lowering: Optimize lowering of tryfinally with trivial finally block #52593
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
Keno
added a commit
that referenced
this pull request
Dec 22, 2023
Third time's the charm hopefully. This builds on #52527 and addresses the TODO left there by keeping track of the appropriate :scope in sroa as we go along in order to ensure correct current_scope folding, even when the try body is nothrow (which was left as a todo). As a result, we also no longer result on lowering assumptions for this transformation, which enables us to enable the lowering optimization in #52593.
Keno
added a commit
that referenced
this pull request
Dec 23, 2023
Third time's the charm hopefully. This builds on #52527 and addresses the TODO left there by keeping track of the appropriate :scope in sroa as we go along in order to ensure correct current_scope folding, even when the try body is nothrow (which was left as a todo). As a result, we also no longer result on lowering assumptions for this transformation, which enables us to enable the lowering optimization in #52593.
Keno
force-pushed
the
kf/tryfinallylowering
branch
from
e3f2715
to
23d0318
Compare
This optimizes the lowering of a tryfinally block with empty finally block to instead use the try/catch lowering, where the catch block is given as simply `rethrow()`. This is equivalent semantically to try/finally in this case, but the code structure is a lot simpler with fewer basic blocks and without the auxiliary slot for tracking the finally slot. The motivation here is to help the compiler optimize better when using the `@with` macro, which has an empty `finally` block (but uses the scope argument of 'tryfinally).
Keno
force-pushed
the
kf/tryfinallylowering
branch
from
23d0318
to
3121abf
Compare
aviatesk
reviewed
Dec 24, 2023
| function with_macro_slot_cross() | ||
| a = 1 | ||
| @with sval=>1 begin | ||
| a = sval_float[] |
There was a problem hiding this comment.
This is intended to be a = sval[]? It's a minor issue, so I believe it can be addressed alongside other upcoming changes.
There was a problem hiding this comment.
No, this is as intended, the test is for the value of the slot a across the @with region and I didn't want it to fold
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.
This optimizes the lowering of a tryfinally block with empty finally block to instead use the try/catch lowering, where the catch block is given as simply
rethrow(). This is equivalent semantically to try/finally in this case, but the code structure is a lot simpler with fewer basic blocks and without the auxiliary slot for tracking the finally slot. The motivation here is to help the compiler optimize better when using the@withmacro, which has an emptyfinallyblock (but uses the scope argument of 'tryfinally). The only problem with this is that it violates the lowering assumptions I made in #52527, so we'll probably need to fix that first.