#endif

// R3 = addr of string
// R4 = len of string
// R5 = byte to find
// On exit:
// R3 = return value
TEXT indexbytebody<>(SB),NOSPLIT|NOFRAME,$0-0
	CMPU	R4,$32

#ifndef GOPPC64_power9
	// Load VBPERMQ constant to reduce compare into an ordered bit mask.
	MOVD	$indexbytevbperm<>+00(SB),R16
	LXVD2X	(R16),V0	// Set up swap string
#endif

	MTVRD	R5,V1
	VSPLTB	$7,V1,V1	// Replicate byte across V1

	VADDUDM	V16, V6, V6
	VADDUDM	V17, V7, V7

	CMPU	INP, END
	BLT	loop

#ifdef GOARCH_ppc64le
	VPERM	V0, V1, KI, V0
	VPERM	V2, V3, KI, V2
	VPERM	V4, V5, KI, V4
	VPERM	V6, V7, KI, V6
#else
	VPERM	V1, V0, KI, V0
	VPERM	V3, V2, KI, V2
	VPERM	V5, V4, KI, V4
	VPERM	V7, V6, KI, V6
#endif
	STXVD2X	VS32, (CTX+R_x000)	// v0 = vs32
	STXVD2X	VS34, (CTX+R_x010)	// v2 = vs34

	SHA256MSG2 X2, X11       // 440f38cdda
	SHA256MSG2 X11, X11      // 450f38cddb
	// Test VPERMQ with both uint8 and int8 immediate args
	VPERMQ $-40, Y8, Y8 // c443fd00c0d8
	VPERMQ $216, Y8, Y8 // c443fd00c0d8
	// Test that VPERMPD that shares ytab list with VPERMQ continues to work too.
	VPERMPD $-40, Y7, Y7 // c4e3fd01ffd8
	VPERMPD $216, Y7, Y7 // c4e3fd01ffd8
	// Check that LEAL is permitted to use overflowing offset.

	VX     d3, d1, d1  \
	VERLLF $7, a1, a1  \
	VERLLF $7, b1, b1  \
	VERLLF $7, c1, c1  \
	VERLLF $7, d1, d1

#define PERMUTE(mask, v0, v1, v2, v3) \
	VPERM v0, v0, mask, v0 \
	VPERM v1, v1, mask, v1 \
	VPERM v2, v2, mask, v2 \
	VPERM v3, v3, mask, v3

#define ADDV(x, v0, v1, v2, v3) \
	VAF x, v0, v0 \
	VAF x, v1, v1 \
	VAF x, v2, v2 \
	VAF x, v3, v3

	AVCMPGTSWCC
	AVCMPGTSD
	AVCMPGTSDCC
	AVCMPNEZB
	AVCMPNEZBCC
	AVCMPNEB
	AVCMPNEBCC
	AVCMPNEH
	AVCMPNEHCC
	AVCMPNEW
	AVCMPNEWCC
	AVPERM
	AVPERMXOR
	AVPERMR
	AVBPERMQ
	AVBPERMD
	AVSEL
	AVSPLTB
	AVSPLTH
	AVSPLTW
	AVSPLTISB
	AVSPLTISH
	AVSPLTISW
	AVCIPH
	AVCIPHER
	AVCIPHERLAST
	AVNCIPH
	AVNCIPHER
	AVNCIPHERLAST
	AVSBOX
	AVSHASIGMA

	LXVD2X (CPOOL+R0), PH
	LXVD2X (CPOOL+R16), PL

	// VPERM byte selections
	LXVD2X (CPOOL+R18), SEL2
	LXVD2X (CPOOL+R19), SEL1

	LXVD2X (R16)(x_ptr), T1
	LXVD2X (R0)(x_ptr), T0

	// Put in true little endian order
	XXPERMDI T0, T0, $2, T0
	XXPERMDI T1, T1, $2, T1

	// First round
	VPERM   T1, T0, SEL1, RED2    // d1 d0 d1 d0
	VPERM   ZER, RED2, SEL2, RED1 // 0  d1 d0  0

	XXLOR	VS31, VS31, V17
	VADDUWM	V16, V6, V6
	VADDUWM	V17, V7, V7

	CMPU	INP, END
	BLT	loop

	LVX	(TBL)(R_x000), V8
	VPERM	V0, V1, KI, V0
	LVX	(TBL)(R_x010), V9
	VPERM	V4, V5, KI, V4
	VPERM	V0, V2, V8, V0
	VPERM	V4, V6, V8, V4
	VPERM	V0, V3, V9, V0
	VPERM	V4, V7, V9, V4
	STXVD2X	V0, (CTX+R_x000)
	STXVD2X	V4, (CTX+R_x010)

end:

	// Third round
	VPERM T1, T0, SEL1, RED2    // d1 d0 d1 d0
	VPERM ZER, RED2, SEL2, RED1 // 0  d1 d0  0
	VSQ   RED1, RED2, RED2      // Guaranteed not to underflow

	VSLDB $8, T1, T0, T0
	VSLDB $8, T2, T1, T1

	VACCQ  T0, RED1, CAR1
	VAQ    T0, RED1, T0
	VACCCQ T1, RED2, CAR1, CAR2
	VACQ   T1, RED2, CAR1, T1
	VAQ    T2, CAR2, T2

	// Last round
	VPERM T1, T0, SEL1, RED2    // d1 d0 d1 d0

	VPERMW Z0, Z6, K7, Z1                              // 62f2cd4f8dc8
	VPERMW Z8, Z6, K7, Z1                              // 62d2cd4f8dc8
	VPERMW 15(R8), Z6, K7, Z1                          // 62d2cd4f8d880f000000
	VPERMW (BP), Z6, K7, Z1                            // 62f2cd4f8d4d00
	VPERMW Z0, Z2, K7, Z1                              // 62f2ed4f8dc8
	VPERMW Z8, Z2, K7, Z1                              // 62d2ed4f8dc8

	VPERM2I128 $7, Y11, Y15, Y11    // c4430546db07
	VPERMD (BX), Y15, Y2                    // c4e2053613
	VPERMD (R11), Y15, Y2                   // c4c2053613
	VPERMD Y2, Y15, Y2                      // c4e20536d2
	VPERMD Y11, Y15, Y2                     // c4c20536d3
	VPERMD (BX), Y15, Y11                   // c46205361b
	VPERMD (R11), Y15, Y11                  // c44205361b
	VPERMD Y2, Y15, Y11                     // c4620536da