kernel.git/fs/btrfs, branch linux-5.16.y

btrfs: prevent subvol with swapfile from being deleted

2022-04-13T18:03:21Z

commit 60021bd754c6ca0addc6817994f20290a321d8d6 upstream. A subvolume with an active swapfile must not be deleted otherwise it would not be possible to deactivate it. After the subvolume is deleted, we cannot swapoff the swapfile in this deleted subvolume because the path is unreachable. The swapfile is still active and holding references, the filesystem cannot be unmounted. The test looks like this: mkfs.btrfs -f $dev > /dev/null mount $dev $mnt btrfs sub create $mnt/subvol touch $mnt/subvol/swapfile chmod 600 $mnt/subvol/swapfile chattr +C $mnt/subvol/swapfile dd if=/dev/zero of=$mnt/subvol/swapfile bs=1K count=4096 mkswap $mnt/subvol/swapfile swapon $mnt/subvol/swapfile btrfs sub delete $mnt/subvol swapoff $mnt/subvol/swapfile # failed: No such file or directory swapoff --all unmount $mnt # target is busy. To prevent above issue, we simply check that whether the subvolume contains any active swapfile, and stop the deleting process. This behavior is like snapshot ioctl dealing with a swapfile. CC: stable@vger.kernel.org # 5.4+ Reviewed-by: Robbie Ko Reviewed-by: Qu Wenruo Reviewed-by: Filipe Manana Signed-off-by: Kaiwen Hu Signed-off-by: David Sterba Signed-off-by: Greg Kroah-Hartman

btrfs: avoid defragging extents whose next extents are not targets

2022-04-13T18:03:21Z

commit 75a36a7d3ea904cef2e5b56af0c58cc60dcf947a upstream. [BUG] There is a report that autodefrag is defragging single sector, which is completely waste of IO, and no help for defragging: btrfs-cleaner-808 defrag_one_locked_range: root=256 ino=651122 start=0 len=4096 [CAUSE] In defrag_collect_targets(), we check if the current range (A) can be merged with next one (B). If mergeable, we will add range A into target for defrag. However there is a catch for autodefrag, when checking mergeability against range B, we intentionally pass 0 as @newer_than, hoping to get a higher chance to merge with the next extent. But in the next iteration, range B will looked up by defrag_lookup_extent(), with non-zero @newer_than. And if range B is not really newer, it will rejected directly, causing only range A being defragged, while we expect to defrag both range A and B. [FIX] Since the root cause is the difference in check condition of defrag_check_next_extent() and defrag_collect_targets(), we fix it by: 1. Pass @newer_than to defrag_check_next_extent() 2. Pass @extent_thresh to defrag_check_next_extent() This makes the check between defrag_collect_targets() and defrag_check_next_extent() more consistent. While there is still some minor difference, the remaining checks are focus on runtime flags like writeback/delalloc, which are mostly transient and safe to be checked only in defrag_collect_targets(). Link: https://github.com/btrfs/linux/issues/423#issuecomment-1066981856 CC: stable@vger.kernel.org # 5.16+ Reviewed-by: Filipe Manana Signed-off-by: Qu Wenruo Signed-off-by: David Sterba Signed-off-by: Greg Kroah-Hartman

btrfs: remove device item and update super block in the same transaction

2022-04-13T18:03:21Z

commit bbac58698a55cc0a6f0c0d69a6dcd3f9f3134c11 upstream. [BUG] There is a report that a btrfs has a bad super block num devices. This makes btrfs to reject the fs completely. BTRFS error (device sdd3): super_num_devices 3 mismatch with num_devices 2 found here BTRFS error (device sdd3): failed to read chunk tree: -22 BTRFS error (device sdd3): open_ctree failed [CAUSE] During btrfs device removal, chunk tree and super block num devs are updated in two different transactions: btrfs_rm_device() |- btrfs_rm_dev_item(device) | |- trans = btrfs_start_transaction() | | Now we got transaction X | | | |- btrfs_del_item() | | Now device item is removed from chunk tree | | | |- btrfs_commit_transaction() | Transaction X got committed, super num devs untouched, | but device item removed from chunk tree. | (AKA, super num devs is already incorrect) | |- cur_devices->num_devices--; |- cur_devices->total_devices--; |- btrfs_set_super_num_devices() All those operations are not in transaction X, thus it will only be written back to disk in next transaction. So after the transaction X in btrfs_rm_dev_item() committed, but before transaction X+1 (which can be minutes away), a power loss happen, then we got the super num mismatch. [FIX] Instead of starting and committing a transaction inside btrfs_rm_dev_item(), start a transaction in side btrfs_rm_device() and pass it to btrfs_rm_dev_item(). And only commit the transaction after everything is done. Reported-by: Luca Béla Palkovics Link: https://lore.kernel.org/linux-btrfs/CA+8xDSpvdm_U0QLBAnrH=zqDq_cWCOH5TiV46CKmp3igr44okQ@mail.gmail.com/ CC: stable@vger.kernel.org # 4.14+ Reviewed-by: Anand Jain Signed-off-by: Qu Wenruo Signed-off-by: David Sterba Signed-off-by: Greg Kroah-Hartman

btrfs: zoned: traverse devices under chunk_mutex in btrfs_can_activate_zone

2022-04-13T18:03:20Z

commit 0b9e66762aa0cda2a9c2d5542d64e04dac528fa6 upstream. btrfs_can_activate_zone() can be called with the device_list_mutex already held, which will lead to a deadlock: insert_dev_extents() // Takes device_list_mutex `-> insert_dev_extent() `-> btrfs_insert_empty_item() `-> btrfs_insert_empty_items() `-> btrfs_search_slot() `-> btrfs_cow_block() `-> __btrfs_cow_block() `-> btrfs_alloc_tree_block() `-> btrfs_reserve_extent() `-> find_free_extent() `-> find_free_extent_update_loop() `-> can_allocate_chunk() `-> btrfs_can_activate_zone() // Takes device_list_mutex again Instead of using the RCU on fs_devices->device_list we can use fs_devices->alloc_list, protected by the chunk_mutex to traverse the list of active devices. We are in the chunk allocation thread. The newer chunk allocation happens from the devices in the fs_device->alloc_list protected by the chunk_mutex. btrfs_create_chunk() lockdep_assert_held(&info->chunk_mutex); gather_device_info list_for_each_entry(device, &fs_devices->alloc_list, dev_alloc_list) Also, a device that reappears after the mount won't join the alloc_list yet and, it will be in the dev_list, which we don't want to consider in the context of the chunk alloc. [15.166572] WARNING: possible recursive locking detected [15.167117] 5.17.0-rc6-dennis #79 Not tainted [15.167487] -------------------------------------------- [15.167733] kworker/u8:3/146 is trying to acquire lock: [15.167733] ffff888102962ee0 (&fs_devs->device_list_mutex){+.+.}-{3:3}, at: find_free_extent+0x15a/0x14f0 [btrfs] [15.167733] [15.167733] but task is already holding lock: [15.167733] ffff888102962ee0 (&fs_devs->device_list_mutex){+.+.}-{3:3}, at: btrfs_create_pending_block_groups+0x20a/0x560 [btrfs] [15.167733] [15.167733] other info that might help us debug this: [15.167733] Possible unsafe locking scenario: [15.167733] [15.171834] CPU0 [15.171834] ---- [15.171834] lock(&fs_devs->device_list_mutex); [15.171834] lock(&fs_devs->device_list_mutex); [15.171834] [15.171834] *** DEADLOCK *** [15.171834] [15.171834] May be due to missing lock nesting notation [15.171834] [15.171834] 5 locks held by kworker/u8:3/146: [15.171834] #0: ffff888100050938 ((wq_completion)events_unbound){+.+.}-{0:0}, at: process_one_work+0x1c3/0x5a0 [15.171834] #1: ffffc9000067be80 ((work_completion)(&fs_info->async_data_reclaim_work)){+.+.}-{0:0}, at: process_one_work+0x1c3/0x5a0 [15.176244] #2: ffff88810521e620 (sb_internal){.+.+}-{0:0}, at: flush_space+0x335/0x600 [btrfs] [15.176244] #3: ffff888102962ee0 (&fs_devs->device_list_mutex){+.+.}-{3:3}, at: btrfs_create_pending_block_groups+0x20a/0x560 [btrfs] [15.176244] #4: ffff8881152e4b78 (btrfs-dev-00){++++}-{3:3}, at: __btrfs_tree_lock+0x27/0x130 [btrfs] [15.179641] [15.179641] stack backtrace: [15.179641] CPU: 1 PID: 146 Comm: kworker/u8:3 Not tainted 5.17.0-rc6-dennis #79 [15.179641] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.15.0-1.fc35 04/01/2014 [15.179641] Workqueue: events_unbound btrfs_async_reclaim_data_space [btrfs] [15.179641] Call Trace: [15.179641] [15.179641] dump_stack_lvl+0x45/0x59 [15.179641] __lock_acquire.cold+0x217/0x2b2 [15.179641] lock_acquire+0xbf/0x2b0 [15.183838] ? find_free_extent+0x15a/0x14f0 [btrfs] [15.183838] __mutex_lock+0x8e/0x970 [15.183838] ? find_free_extent+0x15a/0x14f0 [btrfs] [15.183838] ? find_free_extent+0x15a/0x14f0 [btrfs] [15.183838] ? lock_is_held_type+0xd7/0x130 [15.183838] ? find_free_extent+0x15a/0x14f0 [btrfs] [15.183838] find_free_extent+0x15a/0x14f0 [btrfs] [15.183838] ? _raw_spin_unlock+0x24/0x40 [15.183838] ? btrfs_get_alloc_profile+0x106/0x230 [btrfs] [15.187601] btrfs_reserve_extent+0x131/0x260 [btrfs] [15.187601] btrfs_alloc_tree_block+0xb5/0x3b0 [btrfs] [15.187601] __btrfs_cow_block+0x138/0x600 [btrfs] [15.187601] btrfs_cow_block+0x10f/0x230 [btrfs] [15.187601] btrfs_search_slot+0x55f/0xbc0 [btrfs] [15.187601] ? lock_is_held_type+0xd7/0x130 [15.187601] btrfs_insert_empty_items+0x2d/0x60 [btrfs] [15.187601] btrfs_create_pending_block_groups+0x2b3/0x560 [btrfs] [15.187601] __btrfs_end_transaction+0x36/0x2a0 [btrfs] [15.192037] flush_space+0x374/0x600 [btrfs] [15.192037] ? find_held_lock+0x2b/0x80 [15.192037] ? btrfs_async_reclaim_data_space+0x49/0x180 [btrfs] [15.192037] ? lock_release+0x131/0x2b0 [15.192037] btrfs_async_reclaim_data_space+0x70/0x180 [btrfs] [15.192037] process_one_work+0x24c/0x5a0 [15.192037] worker_thread+0x4a/0x3d0 Fixes: a85f05e59bc1 ("btrfs: zoned: avoid chunk allocation if active block group has enough space") CC: stable@vger.kernel.org # 5.16+ Reviewed-by: Anand Jain Signed-off-by: Johannes Thumshirn Reviewed-by: David Sterba Signed-off-by: David Sterba Signed-off-by: Greg Kroah-Hartman

btrfs: fix qgroup reserve overflow the qgroup limit

2022-04-13T18:03:20Z

commit b642b52d0b50f4d398cb4293f64992d0eed2e2ce upstream. We use extent_changeset->bytes_changed in qgroup_reserve_data() to record how many bytes we set for EXTENT_QGROUP_RESERVED state. Currently the bytes_changed is set as "unsigned int", and it will overflow if we try to fallocate a range larger than 4GiB. The result is we reserve less bytes and eventually break the qgroup limit. Unlike regular buffered/direct write, which we use one changeset for each ordered extent, which can never be larger than 256M. For fallocate, we use one changeset for the whole range, thus it no longer respects the 256M per extent limit, and caused the problem. The following example test script reproduces the problem: $ cat qgroup-overflow.sh #!/bin/bash DEV=/dev/sdj MNT=/mnt/sdj mkfs.btrfs -f $DEV mount $DEV $MNT # Set qgroup limit to 2GiB. btrfs quota enable $MNT btrfs qgroup limit 2G $MNT # Try to fallocate a 3GiB file. This should fail. echo echo "Try to fallocate a 3GiB file..." fallocate -l 3G $MNT/3G.file # Try to fallocate a 5GiB file. echo echo "Try to fallocate a 5GiB file..." fallocate -l 5G $MNT/5G.file # See we break the qgroup limit. echo sync btrfs qgroup show -r $MNT umount $MNT When running the test: $ ./qgroup-overflow.sh (...) Try to fallocate a 3GiB file... fallocate: fallocate failed: Disk quota exceeded Try to fallocate a 5GiB file... qgroupid rfer excl max_rfer -------- ---- ---- -------- 0/5 5.00GiB 5.00GiB 2.00GiB Since we have no control of how bytes_changed is used, it's better to set it to u64. CC: stable@vger.kernel.org # 4.14+ Reviewed-by: Qu Wenruo Signed-off-by: Ethan Lien Reviewed-by: David Sterba Signed-off-by: David Sterba Signed-off-by: Greg Kroah-Hartman

btrfs: do not clean up repair bio if submit fails

2022-04-08T12:06:36Z

[ Upstream commit 8cbc3001a3264d998d6b6db3e23f935c158abd4d ] The submit helper will always run bio_endio() on the bio if it fails to submit, so cleaning up the bio just leads to a variety of use-after-free and NULL pointer dereference bugs because we race with the endio function that is cleaning up the bio. Instead just return BLK_STS_OK as the repair function has to continue to process the rest of the pages, and the endio for the repair bio will do the appropriate cleanup for the page that it was given. Reviewed-by: Boris Burkov Signed-off-by: Josef Bacik Signed-off-by: David Sterba Signed-off-by: Sasha Levin

btrfs: do not double complete bio on errors during compressed reads

2022-04-08T12:06:36Z

[ Upstream commit f9f15de85d74e7eef021af059ca53a15f041cdd8 ] I hit some weird panics while fixing up the error handling from btrfs_lookup_bio_sums(). Turns out the compression path will complete the bio we use if we set up any of the compression bios and then return an error, and then btrfs_submit_data_bio() will also call bio_endio() on the bio. Fix this by making btrfs_submit_compressed_read() responsible for calling bio_endio() on the bio if there are any errors. Currently it was only doing it if we created the compression bios, otherwise it was depending on btrfs_submit_data_bio() to do the right thing. This creates the above problem, so fix up btrfs_submit_compressed_read() to always call bio_endio() in case of an error, and then simply return from btrfs_submit_data_bio() if we had to call btrfs_submit_compressed_read(). Signed-off-by: Josef Bacik Reviewed-by: David Sterba Signed-off-by: David Sterba Signed-off-by: Sasha Levin

btrfs: handle csum lookup errors properly on reads

2022-04-08T12:06:36Z

[ Upstream commit 1784b7d502a94b561eae58249adde5f72c26eb3c ] Currently any error we get while trying to lookup csums during reads shows up as a missing csum, and then on the read completion side we print an error saying there was a csum mismatch and we increase the device corruption count. However we could have gotten an EIO from the lookup. We could also be inside of a memory constrained container and gotten a ENOMEM while trying to do the read. In either case we don't want to make this look like a file system corruption problem, we want to make it look like the actual error it is. Capture any negative value, convert it to the appropriate blk_status_t, free the csum array if we have one and bail. Note: a possible improvement would be to make the relocation code look up the owning inode and see if it's marked as NODATASUM and set EXTENT_NODATASUM there, that way if there's corruption and there isn't a checksum when we want it we can fail here rather than later. Signed-off-by: Josef Bacik Signed-off-by: David Sterba Signed-off-by: Sasha Levin

btrfs: make search_csum_tree return 0 if we get -EFBIG

2022-04-08T12:06:36Z

[ Upstream commit 03ddb19d2ea745228879b9334f3b550c88acb10a ] We can either fail to find a csum entry at all and return -ENOENT, or we can find a range that is close, but return -EFBIG. In essence these both mean the same thing when we are doing a lookup for a csum in an existing range, we didn't find a csum. We want to treat both of these errors the same way, complain loudly that there wasn't a csum. This currently happens anyway because we do count = search_csum_tree(); if (count <= 0) { // reloc and error handling } However it forces us to incorrectly treat EIO or ENOMEM errors as on disk corruption. Fix this by returning 0 if we get either -ENOENT or -EFBIG from btrfs_lookup_csum() so we can do proper error handling. Reviewed-by: Boris Burkov Reviewed-by: Johannes Thumshirn Signed-off-by: Josef Bacik Reviewed-by: David Sterba Signed-off-by: David Sterba Signed-off-by: Sasha Levin

btrfs: harden identification of a stale device

2022-04-08T12:06:36Z

[ Upstream commit 770c79fb65506fc7c16459855c3839429f46cb32 ] Identifying and removing the stale device from the fs_uuids list is done by btrfs_free_stale_devices(). btrfs_free_stale_devices() in turn depends on device_path_matched() to check if the device appears in more than one btrfs_device structure. The matching of the device happens by its path, the device path. However, when device mapper is in use, the dm device paths are nothing but a link to the actual block device, which leads to the device_path_matched() failing to match. Fix this by matching the dev_t as provided by lookup_bdev() instead of plain string compare of the device paths. Reported-by: Josef Bacik Signed-off-by: Anand Jain Signed-off-by: David Sterba Signed-off-by: Sasha Levin