StackMap: FastISel: Add an appropriate number of immediate operands to the frame setup instruction.

[arphaman]

Summary

This patch ensures that the stack map lowering code in FastISel adds an appropriate number of immediate operands for the frame setup instruction.

The previous code added just one immediate operand, which was fine for a target like AArch64, but on X86 the ADJUSTCALLSTACKDOWN64 instruction needs two explicit operands, and this caused the machine verifier to report an error when the old code added just one.

[ributzka]

LGTM

Could you also add a test for PPC?

[arphaman]

Thanks, the committed patch also includes the PPC test.

llvm/trunk/lib/CodeGen/SelectionDAG/FastISel.cpp

assert(I->getCalledFunction()->getReturnType()->isVoidTy() &&
          "Stackmap cannot return a value.");
 
   // The stackmap intrinsic only records the live variables (the arguments
   // passed to it) and emits NOPS (if requested). Unlike the patchpoint
   // intrinsic, this won't be lowered to a function call. This means we don't
   // have to worry about calling conventions and target-specific lowering code.
   // Instead we perform the call lowering right here.
   //
-  // CALLSEQ_START(0)
+  // CALLSEQ_START(0...)
   // STACKMAP(id, nbytes, ...)
   // CALLSEQ_END(0, 0)
   //
   SmallVector<MachineOperand, 32> Ops;
 
   // Add the <id> and <numBytes> constants.
   assert(isa<ConstantInt>(I->getOperand(PatchPointOpers::IDPos)) &&
          "Expected a constant integer.");
bool FastISel::selectStackmap(const CallInst *I) {
   const MCPhysReg *ScratchRegs = TLI.getScratchRegisters(CC);
   for (unsigned i = 0; ScratchRegs[i]; ++i)
     Ops.push_back(MachineOperand::CreateReg(
         ScratchRegs[i], /*IsDef=*/true, /*IsImp=*/true, /*IsKill=*/false,
         /*IsDead=*/false, /*IsUndef=*/false, /*IsEarlyClobber=*/true));
 
   // Issue CALLSEQ_START
   unsigned AdjStackDown = TII.getCallFrameSetupOpcode();
-  BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(AdjStackDown))
-      .addImm(0);
+  auto Builder =
+      BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(AdjStackDown));
+  const MCInstrDesc &MCID = Builder.getInstr()->getDesc();
+  for (unsigned I = 0, E = MCID.getNumOperands(); I < E; ++I)
+    Builder.addImm(0);
 
   // Issue STACKMAP.
   MachineInstrBuilder MIB = BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc,
                                     TII.get(TargetOpcode::STACKMAP));
   for (auto const &MO : Ops)
     MIB.addOperand(MO);
 
   // Issue CALLSEQ_END

llvm/trunk/test/CodeGen/AArch64/stackmap-frame-setup.ll

+; RUN: llc -o /dev/null -verify-machineinstrs -mtriple=aarch64-apple-darwin -stop-after machine-sink %s | FileCheck %s --check-prefix=ISEL
+; RUN: llc -o /dev/null -verify-machineinstrs -mtriple=aarch64-apple-darwin -fast-isel -fast-isel-abort=1 -stop-after machine-sink %s | FileCheck %s --check-prefix=FAST-ISEL
+
+define void @caller_meta_leaf() {
+entry:
+  %metadata = alloca i64, i32 3, align 8
+  store i64 11, i64* %metadata
+  store i64 12, i64* %metadata
+  store i64 13, i64* %metadata
+; ISEL:      ADJCALLSTACKDOWN 0, implicit-def
+; ISEL-NEXT: STACKMAP
+; ISEL-NEXT: ADJCALLSTACKUP 0, 0, implicit-def
+  call void (i64, i32, ...) @llvm.experimental.stackmap(i64 4, i32 0, i64* %metadata)
+; FAST-ISEL:      ADJCALLSTACKDOWN 0, implicit-def
+; FAST-ISEL-NEXT: STACKMAP
+; FAST-ISEL-NEXT: ADJCALLSTACKUP 0, 0, implicit-def
+  ret void
+}
+
+declare void @llvm.experimental.stackmap(i64, i32, ...)

llvm/trunk/test/CodeGen/PowerPC/stackmap-frame-setup.ll

+; RUN: llc -o /dev/null -verify-machineinstrs -mtriple=powerpc64-unknown-gnu-linux -stop-after machine-sink %s | FileCheck %s --check-prefix=ISEL
+; RUN: llc -o /dev/null -verify-machineinstrs -mtriple=powerpc64-unknown-gnu-linux -fast-isel -fast-isel-abort=1 -stop-after machine-sink %s | FileCheck %s --check-prefix=FAST-ISEL
+
+define void @caller_meta_leaf() {
+entry:
+  %metadata = alloca i64, i32 3, align 8
+  store i64 11, i64* %metadata
+  store i64 12, i64* %metadata
+  store i64 13, i64* %metadata
+; ISEL:      ADJCALLSTACKDOWN 0, implicit-def
+; ISEL-NEXT: STACKMAP
+; ISEL-NEXT: ADJCALLSTACKUP 0, 0, implicit-def
+  call void (i64, i32, ...) @llvm.experimental.stackmap(i64 4, i32 0, i64* %metadata)
+; FAST-ISEL:      ADJCALLSTACKDOWN 0, implicit-def
+; FAST-ISEL-NEXT: STACKMAP
+; FAST-ISEL-NEXT: ADJCALLSTACKUP 0, 0, implicit-def
+  ret void
+}
+
+declare void @llvm.experimental.stackmap(i64, i32, ...)

llvm/trunk/test/CodeGen/X86/stackmap-frame-setup.ll

+; RUN: llc -o /dev/null -verify-machineinstrs -mtriple=x86_64-apple-darwin -mcpu=corei7 -stop-after machine-sink %s | FileCheck %s --check-prefix=ISEL
+; RUN: llc -o /dev/null -verify-machineinstrs -mtriple=x86_64-apple-darwin -mcpu=corei7 -fast-isel -fast-isel-abort=1 -stop-after machine-sink %s | FileCheck %s --check-prefix=FAST-ISEL
+
+define void @caller_meta_leaf() {
+entry:
+  %metadata = alloca i64, i32 3, align 8
+  store i64 11, i64* %metadata
+  store i64 12, i64* %metadata
+  store i64 13, i64* %metadata
+; ISEL:      ADJCALLSTACKDOWN64 0, 0, implicit-def
+; ISEL-NEXT: STACKMAP
+; ISEL-NEXT: ADJCALLSTACKUP64 0, 0, implicit-def
+  call void (i64, i32, ...) @llvm.experimental.stackmap(i64 4, i32 0, i64* %metadata)
+; FAST-ISEL:      ADJCALLSTACKDOWN64 0, 0, implicit-def
+; FAST-ISEL-NEXT: STACKMAP
+; FAST-ISEL-NEXT: ADJCALLSTACKUP64 0, 0, implicit-def
+  ret void
+}
+
+declare void @llvm.experimental.stackmap(i64, i32, ...)