[runtime/iouring] use submit_with_args for bounded waits #1467
Conversation
|
If you're aware of a better way to handle network-driven ops in |
|
I think the best way to go about it would be to gradually incorporate more of the features that io_uring provides: vectored reads/writes, multishot, registered buffers, zero-copy (thought might not be worth it for small messages), SQPOLL and FAST_POLL. In no particular order. |
I think |
There was a problem hiding this comment.
Pull Request Overview
This PR modernizes the io_uring runtime to use the EXT_ARG feature for bounded waits instead of submitting timeout SQEs. The change reduces overhead by leveraging kernel 5.11+ functionality that allows timeouts to be specified directly in the submission arguments.
- Replaces timeout SQE submissions with
submit_with_argsusing timespec for both force polling and shutdown timeouts - Removes special timeout work IDs and associated handling logic for force polling and shutdown
- Simplifies the drain function by eliminating async timeout management
Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.
Codecov Report✅ All modified and coverable lines are covered by tests. @@ Coverage Diff @@
## main #1467 +/- ##
==========================================
+ Coverage 91.69% 91.71% +0.01%
==========================================
Files 278 278
Lines 70326 70501 +175
==========================================
+ Hits 64487 64658 +171
- Misses 5839 5843 +4 see 5 files with indirect coverage changes Continue to review full report in Codecov by Sentry.
🚀 New features to boost your workflow:
|
Replaces timeout SQEs for
force_pollandshutdown_timeoutwithsubmit_with_argswith a timespec, which uses the kernel's EXT_ARG feature (5.11+) for bounded waits without the overhead of submitting and processing timeout operations.See
IORING_ENTER_EXT_ARGin https://man7.org/linux/man-pages/man2/io_uring_enter.2.html.