[llvm] r220321 - [PBQP] Teach PassConfig to tell if the default register allocator is used.

[Arnaud A. de Grandmaison]

Author: aadg
Date: Tue Oct 21 15:47:22 2014
New Revision: 220321

URL: http://llvm.org/viewvc/llvm-project?rev=220321&view=rev
Log:
[PBQP] Teach PassConfig to tell if the default register allocator is used.

This enables targets to adapt their pass pipeline to the register
allocator in use. For example, with the AArch64 backend, using PBQP
with the cortex-a57, the FPLoadBalancing pass is no longer necessary.

Modified:
    llvm/trunk/include/llvm/CodeGen/Passes.h
    llvm/trunk/lib/CodeGen/Passes.cpp
    llvm/trunk/lib/Target/AArch64/AArch64TargetMachine.cpp
    llvm/trunk/lib/Target/AArch64/AArch64TargetMachine.h
    llvm/trunk/test/CodeGen/AArch64/PBQP.ll

Modified: llvm/trunk/include/llvm/CodeGen/Passes.h
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/include/llvm/CodeGen/Passes.h?rev=220321&r1=220320&r2=220321&view=diff
==============================================================================
--- llvm/trunk/include/llvm/CodeGen/Passes.h (original)
+++ llvm/trunk/include/llvm/CodeGen/Passes.h Tue Oct 21 15:47:22 2014
@@ -178,6 +178,10 @@ public:
   /// Return true if the optimized regalloc pipeline is enabled.
   bool getOptimizeRegAlloc() const;
 
+  /// Return true if the default global register allocator is in use and
+  /// has not be overriden on the command line with '-regalloc=...'
+  bool usingDefaultRegAlloc() const;
+
   /// Add common target configurable passes that perform LLVM IR to IR
   /// transforms following machine independent optimization.
   virtual void addIRPasses();

Modified: llvm/trunk/lib/CodeGen/Passes.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/CodeGen/Passes.cpp?rev=220321&r1=220320&r2=220321&view=diff
==============================================================================
--- llvm/trunk/lib/CodeGen/Passes.cpp (original)
+++ llvm/trunk/lib/CodeGen/Passes.cpp Tue Oct 21 15:47:22 2014
@@ -690,6 +690,12 @@ FunctionPass *TargetPassConfig::createRe
   return createTargetRegisterAllocator(Optimized);
 }
 
+/// Return true if the default global register allocator is in use and
+/// has not be overriden on the command line with '-regalloc=...'
+bool TargetPassConfig::usingDefaultRegAlloc() const {
+  return RegAlloc == &useDefaultRegisterAllocator;
+}
+
 /// Add the minimum set of target-independent passes that are required for
 /// register allocation. No coalescing or scheduling.
 void TargetPassConfig::addFastRegAlloc(FunctionPass *RegAllocPass) {

Modified: llvm/trunk/lib/Target/AArch64/AArch64TargetMachine.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/AArch64/AArch64TargetMachine.cpp?rev=220321&r1=220320&r2=220321&view=diff
==============================================================================
--- llvm/trunk/lib/Target/AArch64/AArch64TargetMachine.cpp (original)
+++ llvm/trunk/lib/Target/AArch64/AArch64TargetMachine.cpp Tue Oct 21 15:47:22 2014
@@ -76,11 +76,6 @@ EnableCondOpt("aarch64-condopt",
               cl::init(true), cl::Hidden);
 
 static cl::opt<bool>
-EnablePBQP("aarch64-pbqp", cl::Hidden,
-           cl::desc("Use PBQP register allocator (experimental)"),
-           cl::init(false));
-
-static cl::opt<bool>
 EnableA53Fix835769("aarch64-fix-cortex-a53-835769", cl::Hidden,
                 cl::desc("Work around Cortex-A53 erratum 835769"),
                 cl::init(false));
@@ -101,14 +96,8 @@ AArch64TargetMachine::AArch64TargetMachi
                                            CodeGenOpt::Level OL,
                                            bool LittleEndian)
     : LLVMTargetMachine(T, TT, CPU, FS, Options, RM, CM, OL),
-      Subtarget(TT, CPU, FS, *this, LittleEndian), isLittle(LittleEndian),
-      usingPBQP(false) {
+      Subtarget(TT, CPU, FS, *this, LittleEndian), isLittle(LittleEndian) {
   initAsmInfo();
-
-  if (EnablePBQP && Subtarget.isCortexA57() && OL != CodeGenOpt::None) {
-    usingPBQP = true;
-    RegisterRegAlloc::setDefault(createDefaultPBQPRegisterAllocator);
-  }
 }
 
 const AArch64Subtarget *
@@ -263,7 +252,7 @@ bool AArch64PassConfig::addPostRegAlloc(
     addPass(createAArch64DeadRegisterDefinitions());
   if (TM->getOptLevel() != CodeGenOpt::None &&
       TM->getSubtarget<AArch64Subtarget>().isCortexA57() &&
-      !static_cast<const AArch64TargetMachine *>(TM)->isPBQPUsed())
+      usingDefaultRegAlloc())
     // Improve performance for some FP/SIMD code for A57.
     addPass(createAArch64A57FPLoadBalancing());
   return true;

Modified: llvm/trunk/lib/Target/AArch64/AArch64TargetMachine.h
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/AArch64/AArch64TargetMachine.h?rev=220321&r1=220320&r2=220321&view=diff
==============================================================================
--- llvm/trunk/lib/Target/AArch64/AArch64TargetMachine.h (original)
+++ llvm/trunk/lib/Target/AArch64/AArch64TargetMachine.h Tue Oct 21 15:47:22 2014
@@ -43,12 +43,8 @@ public:
   /// \brief Register AArch64 analysis passes with a pass manager.
   void addAnalysisPasses(PassManagerBase &PM) override;
 
-  /// \brief Query if the PBQP register allocator is being used
-  bool isPBQPUsed() const { return usingPBQP; }
-
 private:
   bool isLittle;
-  bool usingPBQP;
 };
 
 // AArch64leTargetMachine - AArch64 little endian target machine.

Modified: llvm/trunk/test/CodeGen/AArch64/PBQP.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/AArch64/PBQP.ll?rev=220321&r1=220320&r2=220321&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/AArch64/PBQP.ll (original)
+++ llvm/trunk/test/CodeGen/AArch64/PBQP.ll Tue Oct 21 15:47:22 2014
@@ -1,9 +1,9 @@
-; RUN: llc -mtriple=aarch64-linux-gnu -mcpu=cortex-a57 -aarch64-pbqp -o - %s | FileCheck %s
+; RUN: llc -mtriple=aarch64-linux-gnu -mcpu=cortex-a57 -regalloc=pbqp -pbqp-coalescing -o - %s | FileCheck %s
 
 define i32 @foo(i32 %a) {
 ; CHECK-LABEL: foo:
 ; CHECK: bl bar
-; CHECK-NEXT: bl baz
+; CHECK: bl baz
   %call = call i32 @bar(i32 %a)
   %call1 = call i32 @baz(i32 %call)
   ret i32 %call1