[AArch64CondBrTuning] Ignore debug insts when scanning for NZCV clobbers [10/14]

Summary:
This fixes several instances in which condbr optimization was missed
due to a debug instruction appearing as a bogus NZCV clobber.

Reviewers: aemerson, paquette

Subscribers: kristof.beyls, hiraditya, jfb, danielkiss, aprantl, llvm-commits

Tags: #llvm

Differential Revision: https://reviews.llvm.org/D78264

llvm/lib/Target/AArch64/AArch64CondBrTuning.cpp

@@ -194,12 +194,8 @@ bool AArch64CondBrTuning::tryToTuneBranch(MachineInstr &MI,
 
       // There must not be any instruction between DefMI and MI that clobbers or
       // reads NZCV.
-      MachineBasicBlock::iterator I(DefMI), E(MI);
-      for (I = std::next(I); I != E; ++I) {
-        if (I->modifiesRegister(AArch64::NZCV, TRI) ||
-            I->readsRegister(AArch64::NZCV, TRI))
-          return false;
-      }
+      if (isNZCVTouchedInInstructionRange(DefMI, MI, TRI))
+        return false;
       LLVM_DEBUG(dbgs() << "  Replacing instructions:\n    ");
       LLVM_DEBUG(DefMI.print(dbgs()));
       LLVM_DEBUG(dbgs() << "    ");
@@ -253,12 +249,8 @@ bool AArch64CondBrTuning::tryToTuneBranch(MachineInstr &MI,
         return false;
       // There must not be any instruction between DefMI and MI that clobbers or
       // reads NZCV.
-      MachineBasicBlock::iterator I(DefMI), E(MI);
-      for (I = std::next(I); I != E; ++I) {
-        if (I->modifiesRegister(AArch64::NZCV, TRI) ||
-            I->readsRegister(AArch64::NZCV, TRI))
-          return false;
-      }
+      if (isNZCVTouchedInInstructionRange(DefMI, MI, TRI))
+        return false;
       LLVM_DEBUG(dbgs() << "  Replacing instructions:\n    ");
       LLVM_DEBUG(DefMI.print(dbgs()));
       LLVM_DEBUG(dbgs() << "    ");

llvm/lib/Target/AArch64/AArch64InstrInfo.cpp

@@ -3173,6 +3173,17 @@ void AArch64InstrInfo::loadRegFromStackSlot(
   MI.addMemOperand(MMO);
 }
 
+bool llvm::isNZCVTouchedInInstructionRange(const MachineInstr &DefMI,
+                                           const MachineInstr &UseMI,
+                                           const TargetRegisterInfo *TRI) {
+  return any_of(instructionsWithoutDebug(std::next(DefMI.getIterator()),
+                                         UseMI.getIterator()),
+                [TRI](const MachineInstr &I) {
+                  return I.modifiesRegister(AArch64::NZCV, TRI) ||
+                         I.readsRegister(AArch64::NZCV, TRI);
+                });
+}
+
 // Helper function to emit a frame offset adjustment from a given
 // pointer (SrcReg), stored into DestReg. This function is explicit
 // in that it requires the opcode.

llvm/lib/Target/AArch64/AArch64InstrInfo.h

@@ -316,6 +316,12 @@ private:
   unsigned findRegisterToSaveLRTo(const outliner::Candidate &C) const;
 };
 
+/// Return true if there is an instruction /after/ \p DefMI and before \p UseMI
+/// which either reads or clobbers NZCV.
+bool isNZCVTouchedInInstructionRange(const MachineInstr &DefMI,
+                                     const MachineInstr &UseMI,
+                                     const TargetRegisterInfo *TRI);
+
 /// emitFrameOffset - Emit instructions as needed to set DestReg to SrcReg
 /// plus Offset.  This is intended to be used from within the prolog/epilog
 /// insertion (PEI) pass, where a virtual scratch register may be allocated

llvm/test/CodeGen/AArch64/arm64-early-ifcvt.ll

@@ -1,4 +1,4 @@
-; RUN: llc < %s -stress-early-ifcvt -aarch64-enable-atomic-cfg-tidy=0 | FileCheck %s
+; RUN: llc -debugify-and-strip-all-safe < %s -stress-early-ifcvt -aarch64-enable-atomic-cfg-tidy=0 | FileCheck %s
 target triple = "arm64-apple-macosx"
 
 ; CHECK: mm2

llvm/test/CodeGen/AArch64/arm64-shrink-wrapping.ll

@@ -1,6 +1,6 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc %s -o - -mtriple=arm64-apple-ios -enable-shrink-wrap=true -disable-post-ra -frame-pointer=non-leaf | FileCheck %s --check-prefix=ENABLE
-; RUN: llc %s -o - -enable-shrink-wrap=false -disable-post-ra -frame-pointer=non-leaf | FileCheck %s --check-prefix=DISABLE
+; RUN: llc -debugify-and-strip-all-safe %s -o - -mtriple=arm64-apple-ios -enable-shrink-wrap=true -disable-post-ra -frame-pointer=non-leaf | FileCheck %s --check-prefix=ENABLE
+; RUN: llc -debugify-and-strip-all-safe %s -o - -enable-shrink-wrap=false -disable-post-ra -frame-pointer=non-leaf | FileCheck %s --check-prefix=DISABLE
 target datalayout = "e-m:o-i64:64-i128:128-n32:64-S128"
 target triple = "arm64-apple-ios"
 

llvm/test/CodeGen/AArch64/cond-br-tuning.ll

@@ -1,4 +1,4 @@
-; RUN: llc < %s -O3 -mtriple=aarch64-eabi -verify-machineinstrs | FileCheck %s
+; RUN: llc -debugify-and-strip-all-safe < %s -O3 -mtriple=aarch64-eabi -verify-machineinstrs | FileCheck %s
 
 target datalayout = "e-m:e-i8:8:32-i16:16:32-i64:64-i128:128-n32:64-S128"
 target triple = "aarch64-linaro-linux-gnueabi"