Fix Issue #43: non-reentrant with-gil causes deadlocks with pytorch #64
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Outdated suggestions cannot be applied.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.
Further research python c api (esp. https://docs.python.org/3/c-api/init.html#non-python-created-threads), base on torch developers who point out that PyGILState_Ensure and PyGILState_Release are reentrant (ie. recursive calls are allowed).
I've been using the patch in a nrepl session that runs for 6 hours with multiple clients (vscode calva remote ssh, and emacs cider-connect). All interactive evals are very responsive, no deadlocks. It used to deadlock within a few tools.namespace/refresh-all calls.