On Saturday, July 8, 2017 1:42:41 PM IST Amir Goldstein wrote: > OK, my pseudo patch was mostly correct, but full of brainoes, > push a tested patch to > https://github.com/amir73il/unionmount-testsuite/commits/overlayfs-devel > > Please review and verify. > Hi Amir, './run --ov' on 'next' kernel (i.e. the one containing code which returns pseudo st_devs for files on lowerdir) fails as shown below, $ ./run --ov *** *** ./run --ov --ts=0 open-plain *** TEST open-plain.py:10: Open O_RDONLY /mnt/a/foo100: File not on lower/upper layer The corresponding code from check_layer() is, elif dev != self.upper_fs() and dev != self.lower_fs(): raise TestError(name + ": File not on lower/upper layer") This happens due to open-plain.py's subtest_1() opening a file on lowerdir in read-only mode (i.e. copy-up does not occur). Later, stat(2) invocation from check_layer() would end up returning the pseudo st_dev corresponding to lowerdir. './run --ov --samefs --verify' on 4.12 kernel fails as shown below, $ ./run --ov --samefs --verify *** *** ./run --ov --ts=0 hard-link *** TEST hard-link.py:10: Hard link file ./run --link /mnt/a/foo100 /mnt/a/no_foo100 /mnt/a/no_foo100: inode number wrong (got 39001, want 37581) The failure happens because the dentry representing the new hard link wouldn't have __OVL_PATH_ORIGIN set. -- chandan -- To unsubscribe from this list: send the line "unsubscribe linux-unionfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
