[pcsx_rearmed.git] / frontend / cspace_neon.s

/*
 * (C) Gražvydas "notaz" Ignotas, 2010
 *
 * This work is licensed under the terms of any of these licenses
 * (at your option):
 *  - GNU GPL, version 2 or later.
 *  - GNU LGPL, version 2.1 or later.
 * See the COPYING file in the top-level directory.
 */

.text
.align 2

.global bgr555_to_rgb565
bgr555_to_rgb565:
    pld         [r1]
    mov         r3, #0x07c0
    vdup.16     q15, r3
    subs        r2, r2, #64
    blt         btr16_end64
0:
    pld         [r1, #64*2]
    vldmia      r1!, {q0-q3}
    vshl.u16    q4, q0, #11
    vshl.u16    q5, q1, #11
    vshl.u16    q6, q2, #11
    vshl.u16    q7, q3, #11
    vsri.u16    q4, q0, #10
    vsri.u16    q5, q1, #10
    vsri.u16    q6, q2, #10
    vsri.u16    q7, q3, #10
    vshl.u16    q0, q0, #1
    vshl.u16    q1, q1, #1
    vshl.u16    q2, q2, #1
    vshl.u16    q3, q3, #1
    vbit        q4, q0, q15
    vbit        q5, q1, q15
    vbit        q6, q2, q15
    vbit        q7, q3, q15
    vstmia      r0!, {q4-q7}
    subs        r2, r2, #64
    bge         0b

btr16_end64:
    adds        r2, r2, #64
    bxeq        lr
    subs        r2, r2, #16
    blt         btr16_end16

    @ handle the remainder (reasonably rare)
0:
    vld1.16     {q0}, [r1]!
    vshl.u16    q1, q0, #11
    vshl.u16    q2, q0, #1
    vsri.u16    q1, q0, #10
    vbit        q1, q2, q15
    subs        r2, r2, #16
    vst1.16     {q1}, [r0]!
    bge         0b

btr16_end16:
    adds        r2, r2, #16
    bxeq        lr
    subs        r2, r2, #8
    bxlt        lr

    @ very rare
    vld1.16     d0, [r1]!
    vshl.u16    d1, d0, #11
    vshl.u16    d2, d0, #1
    vsri.u16    d1, d0, #10
    vbit        d1, d2, d30
    vst1.16     d1, [r0]!
    bx          lr


.global bgr888_to_rgb888
bgr888_to_rgb888:
    pld         [r1]
    @ r2 /= 48
    mov         r2, r2, lsr #4
    movw        r3, #0x5556
    movt        r3, #0x5555
    umull       r12,r2, r3, r2
0:
    pld         [r1, #48*3]
    vld3.8      {d0-d2}, [r1, :64]!
    vld3.8      {d3-d5}, [r1, :64]!
    vswp        d0, d2
    vswp        d3, d5
    vst3.8      {d0-d2}, [r0, :64]!
    vst3.8      {d3-d5}, [r0, :64]!
    subs        r2, r2, #1
    bne         0b

    bx          lr


.global bgr888_to_rgb565
bgr888_to_rgb565:
    pld         [r1]
    @ r2 /= 48
    mov         r2, r2, lsr #4
    movw        r3, #0x5556
    movt        r3, #0x5555
    umull       r12,r2, r3, r2

    mov         r3, #0x07e0
    vdup.16     q15, r3
0:
    pld         [r1, #48*3]
    vld3.8      {d1-d3}, [r1, :64]!
    vld3.8      {d5-d7}, [r1, :64]!

    vshll.u8    q8, d2, #3      @ g
    vshll.u8    q9, d6, #3
    vshr.u8     d0, d3, #3      @ b
    vshr.u8     d4, d7, #3
    vzip.8      d0, d1          @ rb
    vzip.8      d4, d5
    vbit        q0, q8, q15
    vbit        q2, q9, q15

    vstmia      r0!, {d0,d1}
    vstmia      r0!, {d4,d5}
    subs        r2, r2, #1
    bne         0b

    bx          lr


.global rgb888_to_rgb565
rgb888_to_rgb565:
    pld         [r1]
    @ r2 /= 48
    mov         r2, r2, lsr #4
    movw        r3, #0x5556
    movt        r3, #0x5555
    umull       r12,r2, r3, r2

    mov         r3, #0x07e0
    vdup.16     q15, r3
0:
    pld         [r1, #48*3]
    vld3.8      {d1-d3}, [r1, :64]!
    vld3.8      {d5-d7}, [r1, :64]!

    vshll.u8    q8, d2, #3      @ g
    vshll.u8    q9, d6, #3
    vshr.u8     d2, d1, #3      @ b
    vshr.u8     d6, d5, #3
    vzip.8      d2, d3          @ rb
    vzip.8      d6, d7
    vbit        q1, q8, q15
    vbit        q3, q9, q15

    vstmia      r0!, {d2,d3}
    vstmia      r0!, {d6,d7}
    subs        r2, r2, #1
    bne         0b

    bx          lr


@ vim:filetype=armasm
Commit	Line	Data
	1	/*
	2	* (C) Gražvydas "notaz" Ignotas, 2010
	3	*
	4	* This work is licensed under the terms of any of these licenses
	5	* (at your option):
	6	* - GNU GPL, version 2 or later.
	7	* - GNU LGPL, version 2.1 or later.
	8	* See the COPYING file in the top-level directory.
	9	*/
	10
	11	.text
	12	.align 2
	13
	14	.global bgr555_to_rgb565
	15	bgr555_to_rgb565:
	16	pld [r1]
	17	mov r3, #0x07c0
	18	vdup.16 q15, r3
	19	subs r2, r2, #64
	20	blt btr16_end64
	21	0:
	22	pld [r1, #64*2]
	23	vldmia r1!, {q0-q3}
	24	vshl.u16 q4, q0, #11
	25	vshl.u16 q5, q1, #11
	26	vshl.u16 q6, q2, #11
	27	vshl.u16 q7, q3, #11
	28	vsri.u16 q4, q0, #10
	29	vsri.u16 q5, q1, #10
	30	vsri.u16 q6, q2, #10
	31	vsri.u16 q7, q3, #10
	32	vshl.u16 q0, q0, #1
	33	vshl.u16 q1, q1, #1
	34	vshl.u16 q2, q2, #1
	35	vshl.u16 q3, q3, #1
	36	vbit q4, q0, q15
	37	vbit q5, q1, q15
	38	vbit q6, q2, q15
	39	vbit q7, q3, q15
	40	vstmia r0!, {q4-q7}
	41	subs r2, r2, #64
	42	bge 0b
	43
	44	btr16_end64:
	45	adds r2, r2, #64
	46	bxeq lr
	47	subs r2, r2, #16
	48	blt btr16_end16
	49
	50	@ handle the remainder (reasonably rare)
	51	0:
	52	vld1.16 {q0}, [r1]!
	53	vshl.u16 q1, q0, #11
	54	vshl.u16 q2, q0, #1
	55	vsri.u16 q1, q0, #10
	56	vbit q1, q2, q15
	57	subs r2, r2, #16
	58	vst1.16 {q1}, [r0]!
	59	bge 0b
	60
	61	btr16_end16:
	62	adds r2, r2, #16
	63	bxeq lr
	64	subs r2, r2, #8
	65	bxlt lr
	66
	67	@ very rare
	68	vld1.16 d0, [r1]!
	69	vshl.u16 d1, d0, #11
	70	vshl.u16 d2, d0, #1
	71	vsri.u16 d1, d0, #10
	72	vbit d1, d2, d30
	73	vst1.16 d1, [r0]!
	74	bx lr
	75
	76
	77	.global bgr888_to_rgb888
	78	bgr888_to_rgb888:
	79	pld [r1]
	80	@ r2 /= 48
	81	mov r2, r2, lsr #4
	82	movw r3, #0x5556
	83	movt r3, #0x5555
	84	umull r12,r2, r3, r2
	85	0:
	86	pld [r1, #48*3]
	87	vld3.8 {d0-d2}, [r1, :64]!
	88	vld3.8 {d3-d5}, [r1, :64]!
	89	vswp d0, d2
	90	vswp d3, d5
	91	vst3.8 {d0-d2}, [r0, :64]!
	92	vst3.8 {d3-d5}, [r0, :64]!
	93	subs r2, r2, #1
	94	bne 0b
	95
	96	bx lr
	97
	98
	99	.global bgr888_to_rgb565
	100	bgr888_to_rgb565:
	101	pld [r1]
	102	@ r2 /= 48
	103	mov r2, r2, lsr #4
	104	movw r3, #0x5556
	105	movt r3, #0x5555
	106	umull r12,r2, r3, r2
	107
	108	mov r3, #0x07e0
	109	vdup.16 q15, r3
	110	0:
	111	pld [r1, #48*3]
	112	vld3.8 {d1-d3}, [r1, :64]!
	113	vld3.8 {d5-d7}, [r1, :64]!
	114
	115	vshll.u8 q8, d2, #3 @ g
	116	vshll.u8 q9, d6, #3
	117	vshr.u8 d0, d3, #3 @ b
	118	vshr.u8 d4, d7, #3
	119	vzip.8 d0, d1 @ rb
	120	vzip.8 d4, d5
	121	vbit q0, q8, q15
	122	vbit q2, q9, q15
	123
	124	vstmia r0!, {d0,d1}
	125	vstmia r0!, {d4,d5}
	126	subs r2, r2, #1
	127	bne 0b
	128
	129	bx lr
	130
	131
	132	.global rgb888_to_rgb565
	133	rgb888_to_rgb565:
	134	pld [r1]
	135	@ r2 /= 48
	136	mov r2, r2, lsr #4
	137	movw r3, #0x5556
	138	movt r3, #0x5555
	139	umull r12,r2, r3, r2
	140
	141	mov r3, #0x07e0
	142	vdup.16 q15, r3
	143	0:
	144	pld [r1, #48*3]
	145	vld3.8 {d1-d3}, [r1, :64]!
	146	vld3.8 {d5-d7}, [r1, :64]!
	147
	148	vshll.u8 q8, d2, #3 @ g
	149	vshll.u8 q9, d6, #3
	150	vshr.u8 d2, d1, #3 @ b
	151	vshr.u8 d6, d5, #3
	152	vzip.8 d2, d3 @ rb
	153	vzip.8 d6, d7
	154	vbit q1, q8, q15
	155	vbit q3, q9, q15
	156
	157	vstmia r0!, {d2,d3}
	158	vstmia r0!, {d6,d7}
	159	subs r2, r2, #1
	160	bne 0b
	161
	162	bx lr
	163
	164
	165	@ vim:filetype=armasm