[Driver][OpenMP] Add logic for offloading-specific argument translation.

[sfantao]

Summary

This patch includes support for argument translation that is specific of a given offloading kind. Additionally, it implements the translation for OpenMP device kinds in the gcc tool chain.

With this patch, it is possible to compile a functional OpenMP application with offloading capabilities with no separate compilation.

[ABataev]

include/clang/Driver/Compilation.h

71

1. 'final'
2. default initializers for fields.

208

'const' function?

lib/Driver/Compilation.cpp

39–40

for(auto &Arg : TCArgs)

[sfantao]

• Add default initializers to toolchain arguments key and use range based iterator in Dtor.

[sfantao]

Hi Alexey,

Thanks for the review!

include/clang/Driver/Compilation.h

71

I added the default initializers but that required me to add a Ctor so that initializer lists work when this struct is built.

208

This one can't be marked const has it potentially changes the cache of translated arguments.

[sfantao]

• Rebase.

[sfantao]

• Rebase.

[sfantao]

• Rebase.

[ABataev]

LG

[hfinkel]

lib/Driver/ToolChains.cpp

2840

regardless the -> regardless of the

2842

requires to load -> is required to load

2860

And also?

case options::OPT_dynamic:

[sfantao]

• Fix typos and check -dynamic when it comes to translating arguments for offloading gcc toolchains.

[sfantao]

Hi Hal,

Thanks for the review!

lib/Driver/ToolChains.cpp

2860

Oh, yes, that one too! Thanks!

[hfinkel]

LGTM

include/clang/Driver/Compilation.h

class Compilation {
 
   /// The list of actions.  This is maintained and modified by consumers, via
   /// getActions().
   ActionList Actions;
 
   /// The root list of jobs.
   JobList Jobs;
 
-  /// Cache of translated arguments for a particular tool chain and bound
-  /// architecture.
-  llvm::DenseMap<std::pair<const ToolChain *, StringRef>,
-                 llvm::opt::DerivedArgList *>
-      TCArgs;
+  /// Cache of translated arguments for a particular tool chain, bound
+  /// architecture, and device offload kind.
+  struct TCArgsKey final {
+    const ToolChain *TC = nullptr;
+    StringRef BoundArch;
+    Action::OffloadKind DeviceOffloadKind = Action::OFK_None;
+    bool operator<(const TCArgsKey &K) const {
+      if (TC < K.TC)
+        return true;
+      else if (TC == K.TC && BoundArch < K.BoundArch)
+        return true;
+      else if (TC == K.TC && BoundArch == K.BoundArch &&
+               DeviceOffloadKind < K.DeviceOffloadKind)
+        return true;
+      return false;
+    }
+    TCArgsKey(const ToolChain *TC, StringRef BoundArch,
+              Action::OffloadKind DeviceOffloadKind)
+        : TC(TC), BoundArch(BoundArch), DeviceOffloadKind(DeviceOffloadKind) {}
+  };
+  std::map<TCArgsKey, llvm::opt::DerivedArgList *> TCArgs;
 
   /// Temporary files which should be removed on exit.
   llvm::opt::ArgStringList TempFiles;
 
   /// Result files which should be removed on failure.
   ArgStringMap ResultFiles;
 
   /// Result files which are generated correctly on failure, and which should
const ArgStringMap &getFailureResultFiles() const {
     return FailureResultFiles;
   }
 
   /// Returns the sysroot path.
   StringRef getSysRoot() const;
 
   /// getArgsForToolChain - Return the derived argument list for the
   /// tool chain \p TC (or the default tool chain, if TC is not specified).
+  /// If a device offloading kind is specified, a translation specific for that
+  /// kind is performed, if any.
   ///
   /// \param BoundArch - The bound architecture name, or 0.
-  const llvm::opt::DerivedArgList &getArgsForToolChain(const ToolChain *TC,
-                                                       StringRef BoundArch);
+  /// \param DeviceOffloadKind - The offload device kind that should be used in
+  /// the translation, if any.
+  const llvm::opt::DerivedArgList &
+  getArgsForToolChain(const ToolChain *TC, StringRef BoundArch,
+                      Action::OffloadKind DeviceOffloadKind);
 
   /// addTempFile - Add a file to remove on exit, and returns its
   /// argument.
   const char *addTempFile(const char *Name) {
     TempFiles.push_back(Name);
     return Name;
   }
 

include/clang/Driver/ToolChain.h

public:
   /// \return A pair of (`target`, `mode-flag`), where one or both may be empty.
   static std::pair<std::string, std::string>
   getTargetAndModeFromProgramName(StringRef ProgName);
 
   // Tool access.
 
   /// TranslateArgs - Create a new derived argument list for any argument
   /// translations this ToolChain may wish to perform, or 0 if no tool chain
-  /// specific translations are needed.
+  /// specific translations are needed. If \p DeviceOffloadKind is specified
+  /// the translation specific for that offload kind is performed.
   ///
   /// \param BoundArch - The bound architecture name, or 0.
+  /// \param DeviceOffloadKind - The device offload kind used for the
+  /// translation.
   virtual llvm::opt::DerivedArgList *
-  TranslateArgs(const llvm::opt::DerivedArgList &Args,
-                StringRef BoundArch) const {
+  TranslateArgs(const llvm::opt::DerivedArgList &Args, StringRef BoundArch,
+                Action::OffloadKind DeviceOffloadKind) const {
     return nullptr;
   }
 
   /// Choose a tool to use to handle the action \p JA.
   ///
   /// This can be overridden when a particular ToolChain needs to use
   /// a compiler other than Clang.
   virtual Tool *SelectTool(const JobAction &JA) const;

lib/Driver/Compilation.cpp

Compilation::Compilation(const Driver &D, const ToolChain &_DefaultToolChain,
   OrderedOffloadingToolchains.insert(
       std::make_pair(Action::OFK_Host, &DefaultToolChain));
 }
 
 Compilation::~Compilation() {
   delete TranslatedArgs;
   delete Args;
 
   // Free any derived arg lists.
   for (auto Arg : TCArgs)
     if (Arg.second != TranslatedArgs)
       delete Arg.second;
 
   // Free redirections of stdout/stderr.
   if (Redirects) {
     delete Redirects[0];
     delete Redirects[1];
     delete Redirects[2];
     delete [] Redirects;
   }
 }
 
-const DerivedArgList &Compilation::getArgsForToolChain(const ToolChain *TC,
-                                                       StringRef BoundArch) {
+const DerivedArgList &
+Compilation::getArgsForToolChain(const ToolChain *TC, StringRef BoundArch,
+                                 Action::OffloadKind DeviceOffloadKind) {
   if (!TC)
     TC = &DefaultToolChain;
 
-  DerivedArgList *&Entry = TCArgs[std::make_pair(TC, BoundArch)];
+  DerivedArgList *&Entry = TCArgs[{TC, BoundArch, DeviceOffloadKind}];
   if (!Entry) {
-    Entry = TC->TranslateArgs(*TranslatedArgs, BoundArch);
+    Entry = TC->TranslateArgs(*TranslatedArgs, BoundArch, DeviceOffloadKind);
     if (!Entry)
       Entry = TranslatedArgs;
   }
 
   return *Entry;
 }
 
 bool Compilation::CleanupFile(const char *File, bool IssueErrors) const {

lib/Driver/Driver.cpp

InputInfo Driver::BuildJobsForActionNoCache(
   // Append outputs of offload device jobs to the input list
   if (!OffloadDependencesInputInfo.empty())
     InputInfos.append(OffloadDependencesInputInfo.begin(),
                       OffloadDependencesInputInfo.end());
 
   // Set the effective triple of the toolchain for the duration of this job.
   llvm::Triple EffectiveTriple;
   const ToolChain &ToolTC = T->getToolChain();
-  const ArgList &Args = C.getArgsForToolChain(TC, BoundArch);
+  const ArgList &Args =
+      C.getArgsForToolChain(TC, BoundArch, A->getOffloadingDeviceKind());
   if (InputInfos.size() != 1) {
     EffectiveTriple = llvm::Triple(ToolTC.ComputeEffectiveClangTriple(Args));
   } else {
     // Pass along the input type if it can be unambiguously determined.
     EffectiveTriple = llvm::Triple(
         ToolTC.ComputeEffectiveClangTriple(Args, InputInfos[0].getType()));
   }
   RegisterEffectiveTriple TripleRAII(ToolTC, EffectiveTriple);
llvm::errs() << "# \"" << T->getToolChain().getTripleString() << '"'
                  << " - \"" << T->getName() << "\", inputs: [";
     for (unsigned i = 0, e = InputInfos.size(); i != e; ++i) {
       llvm::errs() << InputInfos[i].getAsString();
       if (i + 1 != e)
         llvm::errs() << ", ";
     }
     llvm::errs() << "], output: " << Result.getAsString() << "\n";
   } else {
-    T->ConstructJob(C, *JA, Result, InputInfos,
-                    C.getArgsForToolChain(TC, BoundArch), LinkingOutput);
+    T->ConstructJob(
+        C, *JA, Result, InputInfos,
+        C.getArgsForToolChain(TC, BoundArch, JA->getOffloadingDeviceKind()),
+        LinkingOutput);
   }
   return Result;
 }
 
 const char *Driver::getDefaultImageName() const {
   llvm::Triple Target(llvm::Triple::normalize(DefaultTargetTriple));
   return Target.isOSWindows() ? "a.exe" : "a.out";
 }

lib/Driver/MSVCToolChain.cpp

static void TranslateDArg(Arg *A, llvm::opt::DerivedArgList &DAL,
 
   std::string NewVal = Val;
   NewVal[Hash] = '=';
   DAL.AddJoinedArg(A, Opts.getOption(options::OPT_D), NewVal);
 }
 
 llvm::opt::DerivedArgList *
 MSVCToolChain::TranslateArgs(const llvm::opt::DerivedArgList &Args,
-                             StringRef BoundArch) const {
+                             StringRef BoundArch, Action::OffloadKind) const {
   DerivedArgList *DAL = new DerivedArgList(Args.getBaseArgs());
   const OptTable &Opts = getDriver().getOpts();
 
   // /Oy and /Oy- only has an effect under X86-32.
   bool SupportsForcingFramePointer = getArch() == llvm::Triple::x86;
 
   // 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

lib/Driver/ToolChains.h

public:
 
   void printVerboseInfo(raw_ostream &OS) const override;
 
   bool IsUnwindTablesDefault() const override;
   bool isPICDefault() const override;
   bool isPIEDefault() const override;
   bool isPICDefaultForced() const override;
   bool IsIntegratedAssemblerDefault() const override;
+  llvm::opt::DerivedArgList *
+  TranslateArgs(const llvm::opt::DerivedArgList &Args, StringRef BoundArch,
+                Action::OffloadKind DeviceOffloadKind) const override;
 
 protected:
   Tool *getTool(Action::ActionClass AC) const override;
   Tool *buildAssembler() const override;
   Tool *buildLinker() const override;
 
   /// \name ToolChain Implementation Helper Functions
   /// @{
public:
   /// @name ToolChain Implementation
   /// {
 
   types::ID LookupTypeForExtension(StringRef Ext) const override;
 
   bool HasNativeLLVMSupport() const override;
 
   llvm::opt::DerivedArgList *
-  TranslateArgs(const llvm::opt::DerivedArgList &Args,
-                StringRef BoundArch) const override;
+  TranslateArgs(const llvm::opt::DerivedArgList &Args, StringRef BoundArch,
+                Action::OffloadKind DeviceOffloadKind) const override;
 
   bool IsBlocksDefault() const override {
     // Always allow blocks on Apple; users interested in versioning are
     // expected to use /usr/include/Block.h.
     return true;
   }
   bool IsIntegratedAssemblerDefault() const override {
     // Default integrated assembler to on for Apple's MachO targets.
public:
   /// {
 
   // Darwin tools support multiple architecture (e.g., i386 and x86_64) and
   // most development is done against SDKs, so compiling for a different
   // architecture should not get any special treatment.
   bool isCrossCompiling() const override { return false; }
 
   llvm::opt::DerivedArgList *
-  TranslateArgs(const llvm::opt::DerivedArgList &Args,
-                StringRef BoundArch) const override;
+  TranslateArgs(const llvm::opt::DerivedArgList &Args, StringRef BoundArch,
+                Action::OffloadKind DeviceOffloadKind) const override;
 
   CXXStdlibType GetDefaultCXXStdlibType() const override;
   ObjCRuntime getDefaultObjCRuntime(bool isNonFragile) const override;
   bool hasBlocksRuntime() const override;
 
   bool UseObjCMixedDispatch() const override {
     // This is only used with the non-fragile ABI and non-legacy dispatch.
 

 };
 
 class LLVM_LIBRARY_VISIBILITY CudaToolChain : public Linux {
 public:
   CudaToolChain(const Driver &D, const llvm::Triple &Triple,
                 const llvm::opt::ArgList &Args);
 
   llvm::opt::DerivedArgList *
-  TranslateArgs(const llvm::opt::DerivedArgList &Args,
-                StringRef BoundArch) const override;
+  TranslateArgs(const llvm::opt::DerivedArgList &Args, StringRef BoundArch,
+                Action::OffloadKind DeviceOffloadKind) const override;
   void addClangTargetOptions(const llvm::opt::ArgList &DriverArgs,
                              llvm::opt::ArgStringList &CC1Args) const override;
 
   // Never try to use the integrated assembler with CUDA; always fork out to
   // ptxas.
   bool useIntegratedAs() const override { return false; }
 
   void AddCudaIncludeArgs(const llvm::opt::ArgList &DriverArgs,

 };
 
 class LLVM_LIBRARY_VISIBILITY MSVCToolChain : public ToolChain {
 public:
   MSVCToolChain(const Driver &D, const llvm::Triple &Triple,
                 const llvm::opt::ArgList &Args);
 
   llvm::opt::DerivedArgList *
-  TranslateArgs(const llvm::opt::DerivedArgList &Args,
-                StringRef BoundArch) const override;
+  TranslateArgs(const llvm::opt::DerivedArgList &Args, StringRef BoundArch,
+                Action::OffloadKind DeviceOffloadKind) const override;
 
   bool IsIntegratedAssemblerDefault() const override;
   bool IsUnwindTablesDefault() const override;
   bool isPICDefault() const override;
   bool isPIEDefault() const override;
   bool isPICDefaultForced() const override;
 
   void

lib/Driver/ToolChains.cpp

void DarwinClang::AddCCKextLibArgs(const ArgList &Args,
 
   // For now, allow missing resource libraries to support developers who may
   // not have compiler-rt checked out or integrated into their build.
   if (getVFS().exists(P))
     CmdArgs.push_back(Args.MakeArgString(P));
 }
 
 DerivedArgList *MachO::TranslateArgs(const DerivedArgList &Args,
-                                     StringRef BoundArch) const {
+                                     StringRef BoundArch,
+                                     Action::OffloadKind) const {
   DerivedArgList *DAL = new DerivedArgList(Args.getBaseArgs());
   const OptTable &Opts = getDriver().getOpts();
 
   // FIXME: We really want to get out of the tool chain level argument
   // translation business, as it makes the driver functionality much
   // more opaque. For now, we follow gcc closely solely for the
   // purpose of easily achieving feature parity & testability. Once we
   // have something that works, we should reevaluate each translation
CompilerRT +=
       (tools::arm::getARMFloatABI(*this, Args) == tools::arm::FloatABI::Hard)
           ? "hard"
           : "soft";
   CompilerRT += Args.hasArg(options::OPT_fPIC) ? "_pic.a" : "_static.a";
 
   AddLinkRuntimeLib(Args, CmdArgs, CompilerRT, false, true);
 }
 
-DerivedArgList *Darwin::TranslateArgs(const DerivedArgList &Args,
-                                      StringRef BoundArch) const {
+DerivedArgList *
+Darwin::TranslateArgs(const DerivedArgList &Args, StringRef BoundArch,
+                      Action::OffloadKind DeviceOffloadKind) const {
   // First get the generic Apple args, before moving onto Darwin-specific ones.
-  DerivedArgList *DAL = MachO::TranslateArgs(Args, BoundArch);
+  DerivedArgList *DAL =
+      MachO::TranslateArgs(Args, BoundArch, DeviceOffloadKind);
   const OptTable &Opts = getDriver().getOpts();
 
   // If no architecture is bound, none of the translations here are relevant.
   if (BoundArch.empty())
     return DAL;
 
   // Add an explicit version min argument for the deployment target. We do this
   // after argument translation because -Xarch_ arguments may add a version min
if ((GCCMultiarchTriple.empty() && TargetMultiarchTriple.empty()) ||
     addSystemInclude(DriverArgs, CC1Args,
                      Base + "/" + TargetMultiarchTriple + Suffix);
   }
 
   addSystemInclude(DriverArgs, CC1Args, Base + Suffix + "/backward");
   return true;
 }
 
+llvm::opt::DerivedArgList *
+Generic_GCC::TranslateArgs(const llvm::opt::DerivedArgList &Args, StringRef,
+                           Action::OffloadKind DeviceOffloadKind) const {
+
+  // If this tool chain is used for an OpenMP offloading device we have to make
+  // sure we always generate a shared library regardless of the commands the
+  // user passed to the host. This is required because the runtime library
+  // is required to load the device image dynamically at run time.
+  if (DeviceOffloadKind == Action::OFK_OpenMP) {
+    DerivedArgList *DAL = new DerivedArgList(Args.getBaseArgs());
+    const OptTable &Opts = getDriver().getOpts();
+
+    // Request the shared library. Given that these options are decided
+    // implicitly, they do not refer to any base argument.
+    DAL->AddFlagArg(/*BaseArg=*/nullptr, Opts.getOption(options::OPT_shared));
+    DAL->AddFlagArg(/*BaseArg=*/nullptr, Opts.getOption(options::OPT_fPIC));
+
+    // Filter all the arguments we don't care passing to the offloading
+    // toolchain as they can mess up with the creation of a shared library.
+    for (auto *A : Args) {
+      switch ((options::ID)A->getOption().getID()) {
+      default:
+        DAL->append(A);
+        break;
+      case options::OPT_shared:
+      case options::OPT_dynamic:
+      case options::OPT_static:
+      case options::OPT_fPIC:
+      case options::OPT_fno_PIC:
+      case options::OPT_fpic:
+      case options::OPT_fno_pic:
+      case options::OPT_fPIE:
+      case options::OPT_fno_PIE:
+      case options::OPT_fpie:
+      case options::OPT_fno_pie:
+        break;
+      }
+    }
+    return DAL;
+  }
+  return nullptr;
+}
+
 void Generic_ELF::addClangTargetOptions(const ArgList &DriverArgs,
                                         ArgStringList &CC1Args) const {
   const Generic_GCC::GCCVersion &V = GCCInstallation.getVersion();
   bool UseInitArrayDefault =
       getTriple().getArch() == llvm::Triple::aarch64 ||
       getTriple().getArch() == llvm::Triple::aarch64_be ||
       (getTriple().getOS() == llvm::Triple::Linux &&
        (!V.isOlderThan(4, 7, 0) || getTriple().isAndroid())) ||
if (!DriverArgs.hasArg(options::OPT_nocudainc) &&
     assert(!Arch.empty() && "Must have an explicit GPU arch.");
     CudaInstallation.CheckCudaVersionSupportsArch(StringToCudaArch(Arch));
   }
   Linux::AddCudaIncludeArgs(DriverArgs, CC1Args);
 }
 
 llvm::opt::DerivedArgList *
 CudaToolChain::TranslateArgs(const llvm::opt::DerivedArgList &Args,
-                             StringRef BoundArch) const {
+                             StringRef BoundArch, Action::OffloadKind) const {
   DerivedArgList *DAL = new DerivedArgList(Args.getBaseArgs());
   const OptTable &Opts = getDriver().getOpts();
 
   for (Arg *A : Args) {
     if (A->getOption().matches(options::OPT_Xarch__)) {
       // Skip this argument unless the architecture matches BoundArch
       if (BoundArch.empty() || A->getValue(0) != BoundArch)
         continue;

test/Driver/openmp-offload.c

 //
 // CHK-COMMANDS: clang{{.*}}" "-cc1" "-triple" "powerpc64le--linux" "-emit-llvm-bc" {{.*}}"-o" "[[HOSTBC:.+\.bc]]" "-x" "c" "[[INPUT:.+\.c]]" "-fopenmp-targets=powerpc64le-ibm-linux-gnu,x86_64-pc-linux-gnu"
 // CHK-COMMANDS-ST: clang{{.*}}" "-cc1" "-triple" "powerpc64le--linux" "-E" {{.*}}"-fopenmp" {{.*}}"-o" "[[HOSTPP:.+\.i]]" "-x" "c" "[[INPUT:.+\.c]]"
 // CHK-COMMANDS-ST: clang{{.*}}" "-cc1" "-triple" "powerpc64le--linux" "-emit-llvm-bc" {{.*}}"-fopenmp" {{.*}}"-o" "[[HOSTBC:.+\.bc]]" "-x" "cpp-output" "[[HOSTPP]]" "-fopenmp-targets=powerpc64le-ibm-linux-gnu,x86_64-pc-linux-gnu"
 
 //
 // Compile for the powerpc device.
 //
-// CHK-COMMANDS: clang{{.*}}" "-cc1" "-triple" "powerpc64le-ibm-linux-gnu" "-emit-obj" {{.*}}"-fopenmp" {{.*}}"-o" "[[T1OBJ:.+\.o]]" "-x" "c" "[[INPUT]]" "-fopenmp-is-device" "-fopenmp-host-ir-file-path" "[[HOSTBC]]"
+// CHK-COMMANDS: clang{{.*}}" "-cc1" "-triple" "powerpc64le-ibm-linux-gnu" "-emit-obj" {{.*}}"-pic-level" "2" {{.*}}"-fopenmp" {{.*}}"-o" "[[T1OBJ:.+\.o]]" "-x" "c" "[[INPUT]]" "-fopenmp-is-device" "-fopenmp-host-ir-file-path" "[[HOSTBC]]"
 // CHK-COMMANDS: ld" {{.*}}"-o" "[[T1BIN]]" {{.*}}"[[T1OBJ]]"
 // CHK-COMMANDS-ST: clang{{.*}}" "-cc1" "-triple" "powerpc64le-ibm-linux-gnu" "-E" {{.*}}"-fopenmp" {{.*}}"-o" "[[T1PP:.+\.i]]" "-x" "c" "[[INPUT]]"
-// CHK-COMMANDS-ST: clang{{.*}}" "-cc1" "-triple" "powerpc64le-ibm-linux-gnu" "-emit-llvm-bc" {{.*}}"-fopenmp" {{.*}}"-o" "[[T1BC:.+\.bc]]" "-x" "cpp-output" "[[T1PP]]" "-fopenmp-is-device" "-fopenmp-host-ir-file-path" "[[HOSTBC]]"
+// CHK-COMMANDS-ST: clang{{.*}}" "-cc1" "-triple" "powerpc64le-ibm-linux-gnu" "-emit-llvm-bc" {{.*}}"-pic-level" "2" {{.*}}"-fopenmp" {{.*}}"-o" "[[T1BC:.+\.bc]]" "-x" "cpp-output" "[[T1PP]]" "-fopenmp-is-device" "-fopenmp-host-ir-file-path" "[[HOSTBC]]"
 // CHK-COMMANDS-ST: clang{{.*}}" "-cc1" "-triple" "powerpc64le-ibm-linux-gnu" "-S" {{.*}}"-fopenmp" {{.*}}"-o" "[[T1ASM:.+\.s]]" "-x" "ir" "[[T1BC]]"
 // CHK-COMMANDS-ST: clang{{.*}}" "-cc1as" "-triple" "powerpc64le-ibm-linux-gnu" "-filetype" "obj" {{.*}}"-o" "[[T1OBJ:.+\.o]]" "[[T1ASM]]"
-// CHK-COMMANDS-ST: ld" {{.*}}"-o" "[[T1BIN]]" {{.*}}[[T1OBJ]]
+// CHK-COMMANDS-ST: ld" {{.*}}"-shared" {{.*}}"-o" "[[T1BIN]]" {{.*}}[[T1OBJ]]
 
 //
 // Compile for the x86 device.
 //
-// CHK-COMMANDS: clang{{.*}}" "-cc1" "-triple" "x86_64-pc-linux-gnu" "-emit-obj"  {{.*}}"-fopenmp"  {{.*}}"-o" "[[T2OBJ:.+\.o]]" "-x" "c" "[[INPUT]]" "-fopenmp-is-device" "-fopenmp-host-ir-file-path" "[[HOSTBC]]"
+// CHK-COMMANDS: clang{{.*}}" "-cc1" "-triple" "x86_64-pc-linux-gnu" "-emit-obj" {{.*}}"-pic-level" "2" {{.*}}"-fopenmp"  {{.*}}"-o" "[[T2OBJ:.+\.o]]" "-x" "c" "[[INPUT]]" "-fopenmp-is-device" "-fopenmp-host-ir-file-path" "[[HOSTBC]]"
 // CHK-COMMANDS: ld" {{.*}}"-o" "[[T2BIN]]" {{.*}}"[[T2OBJ]]"
 // CHK-COMMANDS-ST: clang{{.*}}" "-cc1" "-triple" "x86_64-pc-linux-gnu" "-E" {{.*}}"-fopenmp" {{.*}}"-o" "[[T2PP:.+\.i]]" "-x" "c" "[[INPUT]]"
-// CHK-COMMANDS-ST: clang{{.*}}" "-cc1" "-triple" "x86_64-pc-linux-gnu" "-emit-llvm-bc" {{.*}}"-fopenmp" {{.*}}"-o" "[[T2BC:.+\.bc]]" "-x" "cpp-output" "[[T2PP]]" "-fopenmp-is-device" "-fopenmp-host-ir-file-path" "[[HOSTBC]]"
+// CHK-COMMANDS-ST: clang{{.*}}" "-cc1" "-triple" "x86_64-pc-linux-gnu" "-emit-llvm-bc" {{.*}}"-pic-level" "2" {{.*}}"-fopenmp" {{.*}}"-o" "[[T2BC:.+\.bc]]" "-x" "cpp-output" "[[T2PP]]" "-fopenmp-is-device" "-fopenmp-host-ir-file-path" "[[HOSTBC]]"
 // CHK-COMMANDS-ST: clang{{.*}}" "-cc1" "-triple" "x86_64-pc-linux-gnu" "-S" {{.*}}"-fopenmp" {{.*}}"-o" "[[T2ASM:.+\.s]]" "-x" "ir" "[[T2BC]]"
 // CHK-COMMANDS-ST: clang{{.*}}" "-cc1as" "-triple" "x86_64-pc-linux-gnu" "-filetype" "obj" {{.*}}"-o" "[[T2OBJ:.+\.o]]" "[[T2ASM]]"
-// CHK-COMMANDS-ST: ld" {{.*}}"-o" "[[T2BIN]]" {{.*}}[[T2OBJ]]
+// CHK-COMMANDS-ST: ld" {{.*}}"-shared" {{.*}}"-o" "[[T2BIN]]" {{.*}}[[T2OBJ]]
 
 //
 // Generate host object from the BC file and link using the linker script.
 //
 // CHK-COMMANDS: clang{{.*}}" "-cc1" "-triple" "powerpc64le--linux" "-emit-obj" {{.*}}"-fopenmp" {{.*}}"-o" "[[HOSTOBJ:.+\.o]]" "-x" "ir" "[[HOSTBC]]"
 // CHK-COMMANDS: ld" {{.*}}"-o" "[[HOSTBIN:.+\.out]]"  {{.*}}"-lomptarget" {{.*}}"-T" "[[HOSTLK:.+\.lk]]"
 // CHK-COMMANDS-ST: clang{{.*}}" "-cc1" "-triple" "powerpc64le--linux" "-S" {{.*}}"-fopenmp" {{.*}}"-o" "[[HOSTASM:.+\.s]]" "-x" "ir" "[[HOSTBC]]"
 // CHK-COMMANDS-ST: clang{{.*}}" "-cc1as" "-triple" "powerpc64le--linux" "-filetype" "obj" {{.*}}"-o" [[HOSTOBJ:.+\.o]]" [[HOSTASM:.+\.s]]
 // CHK-COMMANDS-ST: ld" {{.*}}"-o" "[[HOSTBIN:.+\.out]]"  {{.*}}"-lomptarget" {{.*}}"-T" "[[HOSTLK:.+\.lk]]"