r276979 - [OpenMP][CUDA] Do not forward OpenMP flags for CUDA device actions.

[Samuel Antao via cfe-commits]

Author: sfantao
Date: Thu Jul 28 09:29:18 2016
New Revision: 276979

URL: http://llvm.org/viewvc/llvm-project?rev=276979&view=rev
Log:
[OpenMP][CUDA] Do not forward OpenMP flags for CUDA device actions.

Summary:
This patch prevents OpenMP flags from being forwarded to CUDA device commands. That was causing the CUDA frontend to attempt to emit OpenMP code which is not supported.

This fixes the bug reported in https://llvm.org/bugs/show_bug.cgi?id=28723.

Reviewers: hfinkel, carlo.bertolli, arpith-jacob, kkwli0, tra, ABataev

Subscribers: caomhin, cfe-commits

Differential Revision: https://reviews.llvm.org/D22895

Added:
    cfe/trunk/test/Driver/offloading-interoperability.c
Modified:
    cfe/trunk/lib/Driver/Tools.cpp

Modified: cfe/trunk/lib/Driver/Tools.cpp
URL: http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/Driver/Tools.cpp?rev=276979&r1=276978&r2=276979&view=diff
==============================================================================
--- cfe/trunk/lib/Driver/Tools.cpp (original)
+++ cfe/trunk/lib/Driver/Tools.cpp Thu Jul 28 09:29:18 2016
@@ -5008,9 +5008,13 @@ void Clang::ConstructJob(Compilation &C,
   Args.AddLastArg(CmdArgs, options::OPT_fdiagnostics_show_template_tree);
   Args.AddLastArg(CmdArgs, options::OPT_fno_elide_type);
 
-  // Forward flags for OpenMP
+  // Forward flags for OpenMP. We don't do this if the current action is an
+  // device offloading action.
+  //
+  // TODO: Allow OpenMP offload actions when they become available.
   if (Args.hasFlag(options::OPT_fopenmp, options::OPT_fopenmp_EQ,
-                   options::OPT_fno_openmp, false)) {
+                   options::OPT_fno_openmp, false) &&
+      JA.isDeviceOffloading(Action::OFK_None)) {
     switch (getOpenMPRuntime(getToolChain(), Args)) {
     case OMPRT_OMP:
     case OMPRT_IOMP5:

Added: cfe/trunk/test/Driver/offloading-interoperability.c
URL: http://llvm.org/viewvc/llvm-project/cfe/trunk/test/Driver/offloading-interoperability.c?rev=276979&view=auto
==============================================================================
--- cfe/trunk/test/Driver/offloading-interoperability.c (added)
+++ cfe/trunk/test/Driver/offloading-interoperability.c Thu Jul 28 09:29:18 2016
@@ -0,0 +1,17 @@
+// REQUIRES: clang-driver
+// REQUIRES: powerpc-registered-target
+// REQUIRES: nvptx-registered-target
+
+//
+// Verify that CUDA device commands do not get OpenMP flags.
+//
+// RUN: %clang -### -x cuda -target powerpc64le-linux-gnu -std=c++11 --cuda-gpu-arch=sm_35 -fopenmp %s 2>&1 \
+// RUN: | FileCheck %s --check-prefix NO-OPENMP-FLAGS-FOR-CUDA-DEVICE
+//
+// NO-OPENMP-FLAGS-FOR-CUDA-DEVICE:      clang{{.*}}" "-cc1" "-triple" "nvptx64-nvidia-cuda"
+// NO-OPENMP-FLAGS-FOR-CUDA-DEVICE-NOT:  -fopenmp
+// NO-OPENMP-FLAGS-FOR-CUDA-DEVICE-NEXT: ptxas" "-m64"
+// NO-OPENMP-FLAGS-FOR-CUDA-DEVICE-NEXT: fatbinary" "--cuda" "-64"
+// NO-OPENMP-FLAGS-FOR-CUDA-DEVICE-NEXT: clang{{.*}}" "-cc1" "-triple" "powerpc64le--linux-gnu"
+// NO-OPENMP-FLAGS-FOR-CUDA-DEVICE:      -fopenmp
+// NO-OPENMP-FLAGS-FOR-CUDA-DEVICE-NEXT: ld" "-z" "relro" "--hash-style=gnu" "--eh-frame-hdr" "-m" "elf64lppc"