[HIP] use offload wrapper for non-device-only non-rdc

Currently HIP still uses offload bundler for non-rdc mode for
the new offload driver.

This patch switches to use offload wrapper for non-device-only
non-rdc mode when new offload driver is enabled.

This makes the rdc and non-rdc compilation more consistent
and speeds up compilation since the offload wrapper supports
parallel compilation for different GPU arch's.

It is implemented by adding a linker wrapper action for
each assemble action of input file. Linker wrapper action
differentiates this special type of work vs normal
linker wrapper work by the fle type. This type of work
results in object instead of image. The linker wrapper
adds "-r" for it and only includes the object file
as input, not the host libraries.

For device-only non-RDC mode, the new driver keeps the
original behavior.

Author: yxsamliu

clang/lib/CodeGen/CGCUDANV.cpp

@@ -1281,7 +1281,8 @@ llvm::Function *CGNVCUDARuntime::finalizeModule() {
     return nullptr;
   }
   if (CGM.getLangOpts().OffloadViaLLVM ||
-      (CGM.getLangOpts().OffloadingNewDriver && RelocatableDeviceCode))
+      (CGM.getLangOpts().OffloadingNewDriver &&
+       (CGM.getLangOpts().HIP || RelocatableDeviceCode)))
     createOffloadingEntries();
   else
     return makeModuleCtorFunction();

clang/lib/Driver/Driver.cpp

@@ -4402,6 +4402,10 @@ void Driver::BuildActions(Compilation &C, DerivedArgList &Args,
                    options::OPT_no_offload_new_driver,
                    C.isOffloadingHostKind(Action::OFK_Cuda));
 
+  bool HIPNoRDC =
+      C.isOffloadingHostKind(Action::OFK_HIP) &&
+      !Args.hasFlag(options::OPT_fgpu_rdc, options::OPT_fno_gpu_rdc, false);
+
   // Builder to be used to build offloading actions.
   std::unique_ptr<OffloadingActionBuilder> OffloadBuilder =
       !UseNewOffloadingDriver
@@ -4502,6 +4506,16 @@ void Driver::BuildActions(Compilation &C, DerivedArgList &Args,
                                                                 InputArg))
         break;
 
+      // For HIP non-rdc non-device-only compilation, create a linker wrapper
+      // action for each host object to link, bundle and wrap device files in
+      // it.
+      if (Phase == phases::Assemble && UseNewOffloadingDriver && HIPNoRDC &&
+          !offloadDeviceOnly()) {
+        ActionList AL{Current};
+        Current = C.MakeAction<LinkerWrapperJobAction>(AL, types::TY_Object);
+        Current->propagateHostOffloadInfo(C.getActiveOffloadKinds(),
+                                          /*BoundArch=*/nullptr);
+      }
       if (Current->getType() == types::TY_Nothing)
         break;
     }
@@ -4535,7 +4549,7 @@ void Driver::BuildActions(Compilation &C, DerivedArgList &Args,
     // Check if this Linker Job should emit a static library.
     if (ShouldEmitStaticLibrary(Args)) {
       LA = C.MakeAction<StaticLibJobAction>(LinkerInputs, types::TY_Image);
-    } else if (UseNewOffloadingDriver ||
+    } else if ((UseNewOffloadingDriver && !HIPNoRDC) ||
                Args.hasArg(options::OPT_offload_link)) {
       LA = C.MakeAction<LinkerWrapperJobAction>(LinkerInputs, types::TY_Image);
       LA->propagateHostOffloadInfo(C.getActiveOffloadKinds(),
@@ -4945,12 +4959,15 @@ Action *Driver::BuildOffloadingActions(Compilation &C,
       }
     }
 
-    // Compiling HIP in non-RDC mode requires linking each action individually.
+    // Compiling HIP in device-only non-RDC mode requires linking each action
+    // individually.
     for (Action *&A : DeviceActions) {
       if ((A->getType() != types::TY_Object &&
            A->getType() != types::TY_LTO_BC) ||
           Kind != Action::OFK_HIP ||
-          Args.hasFlag(options::OPT_fgpu_rdc, options::OPT_fno_gpu_rdc, false))
+          Args.hasFlag(options::OPT_fgpu_rdc, options::OPT_fno_gpu_rdc,
+                       false) ||
+          !offloadDeviceOnly())
         continue;
       ActionList LinkerInput = {A};
       A = C.MakeAction<LinkJobAction>(LinkerInput, types::TY_Image);
@@ -4974,9 +4991,10 @@ Action *Driver::BuildOffloadingActions(Compilation &C,
     }
   }
 
-  // HIP code in non-RDC mode will bundle the output if it invoked the linker.
+  // HIP code in device-only non-RDC mode will bundle the output if it invoked
+  // the linker.
   bool ShouldBundleHIP =
-      C.isOffloadingHostKind(Action::OFK_HIP) &&
+      C.isOffloadingHostKind(Action::OFK_HIP) && offloadDeviceOnly() &&
       Args.hasFlag(options::OPT_gpu_bundle_output,
                    options::OPT_no_gpu_bundle_output, true) &&
       !Args.hasFlag(options::OPT_fgpu_rdc, options::OPT_fno_gpu_rdc, false) &&
@@ -4999,11 +5017,11 @@ Action *Driver::BuildOffloadingActions(Compilation &C,
         C.MakeAction<LinkJobAction>(OffloadActions, types::TY_CUDA_FATBIN);
     DDep.add(*FatbinAction, *C.getSingleOffloadToolChain<Action::OFK_Cuda>(),
              nullptr, Action::OFK_Cuda);
-  } else if (C.isOffloadingHostKind(Action::OFK_HIP) &&
+  } else if (C.isOffloadingHostKind(Action::OFK_HIP) && offloadDeviceOnly() &&
              !Args.hasFlag(options::OPT_fgpu_rdc, options::OPT_fno_gpu_rdc,
                            false)) {
-    // If we are not in RDC-mode we just emit the final HIP fatbinary for each
-    // translation unit, linking each input individually.
+    // If we are in device-only non-RDC-mode we just emit the final HIP
+    // fatbinary for each translation unit, linking each input individually.
     Action *FatbinAction =
         C.MakeAction<LinkJobAction>(OffloadActions, types::TY_HIP_FATBIN);
     DDep.add(*FatbinAction, *C.getSingleOffloadToolChain<Action::OFK_HIP>(),
@@ -5156,8 +5174,11 @@ Action *Driver::ConstructPhaseAction(
         (((Input->getOffloadingToolChain() &&
            Input->getOffloadingToolChain()->getTriple().isAMDGPU()) ||
           TargetDeviceOffloadKind == Action::OFK_HIP) &&
-         (Args.hasFlag(options::OPT_fgpu_rdc, options::OPT_fno_gpu_rdc,
-                       false) ||
+         ((Args.hasFlag(options::OPT_fgpu_rdc, options::OPT_fno_gpu_rdc,
+                        false) ||
+           (Args.hasFlag(options::OPT_offload_new_driver,
+                         options::OPT_no_offload_new_driver, false) &&
+            !offloadDeviceOnly())) ||
           TargetDeviceOffloadKind == Action::OFK_OpenMP))) {
       types::ID Output =
           Args.hasArg(options::OPT_S) &&

clang/lib/Driver/ToolChains/Clang.cpp

@@ -7838,7 +7838,7 @@ void Clang::ConstructJob(Compilation &C, const JobAction &JA,
     CmdArgs.push_back("-fcuda-include-gpubinary");
     CmdArgs.push_back(CudaDeviceInput->getFilename());
   } else if (!HostOffloadingInputs.empty()) {
-    if ((IsCuda || IsHIP) && !IsRDCMode) {
+    if (IsCuda && !IsRDCMode) {
       assert(HostOffloadingInputs.size() == 1 && "Only one input expected");
       CmdArgs.push_back("-fcuda-include-gpubinary");
       CmdArgs.push_back(HostOffloadingInputs.front().getFilename());
@@ -9348,11 +9348,22 @@ void LinkerWrapper::ConstructJob(Compilation &C, const JobAction &JA,
   // Add the linker arguments to be forwarded by the wrapper.
   CmdArgs.push_back(Args.MakeArgString(Twine("--linker-path=") +
                                        LinkCommand->getExecutable()));
-  for (const char *LinkArg : LinkCommand->getArguments())
-    CmdArgs.push_back(LinkArg);
 
-  addOffloadCompressArgs(Args, CmdArgs);
+  // We use action type to differentiate two use cases of the linker wrapper.
+  // TY_Image for normal linker wrapper work.
+  // TY_Object for HIP fno-gpu-rdc embedding device binary in a relocatable
+  // object.
+  assert(JA.getType() == types::TY_Object || JA.getType() == types::TY_Image);
+  if (JA.getType() == types::TY_Object) {
+    CmdArgs.append({"-o", Output.getFilename()});
+    for (auto Input : Inputs)
+      CmdArgs.push_back(Input.getFilename());
+    CmdArgs.push_back(Args.MakeArgString("-r"));
+  } else
+    for (const char *LinkArg : LinkCommand->getArguments())
+      CmdArgs.push_back(LinkArg);
 
+  addOffloadCompressArgs(Args, CmdArgs);
   const char *Exec =
       Args.MakeArgString(getToolChain().GetProgramPath("clang-linker-wrapper"));
 

clang/test/Driver/hip-binding.hip

@@ -93,7 +93,7 @@
 // RUN:        -nogpulib -nogpuinc -foffload-lto --offload-arch=gfx90a --offload-arch=gfx908 -c %s 2>&1 \
 // RUN: | FileCheck -check-prefix=LTO-NO-RDC %s
 //      LTO-NO-RDC: # "amdgcn-amd-amdhsa" - "clang", inputs: ["[[INPUT:.+]]"], output: "[[LTO_908:.+]]"
-// LTO-NO-RDC-NEXT: # "amdgcn-amd-amdhsa" - "AMDGCN::Linker", inputs: ["[[LTO_908]]"], output: "[[OBJ_908:.+]]"
 // LTO-NO-RDC-NEXT: # "amdgcn-amd-amdhsa" - "clang", inputs: ["[[INPUT]]"], output: "[[LTO_90A:.+]]"
-// LTO-NO-RDC-NEXT: # "amdgcn-amd-amdhsa" - "AMDGCN::Linker", inputs: ["[[LTO_90A]]"], output: "[[OBJ_90A:.+]]"
-// LTO-NO-RDC-NEXT: # "amdgcn-amd-amdhsa" - "AMDGCN::Linker", inputs: ["[[OBJ_908]]", "[[OBJ_90A]]"], output: "[[HIPFB:.+]]"
+// LTO-NO-RDC-NEXT: # "x86_64-unknown-linux-gnu" - "Offload::Packager", inputs: ["[[LTO_908]]", "[[LTO_90A]]"], output: "[[PKG:.+]]"
+// LTO-NO-RDC-NEXT: # "x86_64-unknown-linux-gnu" - "clang", inputs: ["[[INPUT]]", "[[PKG]]"], output: "[[OBJ:.+]]"
+// LTO-NO-RDC-NEXT: # "x86_64-unknown-linux-gnu" - "Offload::Linker", inputs: ["[[OBJ]]"], output: "hip-binding.o"

clang/test/Driver/hip-phases.hip

@@ -8,39 +8,50 @@
 //
 // RUN: %clang -x hip --target=x86_64-unknown-linux-gnu -ccc-print-phases \
 // RUN: --no-offload-new-driver --cuda-gpu-arch=gfx803 %s 2>&1 \
-// RUN: | FileCheck -check-prefixes=BIN,NRD,OLD %s
+// RUN: | FileCheck -check-prefixes=BIN,OLD,OLDN %s
 // RUN: %clang -x hip --target=x86_64-unknown-linux-gnu -ccc-print-phases \
 // RUN: --offload-new-driver --cuda-gpu-arch=gfx803 %s 2>&1 \
-// RUN: | FileCheck -check-prefixes=BIN,NRD,NEW %s
+// RUN: | FileCheck -check-prefixes=BIN,NEW,NEWN %s
 //
 // RUN: %clang -x hip --target=x86_64-unknown-linux-gnu -ccc-print-phases \
 // RUN: --no-offload-new-driver --cuda-gpu-arch=gfx803 -fgpu-rdc %s 2>&1 \
-// RUN: | FileCheck -check-prefixes=BIN,RDC %s
+// RUN: | FileCheck -check-prefixes=BIN,OLD,OLDR %s
+// RUN: %clang -x hip --target=x86_64-unknown-linux-gnu -ccc-print-phases \
+// RUN: --offload-new-driver --cuda-gpu-arch=gfx803 -fgpu-rdc %s 2>&1 \
+// RUN: | FileCheck -check-prefixes=BIN,NEW,NEWR %s
 //
 // BIN-DAG: [[P0:[0-9]+]]: input, "{{.*}}hip-phases.hip", [[T:hip]], (host-[[T]])
 // BIN-DAG: [[P1:[0-9]+]]: preprocessor, {[[P0]]}, [[T]]-cpp-output, (host-[[T]])
 // BIN-DAG: [[P2:[0-9]+]]: compiler, {[[P1]]}, ir, (host-[[T]])
-// RDC-DAG: [[P12:[0-9]+]]: backend, {[[P2]]}, assembler, (host-[[T]])
-// RDC-DAG: [[P13:[0-9]+]]: assembler, {[[P12]]}, object, (host-[[T]])
+// OLDR-DAG: [[P12:[0-9]+]]: backend, {[[P2]]}, assembler, (host-[[T]])
+// OLDR-DAG: [[P13:[0-9]+]]: assembler, {[[P12]]}, object, (host-[[T]])
 
 // BIN-DAG: [[P3:[0-9]+]]: input, "{{.*}}hip-phases.hip", [[T]], (device-[[T]], [[ARCH:gfx803]])
 // BIN-DAG: [[P4:[0-9]+]]: preprocessor, {[[P3]]}, [[T]]-cpp-output, (device-[[T]], [[ARCH]])
 // BIN-DAG: [[P5:[0-9]+]]: compiler, {[[P4]]}, ir, (device-[[T]], [[ARCH]])
-// NRD-DAG: [[P6:[0-9]+]]: backend, {[[P5]]}, assembler, (device-[[T]], [[ARCH]])
-// NRD-DAG: [[P7:[0-9]+]]: assembler, {[[P6]]}, object, (device-[[T]], [[ARCH]])
-// RDC-DAG: [[P7:[0-9]+]]: backend, {[[P5]]}, ir, (device-[[T]], [[ARCH]])
-// BIN-DAG: [[P8:[0-9]+]]: linker, {[[P7]]}, image, (device-[[T]], [[ARCH]])
-// BIN-DAG: [[P9:[0-9]+]]: offload, "device-[[T]] (amdgcn-amd-amdhsa:[[ARCH]])" {[[P8]]}, image
-// NRD-DAG: [[P10:[0-9]+]]: linker, {[[P9]]}, hip-fatbin, (device-[[T]])
-// RDC-DAG: [[P10:[0-9]+]]: linker, {[[P9]]}, object, (device-[[T]])
-
-// NRD-DAG: [[P11:[0-9]+]]: offload, "host-[[T]] (x86_64-unknown-linux-gnu)" {[[P2]]}, "device-[[T]] (amdgcn-amd-amdhsa)" {[[P10]]}, ir
-// RDC-DAG: [[P11:[0-9]+]]: offload, "device-[[T]] (amdgcn-amd-amdhsa)" {[[P10]]}, object
-// NRD-DAG: [[P12:[0-9]+]]: backend, {[[P11]]}, assembler, (host-[[T]])
-// NRD-DAG: [[P13:[0-9]+]]: assembler, {[[P12]]}, object, (host-[[T]])
-// OLD-DAG: [[P14:[0-9]+]]: linker, {[[P13]]}, image, (host-[[T]])
-// NEW-DAG: [[P14:[0-9]+]]: clang-linker-wrapper, {[[P13]]}, image, (host-[[T]])
-// RDC-DAG: [[P14:[0-9]+]]: linker, {[[P13]], [[P11]]}, image, (host-[[T]])
+// OLDN-DAG: [[P6:[0-9]+]]: backend, {[[P5]]}, assembler, (device-[[T]], [[ARCH]])
+// NEW-DAG: [[P6:[0-9]+]]: backend, {[[P5]]}, ir, (device-[[T]], [[ARCH]])
+// OLDN-DAG: [[P7:[0-9]+]]: assembler, {[[P6]]}, object, (device-[[T]], [[ARCH]])
+// OLDR-DAG: [[P7:[0-9]+]]: backend, {[[P5]]}, ir, (device-[[T]], [[ARCH]])
+// OLD-DAG: [[P8:[0-9]+]]: linker, {[[P7]]}, image, (device-[[T]], [[ARCH]])
+// OLD-DAG: [[P9:[0-9]+]]: offload, "device-[[T]] (amdgcn-amd-amdhsa:[[ARCH]])" {[[P8]]}, image
+// NEW-DAG: [[P9:[0-9]+]]: offload, "device-[[T]] (amdgcn-amd-amdhsa:[[ARCH]])" {[[P6]]}, ir
+// OLDN-DAG: [[P10:[0-9]+]]: linker, {[[P9]]}, hip-fatbin, (device-[[T]])
+// NEW-DAG: [[P10:[0-9]+]]: clang-offload-packager, {[[P9]]}, image, (device-[[T]])
+// OLDR-DAG: [[P10:[0-9]+]]: linker, {[[P9]]}, object, (device-[[T]])
+
+// OLDN-DAG: [[P11:[0-9]+]]: offload, "host-[[T]] (x86_64-unknown-linux-gnu)" {[[P2]]}, "device-[[T]] (amdgcn-amd-amdhsa)" {[[P10]]}, ir
+// NEW-DAG: [[P11:[0-9]+]]: offload, "host-[[T]] (x86_64-unknown-linux-gnu)" {[[P2]]}, "device-[[T]] (x86_64-unknown-linux-gnu)" {[[P10]]}, ir
+// OLDR-DAG: [[P11:[0-9]+]]: offload, "device-[[T]] (amdgcn-amd-amdhsa)" {[[P10]]}, object
+// OLDN-DAG: [[P12:[0-9]+]]: backend, {[[P11]]}, assembler, (host-[[T]])
+// OLDN-DAG: [[P13:[0-9]+]]: assembler, {[[P12]]}, object, (host-[[T]])
+// NEW-DAG: [[P12:[0-9]+]]: backend, {[[P11]]}, assembler, (host-[[T]])
+// NEW-DAG: [[P13:[0-9]+]]: assembler, {[[P12]]}, object, (host-[[T]])
+// OLDN-DAG: [[P14:[0-9]+]]: linker, {[[P13]]}, image, (host-[[T]])
+// NEWN-DAG: [[P14:[0-9]+]]: clang-linker-wrapper, {[[P13]]}, object, (host-[[T]])
+// OLDR-DAG: [[P14:[0-9]+]]: linker, {[[P13]], [[P11]]}, image, (host-[[T]])
+// NEWR-DAG: [[P14:[0-9]+]]: clang-linker-wrapper, {[[P13]]}, image, (host-[[T]])
+// NEWN-DAG: [[P15:[0-9]+]]: linker, {[[P14]]}, image
 
 //
 // Test single gpu architecture up to the assemble phase.