Stanislav Brabec wrote:
Coverage:
I just found a bug in my testcase:
To prevent race bind-mnt has to be created in time of fs creation.
btrfs subvol create s3 >/dev/null
+mkdir -p s3/bind-mnt
popd >/dev/null
But, what is worse, I found a problem in previous patches.
Things are even more complicated than I thought, and patches still not
cover all cases.
Up to now, I have been thinking that device + subvolid is unique
identifier in the mountinfo. But it is not true, at least not when bind
mounts take its role.
Using bind mounts that point to a sub-directory of a subvolume of btrfs
results in two lines in mountinfo that have the same subvolid but
different subvol.
Second "mount -a" then fails badly. Not only that it mounts the bind
mount second time, it also mistakenly mounts another volume over a
previously mounted volume.
dev/loop0 on
/home/sbrabec/VCS/util-linux/tests/output/mount/fstab-btrfs-mnt-subvol
type btrfs
(rw,relatime,space_cache,subvolid=258,subvol=/d0/dd0/ddd0/d2/dd2/ddd2/s3/bind-mnt)
=> totally wrong
We are back on the beginning.
I have to re-evaluate all patches I sent up to now.
TODO:
Check whether it is possible to directly mount sub-directory of
sub-volume that is not a subvolume per self.
If yes:
- We will need to introduce full path evaluation of btrfs sub-volume,
and always compare it with mountinfo.
If not:
- We will have to search the whole mountinfo for a particular subvolid
and and find a shortest subvol string.
David, could you bring some light into it?
--
Best Regards / S pozdravem,
Stanislav Brabec
software developer
---------------------------------------------------------------------
SUSE LINUX, s. r. o. e-mail: sbrabec@xxxxxxxx
Lihovarská 1060/12 tel: +49 911 7405384547
190 00 Praha 9 fax: +420 284 084 001
Czech Republic http://www.suse.cz/
PGP: 830B 40D5 9E05 35D8 5E27 6FA3 717C 209F A04F CD76
--
To unsubscribe from this list: send the line "unsubscribe util-linux" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
