y3 = (y1 + w2mw6*y3) >> 3
		y1 = y4 + y6
		y4 -= y6
		y6 = y5 + y7
		y5 -= y7

		// Stage 3.
		y7 = y8 + y3
		y8 -= y3
		y3 = y0 + y2
		y0 -= y2
		y2 = (r2*(y4+y5) + 128) >> 8
		y4 = (r2*(y4-y5) + 128) >> 8

		// Stage 4.
		s[8*0] = (y7 + y1) >> 14
		s[8*1] = (y3 + y2) >> 14
		s[8*2] = (y0 + y4) >> 14
		s[8*3] = (y8 + y6) >> 14
		s[8*4] = (y8 - y6) >> 14
		s[8*5] = (y0 - y4) >> 14

var y5 = x5

type T5 struct{}

func (T5) m() int {
	_ = y5
	return 0
}

// issue 4847
// simplified test case

var x6 = f6
var y6 /* ERROR "initialization cycle" */ = f6
func f6() { _ = y6 }

// full test case

type (
      E int
      S int
)

type matcher func(s *S) E

func matchList(s *S) E { return matcher(matchAnyFn)(s) }

L6:
	MOVW.P	4(R2), R6
	ORR	R6<<R4, R7
	MOVW.P	R7, 4(R1)
	MOVW	R6>>R3, R7
E6:
	TEQ	R1, R5
	BNE	L6

	MOVW	R7, 0(R1)
	RET

Y6:	// copy loop, because shift 0 == shift 32
	MOVW.P	4(R2), R6
	MOVW.P	R6, 4(R1)
	TEQ R1, R5
	BNE Y6

X6:
	MOVW	$0, R1
	MOVW	R1, c+28(FP)
	RET


// func mulAddVWW(z, x []Word, y, r Word) (c Word)
TEXT ·mulAddVWW(SB),NOSPLIT,$0
	MOVW	$0, R0

	VPGATHERDQ Y2, (BP)(AX*2), Y1   // ERROR "invalid instruction"
	VPGATHERDQ Y5, 664(DX*8), Y6    // ERROR "invalid instruction"
	VPGATHERDQ Y5, (DX), Y0         // ERROR "invalid instruction"
	// VM/X rejects Y index register.
	VPGATHERDQ Y5, 664(Y14*8), Y6   // ERROR "invalid instruction"
	VPGATHERQQ X2, (BP)(Y7*2), X1   // ERROR "invalid instruction"
	// VM/Y rejects X index register.

	PCALIGN $16
hugeloop_avx2:
	CMPQ	BX, $64
	JB	bigloop_avx2
	VMOVDQU	(SI), Y0
	VMOVDQU	(DI), Y1
	VMOVDQU	32(SI), Y2
	VMOVDQU	32(DI), Y3
	VPCMPEQB	Y1, Y0, Y4
	VPCMPEQB	Y2, Y3, Y5
	VPAND	Y4, Y5, Y6
	VPMOVMSKB Y6, DX
	ADDQ	$64, SI
	ADDQ	$64, DI
	SUBQ	$64, BX
	CMPL	DX, $0xffffffff
	JEQ	hugeloop_avx2
	VZEROUPPER
	XORQ	AX, AX	// return 0
	RET

bigloop_avx2:
	VZEROUPPER

big_loop_avx2:
	VMOVDQU	(SI), Y2
	VMOVDQU	(DI), Y3
	VMOVDQU	32(SI), Y4
	VMOVDQU	32(DI), Y5
	VPCMPEQB Y2, Y3, Y0
	VPMOVMSKB Y0, AX
	XORL	$0xffffffff, AX
	JNE	diff32_avx2
	VPCMPEQB Y4, Y5, Y6
	VPMOVMSKB Y6, AX
	XORL	$0xffffffff, AX
	JNE	diff64_avx2

	ADDQ	$64, SI
	ADDQ	$64, DI
	SUBQ	$64, R8
	CMPQ	R8, $64
	JB	big_loop_avx2_exit
	JMP	big_loop_avx2

	"X19",
	"X20",
	"X21",
	"X22",
	"X23",
	"X24",
	"X25",
	"X26",
	"X27",
	"X28",
	"X29",
	"X30",
	"X31",
	"Y0",
	"Y1",
	"Y2",
	"Y3",
	"Y4",
	"Y5",
	"Y6",
	"Y7",
	"Y8",
	"Y9",
	"Y10",
	"Y11",
	"Y12",
	"Y13",
	"Y14",
	"Y15",
	"Y16",
	"Y17",
	"Y18",
	"Y19",
	"Y20",
	"Y21",
	"Y22",
	"Y23",
	"Y24",
	"Y25",