On Tue, Apr 7, 2020 at 12:08 PM Chengguang Xu <cgxu519@xxxxxxxxxxxx> wrote: > > ---- 在 星期五, 2020-04-03 17:18:06 Amir Goldstein <amir73il@xxxxxxxxx> 撰写 ---- > > On Fri, Apr 3, 2020 at 9:45 AM Chengguang Xu <cgxu519@xxxxxxxxxxxx> wrote: > > > > > > Sharing inode with different whiteout files for saving > > > inode and speeding up delete operation. > > > > > > Signed-off-by: Chengguang Xu <cgxu519@xxxxxxxxxxxx> > > > > A few more nits. > > Please wait with v3 until I fix my patches, so you can test in top of > > them. > > Please run the overlay xfstests to test your patch. > > > > I suspect this part of test overlay/031 will fail and will need to fix > > the test to expect at most single whiteout 'residue' in work dir: > > > > # try to remove test dir from overlay dir, trigger ovl_remove_and_whiteout, > > # it will not clean up the dir and lead to residue. > > rm -rf $SCRATCH_MNT/testdir 2>&1 | _filter_scratch > > ls $workdir/work > > It seems no effect to current test case, I passed all test cases in overlay dir > except nfs_export and mmap related cases. > mmap test fails in baseline. did nfs_export tests fail with my recent branch (c1fe7dcb3db8)? because I had a bug that caused nfs_export tests to fail. ... > > > + return whiteout; > > > + } > > > + > > > + dget(whiteout); > > > + ofs->whiteout = whiteout; > > > > Shorter: > > ofs->whiteout = dget(whiteout); > > > > Here, we don't need to grab the dentry again, I think we have already got > the reference in lookup. > True. Thanks, Amir.
