Merge from 'sycl' to 'sycl-web' (intel#5)

clang/include/clang/Basic/AttrDocs.td

@@ -3310,7 +3310,8 @@ must be compiled and executed with a sub-group of size n. The value of n must be
 set to a sub-group size supported by the device, or device compilation will fail.
 
 In addition to device functions, the required sub-group size attribute may also
-be specified in the definition of a named functor object, as in the example below:
+be specified in the definition of a named functor object and lambda functions,
+as in the examples below:
 
 .. code-block:: c++
 
@@ -3322,6 +3323,11 @@ be specified in the definition of a named functor object, as in the example belo
       }
   }
 
+  kernel<class kernel_name>(
+  []() [[cl::intel_reqd_sub_group_size(n)]] {
+       /* kernel code */
+  });
+
 See Sub-groups for NDRange Parallelism proposal in sycl/doc/extensions/sub_group_ndrange/sub_group_ndrange.md
   }];
 }

clang/include/clang/Sema/Sema.h

@@ -12055,6 +12055,11 @@ class Sema final {
   // is associated with.
   std::unique_ptr<SYCLIntegrationHeader> SyclIntHeader;
 
+  // Used to suppress diagnostics during kernel construction, since these were
+  // already emitted earlier. Diagnosing during Kernel emissions also skips the
+  // useful notes that shows where the kernel was called.
+  bool ConstructingOpenCLKernel = false;
+
 public:
   void addSyclDeviceDecl(Decl *d) { SyclDeviceDecls.push_back(d); }
   SmallVectorImpl<Decl *> &syclDeviceDecls() { return SyclDeviceDecls; }
@@ -12081,6 +12086,7 @@ class Sema final {
     KernelCallVariadicFunction
  };
   DeviceDiagBuilder SYCLDiagIfDeviceCode(SourceLocation Loc, unsigned DiagID);
+  bool isKnownGoodSYCLDecl(const Decl *D);
   void ConstructOpenCLKernel(FunctionDecl *KernelCallerFunc, MangleContext &MC);
   void MarkDevice(void);
   bool CheckSYCLCall(SourceLocation Loc, FunctionDecl *Callee);

clang/lib/Driver/Compilation.cpp

@@ -214,10 +214,11 @@ static bool ActionFailed(const Action *A,
   if (FailingCommands.empty())
     return false;
 
-  // CUDA/HIP can have the same input source code compiled multiple times so do
-  // not compiled again if there are already failures. It is OK to abort the
-  // CUDA pipeline on errors.
-  if (A->isOffloading(Action::OFK_Cuda) || A->isOffloading(Action::OFK_HIP))
+  // CUDA/HIP/SYCL can have the same input source code compiled multiple times
+  // so do not compile again if there are already failures. It is OK to abort
+  // the CUDA pipeline on errors.
+  if (A->isOffloading(Action::OFK_Cuda) || A->isOffloading(Action::OFK_HIP) ||
+      A->isOffloading(Action::OFK_SYCL))
     return true;
 
   for (const auto &CI : FailingCommands)

clang/lib/Sema/SemaExpr.cpp

@@ -5404,6 +5404,12 @@ bool Sema::GatherArgumentsForCall(SourceLocation CallLoc, FunctionDecl *FDecl,
 
     // Otherwise do argument promotion, (C99 6.5.2.2p7).
     } else {
+      // Diagnose variadic calls in SYCL.
+      if (getLangOpts().SYCLIsDevice && !isUnevaluatedContext() &&
+          !isKnownGoodSYCLDecl(FDecl))
+        SYCLDiagIfDeviceCode(CallLoc, diag::err_sycl_restrict)
+            << Sema::KernelCallVariadicFunction;
+
       for (Expr *A : Args.slice(ArgIx)) {
         ExprResult Arg = DefaultVariadicArgumentPromotion(A, CallType, FDecl);
         Invalid |= Arg.isInvalid();

clang/lib/Sema/SemaOverload.cpp

@@ -14297,6 +14297,11 @@ Sema::BuildCallToObjectOfClassType(Scope *S, Expr *Obj,
 
   // If this is a variadic call, handle args passed through "...".
   if (Proto->isVariadic()) {
+    // Diagnose variadic calls in SYCL.
+    if (getLangOpts().SYCLIsDevice && !isUnevaluatedContext())
+      SYCLDiagIfDeviceCode(LParenLoc, diag::err_sycl_restrict)
+          << Sema::KernelCallVariadicFunction;
+
     // Promote the arguments (C99 6.5.2.2p7).
     for (unsigned i = NumParams, e = Args.size(); i < e; i++) {
       ExprResult Arg = DefaultVariadicArgumentPromotion(Args[i], VariadicMethod,

clang/lib/Sema/SemaSYCL.cpp

@@ -184,7 +184,7 @@ static bool IsSyclMathFunc(unsigned BuiltinID) {
   return true;
 }
 
-static bool isKnownGoodDecl(const Decl *D) {
+bool Sema::isKnownGoodSYCLDecl(const Decl *D) {
   if (const FunctionDecl *FD = dyn_cast<FunctionDecl>(D)) {
     const IdentifierInfo *II = FD->getIdentifier();
     const DeclContext *DC = FD->getDeclContext();
@@ -326,7 +326,7 @@ class MarkDeviceFunction : public RecursiveASTVisitor<MarkDeviceFunction> {
 
   bool VisitDeclRefExpr(DeclRefExpr *E) {
     Decl* D = E->getDecl();
-    if (isKnownGoodDecl(D))
+    if (SemaRef.isKnownGoodSYCLDecl(D))
       return true;
 
     CheckSYCLType(E->getType(), E->getSourceRange());
@@ -372,18 +372,6 @@ class MarkDeviceFunction : public RecursiveASTVisitor<MarkDeviceFunction> {
     return true;
   }
 
-  bool VisitGCCAsmStmt(GCCAsmStmt *S) {
-    SemaRef.Diag(S->getBeginLoc(), diag::err_sycl_restrict)
-        << Sema::KernelUseAssembly;
-    return true;
-  }
-
-  bool VisitMSAsmStmt(MSAsmStmt *S) {
-    SemaRef.Diag(S->getBeginLoc(), diag::err_sycl_restrict)
-        << Sema::KernelUseAssembly;
-    return true;
-  }
-
   // The call graph for this translation unit.
   CallGraph SYCLCG;
   // The set of functions called by a kernel function.
@@ -549,9 +537,6 @@ class MarkDeviceFunction : public RecursiveASTVisitor<MarkDeviceFunction> {
         }
       }
     } else if (const auto *FPTy = dyn_cast<FunctionProtoType>(Ty)) {
-      if (FPTy->isVariadic() && SemaRef.getLangOpts().SYCLIsDevice)
-        SemaRef.SYCLDiagIfDeviceCode(Loc.getBegin(), diag::err_sycl_restrict)
-            << Sema::KernelCallVariadicFunction;
       for (const auto &ParamTy : FPTy->param_types())
         if (!CheckSYCLType(ParamTy, Loc, Visited))
           return false;
@@ -1308,8 +1293,10 @@ void Sema::ConstructOpenCLKernel(FunctionDecl *KernelCallerFunc,
   // in different translation units.
   OpenCLKernel->setImplicitlyInline(KernelCallerFunc->isInlined());
 
+  ConstructingOpenCLKernel = true;
   CompoundStmt *OpenCLKernelBody =
       CreateOpenCLKernelBody(*this, KernelCallerFunc, OpenCLKernel);
+  ConstructingOpenCLKernel = false;
   OpenCLKernel->setBody(OpenCLKernelBody);
   addSyclDeviceDecl(OpenCLKernel);
 }
@@ -1404,13 +1391,13 @@ static bool isKnownEmitted(Sema &S, FunctionDecl *FD) {
   if (!FD)
     return true; // Seen in LIT testing
 
-  if (FD->hasAttr<SYCLDeviceAttr>() || FD->hasAttr<SYCLKernelAttr>())
-    return true;
-
   // Templates are emitted when they're instantiated.
   if (FD->isDependentContext())
     return false;
 
+  if (FD->hasAttr<SYCLDeviceAttr>() || FD->hasAttr<SYCLKernelAttr>())
+    return true;
+
   // Otherwise, the function is known-emitted if it's in our set of
   // known-emitted functions.
   return S.DeviceKnownEmittedFns.count(FD) > 0;
@@ -1420,18 +1407,20 @@ Sema::DeviceDiagBuilder Sema::SYCLDiagIfDeviceCode(SourceLocation Loc,
                                                    unsigned DiagID) {
   assert(getLangOpts().SYCLIsDevice &&
          "Should only be called during SYCL compilation");
-  DeviceDiagBuilder::Kind DiagKind = [this] {
-    if (isKnownEmitted(*this, dyn_cast<FunctionDecl>(CurContext)))
+  FunctionDecl *FD = dyn_cast<FunctionDecl>(getCurLexicalContext());
+  DeviceDiagBuilder::Kind DiagKind = [this, FD] {
+    if (ConstructingOpenCLKernel || (FD && FD->isDependentContext()))
+      return DeviceDiagBuilder::K_Nop;
+    else if (isKnownEmitted(*this, FD))
       return DeviceDiagBuilder::K_ImmediateWithCallStack;
     return DeviceDiagBuilder::K_Deferred;
   }();
-  return DeviceDiagBuilder(DiagKind, Loc, DiagID,
-                           dyn_cast<FunctionDecl>(CurContext), *this);
+  return DeviceDiagBuilder(DiagKind, Loc, DiagID, FD, *this);
 }
 
 bool Sema::CheckSYCLCall(SourceLocation Loc, FunctionDecl *Callee) {
   assert(Callee && "Callee may not be null.");
-  FunctionDecl *Caller = getCurFunctionDecl();
+  FunctionDecl *Caller = dyn_cast<FunctionDecl>(getCurLexicalContext());
 
   // If the caller is known-emitted, mark the callee as known-emitted.
   // Otherwise, mark the call in our call graph so we can traverse it later.

clang/lib/Sema/SemaType.cpp

@@ -4685,7 +4685,6 @@ static TypeSourceInfo *GetFullTypeForDeclarator(TypeProcessingState &state,
         // OpenCL doesn't support variadic functions and blocks
         // (s6.9.e and s6.12.5 OpenCL v2.0) except for printf.
         // We also allow here any toolchain reserved identifiers.
-        // FIXME: Use deferred diagnostics engine to skip host side issues.
         if (FTI.isVariadic &&
             !LangOpts.SYCLIsDevice &&
             !(D.getIdentifier() &&

clang/test/Driver/sycl-offload-dup-error.cpp

@@ -0,0 +1,12 @@
+// Tests to make sure that there is no duplicate error messaging for host
+// REQUIRES: x86-registered-target
+// RUN: not %clangxx -c -target x86_64-unknown-linux-gnu -fsycl %s \
+// RUN: 2>&1 | FileCheck %s
+
+void foo() {
+  foobar s;
+}
+
+// CHECK: error: unknown type name 'foobar'
+// CHECK-NOT: error: unknown type name 'foobar'
+

clang/test/SemaSYCL/inline-asm.cpp

@@ -21,11 +21,13 @@ void bar() {
 
 template <typename name, typename Func>
 __attribute__((sycl_kernel)) void kernel_single_task(Func kernelFunc) {
+  // expected-note@+1 {{called by 'kernel_single_task<fake_kernel, (lambda}}
   kernelFunc();
 }
 
 int main() {
   foo();
+  // expected-note@+1 {{called by 'operator()'}}
   kernel_single_task<class fake_kernel>([]() { bar(); });
   return 0;
 }

clang/test/SemaSYCL/sycl-cconv.cpp

@@ -18,10 +18,12 @@ __cdecl __attribute__((sycl_kernel)) void kernel_single_task(Func kernelFunc) {
   printf("cannot call from here\n");
   // expected-no-error@+1
   moo();
+  // expected-note@+1{{called by}}
   kernelFunc();
 }
 
 int main() {
+  //expected-note@+2 {{in instantiation of}}
   //expected-error@+1 {{SYCL kernel cannot call a variadic function}}
   kernel_single_task<class fake_kernel>([]() { printf("world\n"); });
   bar();

clang/test/SemaSYCL/sycl-restrict.cpp

@@ -124,6 +124,7 @@ void eh_not_ok(void)
 
 void usage(myFuncDef functionPtr) {
 
+  // expected-note@+1 {{called by 'usage'}}
   eh_not_ok();
 
 #if ALLOW_FP
@@ -169,7 +170,6 @@ int use2 ( a_type ab, a_type *abp ) {
   return ns::glob +
   // expected-error@+1 {{SYCL kernel cannot use a global variable}}
     AnotherNS::moar_globals;
-  // expected-note@+1 {{called by 'use2'}}
   eh_not_ok();
   Check_RTTI_Restriction:: A *a;
   Check_RTTI_Restriction:: isa_B(a);
@@ -180,15 +180,16 @@ int use2 ( a_type ab, a_type *abp ) {
 
 template <typename name, typename Func>
 __attribute__((sycl_kernel)) void kernel_single_task(Func kernelFunc) {
+  // expected-note@+1 {{called by 'kernel_single_task}}
   kernelFunc();
   a_type ab;
   a_type *p;
-  // expected-note@+1 {{called by 'kernel_single_task'}}
   use2(ab, p);
 }
 
 int main() {
   a_type ab;
+  // expected-note@+1 {{called by 'operator()'}}
   kernel_single_task<class fake_kernel>([]() { usage(  &addInt ); });
   return 0;
 }

clang/test/SemaSYCL/sycl-varargs-cconv.cpp

@@ -39,7 +39,7 @@ void bar() {
 
 template <typename name, typename Func>
 __attribute__((sycl_kernel)) void kernel_single_task(Func kernelFunc) {
-  kernelFunc();
+  kernelFunc(); //expected-note 2+ {{called by 'kernel_single_task}}
 }
 
 int main() {

clang/test/SemaSYCL/variadic-func-call.cpp

@@ -0,0 +1,39 @@
+// RUN: %clang_cc1 -triple spir64-unknown-unknown -fsycl-is-device -fsyntax-only -verify %s
+
+void variadic(int, ...);
+namespace NS {
+void variadic(int, ...);
+}
+
+struct S {
+  S(int, ...);
+  void operator()(int, ...);
+};
+
+void foo() {
+  auto x = [](int, ...) {};
+  x(5, 10); //expected-error{{SYCL kernel cannot call a variadic function}}
+}
+
+void overloaded(int, int);
+void overloaded(int, ...);
+template <typename, typename Func>
+__attribute__((sycl_kernel)) void task(Func KF) {
+  KF(); // expected-note 2 {{called by 'task}}
+}
+
+int main() {
+  task<class FK>([]() {
+    variadic(5);        //expected-error{{SYCL kernel cannot call a variadic function}}
+    variadic(5, 2);     //expected-error{{SYCL kernel cannot call a variadic function}}
+    NS::variadic(5, 3); //expected-error{{SYCL kernel cannot call a variadic function}}
+    S s(5, 4);          //expected-error{{SYCL kernel cannot call a variadic function}}
+    S s2(5);            //expected-error{{SYCL kernel cannot call a variadic function}}
+    s(5, 5);            //expected-error{{SYCL kernel cannot call a variadic function}}
+    s2(5);              //expected-error{{SYCL kernel cannot call a variadic function}}
+    foo();              //expected-note{{called by 'operator()'}}
+    overloaded(5, 6);   //expected-no-error
+    overloaded(5, s);   //expected-error{{SYCL kernel cannot call a variadic function}}
+    overloaded(5);      //expected-error{{SYCL kernel cannot call a variadic function}}
+  });
+}

sycl/doc/extensions/SubGroupNDRange/SubGroupNDRange.md

@@ -35,7 +35,7 @@ In keeping with Section 6.7 of the SYCL 1.2.1 specification, attributes are made
 
 The `[[cl::intel_reqd_sub_group_size(n)]]` attribute indicates that the kernel must be compiled and executed with a sub-group of size `n`.  The value of `n` must be set to a sub-group size supported by the device, or device compilation will fail.
 
-In addition to device functions, the required sub-group size attribute may also be specified in the definition of a named functor object, as in the example below:
+In addition to device functions, the required sub-group size attribute may also be specified in the definition of a named functor object and lambda functions, as in the examples below:
 
 ```c++
 class Functor
@@ -45,6 +45,11 @@ class Functor
         /* kernel code */
     }
 }
+
+kernel<class kernel_name>(
+[]() [[cl::intel_reqd_sub_group_size(n)]] {
+     /* kernel code */
+});
 ```
 
 # Sub-group Queries