1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
|
.arch armv7-a
.text
.global csc_ARGB8888_to_YUV420SP_NEON
.type csc_ARGB8888_to_YUV420SP_NEON, %function
csc_ARGB8888_to_YUV420SP_NEON:
.fnstart
@r0 pDstY
@r1 pDstUV
@r2 pSrcRGB
@r3 nWidth
@r4 pDstY2 = pDstY + nWidth
@r5 pSrcRGB2 = pSrcRGB + nWidthx2
@r6 temp7, nWidth16m
@r7 temp6, accumilator
@r8 temp5, nWidthTemp
@r9 temp4, Raw RGB565
@r10 temp3, r,g,b
@r11 temp2, immediate operand
@r12 temp1, nHeight
@r14 temp0, debugging pointer
.equ CACHE_LINE_SIZE, 32
.equ PRE_LOAD_OFFSET, 6
stmfd sp!, {r4-r12,r14} @ backup registers
ldr r12, [sp, #40] @ load nHeight
@ldr r14, [sp, #44] @ load pTest
add r4, r0, r3 @r4: pDstY2 = pDstY + nWidth
add r5, r2, r3, lsl #2 @r5: pSrcRGB2 = tmpSrcRGB + nWidthx4
sub r8, r3, #16 @r8: nWidthTmp = nWidth -16
@q0: temp1, R
@q1: temp2, GB
@q2: R
@q3: G
@q4: B
@q5: temp3, output
vmov.u16 q6, #66 @coefficient assignment
vmov.u16 q7, #129
vmov.u16 q8, #25
vmov.u16 q9, #0x8080 @ 128<<8 + 128
vmov.u16 q10, #0x1000 @ 16<<8 + 128
vorr.u16 q10, #0x0080
vmov.u16 q11, #38 @#-38
vmov.u16 q12, #74 @#-74
vmov.u16 q13, #112
vmov.u16 q14, #94 @#-94
vmov.u16 q15, #18 @#-18
LOOP_NHEIGHT2:
stmfd sp!, {r12} @ backup registers
LOOP_NWIDTH16:
pld [r2, #(CACHE_LINE_SIZE*PRE_LOAD_OFFSET)]
@-------------------------------------------YUV ------------------------------------------
vmov.u16 q14, #94 @#94
vmov.u16 q15, #18 @#18
vld4.8 {d0,d1,d2,d3}, [r2]! @loadRGB interleavely
vld4.8 {d4,d5,d6,d7}, [r2]! @loadRGB interleavely
vmov.u16 d8,d2
vmov.u16 d9,d6
vmov.u16 d10,d1
vmov.u16 d11,d5
vmov.u16 d12,d0
vmov.u16 d13,d4
vand.u16 q4,#0x00FF @R
vand.u16 q5,#0x00FF @G
vand.u16 q6,#0x00FF @B
vmov.u16 q8,q9 @ CalcU()
vmla.u16 q8,q6,q13 @112 * B[k]
vmls.u16 q8,q4,q11 @q0:U -(38 * R[k]) @128<<6+ 32 + u>>2
vmls.u16 q8,q5,q12 @-(74 * G[k])
vshr.u16 q8,q8, #8 @(128<<8+ 128 + u)>>8
vmov.u16 q7,q9 @CalcV()
vmla.u16 q7,q4,q13 @112 * R[k]
vmls.u16 q7,q5,q14 @q0:U -(94 * G[k]) @128<<6+ 32 + v>>2
vmls.u16 q7,q6,q15 @-(18 * B[k])
vshr.u16 q7,q7, #8 @(128<<8+ 128 + v)>>8
vtrn.8 q8,q7
vst1.8 {q8}, [r1]! @write UV component to yuv420_buffer+linear_ylanesiez
@-------------------------------------------Y ------------------------------------------
vmov.u16 q14, #66 @#66
vmov.u16 q15, #129 @#129
vmov.u16 q8, #25 @#25
@CalcY_Y()
vmul.u16 q7,q4,q14 @q0 = 66 *R[k]
vmla.u16 q7,q5,q15 @q0 += 129 *G[k]
vmla.u16 q7,q6,q8 @q0 += 25 *B[k]
vadd.u16 q7,q7,q10
vshr.u16 q7,q7, #8
vmov.u16 d8,d2
vmov.u16 d9,d6
vmov.u16 d10,d1
vmov.u16 d11,d5
vmov.u16 d12,d0
vmov.u16 d13,d4
vshr.u16 q4,q4,#8 @R
vshr.u16 q5,q5,#8 @G
vshr.u16 q6,q6,#8 @B
vmul.u16 q0,q4,q14 @q0 = 66 *R[k]
vmla.u16 q0,q5,q15 @q0 += 129 *G[k]
vmla.u16 q0,q6,q8 @q0 += 25 *B[k]
vadd.u16 q0,q0,q10
vshr.u16 q0,q0, #8
vtrn.8 q7,q0
vst1.8 {q7}, [r0]!@write to Y to yuv420_buffer
@-------------------------------------------Y ------------------------------------------
@---------------------------------------------Y1-------------------------------------------
pld [r5, #(CACHE_LINE_SIZE*PRE_LOAD_OFFSET)]
vld4.8 {d0,d1,d2,d3}, [r5]! @loadRGB interleavely
vld4.8 {d4,d5,d6,d7}, [r5]! @loadRGB interleavely
vmov.u16 d8,d2
vmov.u16 d9,d6
vmov.u16 d10,d1
vmov.u16 d11,d5
vmov.u16 d12,d0
vmov.u16 d13,d4
vand.u16 q4,#0x00FF @R
vand.u16 q5,#0x00FF @G
vand.u16 q6,#0x00FF @B
vmul.u16 q7,q4,q14 @q0 = 66 *R[k]
vmla.u16 q7,q5,q15 @q0 += 129 *G[k]
vmla.u16 q7,q6,q8 @q0 += 25 *B[k]
vadd.u16 q7,q7,q10
vshr.u16 q7,q7, #8
vmov.u16 d8,d2
vmov.u16 d9,d6
vmov.u16 d10,d1
vmov.u16 d11,d5
vmov.u16 d12,d0
vmov.u16 d13,d4
vshr.u16 q4,q4,#8 @R
vshr.u16 q5,q5,#8 @G
vshr.u16 q6,q6,#8 @B
vmul.u16 q0,q4,q14 @q0 = 66 *R[k]
vmla.u16 q0,q5,q15 @q0 += 129 *G[k]
vmla.u16 q0,q6,q8 @q0 += 25 *B[k]
vadd.u16 q0,q0,q10
vshr.u16 q0,q0, #8
vtrn.8 q7,q0
vst1.8 {q7}, [r4]!@write to Y to yuv420_buffer
subs r8,r8,#16 @nWidth16--
BPL LOOP_NWIDTH16 @if nWidth16>0
@-----------------------------------unaligned ---------------------------------------
adds r8,r8,#16 @ + 16 - 2
BEQ NO_UNALIGNED @in case that nWidht is multiple of 16
LOOP_NWIDTH2:
@----------------------------------pDstRGB1--Y------------------------------------------
@stmfd sp!, {r14} @backup r14
ldr r9, [r2], #4 @loadRGB int
ldr r12, [r2], #4 @loadRGB int
mov r10, r9,lsr #16 @copy to r10
mov r14, r12 @copy to r10
ldr r6, =0x000000FF
and r10, r10, r6 @R: (rgbIn[k] & 0xF800) >> 10;
ldr r6, =0x00FF0000
and r14, r14, r6 @R: (rgbIn[k] & 0xF800) >> 10;
add r10,r10,r14
mov r11, #66 @accumilator += R*66
mul r7, r10, r11
mov r10, r9,lsr #8 @copy to r10
mov r14, r12,lsl #8 @copy to r10
ldr r6, =0x000000FF
and r10, r10, r6 @G:
ldr r6, =0x00FF0000
and r14, r14, r6 @G:
add r10,r10,r14
mov r11, #129 @accumilator += G *129
mla r7, r10, r11, r7
mov r10, r9 @copy to r10
mov r14, r12,lsl #16 @copy to r10
ldr r6, =0x000000FF
and r10, r10, r6 @B
ldr r6, =0x00FF0000
and r14, r14, r6 @B
add r10,r10,r14
mov r11, #25 @accumilator 1 -= B *25
mla r7, r10, r11, r7
ldr r6, =0x10801080
add r7, r6
lsr r7, #8
strb r7, [r0],#1
lsr r7,#16
strb r7, [r0],#1
@ldmfd sp!, {r14} @load r14
@----------------------------------pDstRGB2--UV------------------------------------------
mov r10, r9 @copy to r10
ldr r7,=0x00008080
mov r12,r7
ldr r6, =0x000000FF
and r10, r10, r6 @B:
mov r11, #112 @accumilator += B*112
mla r7, r10, r11, r7
mov r11, #18 @accumilator -= B*18
mul r11, r10, r11
sub r12, r12, r11
mov r10, r9, lsr #16 @copy to r10
ldr r6, =0x000000FF
and r10, r10, r6 @R: (rgbIn[k] & 0xF800) >> 10;
mov r11, #38 @accumilator -= R *38
mul r11, r10, r11
sub r7, r7, r11
mov r11, #112 @accumilator = R *112
mla r12, r10, r11, r12
mov r10, r9,lsr #8 @copy to r10
ldr r6, =0x000000FF
and r10, r10, r6 @G: (rgbIn[k] & 0x07E0) >> 5;
mov r11, #74 @accumilator -= G*74
mul r11, r10, r11
sub r7, r7, r11
mov r11, #94 @accumilator -= G*94
mul r11, r10, r11
sub r12, r12, r11
lsr r7, #8 @ >>8
strb r7, [r1],#1
lsr r12, #8 @ >>8
strb r12, [r1],#1
@----------------------------------pDstRGB2--Y------------------------------------------
@stmfd sp!, {r14} @backup r14
ldr r9, [r5], #4 @loadRGB int
ldr r12, [r5], #4 @loadRGB int
mov r10, r9,lsr #16 @copy to r10
mov r14, r12 @copy to r10
ldr r6, =0x000000FF
and r10, r10, r6 @R: (rgbIn[k] & 0xF800) >> 10;
ldr r6, =0x00FF0000
and r14, r14, r6 @R: (rgbIn[k] & 0xF800) >> 10;
add r10,r10,r14
mov r11, #66 @accumilator += R*66
mul r7, r10, r11
mov r10, r9,lsr #8 @copy to r10
mov r14, r12,lsl #8 @copy to r10
ldr r6, =0x000000FF
and r10, r10, r6 @G:
ldr r6, =0x00FF0000
and r14, r14, r6 @G:
add r10,r10,r14
mov r11, #129 @accumilator += G *129
mla r7, r10, r11, r7
mov r10, r9 @copy to r10
mov r14, r12,lsl #16 @copy to r10
ldr r6, =0x000000FF
and r10, r10, r6 @B
ldr r6, =0x00FF0000
and r14, r14, r6 @B
add r10,r10,r14
mov r11, #25 @accumilator 1 -= B *25
mla r7, r10, r11, r7
ldr r6, =0x10801080
add r7, r6
lsr r7, #8
strb r7, [r4],#1
lsr r7,#16
strb r7, [r4],#1
@ldmfd sp!, {r14} @load r14
subs r8,r8,#2 @ nWidth2 -= 2
BGT LOOP_NWIDTH2 @ if nWidth2>0
NO_UNALIGNED: @in case that nWidht is multiple of 16
@-----------------------------------------------------------------------------
sub r8, r3, #16 @r8: nWidthTmp = nWidth -16
add r0, r0, r3 @pDstY + nwidth
add r2, r2, r3, lsl #2 @pSrcRGB + nwidthx4
add r4, r4, r3 @pDstY2 + nwidth
add r5, r5, r3, lsl #2 @pSrcRGB2 + nwidthx4
ldmfd sp!, {r12}
subs r12,r12,#2 @nHeight -=2
BGT LOOP_NHEIGHT2 @if nHeight2>0
ldmfd sp!, {r4-r12,pc} @ backup registers
.fnend
|
