Enable aggressive constant merge in GlobalMerge for AIX (#113956)

syzaara · web-flow · commit ccddd1360243 · 2024-10-31T11:22:48.000-04:00
Enable merging all constants without looking at use in GlobalMerge by
default to replace PPCMergeStringPool pass on AIX.
diff --git a/llvm/include/llvm/CodeGen/GlobalMerge.h b/llvm/include/llvm/CodeGen/GlobalMerge.h
@@ -30,6 +30,9 @@ struct GlobalMergeOptions {
   bool MergeExternal = true;
   /// Whether we should merge constant global variables.
   bool MergeConstantGlobals = false;
+  /// Whether we should merge constant global variables aggressively without
+  /// looking at use.
+  bool MergeConstAggressive = false;
   /// Whether we should try to optimize for size only.
   /// Currently, this applies a dead simple heuristic: only consider globals
   /// used in minsize functions for merging.
diff --git a/llvm/include/llvm/CodeGen/Passes.h b/llvm/include/llvm/CodeGen/Passes.h
@@ -480,7 +480,8 @@ namespace llvm {
   Pass *createGlobalMergePass(const TargetMachine *TM, unsigned MaximalOffset,
                               bool OnlyOptimizeForSize = false,
                               bool MergeExternalByDefault = false,
-                              bool MergeConstantByDefault = false);
+                              bool MergeConstantByDefault = false,
+                              bool MergeConstAggressiveByDefault = false);
 
   /// This pass splits the stack into a safe stack and an unsafe stack to
   /// protect against stack-based overflow vulnerabilities.
diff --git a/llvm/lib/CodeGen/GlobalMerge.cpp b/llvm/lib/CodeGen/GlobalMerge.cpp
@@ -202,12 +202,13 @@ class GlobalMerge : public FunctionPass {
 
   explicit GlobalMerge(const TargetMachine *TM, unsigned MaximalOffset,
                        bool OnlyOptimizeForSize, bool MergeExternalGlobals,
-                       bool MergeConstantGlobals)
+                       bool MergeConstantGlobals, bool MergeConstAggressive)
       : FunctionPass(ID), TM(TM) {
     Opt.MaxOffset = MaximalOffset;
     Opt.SizeOnly = OnlyOptimizeForSize;
     Opt.MergeExternal = MergeExternalGlobals;
     Opt.MergeConstantGlobals = MergeConstantGlobals;
+    Opt.MergeConstAggressive = MergeConstAggressive;
     initializeGlobalMergePass(*PassRegistry::getPassRegistry());
   }
 
@@ -268,7 +269,7 @@ bool GlobalMergeImpl::doMerge(SmallVectorImpl<GlobalVariable *> &Globals,
       });
 
   // If we want to just blindly group all globals together, do so.
-  if (!GlobalMergeGroupByUse || (GlobalMergeAllConst && isConst)) {
+  if (!GlobalMergeGroupByUse || (Opt.MergeConstAggressive && isConst)) {
     BitVector AllGlobals(Globals.size());
     AllGlobals.set();
     return doMerge(Globals, AllGlobals, M, isConst, AddrSpace);
@@ -758,10 +759,14 @@ bool GlobalMergeImpl::run(Module &M) {
 Pass *llvm::createGlobalMergePass(const TargetMachine *TM, unsigned Offset,
                                   bool OnlyOptimizeForSize,
                                   bool MergeExternalByDefault,
-                                  bool MergeConstantByDefault) {
+                                  bool MergeConstantByDefault,
+                                  bool MergeConstAggressiveByDefault) {
   bool MergeExternal = (EnableGlobalMergeOnExternal == cl::BOU_UNSET) ?
     MergeExternalByDefault : (EnableGlobalMergeOnExternal == cl::BOU_TRUE);
   bool MergeConstant = EnableGlobalMergeOnConst || MergeConstantByDefault;
+  bool MergeConstAggressive = GlobalMergeAllConst.getNumOccurrences() > 0
+                                  ? GlobalMergeAllConst
+                                  : MergeConstAggressiveByDefault;
   return new GlobalMerge(TM, Offset, OnlyOptimizeForSize, MergeExternal,
-                         MergeConstant);
+                         MergeConstant, MergeConstAggressive);
 }
diff --git a/llvm/lib/Target/PowerPC/PPCTargetMachine.cpp b/llvm/lib/Target/PowerPC/PPCTargetMachine.cpp
@@ -505,10 +505,13 @@ bool PPCPassConfig::addPreISel() {
           ? EnableGlobalMerge
           : (TM->getTargetTriple().isOSAIX() &&
              getOptLevel() != CodeGenOptLevel::None))
-    addPass(
-        createGlobalMergePass(TM, GlobalMergeMaxOffset, false, false, true));
+    addPass(createGlobalMergePass(TM, GlobalMergeMaxOffset, false, false, true,
+                                  true));
 
-  if (MergeStringPool && getOptLevel() != CodeGenOptLevel::None)
+  if ((MergeStringPool.getNumOccurrences() > 0)
+          ? MergeStringPool
+          : (TM->getTargetTriple().isOSLinux() &&
+             getOptLevel() != CodeGenOptLevel::None))
     addPass(createPPCMergeStringPoolPass());
 
   if (!DisableInstrFormPrep && getOptLevel() != CodeGenOptLevel::None)
diff --git a/llvm/test/CodeGen/PowerPC/O3-pipeline.ll b/llvm/test/CodeGen/PowerPC/O3-pipeline.ll
@@ -70,9 +70,6 @@
 ; CHECK-NEXT:       CodeGen Prepare
 ; CHECK-NEXT:       Dominator Tree Construction
 ; CHECK-NEXT:       Exception handling preparation
-; CHECK-NEXT:       PPC Merge String Pool
-; CHECK-NEXT:       FunctionPass Manager
-; CHECK-NEXT:       Dominator Tree Construction
 ; CHECK-NEXT:       Natural Loop Information
 ; CHECK-NEXT:       Scalar Evolution Analysis
 ; CHECK-NEXT:       Prepare loop for ppc preferred instruction forms
diff --git a/llvm/test/CodeGen/PowerPC/aix-xcoff-mergeable-const.ll b/llvm/test/CodeGen/PowerPC/aix-xcoff-mergeable-const.ll
@@ -1,13 +1,13 @@
 ; This file tests the codegen of mergeable const in AIX assembly.
 ; This file also tests mergeable const in XCOFF object file generation.
 ; RUN: llc -verify-machineinstrs -mcpu=pwr4 -mattr=-altivec -mtriple powerpc-ibm-aix-xcoff \
-; RUN:     -data-sections=false -xcoff-traceback-table=false < %s | \
+; RUN:     -global-merge-all-const=false -data-sections=false -xcoff-traceback-table=false < %s | \
 ; RUN:   FileCheck --check-prefixes=CHECK,CHECK32 %s
 ; RUN: llc -verify-machineinstrs -mcpu=pwr4 -mattr=-altivec -mtriple powerpc64-ibm-aix-xcoff \
-; RUN:     -xcoff-traceback-table=false -data-sections=false < %s | \
+; RUN:     -global-merge-all-const=false -xcoff-traceback-table=false -data-sections=false < %s | \
 ; RUN:   FileCheck --check-prefixes=CHECK,CHECK64 %s
 ; RUN: llc -verify-machineinstrs -mcpu=pwr4 -mattr=-altivec -mtriple powerpc-ibm-aix-xcoff \
-; RUN:     -xcoff-traceback-table=false -data-sections=false -filetype=obj -o %t.o < %s
+; RUN:     -global-merge-all-const=false -xcoff-traceback-table=false -data-sections=false -filetype=obj -o %t.o < %s
 ; RUN: llvm-objdump -D %t.o | FileCheck --check-prefix=CHECKOBJ %s
 ; RUN: llvm-readobj -s %t.o | FileCheck --check-prefix=CHECKSYM %s
 
diff --git a/llvm/test/CodeGen/PowerPC/aix-xcoff-mergeable-str.ll b/llvm/test/CodeGen/PowerPC/aix-xcoff-mergeable-str.ll
@@ -4,12 +4,15 @@
 ; tests for XCOFF object files.
 
 ; RUN: llc -verify-machineinstrs -mcpu=pwr4 -mattr=-altivec -xcoff-traceback-table=false \
-; RUN:     -mtriple powerpc-ibm-aix-xcoff  -data-sections=false -ppc-merge-string-pool=false < %s | FileCheck %s
+; RUN:     -mtriple powerpc-ibm-aix-xcoff  -data-sections=false -ppc-merge-string-pool=false \
+; RUN:     -global-merge-all-const=false < %s | FileCheck %s
 ; RUN: llc -verify-machineinstrs -mcpu=pwr4 -mattr=-altivec -xcoff-traceback-table=false \
-; RUN:     -mtriple powerpc64-ibm-aix-xcoff -data-sections=false -ppc-merge-string-pool=false < %s | FileCheck %s
+; RUN:     -mtriple powerpc64-ibm-aix-xcoff -data-sections=false -ppc-merge-string-pool=false \
+; RUN:     -global-merge-all-const=false < %s | FileCheck %s
 
 ; RUN: llc -verify-machineinstrs -mcpu=pwr4 -mattr=-altivec -mtriple powerpc-ibm-aix-xcoff \
-; RUN:     -xcoff-traceback-table=false -data-sections=false -ppc-merge-string-pool=false -filetype=obj -o %t.o < %s
+; RUN:     -xcoff-traceback-table=false -data-sections=false -ppc-merge-string-pool=false \
+; RUN:     -global-merge-all-const=false -filetype=obj -o %t.o < %s
 ; RUN: llvm-objdump -D %t.o | FileCheck --check-prefix=CHECKOBJ %s
 
 @magic16 = private unnamed_addr constant [4 x i16] [i16 264, i16 272, i16 213, i16 0], align 2
diff --git a/llvm/test/CodeGen/PowerPC/aix-xcoff-used-with-stringpool.ll b/llvm/test/CodeGen/PowerPC/aix-xcoff-used-with-stringpool.ll
@@ -1,10 +1,12 @@
 ;; Test that the string pooling pass does not pool globals that are
 ;; in llvm.used or in llvm.compiler.used.
 
-; RUN: llc -verify-machineinstrs -mcpu=pwr8 -mtriple powerpc-ibm-aix-xcoff -data-sections=false < %s | \
+; RUN: llc -verify-machineinstrs -mcpu=pwr8 -mtriple powerpc-ibm-aix-xcoff \
+; RUN:   -ppc-merge-string-pool=true -global-merge-all-const=false -data-sections=false < %s | \
 ; RUN:   FileCheck %s
 
-; RUN: llc -verify-machineinstrs -mcpu=pwr8 -mtriple powerpc64-ibm-aix-xcoff -data-sections=false < %s | \
+; RUN: llc -verify-machineinstrs -mcpu=pwr8 -mtriple powerpc64-ibm-aix-xcoff \
+; RUN:   -ppc-merge-string-pool=true -global-merge-all-const=false -data-sections=false < %s | \
 ; RUN:   FileCheck %s
 
 @keep_this = internal constant [5 x i8] c"keep1", align 1
diff --git a/llvm/test/CodeGen/PowerPC/merge-private.ll b/llvm/test/CodeGen/PowerPC/merge-private.ll
@@ -1,9 +1,9 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
 ; RUN: llc -verify-machineinstrs -mtriple powerpc64-ibm-aix-xcoff -mcpu=pwr8 \
-; RUN:     -ppc-asm-full-reg-names -ppc-global-merge=true < %s | FileCheck %s \
+; RUN:     -ppc-asm-full-reg-names < %s | FileCheck %s \
 ; RUN:     --check-prefix=AIX64
 ; RUN: llc -verify-machineinstrs -mtriple powerpc-ibm-aix-xcoff -mcpu=pwr8 \
-; RUN:     -ppc-asm-full-reg-names -ppc-global-merge=true < %s | FileCheck %s \
+; RUN:     -ppc-asm-full-reg-names < %s | FileCheck %s \
 ; RUN:     --check-prefix=AIX32
 ; RUN: llc -verify-machineinstrs -mtriple powerpc64le-unknown-linux -mcpu=pwr8 \
 ; RUN:     -ppc-asm-full-reg-names -ppc-global-merge=true < %s | FileCheck %s \
diff --git a/llvm/test/CodeGen/PowerPC/mergeable-string-pool-large.ll b/llvm/test/CodeGen/PowerPC/mergeable-string-pool-large.ll
diff --git a/llvm/test/CodeGen/PowerPC/mergeable-string-pool-tls.ll b/llvm/test/CodeGen/PowerPC/mergeable-string-pool-tls.ll
diff --git a/llvm/test/CodeGen/PowerPC/mergeable-string-pool.ll b/llvm/test/CodeGen/PowerPC/mergeable-string-pool.ll
