Handle up to 2^16 - 1 operands per inst, and error properly on more. #226
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Outdated suggestions cannot be applied.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 addresses an underlying limitation that became a problem for Cranelift recently, after folding return-value loads into call pseudoinsts (thus producing single instructions with many operands): we only supported 255 operands per instruction.
It appears that
Usewas using au8for the "slot index" for an operand as a result of earlier optimization/packing efforts. However, the current shape of the struct leaves a padding byte free, so we should be able to expand to au16with no loss in performance.Furthermore, previously we weren't catching the overflow, but rather were silently wrapping around (eek). This properly returns a
RegAllocError::TooManyOperandsnow if an instruction has too many operands (more thanu16::MAX).