi-threaded environment.

776     if (READ_ONCE(cache->nr_irq) + cache->nr > ALLOC_CACHE_MAX)
777             goto out_free;

However, the increment operation is performed directly in line 792 without
using WRITE_ONCE(), which can lead to potential inconsistencies in a
multi-threaded context.

792     cache->nr_irq++;

To ensure consistent protection against data races, this update
should utilize WRITE_ONCE.

Signed-off-by: Zilin Guan <zilinguan811@gmail.com>
---
 block/bio.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/block/bio.c b/block/bio.c
index ac4d77c88932..e2648ba71bac 100644
--- a/block/bio.c
+++ b/block/bio.c
@@ -789,7 +789,7 @@ static inline void bio_put_percpu_cache(struct bio *bio)
 		bio_uninit(bio);
 		bio->bi_next = cache->free_list_irq;
 		cache->free_list_irq = bio;
-		cache->nr_irq++;
+		WRITE_ONCE(cache->nr_irq, cache->nr_irq + 1);
 	} else {
 		goto out_free;
 	}
-- 
2.34.1


^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2024-11-04 11:37 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2024-11-04 11:37 [PATCH] block: Use WRITE_ONCE for cache->nr_irq update Zilin Guan

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®