Re: [PATCH] btrfs: always scan a single device when mounted

On Mon, Feb 05, 2024 at 06:43:39PM +0100, David Sterba wrote:

There are reports that since version 6.7 update-grub fails to find the device of the root on systems without initrd and on a single device.

This looks like the device name changed in the output of /proc/self/mountinfo:

6.5-rc5 working

18 1 0:16 / / rw,noatime - btrfs /dev/sda8 ...

6.7 not working:

17 1 0:15 / / rw,noatime - btrfs /dev/root ...

and "update-grub" shows this error:

/usr/sbin/grub-probe: error: cannot find a device for / (is /dev mounted?)

This looks like it's related to the device name, but grub-probe recognizes the "/dev/root" path and tries to find the underlying device. However there's a special case for some filesystems, for btrfs in particular.

The generic root device detection heuristic is not done and it all relies on reading the device infos by a btrfs specific ioctl. This ioctl returns the device name as it was saved at the time of device scan (in this case it's /dev/root).

The change in 6.7 for temp_fsid to allow several single device filesystem to exist with the same fsid (and transparently generate a new UUID at mount time) was to skip caching/registering such devices.

This also skipped mounted device. One step of scanning is to check if the device name hasn't changed, and if yes then update the cached value.

This broke the grub-probe as it always read the device /dev/root and couldn't find it in the system. A temporary workaround is to create a symlink but this does not survive reboot.

The right fix is to allow updating the device path of a mounted filesystem even if this is a single device one. This does not affect the temp_fsid feature, the UUID of the mounted filesystem remains the same and the matching is based on device major:minor which is unique per mounted filesystem.

As the main part of device scanning and list update is done in device_list_add() that handles all corner cases and locking, it is extended to take a parameter that tells it to do everything as before, except adding a new device entry.

This covers the path when the device (that exists for all mounted devices) name changes, updating /dev/root to /dev/sdx. Any other single device with filesystem is skipped.

Note that if a system is booted and initial mount is done on the /dev/root device, this will be the cached name of the device. Only after the command "btrfs device rescan" it will change as it triggers the rename.

The fix was verified by users whose systems were affected.

CC: stable@vger.kernel.org # 6.7+ Fixes: bc27d6f0aa0e ("btrfs: scan but don't register device on single device filesystem") Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=218353 Link: https://lore.kernel.org/lkml/CAKLYgeJ1tUuqLcsquwuFqjDXPSJpEiokrWK2gisPKDZLs8... Signed-off-by: David Sterba dsterba@suse.com

Reviewed-by: Boris Burkov boris@bur.io

---

fs/btrfs/volumes.c | 30 ++++++++++++++---------------- 1 file changed, 14 insertions(+), 16 deletions(-)

diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 474ab7ed65ea..f2c2f7ca5c3d 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -738,6 +738,7 @@ static noinline struct btrfs_device *device_list_add(const char *path, bool same_fsid_diff_dev = false; bool has_metadata_uuid = (btrfs_super_incompat_flags(disk_super) & BTRFS_FEATURE_INCOMPAT_METADATA_UUID);

• bool can_create_new = *new_device_added;

It took me quite a while to figure out all the intended logic with the now in/out parameter. I think it's probably too cute? Why not just add another parameter "can_create_new_device"? I think it feels kind of weird on the caller side too, to set "new_device_added" to true, but then still rely on it to actually get set to true.

Once I got past this, the logic made sense, so definitely don't block yourself on this nit.

if (btrfs_super_flags(disk_super) & BTRFS_SUPER_FLAG_CHANGING_FSID_V2) { btrfs_err(NULL, @@ -753,6 +754,7 @@ static noinline struct btrfs_device *device_list_add(const char *path, return ERR_PTR(error); }

• *new_device_added = false; fs_devices = find_fsid_by_device(disk_super, path_devt, &same_fsid_diff_dev);

if (!fs_devices) { @@ -804,6 +806,15 @@ static noinline struct btrfs_device *device_list_add(const char *path, return ERR_PTR(-EBUSY); }

• if (!can_create_new) {
  

• 	pr_info(
  

• "BTRFS: device fsid %pU devid %llu transid %llu %s skip registration scanned by %s (%d)\n",

• 		disk_super->fsid, devid, found_transid, path,
  

• 		current->comm, task_pid_nr(current));
  

• 	mutex_unlock(&fs_devices->device_list_mutex);
  

• 	return NULL;
  

• }
  

• nofs_flag = memalloc_nofs_save(); device = btrfs_alloc_device(NULL, &devid, disk_super->dev_item.uuid, path);

@@ -1355,27 +1366,14 @@ struct btrfs_device *btrfs_scan_one_device(const char *path, blk_mode_t flags, goto error_bdev_put; }

• if (!mount_arg_dev && btrfs_super_num_devices(disk_super) == 1 &&

•    !(btrfs_super_flags(disk_super) & BTRFS_SUPER_FLAG_SEEDING)) {
  

• dev_t devt;
  

• ret = lookup_bdev(path, &devt);
  

• if (ret)
  

• 	btrfs_warn(NULL, "lookup bdev failed for path %s: %d",
  

• 		   path, ret);
  

• else
  

• 	btrfs_free_stale_devices(devt, NULL);
  

• pr_debug("BTRFS: skip registering single non-seed device %s\n", path);
  

• device = NULL;
  

• goto free_disk_super;
  

• }

• if (mount_arg_dev || btrfs_super_num_devices(disk_super) != 1 ||

•    (btrfs_super_flags(disk_super) & BTRFS_SUPER_FLAG_SEEDING))
  

• new_device_added = true;
  

This is the line I was referring to in the comment above, fwiw.

device = device_list_add(path, disk_super, &new_device_added); if (!IS_ERR(device) && new_device_added) btrfs_free_stale_devices(device->devt, device); -free_disk_super: btrfs_release_disk_super(disk_super); error_bdev_put:

-- 2.42.1