sealing sched: Fix deadlock between worker.wndLk / workersLk #3489
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The deadlock could happen between
sh.workersLk.RLock()
in runWorker (sched.go:L577) (which holdsworker.wndLk
at the time), andtrySched
in the main scheduler goroutine, which indirectly tries to acquireworker.wndLk
through task selector while holdingsh.workersLk.RLock()
Normally this doesn't manifest because how could you possibly deadlock when acquiring an RLock from 2 goroutines
Well
Throw in a 3rd goroutine and write-preferred RWlocks to the mix and you get the deadlock
main trySched goroutine:
One of trySched sub-goroutines blocked on acquiring wndLk:
runWorker goroutine:
Some other random goroutine waiting on a WLock on sh.workersLk
Fixes #3460
(I have to give a fair bit of credit to danstark on filecoin slack for helping figure this out)