Add pass to strip debug info from MIR

[dsanders]

Summary

Removes:

• All LLVM-IR level debug info using StripDebugInfo()
• All DEBUG_VALUE MachineInstrs
• All DebugLocs from MachineInstrs

This is a more complete solution than the previous MIRPrinter
option that just causes it to neglect to print debug-locations.

[vsk]

Lgtm, thanks!

[dsanders]

llvm/lib/CodeGen/MachineStripDebug.cpp

56

I have one small fixup to make here which is that we should also remove the debugify metadata too. Otherwise mir-debugify can't be run more than once in a pipeline as you end up with an llvm.debugify with >2 operands

[dsanders]

Fix a few issues that came up when injecting into pipelines:

• Strip debugify metadata as well
• Strip 'Debug Info Version' module flag
• Don't strip certain broken DBG_VALUE instructions. AArch64 (incorrectly) inserts one when debug info is not present and tests depend on it.

[vsk]

Still looks good, with minor inline comments.

llvm/lib/CodeGen/MachineStripDebug.cpp

41

Could you file a PR on bugzilla about this test?

66

This should update Changed (not 100% sure, but StripDebugInfo may be a no-op if the function is empty).

74

We probably should update Changed here as well.

[dsanders]

Thanks

llvm/lib/CodeGen/MachineStripDebug.cpp

41

Sure

llvm/include/llvm/CodeGen/Passes.h

/// MachineDominanaceFrontier - This pass is a machine dominators analysis pass.
   /// Create Hardware Loop pass. \see HardwareLoops.cpp
   FunctionPass *createHardwareLoopsPass();
 
   /// Create IR Type Promotion pass. \see TypePromotion.cpp
   FunctionPass *createTypePromotionPass();
 
   /// Creates MIR Debugify pass. \see MachineDebugify.cpp
   ModulePass *createDebugifyMachineModulePass();
+
+  /// Creates MIR Strip Debug pass. \see MachineStripDebug.cpp
+  ModulePass *createStripDebugMachineModulePass();
 } // End llvm namespace
 
 #endif

llvm/include/llvm/InitializePasses.h

 void initializeStackProtectorPass(PassRegistry&);
 void initializeStackSafetyGlobalInfoWrapperPassPass(PassRegistry &);
 void initializeStackSafetyInfoWrapperPassPass(PassRegistry &);
 void initializeStackSlotColoringPass(PassRegistry&);
 void initializeStraightLineStrengthReducePass(PassRegistry&);
 void initializeStripDeadDebugInfoPass(PassRegistry&);
 void initializeStripDeadPrototypesLegacyPassPass(PassRegistry&);
 void initializeStripDebugDeclarePass(PassRegistry&);
+void initializeStripDebugMachineModulePass(PassRegistry &);
 void initializeStripGCRelocatesPass(PassRegistry&);
 void initializeStripNonDebugSymbolsPass(PassRegistry&);
 void initializeStripNonLineTableDebugInfoPass(PassRegistry&);
 void initializeStripSymbolsPass(PassRegistry&);
 void initializeStructurizeCFGPass(PassRegistry&);
 void initializeTailCallElimPass(PassRegistry&);
 void initializeTailDuplicatePass(PassRegistry&);
 void initializeTargetLibraryInfoWrapperPassPass(PassRegistry&);

llvm/lib/CodeGen/CMakeLists.txt

add_llvm_component_library(LLVMCodeGen
   MachinePipeliner.cpp
   MachinePostDominators.cpp
   MachineRegionInfo.cpp
   MachineRegisterInfo.cpp
   MachineScheduler.cpp
   MachineSink.cpp
   MachineSizeOpts.cpp
   MachineSSAUpdater.cpp
+  MachineStripDebug.cpp
   MachineTraceMetrics.cpp
   MachineVerifier.cpp
   ModuloSchedule.cpp
   PatchableFunction.cpp
   MBFIWrapper.cpp
   MIRPrinter.cpp
   MIRPrintingPass.cpp
   MacroFusion.cpp

llvm/lib/CodeGen/CodeGen.cpp

void llvm::initializeCodeGen(PassRegistry &Registry) {
   initializeScalarizeMaskedMemIntrinPass(Registry);
   initializeShrinkWrapPass(Registry);
   initializeSjLjEHPreparePass(Registry);
   initializeSlotIndexesPass(Registry);
   initializeStackColoringPass(Registry);
   initializeStackMapLivenessPass(Registry);
   initializeStackProtectorPass(Registry);
   initializeStackSlotColoringPass(Registry);
+  initializeStripDebugMachineModulePass(Registry);
   initializeTailDuplicatePass(Registry);
   initializeTargetPassConfigPass(Registry);
   initializeTwoAddressInstructionPassPass(Registry);
   initializeTypePromotionPass(Registry);
   initializeUnpackMachineBundlesPass(Registry);
   initializeUnreachableBlockElimLegacyPassPass(Registry);
   initializeUnreachableMachineBlockElimPass(Registry);
   initializeVirtRegMapPass(Registry);

llvm/lib/CodeGen/MachineStripDebug.cpp

+//===- MachineStripDebug.cpp - Strip debug info ---------------------------===//
+//
+// 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 removes debug info from everything. It can be used to ensure
+/// tests can be debugified without affecting the output MIR.
+//===----------------------------------------------------------------------===//
+
+#include "llvm/CodeGen/MachineFunctionPass.h"
+#include "llvm/CodeGen/MachineModuleInfo.h"
+#include "llvm/CodeGen/Passes.h"
+#include "llvm/IR/DebugInfo.h"
+#include "llvm/InitializePasses.h"
+
+#define DEBUG_TYPE "mir-strip-debug"
+
+using namespace llvm;
+
+namespace {
+
+struct StripDebugMachineModule : public ModulePass {
+  bool runOnModule(Module &M) override {
+    MachineModuleInfo &MMI =
+        getAnalysis<MachineModuleInfoWrapperPass>().getMMI();
+
+    bool Changed = false;
+    for (Function &F : M.functions()) {
+      MachineFunction &MF = MMI.getOrCreateMachineFunction(F);
+      for (MachineBasicBlock &MBB : MF) {
+        for (MachineBasicBlock::iterator I = MBB.begin(), E = MBB.end();
+             I != E;) {
+          if (I->isDebugInstr()) {
+            // FIXME: We should remove all of them. However, AArch64 emits an
+            //        invalid `DBG_VALUE $lr` with only one operand instead of
+            //        the usual three and has a test that depends on it's
+            //        preservation. Preserve it for now.
+            if (I->getNumOperands() > 1) {
+              LLVM_DEBUG(dbgs() << "Removing debug instruction " << *I);
+              I = MBB.erase(I);
+              Changed |= true;
+              continue;
+            }
+          }
+          if (I->getDebugLoc()) {
+            LLVM_DEBUG(dbgs() << "Removing location " << *I);
+            I->setDebugLoc(DebugLoc());
+            Changed |= true;
+            ++I;
+            continue;
+          }
+          LLVM_DEBUG(dbgs() << "Keeping " << *I);
+          ++I;
+        }
+      }
+    }
+
+    Changed |= StripDebugInfo(M);
+
+    NamedMDNode *NMD = M.getNamedMetadata("llvm.debugify");
+    if (NMD) {
+      NMD->eraseFromParent();
+      Changed |= true;
+    }
+
+    NMD = M.getModuleFlagsMetadata();
+    if (NMD) {
+      // There must be an easier way to remove an operand from a NamedMDNode.
+      SmallVector<MDNode *, 4> Flags;
+      for (MDNode *Flag : NMD->operands())
+        Flags.push_back(Flag);
+      NMD->clearOperands();
+      for (MDNode *Flag : Flags) {
+        MDString *Key = dyn_cast_or_null<MDString>(Flag->getOperand(1));
+        if (Key->getString() == "Debug Info Version") {
+          Changed |= true;
+          continue;
+        }
+        NMD->addOperand(Flag);
+      }
+      // If we left it empty we might as well remove it.
+      if (NMD->getNumOperands() == 0)
+        NMD->eraseFromParent();
+    }
+
+    return Changed;
+  }
+
+  StripDebugMachineModule() : ModulePass(ID) {}
+
+  void getAnalysisUsage(AnalysisUsage &AU) const override {
+    AU.addRequired<MachineModuleInfoWrapperPass>();
+    AU.addPreserved<MachineModuleInfoWrapperPass>();
+  }
+
+  static char ID; // Pass identification.
+};
+char StripDebugMachineModule::ID = 0;
+
+} // end anonymous namespace
+
+INITIALIZE_PASS_BEGIN(StripDebugMachineModule, DEBUG_TYPE,
+                      "Machine Strip Debug Module", false, false)
+INITIALIZE_PASS_END(StripDebugMachineModule, DEBUG_TYPE,
+                    "Machine Strip Debug Module", false, false)
+
+ModulePass *createStripDebugMachineModulePass() {
+  return new StripDebugMachineModule();
+}

llvm/test/CodeGen/Generic/MIRDebugify/locations.mir

 # RUN: llc -run-pass=mir-debugify -o - %s | FileCheck --check-prefixes=ALL,VALUE %s
 # RUN: llc -run-pass=mir-debugify -debugify-level=locations -o - %s | FileCheck --check-prefixes=ALL --implicit-check-not=dbg.value %s
+# RUN: llc -run-pass=mir-debugify,mir-strip-debug,mir-debugify -o - %s | FileCheck --check-prefixes=ALL,VALUE %s
 --- |
   ; ModuleID = 'loc-only.ll'
   source_filename = "loc-only.ll"
   
   ; ALL-LABEL: @test
   define i32 @test(i32 %a, i32 %b) {
     %add = add i32 %a, 2
   ; ALL-NEXT:  %add = add i32 %a, 2, !dbg [[L1:![0-9]+]]
   ; VALUE-NEXT: call void @llvm.dbg.value(metadata i32 %add, metadata [[add:![0-9]+]], metadata !DIExpression()), !dbg [[L1]]
     %sub = sub i32 %add, %b
   ; ALL-NEXT: %sub = sub i32 %add, %b, !dbg [[L2:![0-9]+]]
   ; VALUE-NEXT: call void @llvm.dbg.value(metadata i32 %sub, metadata [[sub:![0-9]+]], metadata !DIExpression()), !dbg [[L2]]
   ; ALL-NEXT: ret i32 %sub, !dbg [[L3:![0-9]+]]
     ret i32 %sub
   }
 
+  ; CHECK: !llvm.dbg.cu = !{!0}
+  ; CHECK: !llvm.debugify =
+  ; CHECK: !llvm.module.flags = !{![[VERSION:[0-9]+]]}
+  ; CHECK: !0 = distinct !DICompileUnit(
+  ; CHECK: ![[VERSION]] = !{i32 2, !"Debug Info Version", i32 3}
+
 ...
 ---
 name:            test
 body:             |
   bb.1 (%ir-block.0):
     %0:_(s32) = IMPLICIT_DEF
     %1:_(s32) = IMPLICIT_DEF
     %2:_(s32) = G_CONSTANT i32 2

llvm/test/CodeGen/Generic/MIRStripDebug/all.mir

+# RUN: llc -run-pass=mir-strip-debug -o - %s | FileCheck %s
+# RUN: llc -run-pass=mir-strip-debug,mir-debugify,mir-strip-debug -o - %s | FileCheck %s
+--- |
+  ; ModuleID = 'loc-only.ll'
+  source_filename = "loc-only.ll"
+  
+  define i32 @test(i32 %a, i32 %b) !dbg !6 {
+    %add = add i32 %a, 2, !dbg !12
+    call void @llvm.dbg.value(metadata i32 %add, metadata !9, metadata !DIExpression()), !dbg !12
+    %sub = sub i32 %add, %b, !dbg !13
+    call void @llvm.dbg.value(metadata i32 %sub, metadata !11, metadata !DIExpression()), !dbg !13
+    ret i32 %sub, !dbg !14
+  }
+  ; CHECK-LABEL: define i32 @test(i32 %a, i32 %b) {
+  ; CHECK-NEXT:    %add = add i32 %a, 2
+  ; CHECK-NEXT:    %sub = sub i32 %add, %b
+  ; CHECK-NEXT:    ret i32 %sub
+  ; CHECK-NEXT:  }
+  
+  ; Function Attrs: nounwind readnone speculatable willreturn
+  declare void @llvm.dbg.value(metadata, metadata, metadata) #0
+  
+  ; Function Attrs: nounwind
+  declare void @llvm.stackprotector(i8*, i8**) #1
+  
+  attributes #0 = { nounwind readnone speculatable willreturn }
+  attributes #1 = { nounwind }
+  
+  !llvm.dbg.cu = !{!0}
+  ; CHECK-NOT: !llvm.dbg.cu
+  !llvm.debugify = !{!3, !4}
+  ; CHECK-NOT: !llvm.debugify
+  !llvm.module.flags = !{!5}
+  ; CHECK-NOT: !llvm.module.flags
+  
+  ; CHECK-NOT: !DI
+  !0 = distinct !DICompileUnit(language: DW_LANG_C, file: !1, producer: "debugify", isOptimized: true, runtimeVersion: 0, emissionKind: FullDebug, enums: !2)
+  !1 = !DIFile(filename: "<stdin>", directory: "/")
+  !2 = !{}
+  !3 = !{i32 3}
+  !4 = !{i32 2}
+  !5 = !{i32 2, !"Debug Info Version", i32 3}
+  !6 = distinct !DISubprogram(name: "test", linkageName: "test", scope: null, file: !1, line: 1, type: !7, scopeLine: 1, spFlags: DISPFlagDefinition | DISPFlagOptimized, unit: !0, retainedNodes: !8)
+  !7 = !DISubroutineType(types: !2)
+  !8 = !{!9, !11}
+  !9 = !DILocalVariable(name: "1", scope: !6, file: !1, line: 1, type: !10)
+  !10 = !DIBasicType(name: "ty32", size: 32, encoding: DW_ATE_unsigned)
+  !11 = !DILocalVariable(name: "2", scope: !6, file: !1, line: 2, type: !10)
+  !12 = !DILocation(line: 1, column: 1, scope: !6)
+  !13 = !DILocation(line: 2, column: 1, scope: !6)
+  !14 = !DILocation(line: 3, column: 1, scope: !6)
+
+...
+---
+name:            test
+body:             |
+  bb.1 (%ir-block.0):
+    %0:_(s32) = G_IMPLICIT_DEF
+    %1:_(s32) = G_IMPLICIT_DEF
+    %2:_(s32) = G_CONSTANT i32 2, debug-location !DILocation(line: 0, scope: !6)
+    %3:_(s32) = G_ADD %0, %2, debug-location !12
+    DBG_VALUE %3(s32), $noreg, !9, !DIExpression(), debug-location !12
+    %4:_(s32) = G_SUB %3, %1, debug-location !13
+    DBG_VALUE %4(s32), $noreg, !11, !DIExpression(), debug-location !13
+
+    ; CHECK-LABEL: body:
+    ; CHECK-NEXT:    bb
+    ; CHECK-NEXT:      %0:_(s32) = G_IMPLICIT_DEF{{$}}
+    ; CHECK-NEXT:      %1:_(s32) = G_IMPLICIT_DEF{{$}}
+    ; CHECK-NEXT:      %2:_(s32) = G_CONSTANT i32 2{{$}}
+    ; CHECK-NEXT:      %3:_(s32) = G_ADD %0, %2{{$}}
+    ; CHECK-NEXT:      %4:_(s32) = G_SUB %3, %1{{$}}
+...

llvm/test/CodeGen/Generic/MIRStripDebug/multiple-moduleflags.mir

+# RUN: llc -run-pass=mir-strip-debug -o - %s | FileCheck %s
+# RUN: llc -run-pass=mir-strip-debug,mir-debugify,mir-strip-debug -o - %s | FileCheck %s
+--- |
+  ; ModuleID = 'loc-only.ll'
+  source_filename = "loc-only.ll"
+  
+  define i32 @test(i32 %a, i32 %b) !dbg !6 {
+    %add = add i32 %a, 2, !dbg !12
+    call void @llvm.dbg.value(metadata i32 %add, metadata !9, metadata !DIExpression()), !dbg !12
+    %sub = sub i32 %add, %b, !dbg !13
+    call void @llvm.dbg.value(metadata i32 %sub, metadata !11, metadata !DIExpression()), !dbg !13
+    ret i32 %sub, !dbg !14
+  }
+  ; CHECK-LABEL: define i32 @test(i32 %a, i32 %b) {
+  ; CHECK-NEXT:    %add = add i32 %a, 2
+  ; CHECK-NEXT:    %sub = sub i32 %add, %b
+  ; CHECK-NEXT:    ret i32 %sub
+  ; CHECK-NEXT:  }
+  
+  ; Function Attrs: nounwind readnone speculatable willreturn
+  declare void @llvm.dbg.value(metadata, metadata, metadata) #0
+  
+  ; Function Attrs: nounwind
+  declare void @llvm.stackprotector(i8*, i8**) #1
+  
+  attributes #0 = { nounwind readnone speculatable willreturn }
+  attributes #1 = { nounwind }
+  
+  !llvm.dbg.cu = !{!0}
+  ; CHECK-NOT: !llvm.dbg.cu
+  !llvm.debugify = !{!3, !4}
+  ; CHECK-NOT: !llvm.debugify
+  !llvm.module.flags = !{!5, !15}
+  ; CHECK: !llvm.module.flags = !{!0}
+  ; CHECK: !0 = !{i32 2, !"Another Flag", i32 3}
+  
+  ; CHECK-NOT: !DI
+  !0 = distinct !DICompileUnit(language: DW_LANG_C, file: !1, producer: "debugify", isOptimized: true, runtimeVersion: 0, emissionKind: FullDebug, enums: !2)
+  !1 = !DIFile(filename: "<stdin>", directory: "/")
+  !2 = !{}
+  !3 = !{i32 3}
+  !4 = !{i32 2}
+  !5 = !{i32 2, !"Debug Info Version", i32 3}
+  !6 = distinct !DISubprogram(name: "test", linkageName: "test", scope: null, file: !1, line: 1, type: !7, scopeLine: 1, spFlags: DISPFlagDefinition | DISPFlagOptimized, unit: !0, retainedNodes: !8)
+  !7 = !DISubroutineType(types: !2)
+  !8 = !{!9, !11}
+  !9 = !DILocalVariable(name: "1", scope: !6, file: !1, line: 1, type: !10)
+  !10 = !DIBasicType(name: "ty32", size: 32, encoding: DW_ATE_unsigned)
+  !11 = !DILocalVariable(name: "2", scope: !6, file: !1, line: 2, type: !10)
+  !12 = !DILocation(line: 1, column: 1, scope: !6)
+  !13 = !DILocation(line: 2, column: 1, scope: !6)
+  !14 = !DILocation(line: 3, column: 1, scope: !6)
+  !15 = !{i32 2, !"Another Flag", i32 3}
+
+...
+---
+name:            test
+body:             |
+  bb.1 (%ir-block.0):
+    %0:_(s32) = G_IMPLICIT_DEF
+    %1:_(s32) = G_IMPLICIT_DEF
+    %2:_(s32) = G_CONSTANT i32 2, debug-location !DILocation(line: 0, scope: !6)
+    %3:_(s32) = G_ADD %0, %2, debug-location !12
+    DBG_VALUE %3(s32), $noreg, !9, !DIExpression(), debug-location !12
+    %4:_(s32) = G_SUB %3, %1, debug-location !13
+    DBG_VALUE %4(s32), $noreg, !11, !DIExpression(), debug-location !13
+
+    ; CHECK-LABEL: body:
+    ; CHECK-NEXT:    bb
+    ; CHECK-NEXT:      %0:_(s32) = G_IMPLICIT_DEF{{$}}
+    ; CHECK-NEXT:      %1:_(s32) = G_IMPLICIT_DEF{{$}}
+    ; CHECK-NEXT:      %2:_(s32) = G_CONSTANT i32 2{{$}}
+    ; CHECK-NEXT:      %3:_(s32) = G_ADD %0, %2{{$}}
+    ; CHECK-NEXT:      %4:_(s32) = G_SUB %3, %1{{$}}
+...