PXOR T2, ACC1;\
	PSHUFD $78, X, T1;\
	PXOR T1, X;\
	MOVOU (16*(i*2+1))(pTbl), T1;\
	PCLMULQDQ $0x00, X, T1;\
	PXOR T1, ACCM

	MOVQ productTable+0(FP), pTbl
	MOVQ data_base+8(FP), aut
	MOVQ data_len+16(FP), autLen
	MOVQ T+32(FP), tPtr

	PXOR ACC0, ACC0
	MOVOU bswapMask<>(SB), BSWAP
	MOVOU gcmPoly<>(SB), POLY

	TESTQ autLen, autLen
	JEQ dataBail

	CMPQ autLen, $13	// optimize the TLS case

	VPMULL2	X.D2, T1.D2, T3.Q1         \
	VEOR	T3.B16, ACC0.B16, ACC0.B16 \
	VPMULL	T0.D1, T2.D1, T3.Q1        \
	VEOR	T3.B16, ACCM.B16, ACCM.B16

	MOVD	productTable+0(FP), pTbl
	MOVD	data_base+8(FP), aut
	MOVD	data_len+16(FP), autLen
	MOVD	T+32(FP), tPtr

	VEOR	ACC0.B16, ACC0.B16, ACC0.B16
	CBZ	autLen, dataBail

	MOVD	$0xC2, H0
	LSL	$56, H0
	VMOV	H0, POLY.D[0]
	MOVD	$1, H0
	VMOV	H0, POLY.D[1]