[PATCH] D125314: [OpenMP] Don't set device runtime debugging flags if using '-nogpulib'

[Joseph Huber via Phabricator via cfe-commits]

jhuber6 created this revision.
jhuber6 added reviewers: jdoerfert, tianshilei1992, ABataev, JonChesterfield.
Herald added subscribers: guansong, yaxunl.
Herald added a project: All.
jhuber6 requested review of this revision.
Herald added subscribers: cfe-commits, sstefan1, MaskRay.
Herald added a project: clang.

We use globals to configure debugging at compile-time for the device
runtime. Because these are only used by the OpenMP runtime we shouldn't
define them if we aren't using the device runtime. When a user passes in
'-nogpulib' this indicates that we are not using the device runtime, so
we should check for the precense of this flag and not emit these globals
if used.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D125314

Files:
  clang/include/clang/Basic/LangOptions.def
  clang/include/clang/Driver/Options.td
  clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp
  clang/lib/Driver/ToolChains/Clang.cpp
  clang/test/OpenMP/target_globals_codegen.cpp


Index: clang/test/OpenMP/target_globals_codegen.cpp
===================================================================
--- clang/test/OpenMP/target_globals_codegen.cpp
+++ clang/test/OpenMP/target_globals_codegen.cpp
@@ -7,6 +7,7 @@
 // RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-assume-threads-oversubscription -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix=CHECK-THREADS
 // RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-assume-teams-oversubscription -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix=CHECK-TEAMS
 // RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-assume-no-thread-state -fopenmp-is-device -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix=CHECK-STATE
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-is-device -nogpulib -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s --check-prefix=CHECK-RUNTIME
 // RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-assume-teams-oversubscription -fopenmp-is-device -o - | FileCheck %s --check-prefix=CHECK-RUNTIME
 // expected-no-diagnostics
 
Index: clang/lib/Driver/ToolChains/Clang.cpp
===================================================================
--- clang/lib/Driver/ToolChains/Clang.cpp
+++ clang/lib/Driver/ToolChains/Clang.cpp
@@ -6272,6 +6272,10 @@
     }
   }
 
+  // Forward -nogpulib to -cc1.
+  if (Args.hasArg(options::OPT_nogpulib))
+    CmdArgs.push_back("-nogpulib");
+
   if (Arg *A = Args.getLastArg(options::OPT_fcf_protection_EQ)) {
     CmdArgs.push_back(
         Args.MakeArgString(Twine("-fcf-protection=") + A->getValue()));
Index: clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp
===================================================================
--- clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp
+++ clang/lib/CodeGen/CGOpenMPRuntimeGPU.cpp
@@ -1203,16 +1203,17 @@
     llvm_unreachable("OpenMP can only handle device code.");
 
   llvm::OpenMPIRBuilder &OMPBuilder = getOMPBuilder();
-  if (!CGM.getLangOpts().OMPHostIRFile.empty()) {
-    OMPBuilder.createGlobalFlag(CGM.getLangOpts().OpenMPTargetDebug,
-                                "__omp_rtl_debug_kind");
-    OMPBuilder.createGlobalFlag(CGM.getLangOpts().OpenMPTeamSubscription,
-                                "__omp_rtl_assume_teams_oversubscription");
-    OMPBuilder.createGlobalFlag(CGM.getLangOpts().OpenMPThreadSubscription,
-                                "__omp_rtl_assume_threads_oversubscription");
-    OMPBuilder.createGlobalFlag(CGM.getLangOpts().OpenMPNoThreadState,
-                                "__omp_rtl_assume_no_thread_state");
-  }
+  if (CGM.getLangOpts().NoGPULib || CGM.getLangOpts().OMPHostIRFile.empty())
+    return;
+
+  OMPBuilder.createGlobalFlag(CGM.getLangOpts().OpenMPTargetDebug,
+                              "__omp_rtl_debug_kind");
+  OMPBuilder.createGlobalFlag(CGM.getLangOpts().OpenMPTeamSubscription,
+                              "__omp_rtl_assume_teams_oversubscription");
+  OMPBuilder.createGlobalFlag(CGM.getLangOpts().OpenMPThreadSubscription,
+                              "__omp_rtl_assume_threads_oversubscription");
+  OMPBuilder.createGlobalFlag(CGM.getLangOpts().OpenMPNoThreadState,
+                              "__omp_rtl_assume_no_thread_state");
 }
 
 void CGOpenMPRuntimeGPU::emitProcBindClause(CodeGenFunction &CGF,
Index: clang/include/clang/Driver/Options.td
===================================================================
--- clang/include/clang/Driver/Options.td
+++ clang/include/clang/Driver/Options.td
@@ -3886,8 +3886,8 @@
 def nohipwrapperinc : Flag<["-"], "nohipwrapperinc">,
   HelpText<"Do not include the default HIP wrapper headers and include paths">;
 def : Flag<["-"], "nocudainc">, Alias<nogpuinc>;
-def nogpulib : Flag<["-"], "nogpulib">,
-  HelpText<"Do not link device library for CUDA/HIP device compilation">;
+def nogpulib : Flag<["-"], "nogpulib">, MarshallingInfoFlag<LangOpts<"NoGPULib">>,
+  Flags<[CC1Option]>, HelpText<"Do not link device library for CUDA/HIP device compilation">;
 def : Flag<["-"], "nocudalib">, Alias<nogpulib>;
 def nodefaultlibs : Flag<["-"], "nodefaultlibs">;
 def nofixprebinding : Flag<["-"], "nofixprebinding">;
Index: clang/include/clang/Basic/LangOptions.def
===================================================================
--- clang/include/clang/Basic/LangOptions.def
+++ clang/include/clang/Basic/LangOptions.def
@@ -252,6 +252,7 @@
 LANGOPT(OpenMPTeamSubscription  , 1, 0, "Assume distributed loops do not have more iterations than participating teams.")
 LANGOPT(OpenMPNoThreadState  , 1, 0, "Assume that no thread in a parallel region will modify an ICV.")
 LANGOPT(OpenMPOffloadMandatory  , 1, 0, "Assert that offloading is mandatory and do not create a host fallback.")
+LANGOPT(NoGPULib  , 1, 0, "Indicate a build without the standard GPU libraries.")
 LANGOPT(RenderScript      , 1, 0, "RenderScript")
 
 LANGOPT(HLSL, 1, 0, "HLSL")


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D125314.428362.patch
Type: text/x-patch
Size: 5380 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20220510/72bf91ab/attachment.bin>