[clang] [llvm] [Clang][AArch64] Implement widening FMMLA intrinsics (PR #165282)

[Eli Friedman via cfe-commits]

efriedma-quic wrote:

Has there been any further discussion about the semantics of llvm.aarch64.set.fpmr?  It looks like there was some discussion in #114248, but it wasn't ever really resolved.  It seems problematic to allow users to explicitly control the state at the IR level, as opposed to letting the compiler configure it in the backend.

https://github.com/llvm/llvm-project/pull/165282