#define buf	buffer-(8+4*16)(SP)	//16 words temporary buffer
		// 3 words at 4..12(R13) for called routine parameters

TEXT	·block(SB), NOSPLIT, $84-16
	MOVW	p+4(FP), Rdata	// pointer to the data
	MOVW	p_len+8(FP), Rt0	// number of bytes
	ADD	Rdata, Rt0
	MOVW	Rt0, p_end	// pointer to end of data

loop:
	MOVW	Rdata, p_data	// Save Rdata
	AND.S	$3, Rdata, Rt0	// TST $3, Rdata not working see issue 5921

#define SHA512ROUND1(index, a, b, c, d, e, f, g, h) \
	MSGSCHEDULE1(index); \
	SHA512ROUND(index, a, b, c, d, e, f, g, h)

// func block(dig *digest, p []byte)
TEXT ·block(SB),0,$128-32
	MOV	p_base+8(FP), X29
	MOV	p_len+16(FP), X30
	SRL	$7, X30
	SLL	$7, X30

	ADD	X29, X30, X28
	BEQ	X28, X29, end

	MOV	·_K(SB), X18		// const table
	ADD	$8, X2, X19		// message schedule

	MOV	dig+0(FP), X20

	SHA256RNDS2		msg, state1, state0		\
	sha256Msg1		(m,a)

TEXT ·block(SB), 0, $536-32
	CMPB	·useSHA(SB), $1
	JE	sha_ni
	CMPB	·useAVX2(SB), $1
	JE	avx2

	MOVQ p_base+8(FP), SI
	MOVQ p_len+16(FP), DX
	SHRQ $6, DX
	SHLQ $6, DX

	LEAQ (SI)(DX*1), DI
	MOVQ DI, 256(SP)
	CMPQ SI, DI
	JEQ  end

	MOVQ dig+0(FP), BP
	MOVL (0*4)(BP), R8  // a = H0
	MOVL (1*4)(BP), R9  // b = H1

	list        = flag.Bool("l", false, "list files whose formatting differs from gofmt's")
	write       = flag.Bool("w", false, "write result to (source) file instead of stdout")
	rewriteRule = flag.String("r", "", "rewrite rule (e.g., 'a[b:len(a)] -> a[b:]')")
	simplifyAST = flag.Bool("s", false, "simplify code")
	doDiff      = flag.Bool("d", false, "display diffs instead of rewriting files")

#define VPERMLE(va,vb,vc,vt) VPERM va, vb, vc, vt
#else
#define VPERMLE(va,vb,vc,vt)
#endif

// func block(dig *digest, p []byte)
TEXT ·block(SB),0,$0-32
	MOVD	dig+0(FP), CTX
	MOVD	p_base+8(FP), INP
	MOVD	p_len+16(FP), LEN

	SRD	$6, LEN
	SLD	$6, LEN
	ADD	INP, LEN, END

	CMP	INP, END
	BEQ	end

	MOVD	$·kcon(SB), TBL_STRT
	MOVD	$0x10, R_x010

#ifdef GOARCH_ppc64le
	MOVWZ	$8, TEMP

#define ROUND4(a, b, c, d, e, index) \
	SHUFFLE(index); \
	FUNC4(a, b, c, d, e); \
	MIX(a, b, c, d, e, 0xCA62C1D6)

TEXT ·blockAMD64(SB),NOSPLIT,$64-32
	MOVQ	dig+0(FP),	BP
	MOVQ	p_base+8(FP),	SI
	MOVQ	p_len+16(FP),	DX
	SHRQ	$6,		DX
	SHLQ	$6,		DX

	LEAQ	(SI)(DX*1),	DI
	MOVL	(0*4)(BP),	AX
	MOVL	(1*4)(BP),	BX
	MOVL	(2*4)(BP),	CX
	MOVL	(3*4)(BP),	DX
	MOVL	(4*4)(BP),	BP

	CMPQ	SI,		DI
	JEQ	end

#define VPERMLE(va,vb,vc,vt) VPERM va, vb, vc, vt
#else
#define VPERMLE(va,vb,vc,vt)
#endif

// func block(dig *digest, p []byte)
TEXT ·block(SB),0,$0-32
	MOVD	dig+0(FP), CTX
	MOVD	p_base+8(FP), INP
	MOVD	p_len+16(FP), LEN

	SRD	$6, LEN
	SLD	$6, LEN
	ADD	INP, LEN, END

	CMP	INP, END
	BEQ	end

	MOVD	$·kcon(SB), TBL_STRT
	MOVD	$0x10, R_x010

#ifdef GOARCH_ppc64le
	MOVWZ	$8, TEMP

	LVX		(TBL)(idx), KI; \
	VADDUDM		S0, h, h; \
	VADDUDM		s1, xj, xj

// func block(dig *digest, p []byte)
TEXT ·block(SB),0,$0-32
	MOVD	dig+0(FP), CTX
	MOVD	p_base+8(FP), INP
	MOVD	p_len+16(FP), LEN

	SRD	$6, LEN
	SLD	$6, LEN

	ADD	INP, LEN, END

	CMP	INP, END
	BEQ	end

	MOVD	$·kcon(SB), TBL_STRT

	MOVD	R0, CNT
	MOVWZ	$0x010, R_x010
	MOVWZ	$0x020, R_x020

#define SHA256ROUND1(index, const, a, b, c, d, e, f, g, h) \
	MSGSCHEDULE1(index); \
	SHA256ROUND(index, const, a, b, c, d, e, f, g, h)

TEXT ·block(SB), 0, $536-32
	MOVQ p_base+8(FP), SI
	MOVQ p_len+16(FP), DX
	SHRQ $6, DX
	SHLQ $6, DX

	LEAQ (SI)(DX*1), DI
	MOVQ DI, 256(SP)
	CMPQ SI, DI
	JEQ  end

	MOVQ dig+0(FP), BP
	MOVL (0*4)(BP), R8  // a = H0
	MOVL (1*4)(BP), R9  // b = H1

#define SHA512ROUND1(index, const, a, b, c, d, e, f, g, h) \
	MSGSCHEDULE1(index); \
	SHA512ROUND(index, const, a, b, c, d, e, f, g, h)

TEXT ·blockAMD64(SB),0,$648-32
	MOVQ	p_base+8(FP), SI
	MOVQ	p_len+16(FP), DX
	SHRQ	$7, DX
	SHLQ	$7, DX

	LEAQ	(SI)(DX*1), DI
	MOVQ	DI, 640(SP)
	CMPQ	SI, DI
	JEQ	end

	MOVQ	dig+0(FP), BP
	MOVQ	(0*8)(BP), R8		// a = H0
	MOVQ	(1*8)(BP), R9		// b = H1