btrfs: fix false alerts when running with 8K block size and 4K page size
Get series with b4
b4 am [email protected]Versions newest first · linked by author and title
- v2 Latest3 patches 1 replyIn review
- v13 patches 6 repliesSuperseded
Changes in v2 from the cover letter
- Add extra block size check against page size before skipping btrfs/012 and btrfs/136
Patches in v2 No applies recorded yet
| # | Subject and files | Size | Replies | Applied |
|---|---|---|---|---|
| 1/3 | btrfs/012 btrfs/136: skip the test if ext* doesn't support the block sizetests/ tests/ | +6 | 1 reply | |
| 2/3 | btrfs/192 btrfs/30[456]: explicitly specify block size to avoid false alertstests/ tests/ tests/ +1 more | +8 | 0 replies | |
| 3/3 | btrfs/26[67]: update the stale commentstests/ tests/ | +3 | 0 replies |
Reviews trailers per patch
All 3 patches have a Reviewed-by or Acked-by.
| Patch | Zorro Lang |
|---|---|
| 1/3 | R-b |
| 2/3 | R-b |
| 3/3 | R-b |
Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes.
Cover letter [PATCH v2 0/3] btrfs: fix false alerts when running with 8K block size and 4K page size
[CHANGELOG]
v2:
- Add extra block size check against page size before skipping btrfs/012 and btrfs/136
I am developing the bs > ps support for btrfs, and now it can pass most btrfs and generic runs.
Some failures are due to the limitaions of the current bs > ps, including:
- btrfs/023
No RAID56 support, and the test case doesn't respect BTRFS_PROFILE_CONFIGS
- btrfs/131
No v1 cache support just like subpage bs support.
No big deal, as v1 cache is already marked deprecated. When the full deprecation comes, the test case needs some update.
- btrfs/226
No Direct IO support.
- btrfs/267
No Direct IO support thus the read falls back to buffered one. The fallback may change the pid thus some mirror is not properly read from disk and no read repair.
- btrfs/281
No encoded send support
However there are some btrfs failures that are false alerts:
- btrfs/012
- btrfs/136
Those are btrfs-convert tests, however ext* doesn't support bs > ps cases yet.
Fix them by skip the run if the initial ext* mount failed.
- btrfs/192
This one requires 4K nodesize, which implies 4K block size, and conflicts with user specified non-4K block size.
- btrfs/30[456]
Those test cases have strict 4K block size requirement but still follows the user specified block size during mkfs.
Fix btrfs/192 and btrfs/30[456] by explicitly specify 4K block size during mkfs.
There is also a minor comment mismatch in btrfs/267:
- btrfs/267 is verifying direct read repair but comments says buffered Instead it's btrfs/266 verifying the buffered behavior. So it's purely a comment mismatch.
Fix it by explicitly mentioning buffered/direct IO for btrfs/26[67].
Qu Wenruo (3):
btrfs/012 btrfs/136: skip the test if ext* doesn't support the block
size
btrfs/192 btrfs/30[456]: explicitly specify block size to avoid false
alerts
btrfs/26[67]: update the stale comments
tests/btrfs/012 | 3 +++ tests/btrfs/136 | 3 +++ tests/btrfs/192 | 8 ++------ tests/btrfs/266 | 4 ++-- tests/btrfs/267 | 2 +- tests/btrfs/304 | 5 ++--- tests/btrfs/305 | 5 ++--- tests/btrfs/306 | 5 ++--- 8 files changed, 17 insertions(+), 18 deletions(-)
Signature · 1 line
2.51.0