Merge branch 'main' into linearize_bitcast

Author: chencha3

.ci/metrics/metrics.py

@@ -130,34 +130,6 @@ def get_per_workflow_metrics(
         workflow_jobs = workflow_run.jobs()
         if workflow_jobs.totalCount == 0:
             continue
-        if workflow_jobs.totalCount > 1:
-            raise ValueError(
-                f"Encountered an unexpected number of jobs: {workflow_jobs.totalCount}"
-            )
-
-        created_at = workflow_jobs[0].created_at
-        started_at = workflow_jobs[0].started_at
-        completed_at = workflow_jobs[0].completed_at
-
-        job_result = int(workflow_jobs[0].conclusion == "success")
-        if job_result:
-            # We still might want to mark the job as a failure if one of the steps
-            # failed. This is required due to use setting continue-on-error in
-            # the premerge pipeline to prevent sending emails while we are
-            # testing the infrastructure.
-            # TODO(boomanaiden154): Remove this once the premerge pipeline is no
-            # longer in a testing state and we can directly assert the workflow
-            # result.
-            for step in workflow_jobs[0].steps:
-                if step.conclusion != "success":
-                    job_result = 0
-                    break
-
-        queue_time = started_at - created_at
-        run_time = completed_at - started_at
-
-        if run_time.seconds == 0:
-            continue
 
         if (
             workflows_to_track[workflow_run.name] is None
@@ -170,20 +142,45 @@ def get_per_workflow_metrics(
         ):
             break
 
-        # The timestamp associated with the event is expected by Grafana to be
-        # in nanoseconds.
-        created_at_ns = int(created_at.timestamp()) * 10**9
-
-        workflow_metrics.append(
-            JobMetrics(
-                workflow_run.name,
-                queue_time.seconds,
-                run_time.seconds,
-                job_result,
-                created_at_ns,
-                workflow_run.id,
+        for workflow_job in workflow_jobs:
+            created_at = workflow_job.created_at
+            started_at = workflow_job.started_at
+            completed_at = workflow_job.completed_at
+
+            job_result = int(workflow_job.conclusion == "success")
+            if job_result:
+                # We still might want to mark the job as a failure if one of the steps
+                # failed. This is required due to use setting continue-on-error in
+                # the premerge pipeline to prevent sending emails while we are
+                # testing the infrastructure.
+                # TODO(boomanaiden154): Remove this once the premerge pipeline is no
+                # longer in a testing state and we can directly assert the workflow
+                # result.
+                for step in workflow_job.steps:
+                    if step.conclusion != "success":
+                        job_result = 0
+                        break
+
+            queue_time = started_at - created_at
+            run_time = completed_at - started_at
+
+            if run_time.seconds == 0:
+                continue
+
+            # The timestamp associated with the event is expected by Grafana to be
+            # in nanoseconds.
+            created_at_ns = int(created_at.timestamp()) * 10**9
+
+            workflow_metrics.append(
+                JobMetrics(
+                    workflow_run.name + "-" + workflow_job.name,
+                    queue_time.seconds,
+                    run_time.seconds,
+                    job_result,
+                    created_at_ns,
+                    workflow_run.id,
+                )
             )
-        )
 
     return workflow_metrics
 

clang/lib/CodeGen/CodeGenFunction.h

@@ -451,7 +451,7 @@ class CodeGenFunction : public CodeGenTypeCache {
              "EBB should be entry block of the current code gen function");
       PostAllocaInsertPt = AllocaInsertPt->clone();
       PostAllocaInsertPt->setName("postallocapt");
-      PostAllocaInsertPt->insertAfter(AllocaInsertPt);
+      PostAllocaInsertPt->insertAfter(AllocaInsertPt->getIterator());
     }
 
     return PostAllocaInsertPt;

clang/lib/Driver/Driver.cpp

@@ -6669,9 +6669,21 @@ const ToolChain &Driver::getToolChain(const ArgList &Args,
     case llvm::Triple::CUDA:
       TC = std::make_unique<toolchains::NVPTXToolChain>(*this, Target, Args);
       break;
-    case llvm::Triple::AMDHSA:
-      TC = std::make_unique<toolchains::ROCMToolChain>(*this, Target, Args);
+    case llvm::Triple::AMDHSA: {
+      bool CL = llvm::any_of(Args, [&](Arg *A) {
+        return (A->getOption().matches(options::OPT_x) &&
+                types::isOpenCL(
+                    types::lookupTypeForExtension(A->getValue()))) ||
+               (A->getOption().getKind() == Option::InputClass &&
+                StringRef(A->getValue()).rfind('.') != StringRef::npos &&
+                types::isOpenCL(types::lookupTypeForExtension(
+                    &A->getValue()[StringRef(A->getValue()).rfind('.') + 1])));
+      });
+      TC = CL ? std::make_unique<toolchains::ROCMToolChain>(*this, Target, Args)
+              : std::make_unique<toolchains::AMDGPUToolChain>(*this, Target,
+                                                              Args);
       break;
+    }
     case llvm::Triple::AMDPAL:
     case llvm::Triple::Mesa3D:
       TC = std::make_unique<toolchains::AMDGPUToolChain>(*this, Target, Args);

clang/lib/Sema/SemaDecl.cpp

@@ -11380,8 +11380,11 @@ static bool CheckMultiVersionAdditionalRules(Sema &S, const FunctionDecl *OldFD,
     return true;
 
   // Only allow transition to MultiVersion if it hasn't been used.
-  if (OldFD && CausesMV && OldFD->isUsed(false))
-    return S.Diag(NewFD->getLocation(), diag::err_multiversion_after_used);
+  if (OldFD && CausesMV && OldFD->isUsed(false)) {
+    S.Diag(NewFD->getLocation(), diag::err_multiversion_after_used);
+    S.Diag(OldFD->getLocation(), diag::note_previous_declaration);
+    return true;
+  }
 
   return S.areMultiversionVariantFunctionsCompatible(
       OldFD, NewFD, S.PDiag(diag::err_multiversion_noproto),

clang/test/Driver/amdgpu-toolchain.c

@@ -36,3 +36,6 @@
 // RUN: %clang -target amdgcn-amd-amdhsa -march=gfx90a -stdlib -startfiles \
 // RUN:   -nogpulib -nogpuinc -### %s 2>&1 | FileCheck -check-prefix=STARTUP %s
 // STARTUP: ld.lld{{.*}}"-lc" "-lm" "{{.*}}crt1.o"
+//
+// RUN: %clang -### --target=amdgcn-amd-amdhsa -mcpu=gfx906 %s 2>&1 | FileCheck -check-prefix=ROCM %s
+// ROCM-NOT: -mlink-builtin-bitcode

clang/test/Sema/attr-cpuspecific.c

@@ -44,7 +44,8 @@ int allow_fwd_decl2(void);
 void use_fwd_decl(void) {
   allow_fwd_decl2();
 }
-// expected-error@+1 {{function declaration cannot become a multiversioned function after first usage}}
+// expected-error@+2 {{function declaration cannot become a multiversioned function after first usage}}
+// expected-note@-5 {{previous declaration is here}}
 int __attribute__((cpu_dispatch(atom))) allow_fwd_decl2(void) {}
 
 

clang/test/Sema/attr-target-mv.c

@@ -66,7 +66,8 @@ int use3(void) {
   return mv_after_use();
 }
 
-// expected-error@+1 {{function declaration cannot become a multiversioned function after first usage}}
+// expected-error@+2 {{function declaration cannot become a multiversioned function after first usage}}
+// expected-note@-6 {{previous declaration is here}}
 int __attribute__((target("arch=sandybridge")))  mv_after_use(void) { return 2; }
 
 int __attribute__((target("sse4.2,arch=sandybridge"))) mangle(void) { return 1; }

clang/test/Sema/attr-target-version.c

@@ -88,7 +88,8 @@ int bar() {
   nodef();
   return def();
 }
-// expected-error@+1 {{function declaration cannot become a multiversioned function after first usage}}
+// expected-error@+2 {{function declaration cannot become a multiversioned function after first usage}}
+// expected-note@-13 {{previous declaration is here}}
 int __attribute__((target_version("sha2"))) def(void) { return 1; }
 
 int __attribute__((target_version("sve"))) prot();

flang/docs/ModFiles.md

@@ -164,3 +164,13 @@ a diagnostic but we still wouldn't have line numbers.
 To provide line numbers and character positions or source lines as the user
 wrote them we would have to save some amount of provenance information in the
 module file as well.
+
+## Hermetic modules files
+
+Top-level module files for libraries can be build with `-fhermetic-module-files`.
+This option causes these module files to contain copies of all of the non-intrinsic
+modules on which they depend, so that non-top-level local modules and the
+modules of dependent libraries need not also be packaged with the library.
+When the compiler reads a hermetic module file, the copies of the dependent
+modules are read into their own scope, and will not conflict with other modules
+of the same name that client code might `USE`.

flang/include/flang/Common/Fortran-features.h

@@ -73,7 +73,8 @@ ENUM_CLASS(UsageWarning, Portability, PointerToUndefinable,
     PreviousScalarUse, RedeclaredInaccessibleComponent, ImplicitShared,
     IndexVarRedefinition, IncompatibleImplicitInterfaces, BadTypeForTarget,
     VectorSubscriptFinalization, UndefinedFunctionResult, UselessIomsg,
-    MismatchingDummyProcedure, SubscriptedEmptyArray, UnsignedLiteralTruncation)
+    MismatchingDummyProcedure, SubscriptedEmptyArray, UnsignedLiteralTruncation,
+    CompatibleDeclarationsFromDistinctModules)
 
 using LanguageFeatures = EnumSet<LanguageFeature, LanguageFeature_enumSize>;
 using UsageWarnings = EnumSet<UsageWarning, UsageWarning_enumSize>;