diff --git a/lib/Target/ARM/ARMLoadStoreOptimizer.cpp b/lib/Target/ARM/ARMLoadStoreOptimizer.cpp index ee7df5476c7..62fd0596569 100644 --- a/lib/Target/ARM/ARMLoadStoreOptimizer.cpp +++ b/lib/Target/ARM/ARMLoadStoreOptimizer.cpp @@ -505,7 +505,7 @@ ARMLoadStoreOpt::MergeOps(MachineBasicBlock &MBB, // Exception: If the base register is in the input reglist, Thumb1 LDM is // non-writeback. Check for this. - if (Opcode == ARM::tLDRi && isThumb1) + if (Opcode == ARM::tLDMIA && isThumb1) for (unsigned I = 0; I < NumRegs; ++I) if (Base == Regs[I].first) { Writeback = false; @@ -519,17 +519,17 @@ ARMLoadStoreOpt::MergeOps(MachineBasicBlock &MBB, // Update tLDMIA with writeback if necessary. Opcode = ARM::tLDMIA_UPD; - // The base isn't dead after a merged instruction with writeback. Update - // future uses of the base with the added offset (if possible), or reset - // the base register as necessary. - if (!BaseKill) - UpdateBaseRegUses(MBB, MBBI, dl, Base, NumRegs, Pred, PredReg); - MIB = BuildMI(MBB, MBBI, dl, TII->get(Opcode)); // Thumb1: we might need to set base writeback when building the MI. MIB.addReg(Base, getDefRegState(true)) .addReg(Base, getKillRegState(BaseKill)); + + // The base isn't dead after a merged instruction with writeback. Update + // future uses of the base with the added offset (if possible), or reset + // the base register as necessary. + if (!BaseKill) + UpdateBaseRegUses(MBB, MBBI, dl, Base, NumRegs, Pred, PredReg); } else { // No writeback, simply build the MachineInstr. MIB = BuildMI(MBB, MBBI, dl, TII->get(Opcode)); diff --git a/test/CodeGen/Thumb/2014-06-10-thumb1-ldst-opt-bug.ll b/test/CodeGen/Thumb/2014-06-10-thumb1-ldst-opt-bug.ll new file mode 100644 index 00000000000..3c43383d7a2 --- /dev/null +++ b/test/CodeGen/Thumb/2014-06-10-thumb1-ldst-opt-bug.ll @@ -0,0 +1,17 @@ +; RUN: llc < %s -mtriple=thumbv6m-eabi -o - | FileCheck %s + +define void @foo(i32* %A) #0 { +entry: +; CHECK-LABEL: foo: +; CHECK: push {r7, lr} +; CHECK: ldm [[REG0:r[0-9]]]!, +; CHECK-NEXT: subs [[REG0]] +; CHECK-NEXT: bl + %0 = load i32* %A, align 4 + %arrayidx1 = getelementptr inbounds i32* %A, i32 1 + %1 = load i32* %arrayidx1, align 4 + tail call void @bar(i32* %A, i32 %0, i32 %1) #2 + ret void +} + +declare void @bar(i32*, i32, i32) #1