Fix optimization of seqcst RMWs in OptimizeInstructions #7280
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
We previously reasoned that RMWs that do not change in-memory values could be optimized to just atomic gets because they did not necessarily synchronize with subsequent atomic reads of the same location. This is not correct for seqcst RMWs, though, since subsequent seqcst reads in the total order of seqcst operations are _required_ to read from and synchronize with these RMWs, even though they do not change the in-memory values. Fix the optimization to only work on acquire-release RMWs and RMWs to unshared memory.
kripken
approved these changes
Feb 7, 2025
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Outdated suggestions cannot be applied.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.
We previously reasoned that RMWs that do not change in-memory values
could be optimized to just atomic gets because they did not necessarily
synchronize with subsequent atomic reads of the same location. This is
not correct for seqcst RMWs, though, since subsequent seqcst reads in
the total order of seqcst operations are required to read from and
synchronize with these RMWs, even though they do not change the
in-memory values. Fix the optimization to only work on acquire-release
RMWs and RMWs to unshared memory.