Make fanotify10 test yet more reliable
Get series with b4
b4 am [email protected]Versions newest first
- v1 Latest3 patches 20 repliesIn review
Patches in v1 No applies recorded yet
| # | Subject and files | Size | Replies | Applied |
|---|---|---|---|---|
| 1/3 | fanotify10: Use named initializerstestcases/ | +249 | 0 replies | |
| 2/3 | fanotify10: Add support for multiple event filestestcases/ | +245 | 13 replies | |
| 3/3 | fanotify10: Make evictable marks tests more reliabletestcases/ | +52 | 7 replies |
Reviews trailers per patch
2 of 3 patches have a Reviewed-by or Acked-by.
| Patch | Pengfei Xu | Petr Vorel |
|---|---|---|
| 1/3 | ||
| 2/3 | R-b | |
| 3/3 | T-b | R-b |
Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes.
Cover letter [LTP] [PATCH 0/3] Make fanotify10 test yet more reliable
Hello!
I was debugging with Pengfei Xu why fanotify10 testcase still occasionally fails in his test setup. After a lot of back and forth we have identified two causes. One lies within the kernel slab reclaim itself (fix submitted), the other one is the inherent problem that slab reclaim needs to first reclaim dentries (which means going through round of LRU aging before dentry is reclaimed) and then inodes have to go through LRU aging before they are reclaimed. As a result code dropping slab caches can decide there's not enough forward progress and stop before the inodes we are interested in are evicted.
This patch modifies fanotify10 testcase to create multiple files / dirs with ignore marks and return success if at least half of ignore marks got reclaimed. This both gives slab reclaim code better feel of forward progress as well as provides some robustness against some inode not being reclaimed for some random reason.
With the kernel fix and this modification to fanotify10 testcase, Xu cannot trigger the failure anymore.
Honza
Signature · 1 line
Mailing list info: https://lists.linux.it/listinfo/ltp