MachineMemOperand *MMO = nullptr);
bool emitStore(MVT VT, unsigned SrcReg, Address Addr,
MachineMemOperand *MMO = nullptr);
+ bool emitStoreRelease(MVT VT, unsigned SrcReg, unsigned AddrReg,
+ MachineMemOperand *MMO = nullptr);
unsigned emitIntExt(MVT SrcVT, unsigned SrcReg, MVT DestVT, bool isZExt);
unsigned emiti1Ext(unsigned SrcReg, MVT DestVT, bool isZExt);
unsigned emitAdd(MVT RetVT, const Value *LHS, const Value *RHS,
return true;
}
+bool AArch64FastISel::emitStoreRelease(MVT VT, unsigned SrcReg,
+ unsigned AddrReg,
+ MachineMemOperand *MMO) {
+ unsigned Opc;
+ switch (VT.SimpleTy) {
+ default: return false;
+ case MVT::i8: Opc = AArch64::STLRB; break;
+ case MVT::i16: Opc = AArch64::STLRH; break;
+ case MVT::i32: Opc = AArch64::STLRW; break;
+ case MVT::i64: Opc = AArch64::STLRX; break;
+ }
+
+ const MCInstrDesc &II = TII.get(Opc);
+ SrcReg = constrainOperandRegClass(II, SrcReg, 0);
+ AddrReg = constrainOperandRegClass(II, AddrReg, 1);
+ BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, II)
+ .addReg(SrcReg)
+ .addReg(AddrReg)
+ .addMemOperand(MMO);
+ return true;
+}
+
bool AArch64FastISel::emitStore(MVT VT, unsigned SrcReg, Address Addr,
MachineMemOperand *MMO) {
if (!TLI.allowsMisalignedMemoryAccesses(VT))
// Verify we have a legal type before going any further. Currently, we handle
// simple types that will directly fit in a register (i32/f32/i64/f64) or
// those that can be sign or zero-extended to a basic operation (i1/i8/i16).
- if (!isTypeSupported(Op0->getType(), VT, /*IsVectorAllowed=*/true) ||
- cast<StoreInst>(I)->isAtomic())
+ if (!isTypeSupported(Op0->getType(), VT, /*IsVectorAllowed=*/true))
return false;
const Value *PtrV = I->getOperand(1);
if (!SrcReg)
return false;
+ auto *SI = cast<StoreInst>(I);
+
+ // Try to emit a STLR for seq_cst/release.
+ if (SI->isAtomic()) {
+ AtomicOrdering Ord = SI->getOrdering();
+ // The non-atomic instructions are sufficient for relaxed stores.
+ if (isReleaseOrStronger(Ord)) {
+ // The STLR addressing mode only supports a base reg; pass that directly.
+ unsigned AddrReg = getRegForValue(PtrV);
+ return emitStoreRelease(VT, SrcReg, AddrReg,
+ createMachineMemOperandFor(I));
+ }
+ }
+
// See if we can handle this address.
Address Addr;
- if (!computeAddress(I->getOperand(1), Addr, I->getOperand(0)->getType()))
+ if (!computeAddress(PtrV, Addr, Op0->getType()))
return false;
if (!emitStore(VT, SrcReg, Addr, createMachineMemOperandFor(I)))
--- /dev/null
+; RUN: llc -mtriple=aarch64-- -O0 -fast-isel -fast-isel-abort=4 -verify-machineinstrs < %s | FileCheck %s
+; RUN: llc -mtriple=aarch64-- -O0 -fast-isel=0 -verify-machineinstrs < %s | FileCheck %s
+
+; Note that checking SelectionDAG output isn't strictly necessary, but they
+; currently match, so we might as well check both! Feel free to remove SDAG.
+
+; CHECK-LABEL: atomic_store_monotonic_8:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: strb w1, [x0]
+; CHECK-NEXT: ret
+define void @atomic_store_monotonic_8(i8* %p, i8 %val) #0 {
+ store atomic i8 %val, i8* %p monotonic, align 1
+ ret void
+}
+
+; CHECK-LABEL: atomic_store_monotonic_8_off:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: strb w1, [x0, #1]
+; CHECK-NEXT: ret
+define void @atomic_store_monotonic_8_off(i8* %p, i8 %val) #0 {
+ %tmp0 = getelementptr i8, i8* %p, i32 1
+ store atomic i8 %val, i8* %tmp0 monotonic, align 1
+ ret void
+}
+
+; CHECK-LABEL: atomic_store_monotonic_16:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: strh w1, [x0]
+; CHECK-NEXT: ret
+define void @atomic_store_monotonic_16(i16* %p, i16 %val) #0 {
+ store atomic i16 %val, i16* %p monotonic, align 2
+ ret void
+}
+
+; CHECK-LABEL: atomic_store_monotonic_16_off:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: strh w1, [x0, #2]
+; CHECK-NEXT: ret
+define void @atomic_store_monotonic_16_off(i16* %p, i16 %val) #0 {
+ %tmp0 = getelementptr i16, i16* %p, i32 1
+ store atomic i16 %val, i16* %tmp0 monotonic, align 2
+ ret void
+}
+
+; CHECK-LABEL: atomic_store_monotonic_32:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: str w1, [x0]
+; CHECK-NEXT: ret
+define void @atomic_store_monotonic_32(i32* %p, i32 %val) #0 {
+ store atomic i32 %val, i32* %p monotonic, align 4
+ ret void
+}
+
+; CHECK-LABEL: atomic_store_monotonic_32_off:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: str w1, [x0, #4]
+; CHECK-NEXT: ret
+define void @atomic_store_monotonic_32_off(i32* %p, i32 %val) #0 {
+ %tmp0 = getelementptr i32, i32* %p, i32 1
+ store atomic i32 %val, i32* %tmp0 monotonic, align 4
+ ret void
+}
+
+; CHECK-LABEL: atomic_store_monotonic_64:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: str x1, [x0]
+; CHECK-NEXT: ret
+define void @atomic_store_monotonic_64(i64* %p, i64 %val) #0 {
+ store atomic i64 %val, i64* %p monotonic, align 8
+ ret void
+}
+
+; CHECK-LABEL: atomic_store_monotonic_64_off:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: str x1, [x0, #8]
+; CHECK-NEXT: ret
+define void @atomic_store_monotonic_64_off(i64* %p, i64 %val) #0 {
+ %tmp0 = getelementptr i64, i64* %p, i32 1
+ store atomic i64 %val, i64* %tmp0 monotonic, align 8
+ ret void
+}
+
+; CHECK-LABEL: atomic_store_release_8:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: stlrb w1, [x0]
+; CHECK-NEXT: ret
+define void @atomic_store_release_8(i8* %p, i8 %val) #0 {
+ store atomic i8 %val, i8* %p release, align 1
+ ret void
+}
+
+; CHECK-LABEL: atomic_store_release_8_off:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: add x0, x0, #1
+; CHECK-NEXT: stlrb w1, [x0]
+; CHECK-NEXT: ret
+define void @atomic_store_release_8_off(i8* %p, i8 %val) #0 {
+ %tmp0 = getelementptr i8, i8* %p, i32 1
+ store atomic i8 %val, i8* %tmp0 release, align 1
+ ret void
+}
+
+; CHECK-LABEL: atomic_store_release_16:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: stlrh w1, [x0]
+; CHECK-NEXT: ret
+define void @atomic_store_release_16(i16* %p, i16 %val) #0 {
+ store atomic i16 %val, i16* %p release, align 2
+ ret void
+}
+
+; CHECK-LABEL: atomic_store_release_16_off:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: add x0, x0, #2
+; CHECK-NEXT: stlrh w1, [x0]
+; CHECK-NEXT: ret
+define void @atomic_store_release_16_off(i16* %p, i16 %val) #0 {
+ %tmp0 = getelementptr i16, i16* %p, i32 1
+ store atomic i16 %val, i16* %tmp0 release, align 2
+ ret void
+}
+
+; CHECK-LABEL: atomic_store_release_32:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: stlr w1, [x0]
+; CHECK-NEXT: ret
+define void @atomic_store_release_32(i32* %p, i32 %val) #0 {
+ store atomic i32 %val, i32* %p release, align 4
+ ret void
+}
+
+; CHECK-LABEL: atomic_store_release_32_off:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: add x0, x0, #4
+; CHECK-NEXT: stlr w1, [x0]
+; CHECK-NEXT: ret
+define void @atomic_store_release_32_off(i32* %p, i32 %val) #0 {
+ %tmp0 = getelementptr i32, i32* %p, i32 1
+ store atomic i32 %val, i32* %tmp0 release, align 4
+ ret void
+}
+
+; CHECK-LABEL: atomic_store_release_64:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: stlr x1, [x0]
+; CHECK-NEXT: ret
+define void @atomic_store_release_64(i64* %p, i64 %val) #0 {
+ store atomic i64 %val, i64* %p release, align 8
+ ret void
+}
+
+; CHECK-LABEL: atomic_store_release_64_off:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: add x0, x0, #8
+; CHECK-NEXT: stlr x1, [x0]
+; CHECK-NEXT: ret
+define void @atomic_store_release_64_off(i64* %p, i64 %val) #0 {
+ %tmp0 = getelementptr i64, i64* %p, i32 1
+ store atomic i64 %val, i64* %tmp0 release, align 8
+ ret void
+}
+
+
+; CHECK-LABEL: atomic_store_seq_cst_8:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: stlrb w1, [x0]
+; CHECK-NEXT: ret
+define void @atomic_store_seq_cst_8(i8* %p, i8 %val) #0 {
+ store atomic i8 %val, i8* %p seq_cst, align 1
+ ret void
+}
+
+; CHECK-LABEL: atomic_store_seq_cst_8_off:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: add x0, x0, #1
+; CHECK-NEXT: stlrb w1, [x0]
+; CHECK-NEXT: ret
+define void @atomic_store_seq_cst_8_off(i8* %p, i8 %val) #0 {
+ %tmp0 = getelementptr i8, i8* %p, i32 1
+ store atomic i8 %val, i8* %tmp0 seq_cst, align 1
+ ret void
+}
+
+; CHECK-LABEL: atomic_store_seq_cst_16:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: stlrh w1, [x0]
+; CHECK-NEXT: ret
+define void @atomic_store_seq_cst_16(i16* %p, i16 %val) #0 {
+ store atomic i16 %val, i16* %p seq_cst, align 2
+ ret void
+}
+
+; CHECK-LABEL: atomic_store_seq_cst_16_off:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: add x0, x0, #2
+; CHECK-NEXT: stlrh w1, [x0]
+; CHECK-NEXT: ret
+define void @atomic_store_seq_cst_16_off(i16* %p, i16 %val) #0 {
+ %tmp0 = getelementptr i16, i16* %p, i32 1
+ store atomic i16 %val, i16* %tmp0 seq_cst, align 2
+ ret void
+}
+
+; CHECK-LABEL: atomic_store_seq_cst_32:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: stlr w1, [x0]
+; CHECK-NEXT: ret
+define void @atomic_store_seq_cst_32(i32* %p, i32 %val) #0 {
+ store atomic i32 %val, i32* %p seq_cst, align 4
+ ret void
+}
+
+; CHECK-LABEL: atomic_store_seq_cst_32_off:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: add x0, x0, #4
+; CHECK-NEXT: stlr w1, [x0]
+; CHECK-NEXT: ret
+define void @atomic_store_seq_cst_32_off(i32* %p, i32 %val) #0 {
+ %tmp0 = getelementptr i32, i32* %p, i32 1
+ store atomic i32 %val, i32* %tmp0 seq_cst, align 4
+ ret void
+}
+
+; CHECK-LABEL: atomic_store_seq_cst_64:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: stlr x1, [x0]
+; CHECK-NEXT: ret
+define void @atomic_store_seq_cst_64(i64* %p, i64 %val) #0 {
+ store atomic i64 %val, i64* %p seq_cst, align 8
+ ret void
+}
+
+; CHECK-LABEL: atomic_store_seq_cst_64_off:
+; CHECK-NEXT: // BB#0:
+; CHECK-NEXT: add x0, x0, #8
+; CHECK-NEXT: stlr x1, [x0]
+; CHECK-NEXT: ret
+define void @atomic_store_seq_cst_64_off(i64* %p, i64 %val) #0 {
+ %tmp0 = getelementptr i64, i64* %p, i32 1
+ store atomic i64 %val, i64* %tmp0 seq_cst, align 8
+ ret void
+}
+
+attributes #0 = { nounwind }