Syscallbuf desched scheme does not work if desched is uninterruptible
Nobody has claimed this yet.
- Dominant language
- C++
- Stars
- 10.7k
- Forks
- 662
- Avg merge
- 2d 3h
- Merged PRs (30d)
- 2
Description
When running on a network or FUSE file system, many operations such as read/write will result in uninterruptible disk sleep. As a result, our syscallbuf desched mechanism does not work, because the desched signal cannot interrupt the system call. This is observable as either performance degradation or a hang (e.g. if the FUSE server is in the same rr session). I'm thinking it might be possible to have the desched signal be delivered to some other task instead, either the tracer (with some carefulness to make sure they're not lost around calls to waitpid) or some dummy ptraced task (that would then cause waitpid to exit in the relevant cases). Thoughts?
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
The issue names no files, tests, or entry points. Start by reproducing the syscallbuf desched failure on a network or FUSE filesystem and trace how the desched signal and waitpid interact; done means uninterruptible operations no longer cause the reported performance degradation or hang.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp, linux
- Domain
- devtools, operating-systems
- Issue type
- Bug
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100