AMDGPU: Really delete AMDGPUAnnotateKernelFeatures

[arsenm]

31f39c8 claimed to have removed it,
but it's still here so really remove it.

[arsenm]

• AMDGPU: Really delete AMDGPUAnnotateKernelFeatures #145136 👈 (View in Graphite)
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[llvmbot]

@llvm/pr-subscribers-backend-amdgpu

Author: Matt Arsenault (arsenm)

Changes

31f39c8 claimed to have removed it,
but it's still here so really remove it.

---

Full diff: https://github.com/llvm/llvm-project/pull/145136.diff

1 Files Affected:

• (removed) llvm/lib/Target/AMDGPU/AMDGPUAnnotateKernelFeatures.cpp (-128)

diff --git a/llvm/lib/Target/AMDGPU/AMDGPUAnnotateKernelFeatures.cpp b/llvm/lib/Target/AMDGPU/AMDGPUAnnotateKernelFeatures.cpp
deleted file mode 100644
index 9c9fa5c6e2f0f..0000000000000
--- a/llvm/lib/Target/AMDGPU/AMDGPUAnnotateKernelFeatures.cpp
+++ /dev/null
@@ -1,128 +0,0 @@
-//===- AMDGPUAnnotateKernelFeaturesPass.cpp -------------------------------===//
-//
-// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
-// See https://llvm.org/LICENSE.txt for license information.
-// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
-//
-//===----------------------------------------------------------------------===//
-//
-/// \file This pass propagates the uniform-work-group-size attribute from
-/// kernels to leaf functions when possible. It also adds additional attributes
-/// to hint ABI lowering optimizations later.
-//
-//===----------------------------------------------------------------------===//
-
-#include "AMDGPU.h"
-#include "GCNSubtarget.h"
-#include "llvm/Analysis/CallGraph.h"
-#include "llvm/Analysis/CallGraphSCCPass.h"
-#include "llvm/CodeGen/TargetPassConfig.h"
-#include "llvm/Target/TargetMachine.h"
-
-#define DEBUG_TYPE "amdgpu-annotate-kernel-features"
-
-using namespace llvm;
-
-namespace {
-class AMDGPUAnnotateKernelFeatures : public CallGraphSCCPass {
-private:
-  const TargetMachine *TM = nullptr;
-
-  bool addFeatureAttributes(Function &F);
-
-public:
-  static char ID;
-
-  AMDGPUAnnotateKernelFeatures() : CallGraphSCCPass(ID) {}
-
-  bool doInitialization(CallGraph &CG) override;
-  bool runOnSCC(CallGraphSCC &SCC) override;
-
-  StringRef getPassName() const override {
-    return "AMDGPU Annotate Kernel Features";
-  }
-
-  void getAnalysisUsage(AnalysisUsage &AU) const override {
-    AU.setPreservesAll();
-    CallGraphSCCPass::getAnalysisUsage(AU);
-  }
-};
-
-} // end anonymous namespace
-
-char AMDGPUAnnotateKernelFeatures::ID = 0;
-
-bool AMDGPUAnnotateKernelFeatures::addFeatureAttributes(Function &F) {
-  bool HaveStackObjects = false;
-  bool Changed = false;
-  bool HaveCall = false;
-  bool IsFunc = !AMDGPU::isEntryFunctionCC(F.getCallingConv());
-
-  for (BasicBlock &BB : F) {
-    for (Instruction &I : BB) {
-      if (isa<AllocaInst>(I)) {
-        HaveStackObjects = true;
-        continue;
-      }
-
-      if (auto *CB = dyn_cast<CallBase>(&I)) {
-        const Function *Callee =
-            dyn_cast<Function>(CB->getCalledOperand()->stripPointerCasts());
-
-        // Note the occurrence of indirect call.
-        if (!Callee) {
-          if (!CB->isInlineAsm())
-            HaveCall = true;
-
-          continue;
-        }
-
-        Intrinsic::ID IID = Callee->getIntrinsicID();
-        if (IID == Intrinsic::not_intrinsic) {
-          HaveCall = true;
-          Changed = true;
-        }
-      }
-    }
-  }
-
-  // TODO: We could refine this to captured pointers that could possibly be
-  // accessed by flat instructions. For now this is mostly a poor way of
-  // estimating whether there are calls before argument lowering.
-  if (!IsFunc && HaveCall) {
-    F.addFnAttr("amdgpu-calls");
-    Changed = true;
-  }
-
-  if (HaveStackObjects) {
-    F.addFnAttr("amdgpu-stack-objects");
-    Changed = true;
-  }
-
-  return Changed;
-}
-
-bool AMDGPUAnnotateKernelFeatures::runOnSCC(CallGraphSCC &SCC) {
-  bool Changed = false;
-
-  for (CallGraphNode *I : SCC) {
-    Function *F = I->getFunction();
-    // Ignore functions with graphics calling conventions, these are currently
-    // not allowed to have kernel arguments.
-    if (!F || F->isDeclaration() || AMDGPU::isGraphics(F->getCallingConv()))
-      continue;
-    // Add feature attributes
-    Changed |= addFeatureAttributes(*F);
-  }
-
-  return Changed;
-}
-
-bool AMDGPUAnnotateKernelFeatures::doInitialization(CallGraph &CG) {
-  auto *TPC = getAnalysisIfAvailable<TargetPassConfig>();
-  if (!TPC)
-    report_fatal_error("TargetMachine is required");
-
-  TM = &TPC->getTM<TargetMachine>();
-  return false;
-}

[rampitec]

Seems to be already disabled. LGTM.

[llvm-ci]

LLVM Buildbot has detected a new failure on builder llvm-clang-key-instructions running on sie-linux-worker5 while building llvm at step 6 "test-build-unified-tree-check-all".

Full details are available at: https://lab.llvm.org/buildbot/#/builders/208/builds/2167

Here is the relevant piece of the build log for the reference

Step 6 (test-build-unified-tree-check-all) failure: test (failure)
******************** TEST 'Clang-Unit :: ./AllClangUnitTests/33/48' FAILED ********************
Script(shard):
--
GTEST_OUTPUT=json:/home/buildbot/buildbot-root/llvm-ki/build/tools/clang/unittests/./AllClangUnitTests-Clang-Unit-3710339-33-48.json GTEST_SHUFFLE=0 GTEST_TOTAL_SHARDS=48 GTEST_SHARD_INDEX=33 /home/buildbot/buildbot-root/llvm-ki/build/tools/clang/unittests/./AllClangUnitTests
--

Note: This is test shard 34 of 48.
[==========] Running 510 tests from 104 test suites.
[----------] Global test environment set-up.
[----------] 1 test from MinimizeSourceToDependencyDirectivesTest
[ RUN      ] MinimizeSourceToDependencyDirectivesTest.RawStringLiteral
[       OK ] MinimizeSourceToDependencyDirectivesTest.RawStringLiteral (0 ms)
[----------] 1 test from MinimizeSourceToDependencyDirectivesTest (0 ms total)

[----------] 1 test from LexerTest
[ RUN      ] LexerTest.GetSourceTextExpandsAcrossMultipleMacroCalls
[       OK ] LexerTest.GetSourceTextExpandsAcrossMultipleMacroCalls (0 ms)
[----------] 1 test from LexerTest (0 ms total)

[----------] 1 test from PPCallbacksTest
[ RUN      ] PPCallbacksTest.OpenCLExtensionPragmaDisabled
[       OK ] PPCallbacksTest.OpenCLExtensionPragmaDisabled (0 ms)
[----------] 1 test from PPCallbacksTest (0 ms total)

[----------] 1 test from ToolChainTest
[ RUN      ] ToolChainTest.VFSGnuLibcxxPathNoSysroot
[       OK ] ToolChainTest.VFSGnuLibcxxPathNoSysroot (0 ms)
[----------] 1 test from ToolChainTest (0 ms total)

[----------] 1 test from MultilibTest
[ RUN      ] MultilibTest.SelectHard
[       OK ] MultilibTest.SelectHard (0 ms)
[----------] 1 test from MultilibTest (0 ms total)

[----------] 2 tests from ExprMutationAnalyzerTest
[ RUN      ] ExprMutationAnalyzerTest.ArrayToPointerDecay
[       OK ] ExprMutationAnalyzerTest.ArrayToPointerDecay (4 ms)
[ RUN      ] ExprMutationAnalyzerTest.PointeeMutatedByAssign
[       OK ] ExprMutationAnalyzerTest.PointeeMutatedByAssign (9 ms)
[----------] 2 tests from ExprMutationAnalyzerTest (14 ms total)

[----------] 1 test from ArenaTest
[ RUN      ] ArenaTest.GetOrCreateDisjunctionReturnsSameExprOnSubsequentCalls
[       OK ] ArenaTest.GetOrCreateDisjunctionReturnsSameExprOnSubsequentCalls (0 ms)
[----------] 1 test from ArenaTest (0 ms total)

[----------] 1 test from EnvironmentTest
[ RUN      ] EnvironmentTest.LambdaCapturingThisInFieldInitializer
[       OK ] EnvironmentTest.LambdaCapturingThisInFieldInitializer (3 ms)
[----------] 1 test from EnvironmentTest (3 ms total)
...