diff options
author | Stanislav Brabec | 2016-04-12 20:22:56 +0200 |
---|---|---|
committer | Karel Zak | 2016-04-22 12:50:14 +0200 |
commit | bdf46c4df9aa9201afce6ae29f72268c54293bd7 (patch) | |
tree | 3032bf4e4caba758e27fbd19bc82031a5e7ab9b3 /libmount/src/context_loopdev.c | |
parent | libmount: Re-organize is_mounted_same_loopfile() (diff) | |
download | kernel-qcow2-util-linux-bdf46c4df9aa9201afce6ae29f72268c54293bd7.tar.gz kernel-qcow2-util-linux-bdf46c4df9aa9201afce6ae29f72268c54293bd7.tar.xz kernel-qcow2-util-linux-bdf46c4df9aa9201afce6ae29f72268c54293bd7.zip |
libmount: reuse existing loop device
According to the Al Viro[1], kernel has no way to detect that a single file is
used by multiple loop devices, and multiple mounts of the same file using
different loop devices will result in a data corruption. Exactly this now
happens, if multiple btrfs sub-volumes in one file are mounted with "-oloop".
Make use of multiple -oloop mounting the same file safe: Do a loop devices
lookup, and if a loop device is already initialized, use it.
Hopefully it is possible, as "losetup -d" will return OK, even if the device
itself is in use, and is not released.
Problems:
There is a risk of race condition between the lookup and real mount.
Once loop device is initialized read-only, kernel offers no way to turn it to
read-write. It has to fail.
References:
https://lkml.org/lkml/2016/2/26/897
Signed-off-by: Stanislav Brabec <sbrabec@suse.cz>
Diffstat (limited to 'libmount/src/context_loopdev.c')
-rw-r--r-- | libmount/src/context_loopdev.c | 24 |
1 files changed, 24 insertions, 0 deletions
diff --git a/libmount/src/context_loopdev.c b/libmount/src/context_loopdev.c index 47d6294c1..97435619e 100644 --- a/libmount/src/context_loopdev.c +++ b/libmount/src/context_loopdev.c @@ -210,6 +210,29 @@ int mnt_context_setup_loopdev(struct libmnt_context *cxt) if (rc) goto done_no_deinit; + /* It is possible to mount the same file more times. If we set more + * than one loop device referring to the same file, kernel has no + * mechanism to detect it. To prevent data corruption, the same loop + * device has to be recycled. + */ + rc = loopcxt_init(&lc, 0); + if (rc) + goto done; + if (backing_file && !(loopcxt_find_by_backing_file(&lc, + backing_file, offset, LOOPDEV_FL_OFFSET))) { + DBG(LOOP, ul_debugobj(cxt, "using existing loop device %s", + loopcxt_get_device(&lc))); + /* Once a loop is initialized RO, there is no way to safely + mount that file in R/W mode. */ + if (loopcxt_is_readonly(&lc) && !(lo_flags & LO_FLAGS_READ_ONLY)) { + rc = -EROFS; + goto done; + } + + goto success; + } + loopcxt_deinit(&lc); + rc = loopcxt_init(&lc, 0); if (rc == 0 && loopval) { rc = loopcxt_set_device(&lc, loopval); @@ -267,6 +290,7 @@ int mnt_context_setup_loopdev(struct libmnt_context *cxt) DBG(LOOP, ul_debugobj(cxt, "device stolen...trying again")); } while (1); +success: if (!rc) rc = mnt_fs_set_source(cxt->fs, loopcxt_get_device(&lc)); |