[clang] [llvm] [PowerPC] fix bug affecting float to int32 conversion on LE PowerPC (PR #150194)

[Nikita Popov via llvm-commits]

nikic wrote:

> Found another related bug: soft float ABI selection was not taking effect on little-endian powerPC with embedded vectors (e.g. e500v2) leading to errors. (embedded vectors use "extended" GPRs to store floating-point values, and this caused issues with variadic arguments assuming dedicated floating-point registers with hard-float ABI)

Please create a separate PR for this, as it seems to be completely unrelated to the original issue this PR is solving.

https://github.com/llvm/llvm-project/pull/150194