// TODO: cleanup inconsistency of printing CMPx opcodes with explicit CR arguments.
	CMP R3, R4                      // 7c232000
	CMP R3, R0                      // 7c230000
	CMP R3, R0, CR1                 // CMP R3,CR1,R0   // 7ca30000
	CMPU R3, R4                     // 7c232040
	CMPU R3, R0                     // 7c230040
	CMPU R3, R0, CR2                // CMPU R3,CR2,R0  // 7d230040
	CMPW R3, R4                     // 7c032000

	MOVQ	g_racectx(R14), RARG0	// goroutine context
	// Check that addr is within [arenastart, arenaend) or within [racedatastart, racedataend).
	CMPQ	RARG1, runtime·racearenastart(SB)
	JB	data
	CMPQ	RARG1, runtime·racearenaend(SB)
	JB	call
data:
	CMPQ	RARG1, runtime·racedatastart(SB)
	JB	ret
	CMPQ	RARG1, runtime·racedataend(SB)
	JAE	ret
call:
	MOVQ	AX, AX		// w/o this 6a miscompiles this function
	JMP	racecall<>(SB)
ret:

// to loops with a check-loop-condition-at-end.
// This helps loops avoid extra unnecessary jumps.
//
//	 loop:
//	   CMPQ ...
//	   JGE exit
//	   ...
//	   JMP loop
//	 exit:
//
//	  JMP entry
//	loop:
//	  ...
//	entry:
//	  CMPQ ...
//	  JLT loop
func loopRotate(f *Func) {
	loopnest := f.loopnest()
	if loopnest.hasIrreducible {
		return
	}
	if len(loopnest.loops) == 0 {

	case C_U8CON:
		return cmp(C_U5CON, b)
	case C_U15CON:
		return cmp(C_U8CON, b)
	case C_S16CON:
		return cmp(C_U15CON, b)
	case C_U16CON:
		return cmp(C_U15CON, b)
	case C_16CON:
		return cmp(C_S16CON, b) || cmp(C_U16CON, b)
	case C_U31CON:
		return cmp(C_U16CON, b)
	case C_U32CON:
		return cmp(C_U31CON, b)
	case C_S32CON:
		return cmp(C_U31CON, b) || cmp(C_S16CON, b)
	case C_32CON:

		}
		if p == nil {
			return false, 0
		}
	}
	cmp := p.As == ACMPB || p.As == ACMPL || p.As == ACMPQ || p.As == ACMPW

	cmpAddSub := p.As == AADDB || p.As == AADDL || p.As == AADDW || p.As == AADDQ ||
		p.As == ASUBB || p.As == ASUBL || p.As == ASUBW || p.As == ASUBQ || cmp

	testAnd := p.As == ATESTB || p.As == ATESTL || p.As == ATESTQ || p.As == ATESTW ||

	return x | ^y
}

// check bitsets
func bitSetPowerOf2Test(x int) bool {
	// amd64:"BTL\t[$]3"
	return x&8 == 8
}

func bitSetTest(x int) bool {
	// amd64:"ANDL\t[$]9, AX"
	// amd64:"CMPQ\tAX, [$]9"
	return x&9 == 9
}

// mask contiguous one bits
func cont1Mask64U(x uint64) uint64 {
	// s390x:"RISBGZ\t[$]16, [$]47, [$]0,"
	return x & 0x0000ffffffff0000
}

	frame = function's stack frame size
	argsize = size of function arguments (call + return)

	stack frame size <= StackSmall:
		CMPQ guard, SP
		JHI 3(PC)
		MOVQ m->morearg, $(argsize << 32)
		CALL morestack(SB)

	stack frame size > StackSmall but < StackBig
		LEAQ (frame-StackSmall)(SP), R0
		CMPQ guard, R0
		JHI 3(PC)
		MOVQ m->morearg, $(argsize << 32)
		CALL morestack(SB)

	stack frame size >= StackBig:

			},
		},
	},
	{
		name:   "CMPW",
		argLen: 2,
		asm:    x86.ACMPW,
		reg: regInfo{
			inputs: []inputInfo{
				{0, 49151}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R15
				{1, 49151}, // AX CX DX BX SP BP SI DI R8 R9 R10 R11 R12 R13 R15
			},
		},
	},
	{
		name:   "CMPB",
		argLen: 2,
		asm:    x86.ACMPB,
		reg: regInfo{
			inputs: []inputInfo{

			// before its corresponding VarDef.
			// See issue 28445.
			//   v1 = LOAD ...
			//   v2 = VARDEF
			//   v3 = LEAQ
			//   v4 = CMPQ v1 v3
			// We don't want to combine the CMPQ with the load, because
			// that would force the CMPQ to schedule before the VARDEF, which
			// in turn requires the LEAQ to schedule before the VARDEF.
			return false
		}
		if v.Type.IsMemory() {

			kind = ssa.BoundsSlice3CU
		}
	}

	var cmp *ssa.Value
	if kind == ssa.BoundsIndex || kind == ssa.BoundsIndexU {
		cmp = s.newValue2(ssa.OpIsInBounds, types.Types[types.TBOOL], idx, len)
	} else {
		cmp = s.newValue2(ssa.OpIsSliceInBounds, types.Types[types.TBOOL], idx, len)
	}
	b := s.endBlock()
	b.Kind = ssa.BlockIf
	b.SetControl(cmp)
	b.Likely = ssa.BranchLikely
	b.AddEdgeTo(bNext)