[PATCH] D91732: [compiler-rt] [builtins] Support conversion between fp16 and fp128

[Adhemerval Zanella via Phabricator via llvm-commits]

zatrazz added a comment.

In D91732#2404933 <https://reviews.llvm.org/D91732#2404933>, @ostannard wrote:

> LGTM
>
>> Also, __fp16 is a storage format and promoted to 'float' for argument passing
>
> It doesn't affect this patch, but for all of the ARM and AArch64 ABIs, __fp16 arguments are passed in the bottom half of an integer or FP register, not promoted to float. There was an old version of the AArch32 ABI (I think) which did that, but I don't think any compiler implemented it before it was changed.

Right, I think I read on an older ACLE about it, but it does seem the newer version change it.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D91732/new/

https://reviews.llvm.org/D91732