On 3/10/13 6:03 PM, Dave Chinner wrote: > On Sat, Mar 09, 2013 at 06:24:47PM -0600, Eric Sandeen wrote: >> On 1/18/13 3:48 PM, Koen De Wit wrote: >>> +} >>> + >>> +_scratch_mount >>> +_create_reflinks_to $TESTDIR2 >>> +_scratch_unmount >>> + >>> +mount $TEST_DEV $SCRATCH_MNT >>> +_create_reflinks_to $TESTDIR3 >>> +umount $SCRATCH_MNT >> >> TBH this confuses me, not that it's necessarily wrong (?) >> You mount TEST_DEV on $SCRATCH_MNT which makes my brain hurt a little. >> Then _create_reflinks_to $TESTDIR3 and at that point, um, what's going on, >> what's linking what to where? > > Mounting the TEST_DEV on SCRATCH_MNT is almost always a bad thing to > do. The test harness expects TEST_DEV to be mounted on TEST_DIR, not > anywhere else. > > If you need multiple scratch filesystems to test cross-device > linkage errors, use loopback devices or make use of the btrfs > scratch device pool...
Actually, looking at it again - does this wind up with TEST_DEV mounted on both TEST_DIR and SCRATCH_MNT? Maybe what the test wants is more mountpoints, not more devices? -Eric > Cheers, > > Dave. > -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
