Use atomics for call cache #14338
Open
Use atomics for call cache #14338
+53
−32
Conversation
This file contains hidden or 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
The same CME may be reused in multiple caches, so once the CACHED flag is set, other Ractors may be checking for it being invalidated and we should avoid writing to the field again.
❌ Tests Failed✖️no tests failed ✔️62204 tests passed(1 flake) |
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Outdated suggestions cannot be applied.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.
The stores and loads of inline call caches and the GCCCT need to be atomic, as they are written and read concurrently by different Ractors.
Previously we were relying on just doing pointer-sized loads, which will be essentially sufficient on x86's memory model, but on other architectures (arm, powerpc) the writes to the CC could be reordered (the writes to klass,
cme_,call_, etc could return previous or uninitialized values).To solve this we should use atomics with release/acquire semantics when writing the CC object to the cache. This will emit identical bytecode on x86, but will be different on ex. ARM (
ldarinstead ofldrhttps://godbolt.org/z/deheaxMsx).This also changes updates to
cc->call_to atomics. For these I used "relaxed" ordering since the value is a function pointer, which should not need "publishing".