diff mbox series

[v4,2/2] block: warn once for each partition in bio_check_ro()

Message ID 20231128123027.971610-3-yukuai1@huaweicloud.com (mailing list archive)
State New, archived
Headers show
Series block: warn once for each partition in bio_check_ro() | expand

Commit Message

Yu Kuai Nov. 28, 2023, 12:30 p.m. UTC 
From: Yu Kuai <yukuai3@huawei.com>

Commit 1b0a151c10a6 ("blk-core: use pr_warn_ratelimited() in
bio_check_ro()") fix message storm by limit the rate, however, there
will still be lots of message in the long term. Fix it better by warn
once for each partition.

Signed-off-by: Yu Kuai <yukuai3@huawei.com>
---
 block/blk-core.c          | 14 +++++++++++---
 include/linux/blk_types.h |  1 +
 2 files changed, 12 insertions(+), 3 deletions(-)

Comments

Christoph Hellwig Nov. 28, 2023, 1 p.m. UTC | #1 
On Tue, Nov 28, 2023 at 08:30:27PM +0800, Yu Kuai wrote:
> From: Yu Kuai <yukuai3@huawei.com>
> 
> Commit 1b0a151c10a6 ("blk-core: use pr_warn_ratelimited() in
> bio_check_ro()") fix message storm by limit the rate, however, there
> will still be lots of message in the long term. Fix it better by warn
> once for each partition.

The new field is in the same dword alignment as bd_make_it_fail and
could in theory corrupt it, at least on alpha.  I guess we're fine,
because if you enable CONFIG_FAIL_MAKE_REQUEST on alpha you're asking
for this.  I still hope we can clean up these non-atomic bools and
replace them with bitops soon.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Yu Kuai Nov. 29, 2023, 1:13 a.m. UTC | #2 
Hi,

在 2023/11/28 21:00, Christoph Hellwig 写道:
> On Tue, Nov 28, 2023 at 08:30:27PM +0800, Yu Kuai wrote:
>> From: Yu Kuai <yukuai3@huawei.com>
>>
>> Commit 1b0a151c10a6 ("blk-core: use pr_warn_ratelimited() in
>> bio_check_ro()") fix message storm by limit the rate, however, there
>> will still be lots of message in the long term. Fix it better by warn
>> once for each partition.
> 
> The new field is in the same dword alignment as bd_make_it_fail and
> could in theory corrupt it, at least on alpha.  I guess we're fine,
> because if you enable CONFIG_FAIL_MAKE_REQUEST on alpha you're asking
> for this.  I still hope we can clean up these non-atomic bools and
> replace them with bitops soon.

Yes, I'm working on this, and thanks for the review!

Kuai
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> .
>
diff mbox series

Patch

diff --git a/block/blk-core.c b/block/blk-core.c
index fdf25b8d6e78..2eca76ccf4ee 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -501,9 +501,17 @@  static inline void bio_check_ro(struct bio *bio)
 	if (op_is_write(bio_op(bio)) && bdev_read_only(bio->bi_bdev)) {
 		if (op_is_flush(bio->bi_opf) && !bio_sectors(bio))
 			return;
-		pr_warn_ratelimited("Trying to write to read-only block-device %pg\n",
-				    bio->bi_bdev);
-		/* Older lvm-tools actually trigger this */
+
+		if (bio->bi_bdev->bd_ro_warned)
+			return;
+
+		bio->bi_bdev->bd_ro_warned = true;
+		/*
+		 * Use ioctl to set underlying disk of raid/dm to read-only
+		 * will trigger this.
+		 */
+		pr_warn("Trying to write to read-only block-device %pg\n",
+			bio->bi_bdev);
 	}
 }
 
diff --git a/include/linux/blk_types.h b/include/linux/blk_types.h
index f7d40692dd94..b29ebd53417d 100644
--- a/include/linux/blk_types.h
+++ b/include/linux/blk_types.h
@@ -70,6 +70,7 @@  struct block_device {
 #ifdef CONFIG_FAIL_MAKE_REQUEST
 	bool			bd_make_it_fail;
 #endif
+	bool			bd_ro_warned;
 	/*
 	 * keep this out-of-line as it's both big and not needed in the fast
 	 * path