[PATCH] D43106: [RISCV] Enable -fforce-enable-int128 through cmake flag COMPILER_RT_HAS_FINT128_FLAG

[Mandeep Singh Grang via Phabricator via llvm-commits]

mgrang added a comment.

In https://reviews.llvm.org/D43106#1017081, @kito-cheng wrote:

> Hi all:
>
> I propose it should check  in the `foreach (arch ${BUILTIN_SUPPORTED_ARCH})`, because COMPILER_RT_DEFAULT_TARGET_ARCH is not always set (I only set COMPILER_RT_DEFAULT_TARGET_TRIPLE instead of COMPILER_RT_DEFAULT_TARGET_ARCH in my build).
>
> And that place seems more like right place to put target specific compilation flags.
>
>   diff --git a/lib/builtins/CMakeLists.txt b/lib/builtins/CMakeLists.txt
>   index 991108b..364a4f6 100644
>   --- a/lib/builtins/CMakeLists.txt
>   +++ b/lib/builtins/CMakeLists.txt
>   @@ -546,6 +546,12 @@ else ()
>            list(APPEND BUILTIN_CFLAGS -fomit-frame-pointer -DCOMPILER_RT_ARMHF_TARGET)
>          endif()
>   
>   +      # For RISC-V 32 bits, we must force enable int128 for compiling long
>   +      # double routines.
>   +      if("${arch}" STREQUAL "riscv32")
>   +        list(APPEND BUILTIN_CFLAGS -fforce-enable-int128)
>   +      endif()
>   +
>          add_compiler_rt_runtime(clang_rt.builtins
>                                  STATIC
>                                  ARCHS ${arch}
>


So, do we want to enable this flag only based on a CMake macro like COMPILER_RT_HAS_FINT128_FLAG, or always add -fforce-enable-int128 to compiler-rt builds for RISCV32?


Repository:
  rCRT Compiler Runtime

https://reviews.llvm.org/D43106