[PATCH] D58233: Allow replacing intrinsic operands with variables

Matt Arsenault via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Fri Feb 15 06:57:16 PST 2019

arsenm added a comment.

In D58233#1399368 <https://reviews.llvm.org/D58233#1399368>, @arsenm wrote:

> In D58233#1399047 <https://reviews.llvm.org/D58233#1399047>, @craig.topper wrote:
> > What protects target intrinsics that haven't been annotated with immarg yet?
> Nothing, these need to be accurate. I looked through the current set of in-tree target intrinsics, but don't see any comments or documentation stating certain arguments need to be constant so I hope we are OK there.

It looks like this is sort of indirectly documented for the cases that have GCCBuiltins handled in clang. It's probably possible to put an assert somewhere in clang to find any of the intrinsics using the "I" modifier that don't have the matching immarg attribute. This of course won't really give 100% coverage though



More information about the llvm-commits mailing list