diff options
| author | Christian Brauner <brauner@kernel.org> | 2026-06-16 13:58:15 +0200 |
|---|---|---|
| committer | Christian Brauner <brauner@kernel.org> | 2026-06-29 10:31:49 +0200 |
| commit | 822d87bc520fee8d95448c0aa3c728a4c1a595af (patch) | |
| tree | 580647067f343da2de95e698d161b83853481eb8 /block/bdev.c | |
| parent | ea4e4cc263011910eb7c62f3bb4fa094a1573c61 (diff) | |
block: split bdev_yield_claim() out of bdev_fput()
bdev_fput() yields the holder claim and then closes the file, which is a
deferred operation. Split the yield half into bdev_yield_claim() so a caller
can give up the holder while the file - and therefore the block device - is
still open, act on the device, and only then bdev_fput().
A filesystem that made a device unfreezable for a membership change with
bdev_deny_freeze() undoes the deny on release with
bdev_yield_claim(bdev_file);
bdev_allow_freeze(file_bdev(bdev_file));
bdev_fput(bdev_file);
Re-allowing only after the holder is yielded avoids stranding the filesystem
on a racing freeze, and doing it while the file is still open avoids touching
the block device after bdev_fput(). bdev_fput() yields again, which is a
no-op once the claim has already been given up.
Link: https://patch.msgid.link/20260616-work-super-freeze_deny_upstream-v2-2-b3567c7f994b@kernel.org
Reviewed-by: Jan Kara <jack@suse.cz>
Reviewd-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Christian Brauner (Amutable) <brauner@kernel.org>
Diffstat (limited to 'block/bdev.c')
| -rw-r--r-- | block/bdev.c | 50 |
1 files changed, 34 insertions, 16 deletions
diff --git a/block/bdev.c b/block/bdev.c index 9b73487a91ca..28b0d40c362f 100644 --- a/block/bdev.c +++ b/block/bdev.c @@ -1196,6 +1196,39 @@ put_no_open: } /** + * bdev_yield_claim - give up the holder claim on an open block device + * @bdev_file: open block device + * + * Yield the holder and any write access for @bdev_file without closing it, so + * the caller can still act on the device - e.g. bdev_allow_freeze() it - before + * the final bdev_fput(). bdev_fput() yields too, so calling it afterwards is + * safe. + */ +void bdev_yield_claim(struct file *bdev_file) +{ + struct block_device *bdev; + struct gendisk *disk; + + if (!bdev_file->private_data) + return; + + bdev = file_bdev(bdev_file); + disk = bdev->bd_disk; + + mutex_lock(&disk->open_mutex); + bdev_yield_write_access(bdev_file); + bd_yield_claim(bdev_file); + /* + * Tell release we already gave up our hold on the + * device and if write restrictions are available that + * we already gave up write access to the device. + */ + bdev_file->private_data = BDEV_I(bdev_file->f_mapping->host); + mutex_unlock(&disk->open_mutex); +} +EXPORT_SYMBOL_GPL(bdev_yield_claim); + +/** * bdev_fput - yield claim to the block device and put the file * @bdev_file: open block device * @@ -1208,22 +1241,7 @@ void bdev_fput(struct file *bdev_file) if (WARN_ON_ONCE(bdev_file->f_op != &def_blk_fops)) return; - if (bdev_file->private_data) { - struct block_device *bdev = file_bdev(bdev_file); - struct gendisk *disk = bdev->bd_disk; - - mutex_lock(&disk->open_mutex); - bdev_yield_write_access(bdev_file); - bd_yield_claim(bdev_file); - /* - * Tell release we already gave up our hold on the - * device and if write restrictions are available that - * we already gave up write access to the device. - */ - bdev_file->private_data = BDEV_I(bdev_file->f_mapping->host); - mutex_unlock(&disk->open_mutex); - } - + bdev_yield_claim(bdev_file); fput(bdev_file); } EXPORT_SYMBOL(bdev_fput); |
