[RISCV] Support -mno-implicit-float.

This can be used to disable vectorization and memcpy/memset
expansion for things like OS kernels. It also disables implicit
uses of scalar FP, but I don't know if we have any of those for
RISC-V.

NOTE: Without this patch you can still do -Xclang -no-implicit-float

Reviewed By: rui.zhang

Differential Revision: https://reviews.llvm.org/D134077

diff --git a/clang/lib/Driver/ToolChains/Clang.cpp b/clang/lib/Driver/ToolChains/Clang.cpp
index 7af0f99..fb64722 100644 (file)
--- a/clang/lib/Driver/ToolChains/Clang.cpp
+++ b/clang/lib/Driver/ToolChains/Clang.cpp
@@ -2194,6 +2194,10 @@ void Clang::AddRISCVTargetArgs(const ArgList &Args,
 
   SetRISCVSmallDataLimit(getToolChain(), Args, CmdArgs);
 
+  if (!Args.hasFlag(options::OPT_mimplicit_float,
+                    options::OPT_mno_implicit_float, true))
+    CmdArgs.push_back("-no-implicit-float");
+
   if (const Arg *A = Args.getLastArg(options::OPT_mtune_EQ)) {
     CmdArgs.push_back("-tune-cpu");
     CmdArgs.push_back(A->getValue());

diff --git a/clang/test/Driver/flags.c b/clang/test/Driver/flags.c
index 656ac41..da25a5c 100644 (file)
--- a/clang/test/Driver/flags.c
+++ b/clang/test/Driver/flags.c
@@ -27,3 +27,9 @@
 
 // RUN: %clang -target x86_64-linux-gnu -### -c -fclang-abi-compat=3.2 %s 2>&1 | FileCheck -check-prefix=TEST9 %s
 // TEST9: "-fclang-abi-compat=3.2"
+//
+// RUN: %clang -target riscv32 -### -S -mno-implicit-float %s 2>&1 | FileCheck -check-prefix=TEST10 %s
+// TEST10: "-no-implicit-float"
+//
+// RUN: %clang -target riscv64 -### -S -mno-implicit-float %s 2>&1 | FileCheck -check-prefix=TEST11 %s
+// TEST11: "-no-implicit-float"