Book a Demo!
CoCalc Logo Icon
StoreFeaturesDocsShareSupportNewsAboutPoliciesSign UpSign In
alexbevi
GitHub Repository: alexbevi/BizHawk
 Path: blob/master/libmupen64plus/mupen64plus-video-glide64/src/TexConv.h
2 views
1
/*

2
*   Glide64 - Glide video plugin for Nintendo 64 emulators.

3
*   Copyright (c) 2002  Dave2001

4
*   Copyright (c) 2008  Günther <[email protected]>

5
*

6
*   This program is free software; you can redistribute it and/or modify

7
*   it under the terms of the GNU General Public License as published by

8
*   the Free Software Foundation; either version 2 of the License, or

9
*   any later version.

10
*

11
*   This program is distributed in the hope that it will be useful,

12
*   but WITHOUT ANY WARRANTY; without even the implied warranty of

13
*   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the

14
*   GNU General Public License for more details.

15
*

16
*   You should have received a copy of the GNU General Public

17
*   Licence along with this program; if not, write to the Free

18
*   Software Foundation, Inc., 51 Franklin Street, Fifth Floor, 

19
*   Boston, MA  02110-1301, USA

20
*/

21


22
//****************************************************************

23
//

24
// Glide64 - Glide Plugin for Nintendo 64 emulators (tested mostly with Project64)

25
// Project started on December 29th, 2001

26
//

27
// To modify Glide64:

28
// * Write your name and (optional)email, commented by your work, so I know who did it, and so that you can find which parts you modified when it comes time to send it to me.

29
// * Do NOT send me the whole project or file that you modified.  Take out your modified code sections, and tell me where to put them.  If people sent the whole thing, I would have many different versions, but no idea how to combine them all.

30
//

31
// Official Glide64 development channel: #Glide64 on EFnet

32
//

33
// Original author: Dave2001 ([email protected])

34
// Other authors: Gonetz, Gugaman

35
//

36
//****************************************************************

37


38
void TexConv_ARGB1555_ARGB4444 (unsigned char * _src, unsigned char * _dst, int width, int height)

39
{

40
    int _size = (width * height) << 1;

41
#if !defined(__GNUC__) && !defined(NO_ASM)

42
    __asm {

43
        mov esi,dword ptr [_src]

44
        mov edi,dword ptr [_dst]

45
        mov ecx,dword ptr [_size]

46


47
tc1_loop:

48
        mov eax,dword ptr [esi]

49
        add esi,4

50


51
        // arrr rrgg gggb bbbb

52
        // aaaa rrrr gggg bbbb

53
        mov edx,eax

54
        and eax,0x80008000

55
        mov ebx,eax             // ebx = 0xa000000000000000

56
        shr eax,1

57
        or ebx,eax              // ebx = 0xaa00000000000000

58
        shr eax,1

59
        or ebx,eax              // ebx = 0xaaa0000000000000

60
        shr eax,1

61
        or ebx,eax              // ebx = 0xaaaa000000000000

62


63
        mov eax,edx

64
        and eax,0x78007800      // eax = 0x0rrrr00000000000

65
        shr eax,3               // eax = 0x0000rrrr00000000

66
        or ebx,eax              // ebx = 0xaaaarrrr00000000

67


68
        mov eax,edx

69
        and eax,0x03c003c0      // eax = 0x000000gggg000000

70
        shr eax,2               // eax = 0x00000000gggg0000

71
        or ebx,eax              // ebx = 0xaaaarrrrgggg0000

72


73
        and edx,0x001e001e      // edx = 0x00000000000bbbb0

74
        shr edx,1               // edx = 0x000000000000bbbb

75
        or ebx,edx              // ebx = 0xaaaarrrrggggbbbb

76


77
        mov dword ptr [edi],ebx

78
        add edi,4

79


80
        dec ecx

81
        jnz tc1_loop

82
    }

83
#elif !defined(NO_ASM)

84
   //printf("TexConv_ARGB1555_ARGB4444\n");

85
   asm volatile (

86
         //"tc1_loop2:             \n"

87
         "0: \n"

88
         "mov (%[_src]), %%eax     \n"

89
         "add $4, %[_src]          \n"

90


91
        // arrr rrgg gggb bbbb

92
        // aaaa rrrr gggg bbbb

93
         "mov %%eax, %%edx       \n"

94
         "and $0x80008000, %%eax \n"

95
         "mov %%eax, %%ecx       \n"                // ecx = 0xa000000000000000

96
         "shr $1, %%eax          \n"

97
         "or %%eax, %%ecx        \n"                // ecx = 0xaa00000000000000

98
         "shr $1, %%eax          \n"

99
         "or %%eax, %%ecx        \n"                // ecx = 0xaaa0000000000000

100
         "shr $1, %%eax          \n"

101
         "or %%eax, %%ecx        \n"                // ecx = 0xaaaa000000000000

102


103
         "mov %%edx, %%eax       \n"

104
         "and $0x78007800, %%eax \n"        // eax = 0x0rrrr00000000000

105
         "shr $3, %%eax          \n"                // eax = 0x0000rrrr00000000

106
         "or %%eax, %%ecx        \n"                // ecx = 0xaaaarrrr00000000

107
         

108
         "mov %%edx, %%eax       \n"

109
         "and $0x03c003c0, %%eax \n"        // eax = 0x000000gggg000000

110
         "shr $2, %%eax          \n"                // eax = 0x00000000gggg0000

111
         "or %%eax, %%ecx        \n"                // ecx = 0xaaaarrrrgggg0000

112
         

113
         "and $0x001e001e, %%edx \n"        // edx = 0x00000000000bbbb0

114
         "shr $1, %%edx          \n"                // edx = 0x000000000000bbbb

115
         "or %%edx, %%ecx        \n"                // ecx = 0xaaaarrrrggggbbbb

116
         

117
         "mov %%ecx, (%[_dst])     \n"

118
         "add $4, %[_dst]          \n"

119
         

120
         "decl %[_size]            \n"

121
         "jnz 0b \n"

122
         : [_src]"+S"(_src), [_dst]"+D"(_dst), [_size]"+g"(_size)

123
         :

124
         : "memory", "cc", "eax", "edx", "ecx"

125
         );

126
#endif

127
}

128


129
void TexConv_AI88_ARGB4444 (unsigned char * _src, unsigned char * _dst, int width, int height)

130
{

131
    int _size = (width * height) << 1;

132
#if !defined(__GNUC__) && !defined(NO_ASM)

133
    __asm {

134
        mov esi,dword ptr [_src]

135
        mov edi,dword ptr [_dst]

136
        mov ecx,dword ptr [_size]

137


138
tc1_loop:

139
        mov eax,dword ptr [esi]

140
        add esi,4

141


142
        // aaaa aaaa iiii iiii

143
        // aaaa rrrr gggg bbbb

144
        mov edx,eax

145
        and eax,0xF000F000      // eax = 0xaaaa000000000000

146
        mov ebx,eax             // ebx = 0xaaaa000000000000

147


148
        and edx,0x00F000F0      // edx = 0x00000000iiii0000

149
        shl edx,4               // edx = 0x0000iiii00000000

150
        or ebx,edx              // ebx = 0xaaaaiiii00000000

151
        shr edx,4               // edx = 0x00000000iiii0000

152
        or ebx,edx              // ebx = 0xaaaaiiiiiiii0000

153
        shr edx,4               // edx = 0x000000000000iiii

154
        or ebx,edx              // ebx = 0xaaaaiiiiiiiiiiii

155


156
        mov dword ptr [edi],ebx

157
        add edi,4

158


159
        dec ecx

160
        jnz tc1_loop

161
    }

162
#elif !defined(NO_ASM)

163
   //printf("TexConv_AI88_ARGB4444\n");

164
   asm volatile (

165
         //"tc1_loop3:              \n"

166
         "0: \n"

167
         "mov (%[_src]), %%eax     \n"

168
         "add $4, %[_src]          \n"

169
         

170
         // aaaa aaaa iiii iiii

171
         // aaaa rrrr gggg bbbb

172
         "mov %%eax, %%edx       \n"

173
         "and $0xF000F000, %%eax \n"        // eax = 0xaaaa000000000000

174
         "mov %%eax, %%ecx       \n"                // ecx = 0xaaaa000000000000

175
         

176
         "and $0x00F000F0, %%edx \n"        // edx = 0x00000000iiii0000

177
         "shl $4, %%edx          \n"                // edx = 0x0000iiii00000000

178
         "or %%edx, %%ecx        \n"                // ecx = 0xaaaaiiii00000000

179
         "shr $4, %%edx          \n"                // edx = 0x00000000iiii0000

180
         "or %%edx, %%ecx        \n"                // ecx = 0xaaaaiiiiiiii0000

181
         "shr $4, %%edx          \n"                // edx = 0x000000000000iiii

182
         "or %%edx, %%ecx        \n"                // ecx = 0xaaaaiiiiiiiiiiii

183
         

184
         "mov %%ecx, (%[_dst])     \n"

185
         "add $4, %[_dst]          \n"

186
         

187
         "decl %[_size]            \n"

188
         "jnz 0b \n"

189
         : [_src]"+S"(_src), [_dst]"+D"(_dst), [_size]"+g"(_size)

190
         :

191
         : "memory", "cc", "eax", "edx", "ecx"

192
         );

193
#endif

194
}

195


196
void TexConv_AI44_ARGB4444 (unsigned char * _src, unsigned char * _dst, int width, int height)

197
{

198
    int _size = width * height;

199
#if !defined(__GNUC__) &&  !defined(NO_ASM)

200
    __asm {

201
        mov esi,dword ptr [_src]

202
        mov edi,dword ptr [_dst]

203
        mov ecx,dword ptr [_size]

204


205
tc1_loop:

206
        mov eax,dword ptr [esi]

207
        add esi,4

208


209
        // aaaa3 iiii3 aaaa2 iiii2 aaaa1 iiii1 aaaa0 iiii0

210
        // aaaa1 rrrr1 gggg1 bbbb1 aaaa0 rrrr0 gggg0 bbbb0

211
        // aaaa3 rrrr3 gggg3 bbbb3 aaaa2 rrrr2 gggg2 bbbb2

212
        mov edx,eax             // eax = aaaa3 iiii3 aaaa2 iiii2 aaaa1 iiii1 aaaa0 iiii0

213
        shl eax,16              // eax = aaaa1 iiii1 aaaa0 iiii0 0000  0000  0000  0000

214
        and eax,0xFF000000      // eax = aaaa1 iiii1 0000  0000  0000  0000  0000  0000

215
        mov ebx,eax             // ebx = aaaa1 iiii1 0000  0000  0000  0000  0000  0000

216
        and eax,0x0F000000      // eax = 0000  iiii1 0000  0000  0000  0000  0000  0000

217
        shr eax,4               // eax = 0000  0000  iiii1 0000  0000  0000  0000  0000

218
        or ebx,eax              // ebx = aaaa1 iiii1 iiii1 0000  0000  0000  0000  0000

219
        shr eax,4               // eax = 0000  0000  0000  iiii1 0000  0000  0000  0000

220
        or ebx,eax              // ebx = aaaa1 iiii1 iiii1 iiii1 0000  0000  0000  0000

221


222
        mov eax,edx             // eax = aaaa3 iiii3 aaaa2 iiii2 aaaa1 iiii1 aaaa0 iiii0

223
        shl eax,8               // eax = aaaa2 iiii2 aaaa1 iiii1 aaaa0 iiii0 0000  0000

224
        and eax,0x0000FF00      // eax = 0000  0000  0000  0000  aaaa0 iiii0 0000  0000

225
        or ebx,eax              // ebx = aaaa1 iiii1 iiii1 iiii1 aaaa0 iiii0 0000  0000

226
        and eax,0x00000F00      // eax = 0000  0000  0000  0000  0000  iiii0 0000  0000

227
        shr eax,4               // eax = 0000  0000  0000  0000  0000  0000  iiii0 0000

228
        or ebx,eax              // ebx = aaaa1 iiii1 iiii1 iiii1 aaaa0 iiii0 iiii0 0000

229
        shr eax,4               // eax = 0000  0000  0000  0000  0000  0000  0000  iiii0

230
        or ebx,eax              // ebx = aaaa1 iiii1 iiii1 iiii1 aaaa0 iiii0 iiii0 iiii0

231


232
        mov dword ptr [edi],ebx

233
        add edi,4

234


235
        mov eax,edx             // eax = aaaa3 iiii3 aaaa2 iiii2 aaaa1 iiii1 aaaa0 iiii0

236
        and eax,0xFF000000      // eax = aaaa3 iiii3 0000  0000  0000  0000  0000  0000

237
        mov ebx,eax             // ebx = aaaa3 iiii3 0000  0000  0000  0000  0000  0000

238
        and eax,0x0F000000      // eax = 0000  iiii3 0000  0000  0000  0000  0000  0000

239
        shr eax,4               // eax = 0000  0000  iiii3 0000  0000  0000  0000  0000

240
        or ebx,eax              // ebx = aaaa3 iiii3 iiii3 0000  0000  0000  0000  0000

241
        shr eax,4               // eax = 0000  0000  0000  iiii3 0000  0000  0000  0000

242
        or ebx,eax              // ebx = aaaa3 iiii3 iiii3 iiii3 0000  0000  0000  0000

243


244
                                // edx = aaaa3 iiii3 aaaa2 iiii2 aaaa1 iiii1 aaaa0 iiii0

245
        shr edx,8               // edx = 0000  0000  aaaa3 aaaa3 aaaa2 iiii2 aaaa1 iiii1

246
        and edx,0x0000FF00      // edx = 0000  0000  0000  0000  aaaa2 iiii2 0000  0000

247
        or ebx,edx              // ebx = aaaa3 iiii3 iiii3 iiii3 aaaa2 iiii2 0000  0000

248
        and edx,0x00000F00      // edx = 0000  0000  0000  0000  0000  iiii2 0000  0000

249
        shr edx,4               // edx = 0000  0000  0000  0000  0000  0000  iiii2 0000

250
        or ebx,edx              // ebx = aaaa3 iiii3 iiii3 iiii3 aaaa2 iiii2 iiii2 0000

251
        shr edx,4               // edx = 0000  0000  0000  0000  0000  0000  0000  iiii2

252
        or ebx,edx              // ebx = aaaa3 iiii3 iiii3 iiii3 aaaa2 iiii2 iiii2 iiii2

253


254
        mov dword ptr [edi],ebx

255
        add edi,4

256


257
        dec ecx

258
        jnz tc1_loop

259
    }

260
#elif !defined(NO_ASM)

261
   //printf("TexConv_AI44_ARGB4444\n");

262
   asm volatile (

263
         //"tc1_loop4:             \n"

264
         "0: \n"

265
         "mov (%[_src]), %%eax     \n"

266
         "add $4, %[_src]          \n"

267
         

268
         // aaaa3 iiii3 aaaa2 iiii2 aaaa1 iiii1 aaaa0 iiii0

269
         // aaaa1 rrrr1 gggg1 bbbb1 aaaa0 rrrr0 gggg0 bbbb0

270
         // aaaa3 rrrr3 gggg3 bbbb3 aaaa2 rrrr2 gggg2 bbbb2

271
         "mov %%eax, %%edx       \n"                // eax = aaaa3 iiii3 aaaa2 iiii2 aaaa1 iiii1 aaaa0 iiii0

272
         "shl $16, %%eax         \n"                // eax = aaaa1 iiii1 aaaa0 iiii0 0000  0000  0000  0000

273
         "and $0xFF000000, %%eax \n"        // eax = aaaa1 iiii1 0000  0000  0000  0000  0000  0000

274
         "mov %%eax, %%ecx       \n"                // ecx = aaaa1 iiii1 0000  0000  0000  0000  0000  0000

275
         "and $0x0F000000, %%eax \n"        // eax = 0000  iiii1 0000  0000  0000  0000  0000  0000

276
         "shr $4, %%eax          \n"                // eax = 0000  0000  iiii1 0000  0000  0000  0000  0000

277
         "or %%eax, %%ecx        \n"                // ecx = aaaa1 iiii1 iiii1 0000  0000  0000  0000  0000

278
         "shr $4, %%eax          \n"                // eax = 0000  0000  0000  iiii1 0000  0000  0000  0000

279
         "or %%eax, %%ecx        \n"                // ecx = aaaa1 iiii1 iiii1 iiii1 0000  0000  0000  0000

280
         

281
         "mov %%edx, %%eax       \n"                // eax = aaaa3 iiii3 aaaa2 iiii2 aaaa1 iiii1 aaaa0 iiii0

282
         "shl $8, %%eax          \n"                // eax = aaaa2 iiii2 aaaa1 iiii1 aaaa0 iiii0 0000  0000

283
         "and $0x0000FF00, %%eax \n"        // eax = 0000  0000  0000  0000  aaaa0 iiii0 0000  0000

284
         "or %%eax, %%ecx        \n"                // ecx = aaaa1 iiii1 iiii1 iiii1 aaaa0 iiii0 0000  0000

285
         "and $0x00000F00, %%eax \n"        // eax = 0000  0000  0000  0000  0000  iiii0 0000  0000

286
         "shr $4, %%eax          \n"                // eax = 0000  0000  0000  0000  0000  0000  iiii0 0000

287
         "or %%eax, %%ecx        \n"                // ecx = aaaa1 iiii1 iiii1 iiii1 aaaa0 iiii0 iiii0 0000

288
         "shr $4, %%eax          \n"                // eax = 0000  0000  0000  0000  0000  0000  0000  iiii0

289
         "or %%eax, %%ecx        \n"                // ecx = aaaa1 iiii1 iiii1 iiii1 aaaa0 iiii0 iiii0 iiii0

290
         

291
         "mov %%ecx, (%[_dst])     \n"

292
         "add $4, %[_dst]          \n"

293
         

294
         "mov %%edx, %%eax       \n"                // eax = aaaa3 iiii3 aaaa2 iiii2 aaaa1 iiii1 aaaa0 iiii0

295
         "and $0xFF000000, %%eax \n"        // eax = aaaa3 iiii3 0000  0000  0000  0000  0000  0000

296
         "mov %%eax, %%ecx       \n"                // ecx = aaaa3 iiii3 0000  0000  0000  0000  0000  0000

297
         "and $0x0F000000, %%eax \n"        // eax = 0000  iiii3 0000  0000  0000  0000  0000  0000

298
         "shr $4, %%eax          \n"                // eax = 0000  0000  iiii3 0000  0000  0000  0000  0000

299
         "or %%eax, %%ecx        \n"                // ecx = aaaa3 iiii3 iiii3 0000  0000  0000  0000  0000

300
         "shr $4, %%eax          \n"                // eax = 0000  0000  0000  iiii3 0000  0000  0000  0000

301
         "or %%eax, %%ecx        \n"                // ecx = aaaa3 iiii3 iiii3 iiii3 0000  0000  0000  0000

302
         

303
                                                // edx = aaaa3 iiii3 aaaa2 iiii2 aaaa1 iiii1 aaaa0 iiii0

304
         "shr $8, %%edx          \n"                // edx = 0000  0000  aaaa3 aaaa3 aaaa2 iiii2 aaaa1 iiii1

305
         "and $0x0000FF00, %%edx \n"        // edx = 0000  0000  0000  0000  aaaa2 iiii2 0000  0000

306
         "or %%edx, %%ecx        \n"                // ecx = aaaa3 iiii3 iiii3 iiii3 aaaa2 iiii2 0000  0000

307
         "and $0x00000F00, %%edx \n"        // edx = 0000  0000  0000  0000  0000  iiii2 0000  0000

308
         "shr $4, %%edx          \n"                // edx = 0000  0000  0000  0000  0000  0000  iiii2 0000

309
         "or %%edx, %%ecx        \n"                // ecx = aaaa3 iiii3 iiii3 iiii3 aaaa2 iiii2 iiii2 0000

310
         "shr $4, %%edx          \n"                // edx = 0000  0000  0000  0000  0000  0000  0000  iiii2

311
         "or %%edx, %%ecx        \n"                // ecx = aaaa3 iiii3 iiii3 iiii3 aaaa2 iiii2 iiii2 iiii2

312
         

313
         "mov %%ecx, (%[_dst])     \n"

314
         "add $4, %[_dst]          \n"

315
         

316
         "decl %[_size]            \n"

317
         "jnz 0b \n"

318
         : [_src]"+S"(_src), [_dst]"+D"(_dst), [_size]"+g"(_size)

319
         :

320
         : "memory", "cc", "eax", "edx", "ecx"

321
         );

322
#endif

323
}

324


325
void TexConv_A8_ARGB4444 (unsigned char * _src, unsigned char * _dst, int width, int height)

326
{

327
    int _size = (width * height) << 1;

328
#if !defined(__GNUC__) && !defined(NO_ASM)

329
    __asm {

330
        mov esi,dword ptr [_src]

331
        mov edi,dword ptr [_dst]

332
        mov ecx,dword ptr [_size]

333


334
tc1_loop:

335
        mov eax,dword ptr [esi]

336
        add esi,4

337


338
        // aaaa3 aaaa3 aaaa2 aaaa2 aaaa1 aaaa1 aaaa0 aaaa0

339
        // aaaa1 rrrr1 gggg1 bbbb1 aaaa0 rrrr0 gggg0 bbbb0

340
        // aaaa3 rrrr3 gggg3 bbbb3 aaaa2 rrrr2 gggg2 bbbb2

341
        mov edx,eax

342
        and eax,0x0000F000      // eax = 00 00 00 00 a1 00 00 00

343
        shl eax,16              // eax = a1 00 00 00 00 00 00 00

344
        mov ebx,eax             // ebx = a1 00 00 00 00 00 00 00

345
        shr eax,4

346
        or ebx,eax              // ebx = a1 a1 00 00 00 00 00 00

347
        shr eax,4

348
        or ebx,eax              // ebx = a1 a1 a1 00 00 00 00 00

349
        shr eax,4

350
        or ebx,eax              // ebx = a1 a1 a1 a1 00 00 00 00

351


352
        mov eax,edx

353
        and eax,0x000000F0      // eax = 00 00 00 00 00 00 a0 00

354
        shl eax,8               // eax = 00 00 00 00 a0 00 00 00

355
        or ebx,eax

356
        shr eax,4

357
        or ebx,eax

358
        shr eax,4

359
        or ebx,eax

360
        shr eax,4

361
        or ebx,eax              // ebx = a1 a1 a1 a1 a0 a0 a0 a0

362


363
        mov dword ptr [edi],ebx

364
        add edi,4

365


366
        mov eax,edx             // eax = a3 a3 a2 a2 a1 a1 a0 a0

367
        and eax,0xF0000000      // eax = a3 00 00 00 00 00 00 00

368
        mov ebx,eax             // ebx = a3 00 00 00 00 00 00 00

369
        shr eax,4

370
        or ebx,eax              // ebx = a3 a3 00 00 00 00 00 00

371
        shr eax,4

372
        or ebx,eax              // ebx = a3 a3 a3 00 00 00 00 00

373
        shr eax,4

374
        or ebx,eax              // ebx = a3 a3 a3 a3 00 00 00 00

375


376
        and edx,0x00F00000      // eax = 00 00 a2 00 00 00 00 00

377
        shr edx,8               // eax = 00 00 00 00 a2 00 00 00

378
        or ebx,edx

379
        shr edx,4

380
        or ebx,edx

381
        shr edx,4

382
        or ebx,edx

383
        shr edx,4

384
        or ebx,edx              // ebx = a3 a3 a3 a3 a2 a2 a2 a2

385


386
        mov dword ptr [edi],ebx

387
        add edi,4

388


389
        dec ecx

390
        jnz tc1_loop

391
    }

392
#elif !defined(NO_ASM)

393
   //printf("TexConv_A8_ARGB4444\n");

394
   asm volatile (

395
         //"tc1_loop:              \n"

396
         "0: \n"

397
         "mov (%[src]), %%eax     \n"

398
         "add $4, %[src]          \n"

399
         

400
         // aaaa3 aaaa3 aaaa2 aaaa2 aaaa1 aaaa1 aaaa0 aaaa0

401
         // aaaa1 rrrr1 gggg1 bbbb1 aaaa0 rrrr0 gggg0 bbbb0

402
         // aaaa3 rrrr3 gggg3 bbbb3 aaaa2 rrrr2 gggg2 bbbb2

403
         "mov %%eax, %%edx       \n"

404
         "and $0x0000F000, %%eax \n"        // eax = 00 00 00 00 a1 00 00 00

405
         "shl $16, %%eax         \n"                // eax = a1 00 00 00 00 00 00 00

406
         "mov %%eax, %%ecx       \n"                // ecx = a1 00 00 00 00 00 00 00

407
         "shr $4, %%eax          \n"

408
         "or %%eax, %%ecx        \n"                // ecx = a1 a1 00 00 00 00 00 00

409
         "shr $4, %%eax          \n"

410
         "or %%eax, %%ecx        \n"                // ecx = a1 a1 a1 00 00 00 00 00

411
         "shr $4, %%eax          \n"

412
         "or %%eax, %%ecx        \n"                // ecx = a1 a1 a1 a1 00 00 00 00

413
         

414
         "mov %%edx, %%eax       \n"

415
         "and $0x000000F0, %%eax \n"        // eax = 00 00 00 00 00 00 a0 00

416
         "shl $8, %%eax          \n"                // eax = 00 00 00 00 a0 00 00 00

417
         "or %%eax, %%ecx        \n"

418
         "shr $4, %%eax          \n"

419
         "or %%eax, %%ecx        \n"

420
         "shr $4, %%eax          \n"

421
         "or %%eax, %%ecx        \n"

422
         "shr $4, %%eax          \n"

423
         "or %%eax, %%ecx        \n"                // ecx = a1 a1 a1 a1 a0 a0 a0 a0

424
         

425
         "mov %%ecx, (%[_dst])     \n"

426
         "add $4, %[_dst]          \n"

427


428
         "mov %%edx, %%eax       \n"                // eax = a3 a3 a2 a2 a1 a1 a0 a0

429
         "and $0xF0000000, %%eax \n"        // eax = a3 00 00 00 00 00 00 00

430
         "mov %%eax, %%ecx       \n"                // ecx = a3 00 00 00 00 00 00 00

431
         "shr $4, %%eax          \n"

432
         "or %%eax, %%ecx        \n"                // ecx = a3 a3 00 00 00 00 00 00

433
         "shr $4, %%eax          \n"

434
         "or %%eax, %%ecx        \n"                // ecx = a3 a3 a3 00 00 00 00 00

435
         "shr $4, %%eax          \n"

436
         "or %%eax, %%ecx        \n"                // ecx = a3 a3 a3 a3 00 00 00 00

437
         

438
         "and $0x00F00000, %%edx \n"        // eax = 00 00 a2 00 00 00 00 00

439
         "shr $8, %%edx          \n"                // eax = 00 00 00 00 a2 00 00 00

440
         "or %%edx, %%ecx        \n"

441
         "shr $4, %%edx          \n"

442
         "or %%edx, %%ecx        \n"

443
         "shr $4, %%edx          \n"

444
         "or %%edx, %%ecx        \n"

445
         "shr $4, %%edx          \n"

446
         "or %%edx, %%ecx        \n"                // ecx = a3 a3 a3 a3 a2 a2 a2 a2

447
         

448
         "mov %%ecx, (%[_dst])     \n"

449
         "add $4, %[_dst]          \n"

450
         

451
         "decl %[_size]            \n"

452
         "jnz 0b \n"

453
         : [src]"+S"(_src), [_dst]"+D"(_dst), [_size]"+g"(_size)

454
         :

455
         : "memory", "cc", "eax", "ecx", "edx"

456
         );

457
#endif

458
}

459


460


461