// license that can be found in the LICENSE file.

//go:build !math_big_pure_go

#include "textflag.h"

// This file provides fast assembly versions for the elementary
// arithmetic operations on vectors implemented in arith.go.

// func addVV(z, x, y []Word) (c Word)
TEXT ·addVV(SB),NOSPLIT,$0
	ADD.S	$0, R0		// clear carry flag
	MOVW	z+0(FP), R1
	MOVW	z_len+4(FP), R4
	MOVW	x+12(FP), R2
	MOVW	y+24(FP), R3

	// Create a struct callbackArgs on our stack.
	MOVD	$cbargs-(18*8+callbackArgs__size)(SP), R13
	MOVD	R12, callbackArgs_index(R13)	// callback index
	MOVD	R14, R0
	MOVD	R0, callbackArgs_args(R13)		// address of args vector
	MOVD	$0, R0
	MOVD	R0, callbackArgs_result(R13)	// result

	// Call cgocallback, which will call callbackWrap(frame).

// equivalent function from the corresponding s390x
// instruction for vector multiply high, low, and add,
// since there aren't exact equivalent instructions.
// The corresponding s390x instructions appear in the
// comments.
// Implementation for big endian would have to be
// investigated, I think it would be different.
//
//
// Vector multiply word
//
//	VMLF  x0, x1, out_low
//	VMLHF x0, x1, out_hi

	// Create a struct callbackArgs on our stack.
	MOVW	R12, (16+callbackArgs_index)(R13)	// callback index
	MOVW	$(16+callbackArgs__size+4*9)(R13), R0
	MOVW	R0, (16+callbackArgs_args)(R13)		// address of args vector
	MOVW	$0, R0
	MOVW	R0, (16+callbackArgs_result)(R13)	// result

	// Prepare for entry to Go.
	BL	runtime·load_g(SB)

	// Call cgocallback, which will call callbackWrap(frame).
	MOVW	$0, R0

// license that can be found in the LICENSE file.

//go:build !math_big_pure_go && loong64

#include "textflag.h"

// This file provides fast assembly versions for the elementary
// arithmetic operations on vectors implemented in arith.go.

TEXT ·addVV(SB),NOSPLIT,$0
	JMP ·addVV_g(SB)

TEXT ·subVV(SB),NOSPLIT,$0
	JMP ·subVV_g(SB)

TEXT ·addVW(SB),NOSPLIT,$0
	JMP ·addVW_g(SB)

TEXT ·subVW(SB),NOSPLIT,$0

// license that can be found in the LICENSE file.

//go:build !math_big_pure_go && (mips64 || mips64le)

#include "textflag.h"

// This file provides fast assembly versions for the elementary
// arithmetic operations on vectors implemented in arith.go.

TEXT ·addVV(SB),NOSPLIT,$0
	JMP ·addVV_g(SB)

TEXT ·subVV(SB),NOSPLIT,$0
	JMP ·subVV_g(SB)

TEXT ·addVW(SB),NOSPLIT,$0
	JMP ·addVW_g(SB)

TEXT ·subVW(SB),NOSPLIT,$0

// license that can be found in the LICENSE file.

//go:build !math_big_pure_go && riscv64

#include "textflag.h"

// This file provides fast assembly versions for the elementary
// arithmetic operations on vectors implemented in arith.go.

TEXT ·addVV(SB),NOSPLIT,$0
	JMP ·addVV_g(SB)

TEXT ·subVV(SB),NOSPLIT,$0
	JMP ·subVV_g(SB)

TEXT ·addVW(SB),NOSPLIT,$0
	JMP ·addVW_g(SB)

TEXT ·subVW(SB),NOSPLIT,$0

	MOVD	$16, R10
	MOVD	$32, R14
	MOVD	$48, R15
	ANDCC	$63, R6, R9	// Check for tailing bytes for later
	PCALIGN $16
	// Case for >= 64 bytes
	// Process 64 bytes per iteration
	// Load 4 vectors of a and b
	// XOR the corresponding vectors
	// from a and b and store the result
loop64:
	LXVD2X	(R4)(R8), VS32
	LXVD2X	(R4)(R10), VS34
	LXVD2X	(R4)(R14), VS36
	LXVD2X	(R4)(R15), VS38
	LXVD2X	(R5)(R8), VS33
	LXVD2X	(R5)(R10), VS35

// license that can be found in the LICENSE file.

//go:build !math_big_pure_go

#include "textflag.h"

// This file provides fast assembly versions for the elementary
// arithmetic operations on vectors implemented in arith.go.

// The carry bit is saved with SBBQ Rx, Rx: if the carry was set, Rx is -1, otherwise it is 0.
// It is restored with ADDQ Rx, Rx: if Rx was -1 the carry is set, otherwise it is cleared.

// license that can be found in the LICENSE file.

//go:build !math_big_pure_go

#include "textflag.h"

// This file provides fast assembly versions for the elementary
// arithmetic operations on vectors implemented in arith.go.

// TODO: Consider re-implementing using Advanced SIMD
// once the assembler supports those instructions.

// func addVV(z, x, y []Word) (c Word)
TEXT ·addVV(SB),NOSPLIT,$0