This repository has been archived by the owner on Jul 1, 2023. It is now read-only.
Add test that stresses races during transport shutdown #250
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
|
This pull request was exported from Phabricator. Differential Revision: D25495683 |
lw
added a commit
to lw/tensorpipe
that referenced
this pull request
Dec 12, 2020
Summary: Pull Request resolved: pytorch#250 This test wants to do two things: defer functions to the loop while the context is shutting down, and create new objects (connections and listeners) before and/or just after the closing. Both these operations interfere with a correct shutdown, and are tricky to handle. Transports however don't offer a way to directly defer functions, so we achieve it by attaching a read callback to the connection, and causing it to be called immediately by closing the connection. Transports also don't really allow to control timing of when something that is deferred to the loop will really run. To "work around" it we simply jam the context by creating new connections over and over and over at an insane rate. This test proved its worth by finding many issues in transport shutdown. Differential Revision: D25495683 fbshipit-source-id: 9cfe593c27681952a32bced2e7afa073c5668980
Differential Revision: D25495685 fbshipit-source-id: 8ef974338cf363fc69189356e886d787547693ee
Differential Revision: D25495682 fbshipit-source-id: 00e62ba390eb2e7d88ff92392d8fa645b0d7a1f3
…ers initing Differential Revision: D25495684 fbshipit-source-id: e074c2f45c046b791da23b9eeb256495b316db6b
Differential Revision: D25495681 fbshipit-source-id: c941fb1bb1a1d49a12c750c0c7b874b09dc6a5b2
Summary: And thus retire the ClosingEmitter/Receiver from transports. Differential Revision: D25495914 fbshipit-source-id: 0075c2db8fa804e6584c23b9519bfc4d373ac307
Summary: Pull Request resolved: pytorch#250 This test wants to do two things: defer functions to the loop while the context is shutting down, and create new objects (connections and listeners) before and/or just after the closing. Both these operations interfere with a correct shutdown, and are tricky to handle. Transports however don't offer a way to directly defer functions, so we achieve it by attaching a read callback to the connection, and causing it to be called immediately by closing the connection. Transports also don't really allow to control timing of when something that is deferred to the loop will really run. To "work around" it we simply jam the context by creating new connections over and over and over at an insane rate. This test proved its worth by finding many issues in transport shutdown. Differential Revision: D25495683 fbshipit-source-id: 6cb4a01385bd23811457a99afab2915425d5ecd8
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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.
Summary:
This test wants to do two things: defer functions to the loop while the context is shutting down, and create new objects (connections and listeners) before and/or just after the closing. Both these operations interfere with a correct shutdown, and are tricky to handle.
Transports however don't offer a way to directly defer functions, so we achieve it by attaching a read callback to the connection, and causing it to be called immediately by closing the connection.
Transports also don't really allow to control timing of when something that is deferred to the loop will really run. To "work around" it we simply jam the context by creating new connections over and over and over at an insane rate.
This test proved its worth by finding many issues in transport shutdown.
Differential Revision: D25495683