-
Notifications
You must be signed in to change notification settings - Fork 171
New issue
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
Do Not Cache PID/TID in Logging #994
Conversation
This pull request was exported from Phabricator. Differential Revision: D63668265 |
This pull request was exported from Phabricator. Differential Revision: D63668265 |
Summary: Pull Request resolved: pytorch#994 S451588 was caused by the LOG macro caching values that would then be copied to other processes via forking. In general, we should probably use fork handlers to clear out said variables, but from a hygiene point of view we should also not be changing control flow based on logging. For this reason, the pid/tid retrieval in logging should get the cached variable if it exists, but never do the caching itself. Differential Revision: D63668265
28b7988
to
a2f124e
Compare
This pull request was exported from Phabricator. Differential Revision: D63668265 |
Summary: Pull Request resolved: pytorch#994 S451588 was caused by the LOG macro caching values that would then be copied to other processes via forking. In general, we should probably use fork handlers to clear out said variables, but from a hygiene point of view we should also not be changing control flow based on logging. For this reason, the pid/tid retrieval in logging should get the cached variable if it exists, but never do the caching itself. Differential Revision: D63668265
a2f124e
to
26602c9
Compare
This pull request was exported from Phabricator. Differential Revision: D63668265 |
Summary: Pull Request resolved: pytorch#994 S451588 was caused by the LOG macro caching values that would then be copied to other processes via forking. In general, we should probably use fork handlers to clear out said variables, but from a hygiene point of view we should also not be changing control flow based on logging. For this reason, the pid/tid retrieval in logging should get the cached variable if it exists, but never do the caching itself. Differential Revision: D63668265
26602c9
to
a39cec1
Compare
Summary: Pull Request resolved: pytorch#994 S451588 was caused by the LOG macro caching values that would then be copied to other processes via forking. In general, we should probably use fork handlers to clear out said variables, but from a hygiene point of view we should also not be changing control flow based on logging. For this reason, the pid/tid retrieval in logging should get the cached variable if it exists, but never do the caching itself. Differential Revision: D63668265
This pull request was exported from Phabricator. Differential Revision: D63668265 |
a39cec1
to
a922a48
Compare
This pull request has been merged in 78737f1. |
Summary: S451588 was caused by the LOG macro caching values that would then be copied to other processes via forking. In general, we should probably use fork handlers to clear out said variables, but from a hygiene point of view we should also not be changing control flow based on logging. For this reason, the pid/tid retrieval in logging should get the cached variable if it exists, but never do the caching itself.
Differential Revision: D63668265