xfstests: cleanups and new tests.
Get series with b4
b4 am [email protected]Versions newest first
- v1 Latest5 patches 10 repliesIn review
Patches in v1 No applies recorded yet
| # | Subject and files | Size | Replies | Applied |
|---|---|---|---|---|
| 1/5 | generic: more tests should clean up TESTDIR on successtests/ tests/ tests/ | +22 | 1 reply | |
| 2/5 | check: more tests that shouldn't check the scratch devicetests/ | +1 | 1 reply | |
| 3/5 | generic: add mmap write vs truncate testtests/ tests/ tests/ | +189 | 1 reply | |
| 4/5 | generic: add mmap write vs truncate/remap testtests/ tests/ tests/ | +202 | 6 replies | |
| 5/5 | generic: add write vs fcollapse testtests/ tests/ tests/ | +93 | 2 replies |
Reviews trailers per patch
4 of 5 patches have a Reviewed-by or Acked-by.
| Patch | Eric Sandeen |
|---|---|
| 1/5 | R-b |
| 2/5 | R-b |
| 3/5 | R-b |
| 4/5 | R-b |
| 5/5 |
Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes.
Cover letter [PATCH 0/5] xfstests: cleanups and new tests.
Hi y'all,
The first two patches are minor cleanup patches. The first changes some tests that can leave large deposits on the test fs when they fail and that can lead to ENOSPC issues on small test devices. The second is another test that shouldn't check the scratch device at the end.
The remaining three patches are new generic tests that exercise data corruption issues that we've tripped over in the past week on XFS. The test cases have been isolated by Brian Foster and Eric Sandeen - they deserve the credit for narrowing down the minimal tests to expose the problems. I just added a few more corner cases and packaged them for xfstests.
These failures manifest on block size smaller than page size filesystems, and we have XFS fixes pending for these failures:
http://oss.sgi.com/pipermail/xfs/2014-September/038191.html
http://oss.sgi.com/pipermail/xfs/2014-September/038167.html
Ext4 folks, can you confirm that both the truncate tests fail on ext4 with data corruption on 1k/2k filesystems? I'm pretty sure that I have the corrupt hexdump outputs in the golden output files for the tests and if I have that means ext4 is corrupting data in a similar manner to XFS. Eric is already looking at this but could probably use some more help. ;)
-Dave.