Linux: Fix a rare crash that could occur when shutting down an interpreter running multiple threads #324
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Turns out on Linux with gcc and libstdc++ calling pthread_exit() raises an uncatchable exception. That is called when reacquiring the GIL in a background thread if the main thread has terminated. The use of G_NOEXCEPT in that code path was resulting in the C++ runtime calling std::terminate when that happened.
…to force the termination that we otheriwse can't seem to get.
python3.5 is long EOL'd.
jamadden
force-pushed
the
noexcept-crashing-on-linux
branch
from
a485d9c
to
bb19adf
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
when some of those threads are in greenlets making calls to functions that release the GIL.
When Python notices that the thread would like to obtain the GIL again, but the interpreter is exiting, it calls
pthread_exit. With GCC/libstdc++ at least, this ultimately "raises" what is meant to be an uncatchable exception in order to cause the stack to unwind and destructors to run. But if anoexceptfunction is in the call stack, then attempting to unwind through that function (with this uncatchable exception in place) would trigger the runtime to think that we had violated our dynamic exception specification, and callstd::terminate()— so even though the program had otherwise exited cleanly, this one background thread was ruining it for everyone.The solution is to be more careful with our
noexceptfunctions.Interestingly, the situation with the windows C/C++ runtime is exactly opposite.