ANDSW	$0x100, R13, ZR
		CSELW	NE, R14, R13, R13 // Fake modulo
		SUBS	$1, R8
		VMOV	V2.S[0], R0
		EORW	R0, R4
		EORW	R4, R5
		EORW	R5, R6
		EORW	R6, R7
		STPW.P	(R4, R5), 8(R10)
		STPW.P	(R6, R7), 8(R10)
	BNE	ks128Loop
	CBZ	R11, ksDone       // If dec is nil we are done
	SUB	$176, R10
	// Decryption keys are encryption keys with InverseMixColumns applied

	// R5: &args[8]
	// R6: loop counter, from 0 to (n-8)*8
	// R7: scratch
	// R8: copy of RSP - (R2)(RSP) assembles as (R2)(ZR)
	SUB	$8, R0, R4
	LSL	$3, R4
	ADD	$(8*8), R12, R5
	MOVD	$0, R6
	MOVD	RSP, R8
stackargs:
	MOVD	(R6)(R5), R7
	MOVD	R7, (R6)(R8)
	ADD	$8, R6
	CMP	R6, R4
	BNE	stackargs

_8args:
	MOVD	(7*8)(R12), R7
_7args:
	MOVD	(6*8)(R12), R6
_6args:
	MOVD	(5*8)(R12), R5
_5args:

// performing [R4:R12:R11:R10:R9] = R4 + R3 * [R8:R7:R6:R5] + [R12:R11:R10:R9]
// where R4 is carried from the previous iteration, R8:R7:R6:R5 hold the next
// 4 words of x, R3 is y and R12:R11:R10:R9 are part of the result z.
loop:
	CBZ	R0, done

	LDP.P	16(R2), (R5, R6)
	LDP.P	16(R2), (R7, R8)

	LDP	(R1), (R9, R10)
	ADDS	R4, R9
	MUL	R6, R3, R14
	ADCS	R14, R10
	MUL	R7, R3, R15
	LDP	16(R1), (R11, R12)

	MOVD	callpc+8(FP), R5
	MOVD	pc+16(FP), R6
	// void __tsan_write_pc(ThreadState *thr, void *addr, void *callpc, void *pc);
	MOVD	$__tsan_write_pc(SB), R8
	BR	racecalladdr<>(SB)

// func runtime·RaceReadRange(addr, size uintptr)
// Called from instrumented Go code.
TEXT	runtime·racereadrange<ABIInternal>(SB), NOSPLIT, $0-16
	MOVD	R4, R5 // size
	MOVD	R3, R4 // addr
	MOVD	LR, R6

// fn is the PC of a func(a unsafe.Pointer) function.
TEXT crosscall2(SB),NOSPLIT|NOFRAME,$0
	// Start with standard C stack frame layout and linkage.

	// Save R6-R15 in the register save area of the calling function.
	STMG	R6, R15, 48(R15)

	// Allocate 96 bytes on the stack.
	MOVD	$-96(R15), R15

	// Save F8-F15 in our stack frame.
	FMOVD	F8, 32(R15)
	FMOVD	F9, 40(R15)
	FMOVD	F10, 48(R15)

	LVX	(R3),byteswap
#endif

	CMPU	R5,$256		// length of bytes
	BLT	short

	RLDICR	$0,R5,$56,R6 // chunk to process

	// First step for larger sizes
l1:	MOVD	$32768,R7
	MOVD	R7,R9
	CMP	R6,R7   // compare R6, R7 (MAX SIZE)
	BGT	top	// less than MAX, just do remainder
	MOVD	R6,R7
top:
	SUB	R7,R6,R6

	// mainloop does 128 bytes at a time
	SRD	$7,R7

	MOVW	a+0(FP), R1
	MOVW	b+4(FP), R2
	BEQ	R1, R2, eq
	MOVW	size+8(FP), R3
	ADDU	R1, R3, R4
loop:
	BNE	R1, R4, test
	MOVW	$1, R1
	MOVB	R1, ret+12(FP)
	RET
test:
	MOVBU	(R1), R6
	ADDU	$1, R1
	MOVBU	(R2), R7
	ADDU	$1, R2
	BEQ	R6, R7, loop

	MOVB	R0, ret+12(FP)
	RET
eq:
	MOVW	$1, R1
	MOVB	R1, ret+12(FP)
	RET

// memequal_varlen(a, b unsafe.Pointer) bool
TEXT runtime·memequal_varlen(SB),NOSPLIT,$0-9

	genSize1 := (1 << 16) + 16
	genSize2 := (1 << 21) + 16

	fmt.Fprintln(buf, "TEXT f(SB),0,$0-0")
	fmt.Fprintln(buf, "BEQ R5, R6, label18")
	fmt.Fprintln(buf, "BNE R5, R6, label18")
	fmt.Fprintln(buf, "BGE R5, R6, label18")

	fmt.Fprintln(buf, "BGEU R5, R6, label18")
	fmt.Fprintln(buf, "BLTU R5, R6, label18")

	fmt.Fprintln(buf, "BLEZ R5, label18")
	fmt.Fprintln(buf, "BGEZ R5, label18")
	fmt.Fprintln(buf, "BLTZ R5, label18")

	MOVWHI	R0, 0(R1)	// MOVWHI will write zeros up to next word boundary
	SUBU	R3, R2
	ADDU	R3, R1

setup:
	AND	$31, R2, R6
	AND	$3, R2, R5
	SUBU	R6, R4, R6	// end pointer for 32-byte chunks
	SUBU	R5, R4, R5	// end pointer for 4-byte chunks

large:
	BEQ	R1, R6, words
	MOVW	R0, 0(R1)
	MOVW	R0, 4(R1)
	MOVW	R0, 8(R1)
	MOVW	R0, 12(R1)
	MOVW	R0, 16(R1)
	MOVW	R0, 20(R1)
	MOVW	R0, 24(R1)

//go:build ppc64le || ppc64

#include "go_asm.h"
#include "textflag.h"

TEXT ·Count<ABIInternal>(SB),NOSPLIT|NOFRAME,$0-40
	// R3 = byte array pointer
	// R4 = length
	// R6 = byte to count
	MTVRD	R6, V1		// move compare byte
	MOVD	R6, R5
	VSPLTB	$7, V1, V1	// replicate byte across V1
	BR	countbytebody<>(SB)

TEXT ·CountString<ABIInternal>(SB), NOSPLIT|NOFRAME, $0-32
	// R3 = byte array pointer
	// R4 = length