[OpenMP] Make OpenMP generated code for the NVIDIA device relocatable by default

Original Diff: D29642

This patch was previously reverted due to an error with patch D29654
that this depends on.

llvm-svn: 310479

diff --git a/clang/lib/Driver/ToolChains/Cuda.cpp b/clang/lib/Driver/ToolChains/Cuda.cpp
index 70f472f..927704a 100644 (file)
--- a/clang/lib/Driver/ToolChains/Cuda.cpp
+++ b/clang/lib/Driver/ToolChains/Cuda.cpp
@@ -291,6 +291,10 @@ void NVPTX::Assembler::ConstructJob(Compilation &C, const JobAction &JA,
   for (const auto& A : Args.getAllArgValues(options::OPT_Xcuda_ptxas))
     CmdArgs.push_back(Args.MakeArgString(A));
 
+  // In OpenMP we need to generate relocatable code.
+  if (JA.isOffloading(Action::OFK_OpenMP))
+    CmdArgs.push_back("-c");
+
   const char *Exec;
   if (Arg *A = Args.getLastArg(options::OPT_ptxas_path_EQ))
     Exec = A->getValue();

diff --git a/clang/test/Driver/openmp-offload.c b/clang/test/Driver/openmp-offload.c
index e9ad2dc..af50d15 100644 (file)
--- a/clang/test/Driver/openmp-offload.c
+++ b/clang/test/Driver/openmp-offload.c
@@ -687,3 +687,11 @@
 // RUN:   | FileCheck -check-prefix=CHK-TWOCUBIN-DARWIN %s
 
 // CHK-TWOCUBIN-DARWIN: nvlink"{{.*}}"openmp-offload-{{.*}}.cubin" "openmp-offload-{{.*}}.cubin"
+
+/// ###########################################################################
+
+/// Check PTXAS is passed -c flag when offloading to an NVIDIA device using OpenMP.
+// RUN:   %clang -### -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -no-canonical-prefixes %s 2>&1 \
+// RUN:   | FileCheck -check-prefix=CHK-PTXAS-DEFAULT %s
+
+// CHK-PTXAS-DEFAULT: ptxas{{.*}}" "-c"