From b1d72c26f1117300707b0f9a8624dc7b749baae1 Mon Sep 17 00:00:00 2001 From: Oliver Stannard Date: Fri, 23 Feb 2024 09:31:30 +0000 Subject: [PATCH] [ARM] Update IsRestored for LR based on all returns PR #75527 fixed ARMFrameLowering to set the IsRestored flag for LR based on all of the return instructions in the function, not just one. However, there is also code in ARMLoadStoreOptimizer which changes return instructions, but it set IsRestored based on the one instruction it changed, not the whole function. The fix is to factor out the code added in #75527, and also call it from ARMLoadStoreOptimizer if it made a change to return instructions. Fixes #80287. --- llvm/lib/Target/ARM/ARMFrameLowering.cpp | 11 +++++---- llvm/lib/Target/ARM/ARMFrameLowering.h | 1 + llvm/lib/Target/ARM/ARMLoadStoreOptimizer.cpp | 23 ++++++++----------- llvm/test/CodeGen/ARM/ldst-opt-lr-restored.ll | 11 +++++---- 4 files changed, 24 insertions(+), 22 deletions(-) diff --git a/llvm/lib/Target/ARM/ARMFrameLowering.cpp b/llvm/lib/Target/ARM/ARMFrameLowering.cpp index eeb7f64aa5810e..7f6674934aae31 100644 --- a/llvm/lib/Target/ARM/ARMFrameLowering.cpp +++ b/llvm/lib/Target/ARM/ARMFrameLowering.cpp @@ -2781,10 +2781,7 @@ void ARMFrameLowering::determineCalleeSaves(MachineFunction &MF, AFI->setLRIsSpilled(SavedRegs.test(ARM::LR)); } -void ARMFrameLowering::processFunctionBeforeFrameFinalized( - MachineFunction &MF, RegScavenger *RS) const { - TargetFrameLowering::processFunctionBeforeFrameFinalized(MF, RS); - +void ARMFrameLowering::updateLRRestored(MachineFunction &MF) const { MachineFrameInfo &MFI = MF.getFrameInfo(); if (!MFI.isCalleeSavedInfoValid()) return; @@ -2808,6 +2805,12 @@ void ARMFrameLowering::processFunctionBeforeFrameFinalized( } } +void ARMFrameLowering::processFunctionBeforeFrameFinalized( + MachineFunction &MF, RegScavenger *RS) const { + TargetFrameLowering::processFunctionBeforeFrameFinalized(MF, RS); + updateLRRestored(MF); +} + void ARMFrameLowering::getCalleeSaves(const MachineFunction &MF, BitVector &SavedRegs) const { TargetFrameLowering::getCalleeSaves(MF, SavedRegs); diff --git a/llvm/lib/Target/ARM/ARMFrameLowering.h b/llvm/lib/Target/ARM/ARMFrameLowering.h index 8d2b8beb9a58fb..13c48f5dc61743 100644 --- a/llvm/lib/Target/ARM/ARMFrameLowering.h +++ b/llvm/lib/Target/ARM/ARMFrameLowering.h @@ -59,6 +59,7 @@ class ARMFrameLowering : public TargetFrameLowering { void determineCalleeSaves(MachineFunction &MF, BitVector &SavedRegs, RegScavenger *RS) const override; + void updateLRRestored(MachineFunction &MF) const; void processFunctionBeforeFrameFinalized( MachineFunction &MF, RegScavenger *RS = nullptr) const override; diff --git a/llvm/lib/Target/ARM/ARMLoadStoreOptimizer.cpp b/llvm/lib/Target/ARM/ARMLoadStoreOptimizer.cpp index ed9d30c3c3ab90..5716e022e6711d 100644 --- a/llvm/lib/Target/ARM/ARMLoadStoreOptimizer.cpp +++ b/llvm/lib/Target/ARM/ARMLoadStoreOptimizer.cpp @@ -2062,17 +2062,6 @@ bool ARMLoadStoreOpt::MergeReturnIntoLDM(MachineBasicBlock &MBB) { MO.setReg(ARM::PC); PrevMI.copyImplicitOps(*MBB.getParent(), *MBBI); MBB.erase(MBBI); - // We now restore LR into PC so it is not live-out of the return block - // anymore: Clear the CSI Restored bit. - MachineFrameInfo &MFI = MBB.getParent()->getFrameInfo(); - // CSI should be fixed after PrologEpilog Insertion - assert(MFI.isCalleeSavedInfoValid() && "CSI should be valid"); - for (CalleeSavedInfo &Info : MFI.getCalleeSavedInfo()) { - if (Info.getReg() == ARM::LR) { - Info.setRestored(false); - break; - } - } return true; } } @@ -2120,14 +2109,22 @@ bool ARMLoadStoreOpt::runOnMachineFunction(MachineFunction &Fn) { isThumb2 = AFI->isThumb2Function(); isThumb1 = AFI->isThumbFunction() && !isThumb2; - bool Modified = false; + bool Modified = false, ModifiedLDMReturn = false; for (MachineBasicBlock &MBB : Fn) { Modified |= LoadStoreMultipleOpti(MBB); if (STI->hasV5TOps() && !AFI->shouldSignReturnAddress()) - Modified |= MergeReturnIntoLDM(MBB); + ModifiedLDMReturn |= MergeReturnIntoLDM(MBB); if (isThumb1) Modified |= CombineMovBx(MBB); } + Modified |= ModifiedLDMReturn; + + // If we merged a BX instruction into an LDM, we need to re-calculate whether + // LR is restored. This check needs to consider the whole function, not just + // the instruction(s) we changed, because there may be other BX returns which + // still need LR to be restored. + if (ModifiedLDMReturn) + Fn.getSubtarget().getFrameLowering()->updateLRRestored(Fn); Allocator.DestroyAll(); return Modified; diff --git a/llvm/test/CodeGen/ARM/ldst-opt-lr-restored.ll b/llvm/test/CodeGen/ARM/ldst-opt-lr-restored.ll index 883a812139ded1..9494880f990258 100644 --- a/llvm/test/CodeGen/ARM/ldst-opt-lr-restored.ll +++ b/llvm/test/CodeGen/ARM/ldst-opt-lr-restored.ll @@ -10,16 +10,17 @@ define i32 @foo(ptr %ctx) { ; CHECK: @ %bb.0: @ %entry ; CHECK-NEXT: cbz r0, .LBB0_2 ; CHECK-NEXT: @ %bb.1: @ %if.end +; CHECK-NEXT: movw r12, :lower16:val2 +; CHECK-NEXT: movw r3, :lower16:val1 ; CHECK-NEXT: movw r2, :lower16:val0 ; CHECK-NEXT: mov r1, r0 ; CHECK-NEXT: movs r0, #0 -; CHECK-NEXT: movw r12, :lower16:val2 -; CHECK-NEXT: movw r3, :lower16:val1 -; CHECK-NEXT: movt r2, :upper16:val0 -; CHECK-NEXT: add.w lr, r1, #4 ; CHECK-NEXT: movt r12, :upper16:val2 ; CHECK-NEXT: movt r3, :upper16:val1 -; CHECK-NEXT: stm.w lr, {r2, r3, r12} +; CHECK-NEXT: movt r2, :upper16:val0 +; CHECK-NEXT: str r2, [r1, #4] +; CHECK-NEXT: str r3, [r1, #8] +; CHECK-NEXT: str.w r12, [r1, #12] ; CHECK-NEXT: str r0, [r1, #16] ; CHECK-NEXT: bx lr ; CHECK-NEXT: .LBB0_2: @ %if.then