Skip to contentSkip to conversation

dm: core: put device mapper block device synchronously

PATCH New Ming Lei 3 participants 5 messages Apr 16, 2024 – Apr 18, 2024

Subsystems (3)
lore
Share & tools
lore

Ming Lei

author
#0 lore raw

PATCH 1/1 dm: core: put device mapper block device synchronously

'dmsetup remove_all' actually depends on sync bdev release since dm_lock_for_deletion() may return -EBUSY if the open count is > 0, and the open count is dropped in dm_blk_close().

So if dm_blk_close() is delayed because of fput(), this device mapper device is skipped in remove_all, and cause regression.

Fix the issue by using __fput_sync().

Fixes a28d893eb327 ("md: port block device access to file")
Cc Jan Kara <[email protected]>
---
 drivers/md/dm.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/md/dm.c +2 -2
index 56aa2a8b9d71..93f3d28b0f03 100644
@@ -765,7 +765,7 @@ static struct table_device *open_table_device(struct mapped_device *md,
 	return td;
 
 out_blkdev_put:
-	fput(bdev_file);
+	__fput_sync(bdev_file);
 out_free_td:
 	kfree(td);
 	return ERR_PTR(r);
@@ -778,7 +778,7 @@ static void close_table_device(struct table_device *td, struct mapped_device *md
 {
 	if (md->disk->slave_dir)
 		bd_unlink_disk_holder(td->dm_dev.bdev, md->disk);
-	fput(td->dm_dev.bdev_file);
+	__fput_sync(td->dm_dev.bdev_file);
 	put_dax(td->dm_dev.dax_dev);
 	list_del(&td->list);
 	kfree(td);
Signature · 1 line
2.44.0
maintainer
#1 lore raw

PATCH v2 1/1 dm: restore synchronous close of device mapper block device

From: Ming Lei <[email protected]>

'dmsetup remove' and 'dmsetup remove_all' require synchronous bdev release. Otherwise dm_lock_for_deletion() may return -EBUSY if the open count is > 0, because the open count is dropped in dm_blk_close() which occurs after fput() completes.

So if dm_blk_close() is delayed because of asynchronous fput(), this device mapper device is skipped during remove, which is a regression.

Fix the issue by using __fput_sync().

Also: DM device removal has long supported being made asynchronous by setting the DMF_DEFERRED_REMOVE flag on the DM device. So leverage using async fput() in close_table_device() if DMF_DEFERRED_REMOVE flag is set.

Fixes a28d893eb327 ("md: port block device access to file")
[snitzer: editted commit header, use fput() if DMF_DEFERRED_REMOVE set]
---
 drivers/md/dm.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
drivers/md/dm.c +8 -2
index 56aa2a8b9d71..7d0746b37c8e 100644
@@ -765,7 +765,7 @@ static struct table_device *open_table_device(struct mapped_device *md,
 	return td;
 
 out_blkdev_put:
-	fput(bdev_file);
+	__fput_sync(bdev_file);
 out_free_td:
 	kfree(td);
 	return ERR_PTR(r);
@@ -778,7 +778,13 @@ static void close_table_device(struct table_device *td, struct mapped_device *md
 {
 	if (md->disk->slave_dir)
 		bd_unlink_disk_holder(td->dm_dev.bdev, md->disk);
-	fput(td->dm_dev.bdev_file);
+
+	/* Leverage async fput() if DMF_DEFERRED_REMOVE set */
+	if (unlikely(test_bit(DMF_DEFERRED_REMOVE, &md->flags)))
+		fput(td->dm_dev.bdev_file);
+	else
+		__fput_sync(td->dm_dev.bdev_file);
+
 	put_dax(td->dm_dev.dax_dev);
 	list_del(&td->list);
 	kfree(td);
Signature · 1 line
2.40.0
author
#2 lore raw

Re: [PATCH v2] dm: restore synchronous close of device mapper block device

Quoted 10 lines · Mike Snitzer

From: Ming Lei <[email protected]>

'dmsetup remove' and 'dmsetup remove_all' require synchronous bdev release. Otherwise dm_lock_for_deletion() may return -EBUSY if the open count is > 0, because the open count is dropped in dm_blk_close() which occurs after fput() completes.

So if dm_blk_close() is delayed because of asynchronous fput(), this device mapper device is skipped during remove, which is a regression.

Quoted 6 lines

Fix the issue by using __fput_sync().

Also: DM device removal has long supported being made asynchronous by setting the DMF_DEFERRED_REMOVE flag on the DM device. So leverage using async fput() in close_table_device() if DMF_DEFERRED_REMOVE flag is set.

IMO, this way isn't necessary, because the patch is one bug fix, and we are supposed to recover into exact previous behavior before commit a28d893eb327 ("md: port block device access to file") for minimizing regression risk.

But the extra change seems work.

thanks,
Ming

maintainer
#3 lore raw

Re: [PATCH v2] dm: restore synchronous close of device mapper block device

Quoted 18 lines · Ming Lei
Quoted 10 lines · Mike Snitzer

From: Ming Lei <[email protected]>

'dmsetup remove' and 'dmsetup remove_all' require synchronous bdev release. Otherwise dm_lock_for_deletion() may return -EBUSY if the open count is > 0, because the open count is dropped in dm_blk_close() which occurs after fput() completes.

So if dm_blk_close() is delayed because of asynchronous fput(), this device mapper device is skipped during remove, which is a regression.

Quoted 6 lines

Fix the issue by using __fput_sync().

Also: DM device removal has long supported being made asynchronous by setting the DMF_DEFERRED_REMOVE flag on the DM device. So leverage using async fput() in close_table_device() if DMF_DEFERRED_REMOVE flag is set.

Quoted 5 lines

IMO, this way isn't necessary, because the patch is one bug fix, and we are supposed to recover into exact previous behavior before commit a28d893eb327 ("md: port block device access to file") for minimizing regression risk.

But the extra change seems work.

I normally would agree but I see no real reason to avoid leveraging async fput() for the async DM device removal use-case ;)

Mike

#4 lore raw

Re: [PATCH v2] dm: restore synchronous close of device mapper block device

In reply to Mike Snitzer
Quoted 57 lines · Mike Snitzer

From: Ming Lei <[email protected]>

'dmsetup remove' and 'dmsetup remove_all' require synchronous bdev release. Otherwise dm_lock_for_deletion() may return -EBUSY if the open count is > 0, because the open count is dropped in dm_blk_close() which occurs after fput() completes.

So if dm_blk_close() is delayed because of asynchronous fput(), this device mapper device is skipped during remove, which is a regression.

Fix the issue by using __fput_sync().

Also: DM device removal has long supported being made asynchronous by setting the DMF_DEFERRED_REMOVE flag on the DM device. So leverage using async fput() in close_table_device() if DMF_DEFERRED_REMOVE flag is set.

Fixes a28d893eb327 ("md: port block device access to file")
[snitzer: editted commit header, use fput() if DMF_DEFERRED_REMOVE set]
---
 drivers/md/dm.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
drivers/md/dm.c +8 -3
index 56aa2a8b9d71..7d0746b37c8e 100644
@@ -765,7 +765,7 @@ static struct table_device *open_table_device(struct mapped_device *md,
        return td;
 
 out_blkdev_put:
-       fput(bdev_file);
+       __fput_sync(bdev_file);
 out_free_td:
        kfree(td);
        return ERR_PTR(r);
@@ -778,7 +778,13 @@ static void close_table_device(struct table_device *td, struct mapped_device *md
 {
        if (md->disk->slave_dir)
                bd_unlink_disk_holder(td->dm_dev.bdev, md->disk);
-       fput(td->dm_dev.bdev_file);
+
+       /* Leverage async fput() if DMF_DEFERRED_REMOVE set */
+       if (unlikely(test_bit(DMF_DEFERRED_REMOVE, &md->flags)))
+               fput(td->dm_dev.bdev_file);
+       else
+               __fput_sync(td->dm_dev.bdev_file);
+
        put_dax(td->dm_dev.dax_dev);
        list_del(&td->list);
        kfree(td);
--
Quoted 2 lines

2.40.0

I tried to apply this patch and looks this issue has solved by this patch