[IRGenDebugInfo] Look through pointer casts #66409
Closed
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
We decide whether or not to generate shadow copies based on whether an instruction is an Alloca or not. We should ignore all pointer casts when doing so, otherwise we may artificially reach the conclusion that the instruction does not represent an alloca. Furthermore, once we reach the alloca, we should also check if it is a static alloca: in that case, no shadow copy is needed. It may look like this change is not tested, but it is. We were forgetting to do two things (strip ptr casts + static alloca check). Fixing only one of those would introduce test failures. Fixing both maintain the existing test behavior.
When deciding whether to generate dbg.declare or dbg.values, we check whether the variable "storage" is an alloca or not. But this fails to account for no-op pointer casts. Some of the tests I updated were expecting really weird behavior. For example, debug_info_expression was generating an alloca + dbg declare, followed by a debug value for the same variable.
|
@swift-ci test |
aschwaighofer
added a commit
to aschwaighofer/swift
that referenced
this pull request
Jun 14, 2023
Based on swiftlang#66409 With the observation that the pre-opaque world was using bitcast as an indication that the storage type and the type of the variable were different. We can recover this information from the storage type of the alloca and the storage type of the type info.
aschwaighofer
added a commit
to aschwaighofer/swift
that referenced
this pull request
Jun 14, 2023
Based on swiftlang#66409 With the observation that the pre-opaque world was using bitcast as an indication that the storage type and the type of the variable were different. We can recover this information from the storage type of the alloca and the storage type of the type info.
|
@aschwaighofer merged a variant of this: #66077 |
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 PR contains two commits, both fixing the same issue in different parts of the code base: we fail to look through pointer casts when making codegen decisions. These were exposed by the opaque pointer work, since in this mode most pointer casts are gone.