| commit | d7a753a148d4738eef95a3b193b081a9c905399f | |
| author | Eric Blake <eblake@redhat.com> | |
| Wed, 9 Aug 2017 20:38:06 +0000 (9 15:38 -0500) | ||
| committer | Kevin Wolf <kwolf@redhat.com> | |
| Mon, 4 Sep 2017 16:33:00 +0000 (4 18:33 +0200) | ||
| tree | 0ca51f42a2bca13a616270305283a1d375f65403 | treesnapshot (tar.gz zip) |
| parent | 56439e9d55626b65ecb887f1ac3714652555312e | commitdiff |
qcow: Check failure of bdrv_getlength() and bdrv_truncate()
Omitting the check for whether bdrv_getlength() and bdrv_truncate()
failed meant that it was theoretically possible to return an
incorrect offset to the caller. More likely, conditions for either
of these functions to fail would also cause one of our other calls
(such as bdrv_pread() or bdrv_pwrite_sync()) to also fail, but
auditing that we are safe is difficult compared to just patching
things to always forward on the error rather than ignoring it.
Use osdep.h macros instead of open-coded rounding while in the
area.
Reported-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Omitting the check for whether bdrv_getlength() and bdrv_truncate()
failed meant that it was theoretically possible to return an
incorrect offset to the caller. More likely, conditions for either
of these functions to fail would also cause one of our other calls
(such as bdrv_pread() or bdrv_pwrite_sync()) to also fail, but
auditing that we are safe is difficult compared to just patching
things to always forward on the error rather than ignoring it.
Use osdep.h macros instead of open-coded rounding while in the
area.
Reported-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
| block/qcow.c | diffblobblamehistory |