Martin Storsjö

I understand what's happening now, but I don't yet know what to do about it. CC @cjacek who has some experience with the LLD internals. When doing LTO linking, we...

As this issue wasn't quite trivial, I think it would be good to open an issue under https://github.com/llvm/llvm-project/issues where we can bring in more people to have opinions on it....

A lot of fixes for unhandled instructions have been merged quite recently; do you get the same issues still with e.g. the latest nightly builds? https://github.com/mstorsjo/llvm-mingw/releases/nightly

CC @bernhardu

> But as far as I see the instruction `44 0f b6 1a` got added in [this patch](https://github.com/llvm/llvm-project/commit/ce4618a9c405bd8a9c1e096eb45e9ca83d3891f1) to main branch four month ago, which mentions some relation to "the...

> But nowadays llvm-mingw is not gcc-compatible anyway Can you quantify what you refer to here? llvm-mingw is generally ABI compatible with GCC. C++ code that refers to the standard...

> If you want the performance of 64 bit doubles, use `double`, not `long double`. To provide more context on this; e.g. on aarch64 linux, `long double` is actually 128...

> > Can you quantify what you refer to here? llvm-mingw is generally ABI compatible with GCC. C++ code that refers to the standard C++ library isn't compatible across C++...

> The latest MinGW seems to be working with -mlong-double-64. The things that are merged so far is not everything you need for working `-mlong-double-64`, a couple other things are...

> Thanks! Yes, since I'm targeting an x86-64-v3 toolchain, I've always just considered UCRT and built the whole toolchain from scratch. Right, ok. As long as you include the `-mlong-double-64`...