[PATCH] D109175: [openmp] Add clang cc1 option -fopenmp-skip-deferred-diags

[Wei Wang via Phabricator via cfe-commits]

weiwang created this revision.
Herald added subscribers: hoy, dexonsmith, wenlei, dang, guansong, yaxunl.
weiwang requested review of this revision.
Herald added a reviewer: jdoerfert.
Herald added subscribers: cfe-commits, sstefan1.
Herald added a project: clang.

There are cases where no target compilation/offloading would ever happen, but
we are still paying the price for handling deferred diags. The new cc1 option
allows user to explictly skip deferred diags handling in FE.

This results in considerable build time improvement from our internal
workloads.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D109175

Files:
  clang/include/clang/Basic/LangOptions.def
  clang/include/clang/Driver/Options.td
  clang/lib/Frontend/CompilerInvocation.cpp
  clang/lib/Sema/SemaDecl.cpp
  clang/test/OpenMP/declare_target_messages.cpp


Index: clang/test/OpenMP/declare_target_messages.cpp
===================================================================
--- clang/test/OpenMP/declare_target_messages.cpp
+++ clang/test/OpenMP/declare_target_messages.cpp
@@ -6,6 +6,9 @@
 // RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -verify=expected,omp5,host5 -fopenmp-simd -fopenmp-is-device -fnoopenmp-use-tls -ferror-limit 100 -o - %s
 // RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -verify=expected,omp45 -fopenmp-version=45 -fopenmp-simd -fnoopenmp-use-tls -ferror-limit 100 -o - %s
 
+// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -verify=expected,omp5,dev5 -fopenmp -fopenmp-is-device -fopenmp-targets=x86_64-apple-macos10.7.0 -aux-triple x86_64-apple-macos10.7.0 -fnoopenmp-use-tls -fopenmp-skip-deferred-diags -ferror-limit 100 -o - %s
+// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -verify=expected,omp5,host5 -fopenmp-simd -fopenmp-is-device -fnoopenmp-use-tls -fopenmp-skip-deferred-diags -ferror-limit 100 -o - %s
+// RUN: %clang_cc1 -triple x86_64-apple-macos10.7.0 -verify=expected,omp5 -fopenmp -fnoopenmp-use-tls -fopenmp-skip-deferred-diags -ferror-limit 100 -o - %s
 #pragma omp end declare target // expected-error {{unexpected OpenMP directive '#pragma omp end declare target'}}
 
 int a, b, z; // omp5-error {{variable captured in declare target region must appear in a to clause}}
Index: clang/lib/Sema/SemaDecl.cpp
===================================================================
--- clang/lib/Sema/SemaDecl.cpp
+++ clang/lib/Sema/SemaDecl.cpp
@@ -12609,7 +12609,8 @@
     VDecl->setInitStyle(VarDecl::ListInit);
   }
 
-  if (LangOpts.OpenMP && VDecl->isFileVarDecl())
+  if (LangOpts.OpenMP && !LangOpts.OpenMPSkipDeferredDiags &&
+      VDecl->isFileVarDecl())
     DeclsToCheckForDeferredDiags.insert(VDecl);
   CheckCompleteVariableDeclaration(VDecl);
 }
@@ -14839,7 +14840,8 @@
     DiscardCleanupsInEvaluationContext();
   }
 
-  if (FD && (LangOpts.OpenMP || LangOpts.CUDA || LangOpts.SYCLIsDevice)) {
+  if (FD && ((LangOpts.OpenMP && !LangOpts.OpenMPSkipDeferredDiags) ||
+             LangOpts.CUDA || LangOpts.SYCLIsDevice)) {
     auto ES = getEmissionStatus(FD);
     if (ES == Sema::FunctionEmissionStatus::Emitted ||
         ES == Sema::FunctionEmissionStatus::Unknown)
Index: clang/lib/Frontend/CompilerInvocation.cpp
===================================================================
--- clang/lib/Frontend/CompilerInvocation.cpp
+++ clang/lib/Frontend/CompilerInvocation.cpp
@@ -3489,6 +3489,9 @@
   if (Opts.OpenMPCUDAForceFullRuntime)
     GenerateArg(Args, OPT_fopenmp_cuda_force_full_runtime, SA);
 
+  if (Opts.OpenMPSkipDeferredDiags)
+    GenerateArg(Args, OPT_fopenmp_skip_deferred_diags, SA);
+
   // The arguments used to set Optimize, OptimizeSize and NoInlineDefine are
   // generated from CodeGenOptions.
 
@@ -3928,6 +3931,10 @@
       Opts.OpenMPIsDevice && (T.isNVPTX() || T.isAMDGCN()) &&
       Args.hasArg(options::OPT_fopenmp_cuda_force_full_runtime);
 
+  Opts.OpenMPSkipDeferredDiags =
+      Opts.OpenMP && !Opts.OpenMPSimd && !IsTargetSpecified &&
+      Args.hasArg(options::OPT_fopenmp_skip_deferred_diags);
+
   // FIXME: Eliminate this dependency.
   unsigned Opt = getOptimizationLevel(Args, IK, Diags),
        OptSize = getOptimizationLevelSize(Args);
Index: clang/include/clang/Driver/Options.td
===================================================================
--- clang/include/clang/Driver/Options.td
+++ clang/include/clang/Driver/Options.td
@@ -5799,6 +5799,8 @@
   HelpText<"Generate code only for an OpenMP target device.">;
 def fopenmp_host_ir_file_path : Separate<["-"], "fopenmp-host-ir-file-path">,
   HelpText<"Path to the IR file produced by the frontend for the host.">;
+def fopenmp_skip_deferred_diags : Flag<["-"], "fopenmp-skip-deferred-diags">,
+  HelpText<"Do not generate deferred diagnostics for function host/device availability. Use this flag only when no offloading or omp target directive is used.">;
 
 //===----------------------------------------------------------------------===//
 // SYCL Options
Index: clang/include/clang/Basic/LangOptions.def
===================================================================
--- clang/include/clang/Basic/LangOptions.def
+++ clang/include/clang/Basic/LangOptions.def
@@ -243,6 +243,7 @@
 LANGOPT(OpenMPCUDAReductionBufNum , 32, 1024, "Number of the reduction records in the intermediate reduction buffer used for the teams reductions.")
 LANGOPT(OpenMPTargetNewRuntime , 1, 0, "Use the new bitcode library for OpenMP offloading")
 LANGOPT(OpenMPOptimisticCollapse  , 1, 0, "Use at most 32 bits to represent the collapsed loop nest counter.")
+LANGOPT(OpenMPSkipDeferredDiags,  1, 0, "Do not generate deferred diagnostics for function host/device availability.")
 LANGOPT(RenderScript      , 1, 0, "RenderScript")
 
 LANGOPT(CUDAIsDevice      , 1, 0, "compiling for CUDA device")


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D109175.370328.patch
Type: text/x-patch
Size: 4908 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20210902/3c15ee52/attachment-0001.bin>