VPKSHUS:        "vpkshus",
	VPKSWSS:        "vpkswss",
	VPKSWUS:        "vpkswus",
	VPKUHUM:        "vpkuhum",
	VPKUHUS:        "vpkuhus",
	VPKUWUM:        "vpkuwum",
	VPKUWUS:        "vpkuwus",
	VREFP:          "vrefp",
	VRFIM:          "vrfim",
	VRFIN:          "vrfin",
	VRFIP:          "vrfip",
	VRFIZ:          "vrfiz",
	VRLB:           "vrlb",
	VRLH:           "vrlh",
	VRLW:           "vrlw",

	VREPIF $4, INC

chacha:
	VREPF $0, J0, X0
	VREPF $1, J0, X1
	VREPF $2, J0, X2
	VREPF $3, J0, X3
	VREPF $0, KEY0, X4
	VREPF $1, KEY0, X5
	VREPF $2, KEY0, X6
	VREPF $3, KEY0, X7
	VREPF $0, KEY1, X8
	VREPF $1, KEY1, X9
	VREPF $2, KEY1, X10
	VREPF $3, KEY1, X11
	VLR   CTR, X12
	VREPF $1, NONCE, X13
	VREPF $2, NONCE, X14
	VREPF $3, NONCE, X15

	MOVD $(NUM_ROUNDS/2), R1

	"VPKLSG",
	"VPKLSHS",
	"VPKLSFS",
	"VPKLSGS",
	"VPKS",
	"VPKSH",
	"VPKSF",
	"VPKSG",
	"VPKSHS",
	"VPKSFS",
	"VPKSGS",
	"VPERM",
	"VPDI",
	"VPOPCT",
	"VREP",
	"VREPB",
	"VREPH",
	"VREPF",
	"VREPG",
	"VREPI",
	"VREPIB",
	"VREPIH",
	"VREPIF",
	"VREPIG",
	"VSCEF",
	"VSCEG",
	"VSEL",
	"VSL",
	"VSLB",
	"VSLDB",
	"VSRA",
	"VSRAB",
	"VSRL",

	VL   (0*16)(y_ptr), Y0
	VPDI $0x4, Y0, Y0, Y0
	VL   (1*16)(y_ptr), Y1
	VPDI $0x4, Y1, Y1, Y1

	// ---------------------------------------------------------------------------/
	VREPF $3, Y0, YDIG
	VMLF  X0, YDIG, ADD1
	VMLF  ADD1, K0, MK0
	VREPF $3, MK0, MK0

	VMLF  X1, YDIG, ADD2
	VMLHF X0, YDIG, ADD1H
	VMLHF X1, YDIG, ADD2H

	VMALF  M0, MK0, ADD1, RED1
	VMALHF M0, MK0, ADD1, RED1H
	VMALF  M1, MK0, ADD2, RED2

	// replicate r across all 4 vector elements
	VREPF $3, H_0, R_0 // [r₂₆[0], r₂₆[0], r₂₆[0], r₂₆[0]]
	VREPF $3, H_1, R_1 // [r₂₆[1], r₂₆[1], r₂₆[1], r₂₆[1]]
	VREPF $3, H_2, R_2 // [r₂₆[2], r₂₆[2], r₂₆[2], r₂₆[2]]
	VREPF $3, H_3, R_3 // [r₂₆[3], r₂₆[3], r₂₆[3], r₂₆[3]]
	VREPF $3, H_4, R_4 // [r₂₆[4], r₂₆[4], r₂₆[4], r₂₆[4]]

	// zero out lane 1 of h
	VLEIG $1, $0, H_0 // [h₂₆[0], 0]

	// ---------------------------------------------------

	VSPLTW $3, Y0, YDIG // VREPF Y0 is input

	//	VMLHF X0, YDIG, ADD1H
	//	VMLHF X1, YDIG, ADD2H
	//	VMLF  X0, YDIG, ADD1
	//	VMLF  X1, YDIG, ADD2
	//
	VMULT(X0, YDIG, ADD1, ADD1H)
	VMULT(X1, YDIG, ADD2, ADD2H)

	VSPLTISW $1, ONE
	VSPLTW $2, Y0, YDIG // VREPF

	//	VMALF  X0, YDIG, ADD1H, ADD3
	//	VMALF  X1, YDIG, ADD2H, ADD4

	ADD	$15, R7, R9
	CMPBGE	R9, R2, index2to16
	MOVD	$2, R0
	VGBM	$0x8888, V29       // 0xff000000ff000000...
	VGBM	$0x2222, V30       // 0x0000ff000000ff00...
	VGBM	$0xcccc, V31       // 0xffff0000ffff0000...
	VONE	V16
	VREPF	$0, V0, V1
index4loop:
	VL	(R7), V2           // load 16-bytes into V2
	VLL	R0, 16(R7), V3     // load 3-bytes into V3
	VSLDB	$1, V2, V3, V4     // V4=(V2:V3)<<1
	VSLDB	$2, V2, V3, V9     // V9=(V2:V3)<<1