Skip to content

Conversation

RKSimon
Copy link
Collaborator

@RKSimon RKSimon commented Jul 17, 2025

Possible fix for failed fold in #148191

…des can't create undef/poison

Possible fix for failed fold in llvm#148191
case AArch64ISD::MOVIedit:
case AArch64ISD::MOVImsl:
case AArch64ISD::MOVIshift:
case AArch64ISD::MVNIshift:
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

In fact I think these can be put into a AArch64TargetLowering::isGuaranteedNotToBeUndefOrPoisonForTargetNode implementation, but this should have the same effect

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

That sounds OK to me, they should never be poison. Do we have a standard way of creating tests for changes like this?

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It might be that #148191 is the first case where we manage to show the problem (we need a case where the freeze needs removing quite late into lowering) - I haven't had time to investigate though.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants