func Or8(ptr *uint8, val uint8)

// NOTE: Do not add atomicxor8 (XOR is not idempotent).

//go:noescape
func And(ptr *uint32, val uint32)

//go:noescape
func Or(ptr *uint32, val uint32)

//go:noescape
func And32(ptr *uint32, val uint32) uint32

//go:noescape
func Or32(ptr *uint32, val uint32) uint32

//go:noescape
func And64(ptr *uint64, val uint64) uint64

//go:noescape
func Or64(ptr *uint64, val uint64) uint64

//go:noescape
func And8(ptr *uint8, val uint8)

//go:noescape
func Or8(ptr *uint8, val uint8)

//go:noescape
func And(ptr *uint32, val uint32)

//go:noescape
func Or(ptr *uint32, val uint32)

//go:noescape
func And32(ptr *uint32, val uint32) uint32

//go:noescape
func Or32(ptr *uint32, val uint32) uint32

//go:noescape
func And64(ptr *uint64, val uint64) uint64

//go:noescape
func Or64(ptr *uint64, val uint64) uint64

	return new
}

//go:nosplit
//go:noinline
func Xchg(ptr *uint32, new uint32) uint32 {
	old := *ptr
	*ptr = new
	return old
}

//go:nosplit
//go:noinline
func Xchg64(ptr *uint64, new uint64) uint64 {
	old := *ptr
	*ptr = new
	return old
}

//go:nosplit
//go:noinline
func Xchgint32(ptr *int32, new int32) int32 {
	old := *ptr
	*ptr = new
	return old
}

}

// Atomic add and return new value.
//
//go:nosplit
func Xadd(val *uint32, delta int32) uint32 {
	for {
		oval := *val
		nval := oval + uint32(delta)
		if Cas(val, oval, nval) {
			return nval
		}
	}
}

//go:noescape
func Xadduintptr(ptr *uintptr, delta uintptr) uintptr

//go:nosplit
func Xchg(addr *uint32, v uint32) uint32 {
	for {
		old := *addr
		if Cas(addr, old, v) {
			return old
		}

			return old
		}
	}
}

//go:noescape
func Xadd(ptr *uint32, delta int32) uint32

//go:noescape
func Xadduintptr(ptr *uintptr, delta uintptr) uintptr

//go:noescape
func Xchg(ptr *uint32, new uint32) uint32

//go:noescape
func Xchguintptr(ptr *uintptr, new uintptr) uintptr

//go:noescape
func Load(ptr *uint32) uint32

//go:noescape
func Load8(ptr *uint8) uint8

^uint32(0)/416 + 1, ^uint32(0)/448 + 1, ^uint32(0)/480 + 1, ^uint32(0)/512 + 1, ^uint32(0)/576 + 1, ^uint32(0)/640 + 1, ^uint32(0)/704 + 1, ^uint32(0)/768 + 1, ^uint32(0)/896 + 1, ^uint32(0)/1024 + 1, ^uint32(0)/1152 + 1, ^uint32(0)/1280 + 1, ^uint32(0)/1408 + 1, ^uint32(0)/1536 + 1, ^uint32(0)/1792 + 1, ^uint32(0)/2048 + 1, ^uint32(0)/2304 + 1, ^uint32(0)/2688 + 1, ^uint32(0)/3072 + 1, ^uint32(0)/3200 + 1, ^uint32(0)/3456 + 1, ^uint32(0)/4096 + 1, ^uint32(0)/4864 + 1, ^uint32(0)/5376 + 1, ^uint32(0)/6144 +...

// A Uint32 is an atomic uint32. The zero value is zero.
type Uint32 struct {
	_ noCopy
	v uint32
}

// Load atomically loads and returns the value stored in x.
func (x *Uint32) Load() uint32 { return LoadUint32(&x.v) }

// Store atomically stores val into x.
func (x *Uint32) Store(val uint32) { StoreUint32(&x.v, val) }

	s0 = uint32(sbox0[t0>>24])<<24 | uint32(sbox0[t1>>16&0xff])<<16 | uint32(sbox0[t2>>8&0xff])<<8 | uint32(sbox0[t3&0xff])
	s1 = uint32(sbox0[t1>>24])<<24 | uint32(sbox0[t2>>16&0xff])<<16 | uint32(sbox0[t3>>8&0xff])<<8 | uint32(sbox0[t0&0xff])
	s2 = uint32(sbox0[t2>>24])<<24 | uint32(sbox0[t3>>16&0xff])<<16 | uint32(sbox0[t0>>8&0xff])<<8 | uint32(sbox0[t1&0xff])

	SC	R3, (R1)
	BEQ	R3, -4(PC)
	SYNC
	RET

// func Or32(addr *uint32, v uint32) old uint32
TEXT ·Or32(SB), NOSPLIT, $0-12
	MOVW	ptr+0(FP), R1
	MOVW	val+4(FP), R2

	SYNC
	LL	(R1), R3
	OR	R2, R3, R4
	SC	R4, (R1)
	BEQ	R4, -4(PC)
	SYNC
	MOVW	R3, ret+8(FP)
	RET

// func And32(addr *uint32, v uint32) old uint32
TEXT ·And32(SB), NOSPLIT, $0-12
	MOVW	ptr+0(FP), R1
	MOVW	val+4(FP), R2

	return a & (1 << 63)
}

func and_mask_3(a, b uint32) (uint32, uint32) {
	// arm/7:`BIC`,-`AND`
	a &= 0xffffaaaa
	// arm/7:`BFC`,-`AND`,-`BIC`
	b &= 0xffc003ff
	return a, b
}

// Check generation of arm64 BIC/EON/ORN instructions

func op_bic(x, y uint32) uint32 {
	// arm64:`BIC\t`,-`AND`
	return x &^ y
}

func op_eon(x, y, z uint32, a []uint32, n, m uint64) uint64 {
	// arm64:`EON\t`,-`EOR`,-`MVN`