io_uring: use TWA_SIGNAL more carefully
Get series with b4
b4 am [email protected]Versions newest first
- v1 Latest2 patches 38 repliesIn review
Patches in v1 No applies recorded yet
| # | Subject and files | Size | Replies | Applied |
|---|---|---|---|---|
| 1/2 | kernel: split task_work_add() into two separate helpersinclude/ kernel/ | +45 | 5 replies | |
| 2/2 | io_uring: use TWA_SIGNAL for task_work if the task isn't runningfs/ | +14 | 33 replies |
Reviews trailers per patch
All 2 patches have a Reviewed-by or Acked-by.
| Patch | Peter Zijlstra (Intel) |
|---|---|
| 1/2 | A-b |
| 2/2 | A-b |
Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes.
Cover letter [PATCHSET 0/2] io_uring: use TWA_SIGNAL more carefully
Adding io_uring support to Netty, Josef noticed that there's another case where we can fail to process task_work in time. We previously added a "work-around" for eventfd, see:
b7db41c9e03b ("io_uring: fix regression with always ignoring signals in io_cqring_wait()")
but we can run into this dependency issue even without that. See the test case added to liburing:
https://git.kernel.dk/cgit/liburing/tree/test/wakeup-hang.c
for an example of that. This series adds a split way to call task_work_add(), so we can use TWA_SIGNAL if the task is currently not running. This over-reaches a bit since there are definitely cases where we do not need to use it, but better safe than sorry for now.
Signature · 1 line
Jens Axboe