Repository navigation
aarch64: fix the operand order of the data processing (1 source) instructions - #1546
Merged
Redhpm merged 1 commit intoOct 30, 2026
Merged
Conversation
Contributor
Author
|
Closing as a duplicate of #1527, which makes the same change to The only addition here was a set of regression vectors in |
Contributor
|
I prefer adding tests, so I'll merge yours |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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 aarch64 data processing (1 source) instructions (
RBIT,REV16,REV,REV32,CLZ,CLS) are defined without an explicit argument list, so their operands are taken in field order,rn, rd. Miasm therefore shows and lifts them with the source and destination swapped:and the IR writes the result to the source register. The two definitions now list the arguments as
[rd, rn], like the other instructions; the semantics already treat the first argument as the destination.Added
CLZ,CLS,RBIT,REV16,REV(W and X), andREV32vectors totest/arch/aarch64/arch.py, with encodings checked againstllvm-mc. On master the test fails at the first one (CLZ W3, W7is disassembled asCLZ W7, W3); with this change it passes.Also compared the python jitter with Unicorn 2.1.4 on these instructions:
CLZ,RBIT,REV16and 32-bitREVnow give the same results. The aarch64 unit scripts pass with the python and gcc jitters.