ADCXQ BX, R8
	ADOXQ 32(AX), R8
	MOVQ  R8, 32(AX)

	// Iteration 5
	MULXQ 40(CX), R8, BX
	ADCXQ DI, R8
	ADOXQ 40(AX), R8
	MOVQ  R8, 40(AX)

	// Iteration 6
	MULXQ 48(CX), R8, DI
	ADCXQ BX, R8
	ADOXQ 48(AX), R8
	MOVQ  R8, 48(AX)

	// Iteration 7
	MULXQ 56(CX), R8, BX
	ADCXQ DI, R8
	ADOXQ 56(AX), R8
	MOVQ  R8, 56(AX)

	// Iteration 8
	MULXQ 64(CX), R8, DI

	JA   _5_or_more
	MOVL (R8), R8
	LEAQ -3(DI)(DX*1), DX
loop4:
	MOVL (DI), SI
	CMPL SI,R8
	JZ   success
	ADDQ $1,DI
	CMPQ DI,DX
	JB loop4
	JMP fail
_5_or_more:
	CMPQ AX, $7
	JA   _8_or_more
	LEAQ 1(DI)(DX*1), DX
	SUBQ AX, DX
	MOVL -4(R8)(AX*1), BX
	MOVL (R8), R8
loop5to7:
	MOVL (DI), SI
	CMPL SI,R8
	JZ   partial_success5to7
	ADDQ $1,DI
	CMPQ DI,DX
	JB loop5to7
	JMP fail

DATA zosLibVec<>(SB)/8, $0
GLOBL zosLibVec<>(SB), NOPTR, $8

TEXT ·initZosLibVec(SB), NOSPLIT|NOFRAME, $0-0
	MOVW PSALAA, R8
	MOVD LCA64(R8), R8
	MOVD CAA(R8), R8
	MOVD EDCHPXV(R8), R8
	MOVD R8, zosLibVec<>(SB)
	RET

TEXT ·GetZosLibVec(SB), NOSPLIT|NOFRAME, $0-0
	MOVD zosLibVec<>(SB), R8
	MOVD R8, ret+0(FP)
	RET

TEXT ·clearErrno(SB), NOSPLIT, $0-0
	BL   addrerrno<>(SB)
	MOVD $0, 0(R3)
	RET

largeLoop:  // Copying 256 bytes at a time.
	MVC    $256, 0(R6), 0(R8)
	MOVD   $256(R6), R6
	MOVD   $256(R8), R8
	MOVD   $-32(R5), R5
	CMPBGE R5, $32, largeLoop
	BR     mediumLoop

mediumLoopBody:  // Copying 32 bytes at a time
	MVC    $32, 0(R6), 0(R8)
	MOVD   $32(R6), R6
	MOVD   $32(R8), R8
	MOVD   $-4(R5), R5
	CMPBGE R5, $4, mediumLoopBody
	BR     smallLoop

returnC:

chunk16_loop:
	BEQ	R0, R14, byte_loop
	MOVV	(R4), R8
	MOVV	(R6), R9
	BNE	R8, R9, byte_loop
	MOVV	8(R4), R16
	MOVV	8(R6), R17
	ADDV	$16, R4
	ADDV	$16, R6
	SUBVU	$1, R14
	BEQ	R16, R17, chunk16_loop
	SUBV	$8, R4
	SUBV	$8, R6

byte_loop:
	BEQ	R4, R12, samebytes
	MOVBU	(R4), R8
	ADDVU	$1, R4
	MOVBU	(R6), R9
	ADDVU	$1, R6
	BEQ	R8, R9, byte_loop

byte_cmp:
	SGTU	R8, R9, R4 // R12 = 1 if (R8 > R9)

	// alignment requirements.
	ANDCC	$PAGE_OFFSET, R8, R6	// &sX & PAGE_OFFSET
	ANDCC	$PAGE_OFFSET, R4, R9
	SUBC	R5, $8, R12		// 8-len
	SLD	$3, R12, R14		// (8-len)*8
	CMPU	R6, R12, CR1		// Enough bytes lower in the page to load lower?
	CMPU	R9, R12, CR0
	SUB	R12, R8, R6		// compute lower load address
	SUB	R12, R4, R9
	ISEL	CR1LT, R8, R6, R8	// R8 = R6 < 0 ? R8 (&s1) : R6 (&s1 - (8-len))

	MOVWZ	(R5)(R8), R15
	XOR	R14, R15, R16
	MOVW	R16, (R3)(R8)
	ADD	$4,R8
	ADD	$-4,R6
xor2:
	CMP	R6, $2
	BLT	xor1
	MOVHZ	(R4)(R8), R14
	MOVHZ	(R5)(R8), R15
	XOR	R14, R15, R16
	MOVH	R16, (R3)(R8)
	ADD	$2,R8
	ADD	$-2,R6
xor1:
	CMP	R6, $0
	BC	12,2,LR		// BEQLR
	MOVBZ	(R4)(R8), R14	// R14 = a[i]
	MOVBZ	(R5)(R8), R15	// R15 = b[i]
	XOR	R14, R15, R16	// R16 = a[i] ^ b[i]

	MOVW	R0, R8

	MOVW	(0*4)(R8), R7 // fn
	MOVW	(1*4)(R8), R0 // a1
	MOVW	(2*4)(R8), R1 // a2
	MOVW	(3*4)(R8), R2 // a3
	MOVW	(4*4)(R8), R3 // a4
	MOVW	(5*4)(R8), R4 // a5
	MOVW	R4, 0(R13)
	MOVW	(6*4)(R8), R5 // a6
	MOVW	R5, 4(R13)
	MOVW	(7*4)(R8), R6 // a7
	MOVW	R6, 8(R13)
	MOVW	(8*4)(R8), R4 // a8
	MOVW	R4, 12(R13)
	MOVW	(9*4)(R8), R5 // a9
	MOVW	R5, 16(R13)

	SHA256ROUND0(3, 0xe9b5dba5, R13, R14, R15, R8, R9, R10, R11, R12)
	SHA256ROUND0(4, 0x3956c25b, R12, R13, R14, R15, R8, R9, R10, R11)
	SHA256ROUND0(5, 0x59f111f1, R11, R12, R13, R14, R15, R8, R9, R10)
	SHA256ROUND0(6, 0x923f82a4, R10, R11, R12, R13, R14, R15, R8, R9)
	SHA256ROUND0(7, 0xab1c5ed5, R9, R10, R11, R12, R13, R14, R15, R8)
	SHA256ROUND0(8, 0xd807aa98, R8, R9, R10, R11, R12, R13, R14, R15)

	MOVQ (CX), AX
	MULQ 8(BX)
	MOVQ AX, R9
	MOVQ DX, R8

	// r1 += a1×b0
	MOVQ 8(CX), AX
	MULQ (BX)
	ADDQ AX, R9
	ADCQ DX, R8

	// r1 += 19×a2×b4
	MOVQ   16(CX), AX
	IMUL3Q $0x13, AX, AX
	MULQ   32(BX)
	ADDQ   AX, R9
	ADCQ   DX, R8

	// r1 += 19×a3×b3
	MOVQ   24(CX), AX
	IMUL3Q $0x13, AX, AX
	MULQ   24(BX)
	ADDQ   AX, R9
	ADCQ   DX, R8

	// r1 += 19×a4×b2
	MOVQ   32(CX), AX