| FazBrowse GitHub Viewer | Trending | | Home |
| Tools: [Download Repo ZIP] [Original HTTPS Page] |
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.
| Back | FazBrowse Home | New Git URL |
Uh oh!
There was an error while loading. Please reload this page.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Choose a reason Spam Abuse Off Topic Outdated Duplicate Resolved Low QualityIs relaxed strong enough? Do any consumers need to see the updates that are sequenced-before this store? Same comment applies to the below instances of using relaxed stores to update bytecode.
Sorry, something went wrong.
Uh oh!
There was an error while loading. Please reload this page.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Choose a reason Spam Abuse Off Topic Outdated Duplicate Resolved Low QualityIn the deinstrumented case if we race and still see the instrumented opcode the data will still be there. In the instrumented case I think we do need some more synchronization around the tooling metadata, if we see the instrumented opcode before seeing the writes to the data that'll be bad.
I had originally had synchronizeds reads/writes around that in d8ea6bb#diff-adaefb7da847260ef7aff6e66007bc83f5b226c5389a23e1c9ea622c3b02c419 but the thought was that the synchronization around _PyFrame_GetCode(frame)->_co_instrumentation_version would be sufficient (which is in the latest version of #116775).
But that doesn't actually seem to be good enough - there's still some rare TSAN failures around the instrumentation data. But my plan is to push the synchronization there instead of on the opcode because I don't think we want to do acquire loads on the byte code which would be required to make a release write here meaningful.
Sorry, something went wrong.
Uh oh!
There was an error while loading. Please reload this page.