Rewrite StackNesting to use a single-pass algorithm
#84688
Merged
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
Contributor
Author
|
@swift-ci Please test |
eeckstein
reviewed
Oct 6, 2025
rjmccall
force-pushed
the
stack-nesting-2
branch
2 times, most recently
from
19b34ee to
30df880
Compare
Contributor
Author
|
@swift-ci Please test |
Contributor
|
@rjmccall can you please
|
Contributor
Author
I saw your comment before, and I'm intending to write a lot more tests, yes. Right now I'm still just trying to get this thing working. |
Contributor
Author
|
@swift-ci Please test windows |
1 similar comment
Contributor
Author
|
@swift-ci Please test windows |
rjmccall
force-pushed
the
stack-nesting-2
branch
from
94538b0 to
0b5853e
Compare
Contributor
Author
|
@swift-ci Please test |
Contributor
Author
|
Now with the tweaked algorithm and a rather over-pedantic proof. Still needs benchmarking and (mostly) testing |
rjmccall
force-pushed
the
stack-nesting-2
branch
from
0b5853e to
203f0f7
Compare
Contributor
Author
|
@swift-ci Please test |
Contributor
Author
|
@swift-ci Please test compiler performance |
Contributor
Author
|
Well, the performance run seems to show no drastic regressions, and that Windows failure seems unrelated, so I just need to write tests. |
The previous algorithm was doing an iterative forward data flow analysis followed by a reverse data flow analysis. I suspect the history here is that it was a reverse analysis, and that didn't really work for infinite loops, and so complexity accumulated. The new algorithm is quite straightforward and relies on the allocations being properly jointly post-dominated, just not nested. We simply walk forward through the blocks in consistent-with-dominance order, maintaining the stack of active allocations and deferring deallocations that are improperly nested until we deallocate the allocations above it. The only real subtlety is that we have to delay walking into dead-end regions until we've seen all of the edges into them, so that we can know whether we have a coherent stack state in them. If the state is incoherent, we need to remove any deallocations of previous allocations because we cannot talk correctly about what's on top of the stack. The reason I'm doing this, besides it just being a simpler and hopefully faster algorithm, is that modeling some of the uses of the async stack allocator properly requires builtins that cannot just be semantically reordered. That should be somewhat easier to handle with the new approach, although really (1) we should not have runtime functions that need this and (2) we're going to need a conservatively-correct solution that's different from this anyway because hoisting allocations is *also* limited in its own way. I've attached a rather pedantic proof of the correctness of the algorithm. The thing that concerns me most about the rewritten pass is that it isn't actually validating joint post-dominance on input, so if you give it bad input, it might be a little mystifying to debug the verifier failures.
rjmccall
force-pushed
the
stack-nesting-2
branch
from
203f0f7 to
8d231d2
Compare
Contributor
Author
|
@swift-ci Please test |
Contributor
Author
|
@eeckstein Tests added, ready for final review. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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 previous algorithm was doing an iterative forward data flow analysis followed by a reverse data flow analysis. I suspect the history here is that it was a reverse analysis, and that didn't really work for infinite loops, and so complexity kindof accumulated.
The new algorithm is quite straightforward and relies on the allocations being properly jointly post-dominated, just not nested. We simply walk forward through the blocks in consistent-with-dominance order, maintaining the stack of active allocations and deferring deallocations that are improperly nested until we deallocate the allocations above it.
The reason I'm doing this, besides it just being a simpler, faster algorithm, is that modeling some of the uses of the async stack allocator properly requires builtins that cannot just be semantically reordered. That should be somewhat easier to handle with the new approach, although really (1) we should not have runtime functions that need this and (2) we're going to need a conservatively-correct solution that's different from this anyway because hoisting allocations is also limited in its own way.
The test cases that changed are... I don't think the new output is wrong under the current rules that are being enforced, but really we should be enforcing different rules, because it's not really okay to have broken stack nesting in blocks just because they don't lead to an exit. But it's broken coming into StackNesting, and I don't think the rewrite actually makes it worse, so...
The thing that concerns me most about the rewritten pass is that it isn't actually validating joint post-dominance on input, so if you give it bad input, it might be a little mystifying to debug the verifier failures.