On Fri, Aug 5, 2011 at 9:53 AM, Amir Goldstein <amir73il@xxxxxxxxx> wrote: > Hi Anand, > > Can you please post the patch in the body of the message so we can > comment on it inline. > > Aditya is also working on a btrfs snapshots xfstest (currently named > 257) as part of > his Google summer of code project. > > It would be great if you guys can cooperate your efforts. > > BTW, while running Aditya's test, Greg has stumbled upon a btrfs OOPS. > Greg can provide more details about it. > > Cheers, > Amir. > > On Fri, Aug 5, 2011 at 10:59 AM, Anand Jain <Anand.Jain@xxxxxxxxxx> wrote: >> >> Hi, >> >> Attached is the patch for the xfstests, which adds snapshot, >> defragment and volume management test cases for the btrfs >> (257, 258 and 259 respectively). >> >> This introduces a new user variable 'DISK_POOL' which should >> be set to disks for the raid tests. >> >> An example of usage of these tests is as below. >> ------------ >> [root@localhost xfstests]# cat local.config >> TEST_DEV="/dev/sdd" >> TEST_DIR=/btrfs >> SCRATCH_DEV="/dev/sde" >> SCRATCH_MNT=/btrfs1 >> DISK_POOL="/dev/sdf /dev/sdg" >> [root@localhost xfstests]# >> >> [root@localhost xfstests]# ./check 257 258 259 >> FSTYP -- btrfs >> PLATFORM -- Linux/i686 localhost 3.0.0-rc6+ >> MKFS_OPTIONS -- /dev/sde >> MOUNT_OPTIONS -- -o context=system_u:object_r:nfs_t:s0 /dev/sde /btrfs1 >> >> 257 8s >> 258 3s >> 259 33s >> Ran: 257 258 259 >> Passed all 3 tests >> ---------------- >> >> Thank you. >> >> Anand >> Anand, I've only reproduced the oops in 2.6.39 and older using openSUSE 11.4 userspace. I have not tried with 3.0 or newer nor with newer userspace, so it may or may not have been fixed. It happened with a openSUSE kernel, so I put it in their bugzilla. https://bugzilla.novell.com/show_bug.cgi?id=705039 Just yesterday, I put the actual full recreate details in the bugzilla. They are: === I was working the xfstests git tree cloned via: git clone https://github.com/adityadani/xfstests # Then, get an older commit that apparently has a userspace bug that triggers a kernel oops. git checkout 1b712a4512d63502fab6 # Then prep for the test export TEST_DEV=/dev/sda3 export TEST_DIR=/xfstests_mnt export SCRATCH_DEV=/dev/sda1 export SCRATCH_MNT=/xfstests_mnt_scratch export FSTYP=btrfs umount $TEST_DIR umount $SCRATCH_MNT mkfs.btrfs $TEST_DEV mkfs.btrfs $SCRATCH_DEV mount $TEST_DEV $TEST_DIR mount $SCRATCH_DEV $SCRATCH_MNT # Then run the test ./check 255 I didn't keep the oops, but it seems reproducible enough. The oops was also put into the kde-oops catcher. === I don't know what aspect of that test triggers the oops. But the latest version of test does not trigger it. The latest version of that test in Aditya's tree has been changed to test 256. Sounds like that will change to at least 260 before it gets submitted. Hope that helps, Greg -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
