[picodrive.git] / pico / misc_arm.s

/*
 * Generic memory routines.
 * (C) notaz, 2007-2008
 *
 * This work is licensed under the terms of MAME license.
 * See COPYING file in the top-level directory.
 */

.global memcpy16 @ unsigned short *dest, unsigned short *src, int count

memcpy16:
    eor     r3, r0, r1
    tst     r3, #2
    bne     mcp16_cant_align

    tst     r0, #2
    ldrneh  r3, [r1], #2
    subne   r2, r2, #1
    strneh  r3, [r0], #2

    subs    r2, r2, #4
    bmi     mcp16_fin

mcp16_loop:
    ldmia   r1!, {r3,r12}
    subs    r2, r2, #4
    stmia   r0!, {r3,r12}
    bpl     mcp16_loop

mcp16_fin:
    tst     r2, #2
    ldrne   r3, [r1], #4
    strne   r3, [r0], #4
    ands    r2, r2, #1
    bxeq    lr

mcp16_cant_align:
    ldrh    r3, [r1], #2
    subs    r2, r2, #1
    strh    r3, [r0], #2
    bne     mcp16_cant_align

    bx      lr


@ 0x12345678 -> 0x34127856
@ r4=temp, lr=0x00ff00ff
.macro bswap reg
    and     r4,   \reg, lr
    and     \reg, lr,   \reg, lsr #8
    orr     \reg, \reg, r4,   lsl #8
.endm


@ dest must be halfword aligned, src can be unaligned
.global memcpy16bswap @ unsigned short *dest, void *src, int count

memcpy16bswap:
    tst     r1, #1
    bne     mcp16bs_cant_align2

    eor     r3, r0, r1
    tst     r3, #2
    bne     mcp16bs_cant_align

    tst     r0, #2
    beq     mcp16bs_aligned
    ldrh    r3, [r1], #2
    sub     r2, r2, #1
    orr     r3, r3, r3, lsl #16
    mov     r3, r3, lsr #8
    strh    r3, [r0], #2

mcp16bs_aligned:
    stmfd   sp!, {r4,lr}
    mov     lr, #0xff
    orr     lr, lr, lr, lsl #16

    subs    r2, r2, #4
    bmi     mcp16bs_fin4

mcp16bs_loop:
    ldmia   r1!, {r3,r12}
    subs    r2, r2, #4
    bswap   r3
    bswap   r12
    stmia   r0!, {r3,r12}
    bpl     mcp16bs_loop

mcp16bs_fin4:
    tst     r2, #2
    beq     mcp16bs_fin2
    ldr     r3, [r1], #4
    bswap   r3
    str     r3, [r0], #4

mcp16bs_fin2:
    ldmfd   sp!, {r4,lr}
    ands    r2, r2, #1
    bxeq    lr

mcp16bs_cant_align:
    ldrh    r3, [r1], #2
    subs    r2, r2, #1
    orr     r3, r3, r3, lsl #16
    mov     r3, r3, lsr #8
    strh    r3, [r0], #2
    bne     mcp16bs_cant_align
    bx      lr

    @ worst case
mcp16bs_cant_align2:
    ldrb    r3, [r1], #1
    ldrb    r12,[r1], #1
    subs    r2, r2, #1
    mov     r3, r3, lsl #8
    orr     r3, r3, r12
    strh    r3, [r0], #2
    bne     mcp16bs_cant_align2
    bx      lr


.global memcpy32 @ int *dest, int *src, int count

memcpy32:
    stmfd   sp!, {r4,lr}

    subs    r2, r2, #4
    bmi     mcp32_fin

mcp32_loop:
    ldmia   r1!, {r3,r4,r12,lr}
    subs    r2, r2, #4
    stmia   r0!, {r3,r4,r12,lr}
    bpl     mcp32_loop

mcp32_fin:
    tst     r2, #3
    ldmeqfd sp!, {r4,pc}
    tst     r2, #1
    ldrne   r3, [r1], #4
    strne   r3, [r0], #4

mcp32_no_unal1:
    tst     r2, #2
    ldmneia r1!, {r3,r12}
    ldmfd   sp!, {r4,lr}
    stmneia r0!, {r3,r12}
    bx      lr


.global memset32 @ int *dest, int c, int count

memset32:
    stmfd   sp!, {lr}

    mov     r3, r1
    subs    r2, r2, #4
    bmi     mst32_fin

    mov     r12,r1
    mov     lr, r1

mst32_loop:
    subs    r2, r2, #4
    stmia   r0!, {r1,r3,r12,lr}
    bpl     mst32_loop

mst32_fin:
    tst     r2, #1
    strne   r1, [r0], #4

    tst     r2, #2
    stmneia r0!, {r1,r3}

    ldmfd   sp!, {lr}
    bx      lr

@ vim:filetype=armasm
Commit	Line	Data
	1	/*
	2	* Generic memory routines.
	3	* (C) notaz, 2007-2008
	4	*
	5	* This work is licensed under the terms of MAME license.
	6	* See COPYING file in the top-level directory.
	7	*/
	8
	9	.global memcpy16 @ unsigned short dest, unsigned short src, int count
	10
	11	memcpy16:
	12	eor r3, r0, r1
	13	tst r3, #2
	14	bne mcp16_cant_align
	15
	16	tst r0, #2
	17	ldrneh r3, [r1], #2
	18	subne r2, r2, #1
	19	strneh r3, [r0], #2
	20
	21	subs r2, r2, #4
	22	bmi mcp16_fin
	23
	24	mcp16_loop:
	25	ldmia r1!, {r3,r12}
	26	subs r2, r2, #4
	27	stmia r0!, {r3,r12}
	28	bpl mcp16_loop
	29
	30	mcp16_fin:
	31	tst r2, #2
	32	ldrne r3, [r1], #4
	33	strne r3, [r0], #4
	34	ands r2, r2, #1
	35	bxeq lr
	36
	37	mcp16_cant_align:
	38	ldrh r3, [r1], #2
	39	subs r2, r2, #1
	40	strh r3, [r0], #2
	41	bne mcp16_cant_align
	42
	43	bx lr
	44
	45
	46
	47	@ 0x12345678 -> 0x34127856
	48	@ r4=temp, lr=0x00ff00ff
	49	.macro bswap reg
	50	and r4, \reg, lr
	51	and \reg, lr, \reg, lsr #8
	52	orr \reg, \reg, r4, lsl #8
	53	.endm
	54
	55
	56	@ dest must be halfword aligned, src can be unaligned
	57	.global memcpy16bswap @ unsigned short dest, void src, int count
	58
	59	memcpy16bswap:
	60	tst r1, #1
	61	bne mcp16bs_cant_align2
	62
	63	eor r3, r0, r1
	64	tst r3, #2
	65	bne mcp16bs_cant_align
	66
	67	tst r0, #2
	68	beq mcp16bs_aligned
	69	ldrh r3, [r1], #2
	70	sub r2, r2, #1
	71	orr r3, r3, r3, lsl #16
	72	mov r3, r3, lsr #8
	73	strh r3, [r0], #2
	74
	75	mcp16bs_aligned:
	76	stmfd sp!, {r4,lr}
	77	mov lr, #0xff
	78	orr lr, lr, lr, lsl #16
	79
	80	subs r2, r2, #4
	81	bmi mcp16bs_fin4
	82
	83	mcp16bs_loop:
	84	ldmia r1!, {r3,r12}
	85	subs r2, r2, #4
	86	bswap r3
	87	bswap r12
	88	stmia r0!, {r3,r12}
	89	bpl mcp16bs_loop
	90
	91	mcp16bs_fin4:
	92	tst r2, #2
	93	beq mcp16bs_fin2
	94	ldr r3, [r1], #4
	95	bswap r3
	96	str r3, [r0], #4
	97
	98	mcp16bs_fin2:
	99	ldmfd sp!, {r4,lr}
	100	ands r2, r2, #1
	101	bxeq lr
	102
	103	mcp16bs_cant_align:
	104	ldrh r3, [r1], #2
	105	subs r2, r2, #1
	106	orr r3, r3, r3, lsl #16
	107	mov r3, r3, lsr #8
	108	strh r3, [r0], #2
	109	bne mcp16bs_cant_align
	110	bx lr
	111
	112	@ worst case
	113	mcp16bs_cant_align2:
	114	ldrb r3, [r1], #1
	115	ldrb r12,[r1], #1
	116	subs r2, r2, #1
	117	mov r3, r3, lsl #8
	118	orr r3, r3, r12
	119	strh r3, [r0], #2
	120	bne mcp16bs_cant_align2
	121	bx lr
	122
	123
	124
	125	.global memcpy32 @ int dest, int src, int count
	126
	127	memcpy32:
	128	stmfd sp!, {r4,lr}
	129
	130	subs r2, r2, #4
	131	bmi mcp32_fin
	132
	133	mcp32_loop:
	134	ldmia r1!, {r3,r4,r12,lr}
	135	subs r2, r2, #4
	136	stmia r0!, {r3,r4,r12,lr}
	137	bpl mcp32_loop
	138
	139	mcp32_fin:
	140	tst r2, #3
	141	ldmeqfd sp!, {r4,pc}
	142	tst r2, #1
	143	ldrne r3, [r1], #4
	144	strne r3, [r0], #4
	145
	146	mcp32_no_unal1:
	147	tst r2, #2
	148	ldmneia r1!, {r3,r12}
	149	ldmfd sp!, {r4,lr}
	150	stmneia r0!, {r3,r12}
	151	bx lr
	152
	153
	154
	155	.global memset32 @ int *dest, int c, int count
	156
	157	memset32:
	158	stmfd sp!, {lr}
	159
	160	mov r3, r1
	161	subs r2, r2, #4
	162	bmi mst32_fin
	163
	164	mov r12,r1
	165	mov lr, r1
	166
	167	mst32_loop:
	168	subs r2, r2, #4
	169	stmia r0!, {r1,r3,r12,lr}
	170	bpl mst32_loop
	171
	172	mst32_fin:
	173	tst r2, #1
	174	strne r1, [r0], #4
	175
	176	tst r2, #2
	177	stmneia r0!, {r1,r3}
	178
	179	ldmfd sp!, {lr}
	180	bx lr
	181
	182	@ vim:filetype=armasm