[CIR][ABI][Lowering] Fixes calls with union type #1119
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
|
✅ With the latest revision this PR passed the C/C++ code formatter. |
bcardosolopes
requested changes
Nov 13, 2024
clang/lib/CIR/Dialect/Transforms/TargetLowering/LowerFunction.cpp
Outdated
Show resolved
Hide resolved
bcardosolopes
approved these changes
Nov 14, 2024
lanza
pushed a commit
that referenced
this pull request
Mar 18, 2025
This PR handles calls with unions passed by value in the calling convention pass. #### Implementation As one may know, data layout for unions in CIR and in LLVM differ one from another. In CIR we track all the union members, while in LLVM IR only the largest one. And here we need to take this difference into account: we need to find a type of the largest member and treat it as the first (and only) union member in order to preserve all the logic from the original codegen. There is a method `StructType::getLargestMember` - but looks like it produces different results (with the one I implemented or better to say copy-pasted). Maybe it's done intentionally, I don't know. The LLVM IR produced has also some difference from the original one. In the original IR `gep` is emitted - and we can not do the same. If we create `getMemberOp` we may fail on type checking for unions - since the first member type may differ from the largest type. This is why we create `bitcast` instead. Relates to the issue #1061
terapines-osc-cir
pushed a commit
to Terapines/clangir
that referenced
this pull request
Sep 2, 2025
This PR handles calls with unions passed by value in the calling convention pass. #### Implementation As one may know, data layout for unions in CIR and in LLVM differ one from another. In CIR we track all the union members, while in LLVM IR only the largest one. And here we need to take this difference into account: we need to find a type of the largest member and treat it as the first (and only) union member in order to preserve all the logic from the original codegen. There is a method `StructType::getLargestMember` - but looks like it produces different results (with the one I implemented or better to say copy-pasted). Maybe it's done intentionally, I don't know. The LLVM IR produced has also some difference from the original one. In the original IR `gep` is emitted - and we can not do the same. If we create `getMemberOp` we may fail on type checking for unions - since the first member type may differ from the largest type. This is why we create `bitcast` instead. Relates to the issue llvm#1061
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 handles calls with unions passed by value in the calling convention pass.
Implementation
As one may know, data layout for unions in CIR and in LLVM differ one from another. In CIR we track all the union members, while in LLVM IR only the largest one.
And here we need to take this difference into account: we need to find a type of the largest member and treat it as the first (and only) union member in order to preserve all the logic from the original codegen.
There is a method
StructType::getLargestMember- but looks like it produces different results (with the one I implemented or better to say copy-pasted). Maybe it's done intentionally, I don't know.The LLVM IR produced has also some difference from the original one. In the original IR
gepis emitted - and we can not do the same. If we creategetMemberOpwe may fail on type checking for unions - since the first member type may differ from the largest type. This is why we createbitcastinstead. Relates to the issue #1061