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
	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

	ADD  $64, R6
	VPDI $0x4, V9, V9, V9    // flip the doublewords to big-endian order
	VPDI $0x4, V10, V10, V10 // flip the doublewords to big-endian order

	VACCCQ V1, V9, V0, V25
	VACQ   V1, V9, V0, V17
	VACCCQ V2, V10, V25, V26
	VACQ   V2, V10, V25, V18

	VLM 0(R5), V5, V6   // 32-bytes into V1..V8
	VLM 0(R6), V13, V14 // 32-bytes into V9..V16
	ADD $32, R5
	ADD $32, R6

	VADDCUQ  T0, RED1, CAR1     // VACCQ
	VADDUQM  T0, RED1, T0       // VAQ
	VADDECUQ T1, RED2, CAR1, T2 // VACCCQ
	VADDEUQM T1, RED2, CAR1, T1 // VACQ

	VADDCUQ  T0, ADD3, CAR1       // VACCQ
	VADDUQM  T0, ADD3, T0         // VAQ
	VADDECUQ T1, ADD4, CAR1, CAR2 // VACCCQ
	VADDEUQM T1, ADD4, CAR1, T1   // VACQ
	VADDUQM  T2, CAR2, T2         // VAQ

	LXVD2X  (R18)(CPOOL), SEL2

	"KMA",
	"KMCTR",
	"VA",
	"VAB",
	"VAH",
	"VAF",
	"VAG",
	"VAQ",
	"VACC",
	"VACCB",
	"VACCH",
	"VACCF",
	"VACCG",
	"VACCQ",
	"VAC",
	"VACQ",
	"VACCC",
	"VACCCQ",
	"VN",
	"VNC",
	"VAVG",
	"VAVGB",
	"VAVGH",
	"VAVGF",
	"VAVGG",
	"VAVGL",
	"VAVGLB",
	"VAVGLH",
	"VAVGLF",
	"VAVGLG",
	"VCKSM",
	"VCEQ",
	"VCEQB",
	"VCEQH",