[HIP] Fix duplicate clang -cc1 options on MSVC toolchain

[yaxunl]

Summary

HIPToolChain::TranslateArgs call TranslateArgs of host toolchain with the input args to get a list of derived args called DAL, then
go through the input args by itself and append them to DAL.

This assumes that the host toolchain should not append any unchanged args to DAL, otherwise there will be duplicates since
HIPToolChain will append it again.

This works for GNU toolchain since it returns an empty list for DAL.

However, MSVC toolchain will append unchanged args to DAL, which causes duplicate args.

This patch let MSVC toolchain not append unchanged args for HIP offloading kind, which fixes this issue.

[tra]

LGTM, but I'm not familiar with the details of clang-cl. I've added Reid who'd have a better idea.

[yaxunl]

ping

clang/lib/Driver/ToolChains/MSVC.cpp

static void TranslateDArg(Arg *A, llvm::opt::DerivedArgList &DAL,
 
   std::string NewVal = std::string(Val);
   NewVal[Hash] = '=';
   DAL.AddJoinedArg(A, Opts.getOption(options::OPT_D), NewVal);
 }
 
 llvm::opt::DerivedArgList *
 MSVCToolChain::TranslateArgs(const llvm::opt::DerivedArgList &Args,
-                             StringRef BoundArch, Action::OffloadKind) const {
+                             StringRef BoundArch,
+                             Action::OffloadKind OFK) const {
   DerivedArgList *DAL = new DerivedArgList(Args.getBaseArgs());
   const OptTable &Opts = getDriver().getOpts();
 
   // /Oy and /Oy- don't have an effect on X86-64
   bool SupportsForcingFramePointer = getArch() != llvm::Triple::x86_64;
 
   // The -O[12xd] flag actually expands to several flags.  We must desugar the
   // flags so that options embedded can be negated.  For example, the '-O2' flag
MSVCToolChain::TranslateArgs(const llvm::opt::DerivedArgList &Args,
   for (Arg *A : Args) {
     if (A->getOption().matches(options::OPT__SLASH_O)) {
       // The -O flag actually takes an amalgam of other options.  For example,
       // '/Ogyb2' is equivalent to '/Og' '/Oy' '/Ob2'.
       TranslateOptArg(A, *DAL, SupportsForcingFramePointer, ExpandChar, Opts);
     } else if (A->getOption().matches(options::OPT_D)) {
       // Translate -Dfoo#bar into -Dfoo=bar.
       TranslateDArg(A, *DAL, Opts);
-    } else {
+    } else if (OFK != Action::OFK_HIP) {
+      // HIP Toolchain translates input args by itself.
       DAL->append(A);
     }
   }
 
   return DAL;
 }

clang/test/Driver/hip-options.hip

 // REQUIRES: clang-driver
 // REQUIRES: x86-registered-target
 // REQUIRES: amdgpu-registered-target
 
 // RUN: %clang -### -x hip --gpu-max-threads-per-block=1024 %s 2>&1 | FileCheck %s
 
 // Check that there are commands for both host- and device-side compilations.
 //
 // CHECK: clang{{.*}}" "-cc1" {{.*}} "-fcuda-is-device"
 // CHECK-SAME: "--gpu-max-threads-per-block=1024"
+
+// RUN: %clang -### -x hip -target x86_64-pc-windows-msvc -fms-extensions \
+// RUN:   -mllvm -amdgpu-early-inline-all=true  %s 2>&1 | \
+// RUN:   FileCheck -check-prefix=MLLVM %s
+// MLLVM-NOT: "-mllvm"{{.*}}"-amdgpu-early-inline-all=true"{{.*}}"-mllvm"{{.*}}"-amdgpu-early-inline-all=true"