FFmpeg coverage

Line	Branch	Exec	Source
1			/*
2			* Copyright (C) 2001-2011 Michael Niedermayer <michaelni@gmx.at>
3			*
4			* This file is part of FFmpeg.
5			*
6			* FFmpeg is free software; you can redistribute it and/or
7			* modify it under the terms of the GNU Lesser General Public
8			* License as published by the Free Software Foundation; either
9			* version 2.1 of the License, or (at your option) any later version.
10			*
11			* FFmpeg 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 GNU
14			* Lesser General Public License for more details.
15			*
16			* You should have received a copy of the GNU Lesser General Public
17			* License along with FFmpeg; if not, write to the Free Software
18			* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
19			*/
20
21			#include <stdint.h>
22
23			#include "libavutil/x86/asm.h"
24			#include "libswscale/swscale_internal.h"
25
26			#undef REAL_MOVNTQ
27			#undef MOVNTQ
28
29
30			#define REAL_MOVNTQ(a,b) "movntq " #a ", " #b " \n\t"
31			#define MOVNTQ(a,b) REAL_MOVNTQ(a,b)
32
33			#define YSCALEYUV2PACKEDX_UV \
34			__asm__ volatile(\
35			"xor %%"FF_REG_a", %%"FF_REG_a" \n\t"\
36			".p2align 4 \n\t"\
37			"nop \n\t"\
38			"1: \n\t"\
39			"lea "CHR_MMX_FILTER_OFFSET"(%0), %%"FF_REG_d" \n\t"\
40			"mov (%%"FF_REG_d"), %%"FF_REG_S" \n\t"\
41			"movq "VROUNDER_OFFSET"(%0), %%mm3 \n\t"\
42			"movq %%mm3, %%mm4 \n\t"\
43			".p2align 4 \n\t"\
44			"2: \n\t"\
45			"movq 8(%%"FF_REG_d"), %%mm0 \n\t" /* filterCoeff */\
46			"movq (%%"FF_REG_S", %%"FF_REG_a"), %%mm2 \n\t" /* UsrcData */\
47			"add %6, %%"FF_REG_S" \n\t" \
48			"movq (%%"FF_REG_S", %%"FF_REG_a"), %%mm5 \n\t" /* VsrcData */\
49			"add $16, %%"FF_REG_d" \n\t"\
50			"mov (%%"FF_REG_d"), %%"FF_REG_S" \n\t"\
51			"pmulhw %%mm0, %%mm2 \n\t"\
52			"pmulhw %%mm0, %%mm5 \n\t"\
53			"paddw %%mm2, %%mm3 \n\t"\
54			"paddw %%mm5, %%mm4 \n\t"\
55			"test %%"FF_REG_S", %%"FF_REG_S" \n\t"\
56			" jnz 2b \n\t"\
57
58			#define YSCALEYUV2PACKEDX_YA(offset,coeff,src1,src2,dst1,dst2) \
59			"lea "offset"(%0), %%"FF_REG_d" \n\t"\
60			"mov (%%"FF_REG_d"), %%"FF_REG_S" \n\t"\
61			"movq "VROUNDER_OFFSET"(%0), "#dst1" \n\t"\
62			"movq "#dst1", "#dst2" \n\t"\
63			".p2align 4 \n\t"\
64			"2: \n\t"\
65			"movq 8(%%"FF_REG_d"), "#coeff" \n\t" /* filterCoeff */\
66			"movq (%%"FF_REG_S", %%"FF_REG_a", 2), "#src1" \n\t" /* Y1srcData */\
67			"movq 8(%%"FF_REG_S", %%"FF_REG_a", 2), "#src2" \n\t" /* Y2srcData */\
68			"add $16, %%"FF_REG_d" \n\t"\
69			"mov (%%"FF_REG_d"), %%"FF_REG_S" \n\t"\
70			"pmulhw "#coeff", "#src1" \n\t"\
71			"pmulhw "#coeff", "#src2" \n\t"\
72			"paddw "#src1", "#dst1" \n\t"\
73			"paddw "#src2", "#dst2" \n\t"\
74			"test %%"FF_REG_S", %%"FF_REG_S" \n\t"\
75			" jnz 2b \n\t"\
76
77			#define YSCALEYUV2PACKEDX \
78			YSCALEYUV2PACKEDX_UV \
79			YSCALEYUV2PACKEDX_YA(LUM_MMX_FILTER_OFFSET,%%mm0,%%mm2,%%mm5,%%mm1,%%mm7) \
80
81			#define YSCALEYUV2PACKEDX_END \
82			:: "r" (&c->redDither), \
83			"m" (dummy), "m" (dummy), "m" (dummy),\
84			"r" (dest), "m" (dstW_reg), "m"(uv_off) \
85			NAMED_CONSTRAINTS_ADD(bF8,bFC) \
86			: "%"FF_REG_a, "%"FF_REG_d, "%"FF_REG_S \
87			);
88
89			#define YSCALEYUV2PACKEDX_ACCURATE_UV \
90			__asm__ volatile(\
91			"xor %%"FF_REG_a", %%"FF_REG_a" \n\t"\
92			".p2align 4 \n\t"\
93			"nop \n\t"\
94			"1: \n\t"\
95			"lea "CHR_MMX_FILTER_OFFSET"(%0), %%"FF_REG_d" \n\t"\
96			"mov (%%"FF_REG_d"), %%"FF_REG_S" \n\t"\
97			"pxor %%mm4, %%mm4 \n\t"\
98			"pxor %%mm5, %%mm5 \n\t"\
99			"pxor %%mm6, %%mm6 \n\t"\
100			"pxor %%mm7, %%mm7 \n\t"\
101			".p2align 4 \n\t"\
102			"2: \n\t"\
103			"movq (%%"FF_REG_S", %%"FF_REG_a"), %%mm0 \n\t" /* UsrcData */\
104			"add %6, %%"FF_REG_S" \n\t" \
105			"movq (%%"FF_REG_S", %%"FF_REG_a"), %%mm2 \n\t" /* VsrcData */\
106			"mov "STR(APCK_PTR2)"(%%"FF_REG_d"), %%"FF_REG_S" \n\t"\
107			"movq (%%"FF_REG_S", %%"FF_REG_a"), %%mm1 \n\t" /* UsrcData */\
108			"movq %%mm0, %%mm3 \n\t"\
109			"punpcklwd %%mm1, %%mm0 \n\t"\
110			"punpckhwd %%mm1, %%mm3 \n\t"\
111			"movq "STR(APCK_COEF)"(%%"FF_REG_d"),%%mm1 \n\t" /* filterCoeff */\
112			"pmaddwd %%mm1, %%mm0 \n\t"\
113			"pmaddwd %%mm1, %%mm3 \n\t"\
114			"paddd %%mm0, %%mm4 \n\t"\
115			"paddd %%mm3, %%mm5 \n\t"\
116			"add %6, %%"FF_REG_S" \n\t" \
117			"movq (%%"FF_REG_S", %%"FF_REG_a"), %%mm3 \n\t" /* VsrcData */\
118			"mov "STR(APCK_SIZE)"(%%"FF_REG_d"), %%"FF_REG_S" \n\t"\
119			"add $"STR(APCK_SIZE)", %%"FF_REG_d" \n\t"\
120			"test %%"FF_REG_S", %%"FF_REG_S" \n\t"\
121			"movq %%mm2, %%mm0 \n\t"\
122			"punpcklwd %%mm3, %%mm2 \n\t"\
123			"punpckhwd %%mm3, %%mm0 \n\t"\
124			"pmaddwd %%mm1, %%mm2 \n\t"\
125			"pmaddwd %%mm1, %%mm0 \n\t"\
126			"paddd %%mm2, %%mm6 \n\t"\
127			"paddd %%mm0, %%mm7 \n\t"\
128			" jnz 2b \n\t"\
129			"psrad $16, %%mm4 \n\t"\
130			"psrad $16, %%mm5 \n\t"\
131			"psrad $16, %%mm6 \n\t"\
132			"psrad $16, %%mm7 \n\t"\
133			"movq "VROUNDER_OFFSET"(%0), %%mm0 \n\t"\
134			"packssdw %%mm5, %%mm4 \n\t"\
135			"packssdw %%mm7, %%mm6 \n\t"\
136			"paddw %%mm0, %%mm4 \n\t"\
137			"paddw %%mm0, %%mm6 \n\t"\
138			"movq %%mm4, "U_TEMP"(%0) \n\t"\
139			"movq %%mm6, "V_TEMP"(%0) \n\t"\
140
141			#define YSCALEYUV2PACKEDX_ACCURATE_YA(offset) \
142			"lea "offset"(%0), %%"FF_REG_d" \n\t"\
143			"mov (%%"FF_REG_d"), %%"FF_REG_S" \n\t"\
144			"pxor %%mm1, %%mm1 \n\t"\
145			"pxor %%mm5, %%mm5 \n\t"\
146			"pxor %%mm7, %%mm7 \n\t"\
147			"pxor %%mm6, %%mm6 \n\t"\
148			".p2align 4 \n\t"\
149			"2: \n\t"\
150			"movq (%%"FF_REG_S", %%"FF_REG_a", 2), %%mm0 \n\t" /* Y1srcData */\
151			"movq 8(%%"FF_REG_S", %%"FF_REG_a", 2), %%mm2 \n\t" /* Y2srcData */\
152			"mov "STR(APCK_PTR2)"(%%"FF_REG_d"), %%"FF_REG_S" \n\t"\
153			"movq (%%"FF_REG_S", %%"FF_REG_a", 2), %%mm4 \n\t" /* Y1srcData */\
154			"movq %%mm0, %%mm3 \n\t"\
155			"punpcklwd %%mm4, %%mm0 \n\t"\
156			"punpckhwd %%mm4, %%mm3 \n\t"\
157			"movq "STR(APCK_COEF)"(%%"FF_REG_d"), %%mm4 \n\t" /* filterCoeff */\
158			"pmaddwd %%mm4, %%mm0 \n\t"\
159			"pmaddwd %%mm4, %%mm3 \n\t"\
160			"paddd %%mm0, %%mm1 \n\t"\
161			"paddd %%mm3, %%mm5 \n\t"\
162			"movq 8(%%"FF_REG_S", %%"FF_REG_a", 2), %%mm3 \n\t" /* Y2srcData */\
163			"mov "STR(APCK_SIZE)"(%%"FF_REG_d"), %%"FF_REG_S" \n\t"\
164			"add $"STR(APCK_SIZE)", %%"FF_REG_d" \n\t"\
165			"test %%"FF_REG_S", %%"FF_REG_S" \n\t"\
166			"movq %%mm2, %%mm0 \n\t"\
167			"punpcklwd %%mm3, %%mm2 \n\t"\
168			"punpckhwd %%mm3, %%mm0 \n\t"\
169			"pmaddwd %%mm4, %%mm2 \n\t"\
170			"pmaddwd %%mm4, %%mm0 \n\t"\
171			"paddd %%mm2, %%mm7 \n\t"\
172			"paddd %%mm0, %%mm6 \n\t"\
173			" jnz 2b \n\t"\
174			"psrad $16, %%mm1 \n\t"\
175			"psrad $16, %%mm5 \n\t"\
176			"psrad $16, %%mm7 \n\t"\
177			"psrad $16, %%mm6 \n\t"\
178			"movq "VROUNDER_OFFSET"(%0), %%mm0 \n\t"\
179			"packssdw %%mm5, %%mm1 \n\t"\
180			"packssdw %%mm6, %%mm7 \n\t"\
181			"paddw %%mm0, %%mm1 \n\t"\
182			"paddw %%mm0, %%mm7 \n\t"\
183			"movq "U_TEMP"(%0), %%mm3 \n\t"\
184			"movq "V_TEMP"(%0), %%mm4 \n\t"\
185
186			#define YSCALEYUV2PACKEDX_ACCURATE \
187			YSCALEYUV2PACKEDX_ACCURATE_UV \
188			YSCALEYUV2PACKEDX_ACCURATE_YA(LUM_MMX_FILTER_OFFSET)
189
190			#define YSCALEYUV2RGBX \
191			"psubw "U_OFFSET"(%0), %%mm3 \n\t" /* (U-128)8*/\
192			"psubw "V_OFFSET"(%0), %%mm4 \n\t" /* (V-128)8*/\
193			"movq %%mm3, %%mm2 \n\t" /* (U-128)8*/\
194			"movq %%mm4, %%mm5 \n\t" /* (V-128)8*/\
195			"pmulhw "UG_COEFF"(%0), %%mm3 \n\t"\
196			"pmulhw "VG_COEFF"(%0), %%mm4 \n\t"\
197			/* mm2=(U-128)8, mm3=ug, mm4=vg mm5=(V-128)8 */\
198			"pmulhw "UB_COEFF"(%0), %%mm2 \n\t"\
199			"pmulhw "VR_COEFF"(%0), %%mm5 \n\t"\
200			"psubw "Y_OFFSET"(%0), %%mm1 \n\t" /* 8(Y-16)*/\
201			"psubw "Y_OFFSET"(%0), %%mm7 \n\t" /* 8(Y-16)*/\
202			"pmulhw "Y_COEFF"(%0), %%mm1 \n\t"\
203			"pmulhw "Y_COEFF"(%0), %%mm7 \n\t"\
204			/* mm1= Y1, mm2=ub, mm3=ug, mm4=vg mm5=vr, mm7=Y2 */\
205			"paddw %%mm3, %%mm4 \n\t"\
206			"movq %%mm2, %%mm0 \n\t"\
207			"movq %%mm5, %%mm6 \n\t"\
208			"movq %%mm4, %%mm3 \n\t"\
209			"punpcklwd %%mm2, %%mm2 \n\t"\
210			"punpcklwd %%mm5, %%mm5 \n\t"\
211			"punpcklwd %%mm4, %%mm4 \n\t"\
212			"paddw %%mm1, %%mm2 \n\t"\
213			"paddw %%mm1, %%mm5 \n\t"\
214			"paddw %%mm1, %%mm4 \n\t"\
215			"punpckhwd %%mm0, %%mm0 \n\t"\
216			"punpckhwd %%mm6, %%mm6 \n\t"\
217			"punpckhwd %%mm3, %%mm3 \n\t"\
218			"paddw %%mm7, %%mm0 \n\t"\
219			"paddw %%mm7, %%mm6 \n\t"\
220			"paddw %%mm7, %%mm3 \n\t"\
221			/* mm0=B1, mm2=B2, mm3=G2, mm4=G1, mm5=R1, mm6=R2 */\
222			"packuswb %%mm0, %%mm2 \n\t"\
223			"packuswb %%mm6, %%mm5 \n\t"\
224			"packuswb %%mm3, %%mm4 \n\t"\
225
226			#define REAL_WRITEBGR32(dst, dstw, index, b, g, r, a, q0, q2, q3, t) \
227			"movq "#b", "#q2" \n\t" /* B */\
228			"movq "#r", "#t" \n\t" /* R */\
229			"punpcklbw "#g", "#b" \n\t" /* GBGBGBGB 0 */\
230			"punpcklbw "#a", "#r" \n\t" /* ARARARAR 0 */\
231			"punpckhbw "#g", "#q2" \n\t" /* GBGBGBGB 2 */\
232			"punpckhbw "#a", "#t" \n\t" /* ARARARAR 2 */\
233			"movq "#b", "#q0" \n\t" /* GBGBGBGB 0 */\
234			"movq "#q2", "#q3" \n\t" /* GBGBGBGB 2 */\
235			"punpcklwd "#r", "#q0" \n\t" /* ARGBARGB 0 */\
236			"punpckhwd "#r", "#b" \n\t" /* ARGBARGB 1 */\
237			"punpcklwd "#t", "#q2" \n\t" /* ARGBARGB 2 */\
238			"punpckhwd "#t", "#q3" \n\t" /* ARGBARGB 3 */\
239			\
240			MOVNTQ( q0, (dst, index, 4))\
241			MOVNTQ( b, 8(dst, index, 4))\
242			MOVNTQ( q2, 16(dst, index, 4))\
243			MOVNTQ( q3, 24(dst, index, 4))\
244			\
245			"add $8, "#index" \n\t"\
246			"cmp "dstw", "#index" \n\t"\
247			" jb 1b \n\t"
248			#define WRITEBGR32(dst, dstw, index, b, g, r, a, q0, q2, q3, t) REAL_WRITEBGR32(dst, dstw, index, b, g, r, a, q0, q2, q3, t)
249
250		✗	static void RENAME(yuv2rgb32_X_ar)(SwsInternal *c, const int16_t *lumFilter,
251			const int16_t **lumSrc, int lumFilterSize,
252			const int16_t *chrFilter, const int16_t **chrUSrc,
253			const int16_t **chrVSrc,
254			int chrFilterSize, const int16_t **alpSrc,
255			uint8_t *dest, int dstW, int dstY)
256			{
257		✗	x86_reg dummy=0;
258		✗	x86_reg dstW_reg = dstW;
259		✗	x86_reg uv_off = c->uv_offx2;
260
261		✗	if (CONFIG_SWSCALE_ALPHA && c->needAlpha) {
262		✗	YSCALEYUV2PACKEDX_ACCURATE
263			YSCALEYUV2RGBX
264			"movq %%mm2, "U_TEMP"(%0) \n\t"
265			"movq %%mm4, "V_TEMP"(%0) \n\t"
266			"movq %%mm5, "Y_TEMP"(%0) \n\t"
267			YSCALEYUV2PACKEDX_ACCURATE_YA(ALP_MMX_FILTER_OFFSET)
268			"movq "Y_TEMP"(%0), %%mm5 \n\t"
269			"psraw $3, %%mm1 \n\t"
270			"psraw $3, %%mm7 \n\t"
271			"packuswb %%mm7, %%mm1 \n\t"
272			WRITEBGR32(%4, "%5", %%FF_REGa, %%mm3, %%mm4, %%mm5, %%mm1, %%mm0, %%mm7, %%mm2, %%mm6)
273		✗	YSCALEYUV2PACKEDX_END
274			} else {
275		✗	YSCALEYUV2PACKEDX_ACCURATE
276			YSCALEYUV2RGBX
277			"pcmpeqd %%mm7, %%mm7 \n\t"
278			WRITEBGR32(%4, "%5", %%FF_REGa, %%mm2, %%mm4, %%mm5, %%mm7, %%mm0, %%mm1, %%mm3, %%mm6)
279		✗	YSCALEYUV2PACKEDX_END
280			}
281		✗	}
282
283		✗	static void RENAME(yuv2rgb32_X)(SwsInternal *c, const int16_t *lumFilter,
284			const int16_t **lumSrc, int lumFilterSize,
285			const int16_t *chrFilter, const int16_t **chrUSrc,
286			const int16_t **chrVSrc,
287			int chrFilterSize, const int16_t **alpSrc,
288			uint8_t *dest, int dstW, int dstY)
289			{
290		✗	x86_reg dummy=0;
291		✗	x86_reg dstW_reg = dstW;
292		✗	x86_reg uv_off = c->uv_offx2;
293
294		✗	if (CONFIG_SWSCALE_ALPHA && c->needAlpha) {
295		✗	YSCALEYUV2PACKEDX
296			YSCALEYUV2RGBX
297			YSCALEYUV2PACKEDX_YA(ALP_MMX_FILTER_OFFSET, %%mm0, %%mm3, %%mm6, %%mm1, %%mm7)
298			"psraw $3, %%mm1 \n\t"
299			"psraw $3, %%mm7 \n\t"
300			"packuswb %%mm7, %%mm1 \n\t"
301			WRITEBGR32(%4, "%5", %%FF_REGa, %%mm2, %%mm4, %%mm5, %%mm1, %%mm0, %%mm7, %%mm3, %%mm6)
302		✗	YSCALEYUV2PACKEDX_END
303			} else {
304		✗	YSCALEYUV2PACKEDX
305			YSCALEYUV2RGBX
306			"pcmpeqd %%mm7, %%mm7 \n\t"
307			WRITEBGR32(%4, "%5", %%FF_REGa, %%mm2, %%mm4, %%mm5, %%mm7, %%mm0, %%mm1, %%mm3, %%mm6)
308		✗	YSCALEYUV2PACKEDX_END
309			}
310		✗	}
311
312		✗	static void RENAME(yuv2bgr32_X)(SwsInternal *c, const int16_t *lumFilter,
313			const int16_t **lumSrc, int lumFilterSize,
314			const int16_t *chrFilter, const int16_t **chrUSrc,
315			const int16_t **chrVSrc,
316			int chrFilterSize, const int16_t **alpSrc,
317			uint8_t *dest, int dstW, int dstY)
318			{
319		✗	x86_reg dummy=0;
320		✗	x86_reg dstW_reg = dstW;
321		✗	x86_reg uv_off = c->uv_offx2;
322
323		✗	if (CONFIG_SWSCALE_ALPHA && c->needAlpha) {
324		✗	YSCALEYUV2PACKEDX
325			YSCALEYUV2RGBX
326			YSCALEYUV2PACKEDX_YA(ALP_MMX_FILTER_OFFSET, %%mm0, %%mm3, %%mm6, %%mm1, %%mm7)
327			"psraw $3, %%mm1 \n\t"
328			"psraw $3, %%mm7 \n\t"
329			"packuswb %%mm7, %%mm1 \n\t"
330			WRITEBGR32(%4, "%5", %%FF_REGa, %%mm5, %%mm4, %%mm2, %%mm1, %%mm0, %%mm7, %%mm3, %%mm6)
331		✗	YSCALEYUV2PACKEDX_END
332			} else {
333		✗	YSCALEYUV2PACKEDX
334			YSCALEYUV2RGBX
335			"pcmpeqd %%mm7, %%mm7 \n\t"
336			WRITEBGR32(%4, "%5", %%FF_REGa, %%mm5, %%mm4, %%mm2, %%mm7, %%mm0, %%mm1, %%mm3, %%mm6)
337		✗	YSCALEYUV2PACKEDX_END
338			}
339		✗	}
340
341			#define REAL_WRITERGB16(dst, dstw, index) \
342			"pand "MANGLE(bF8)", %%mm2 \n\t" /* B */\
343			"pand "MANGLE(bFC)", %%mm4 \n\t" /* G */\
344			"pand "MANGLE(bF8)", %%mm5 \n\t" /* R */\
345			"psrlq $3, %%mm2 \n\t"\
346			\
347			"movq %%mm2, %%mm1 \n\t"\
348			"movq %%mm4, %%mm3 \n\t"\
349			\
350			"punpcklbw %%mm7, %%mm3 \n\t"\
351			"punpcklbw %%mm5, %%mm2 \n\t"\
352			"punpckhbw %%mm7, %%mm4 \n\t"\
353			"punpckhbw %%mm5, %%mm1 \n\t"\
354			\
355			"psllq $3, %%mm3 \n\t"\
356			"psllq $3, %%mm4 \n\t"\
357			\
358			"por %%mm3, %%mm2 \n\t"\
359			"por %%mm4, %%mm1 \n\t"\
360			\
361			MOVNTQ(%%mm2, (dst, index, 2))\
362			MOVNTQ(%%mm1, 8(dst, index, 2))\
363			\
364			"add $8, "#index" \n\t"\
365			"cmp "dstw", "#index" \n\t"\
366			" jb 1b \n\t"
367			#define WRITERGB16(dst, dstw, index) REAL_WRITERGB16(dst, dstw, index)
368
369		✗	static void RENAME(yuv2rgb565_X_ar)(SwsInternal *c, const int16_t *lumFilter,
370			const int16_t **lumSrc, int lumFilterSize,
371			const int16_t *chrFilter, const int16_t **chrUSrc,
372			const int16_t **chrVSrc,
373			int chrFilterSize, const int16_t **alpSrc,
374			uint8_t *dest, int dstW, int dstY)
375			{
376		✗	x86_reg dummy=0;
377		✗	x86_reg dstW_reg = dstW;
378		✗	x86_reg uv_off = c->uv_offx2;
379
380		✗	YSCALEYUV2PACKEDX_ACCURATE
381			YSCALEYUV2RGBX
382			"pxor %%mm7, %%mm7 \n\t"
383			/* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */
384			"paddusb "BLUE_DITHER"(%0), %%mm2\n\t"
385			"paddusb "GREEN_DITHER"(%0), %%mm4\n\t"
386			"paddusb "RED_DITHER"(%0), %%mm5\n\t"
387			WRITERGB16(%4, "%5", %%FF_REGa)
388		✗	YSCALEYUV2PACKEDX_END
389		✗	}
390
391		✗	static void RENAME(yuv2rgb565_X)(SwsInternal *c, const int16_t *lumFilter,
392			const int16_t **lumSrc, int lumFilterSize,
393			const int16_t *chrFilter, const int16_t **chrUSrc,
394			const int16_t **chrVSrc,
395			int chrFilterSize, const int16_t **alpSrc,
396			uint8_t *dest, int dstW, int dstY)
397			{
398		✗	x86_reg dummy=0;
399		✗	x86_reg dstW_reg = dstW;
400		✗	x86_reg uv_off = c->uv_offx2;
401
402		✗	YSCALEYUV2PACKEDX
403			YSCALEYUV2RGBX
404			"pxor %%mm7, %%mm7 \n\t"
405			/* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */
406			"paddusb "BLUE_DITHER"(%0), %%mm2 \n\t"
407			"paddusb "GREEN_DITHER"(%0), %%mm4 \n\t"
408			"paddusb "RED_DITHER"(%0), %%mm5 \n\t"
409			WRITERGB16(%4, "%5", %%FF_REGa)
410		✗	YSCALEYUV2PACKEDX_END
411		✗	}
412
413			#define REAL_WRITERGB15(dst, dstw, index) \
414			"pand "MANGLE(bF8)", %%mm2 \n\t" /* B */\
415			"pand "MANGLE(bF8)", %%mm4 \n\t" /* G */\
416			"pand "MANGLE(bF8)", %%mm5 \n\t" /* R */\
417			"psrlq $3, %%mm2 \n\t"\
418			"psrlq $1, %%mm5 \n\t"\
419			\
420			"movq %%mm2, %%mm1 \n\t"\
421			"movq %%mm4, %%mm3 \n\t"\
422			\
423			"punpcklbw %%mm7, %%mm3 \n\t"\
424			"punpcklbw %%mm5, %%mm2 \n\t"\
425			"punpckhbw %%mm7, %%mm4 \n\t"\
426			"punpckhbw %%mm5, %%mm1 \n\t"\
427			\
428			"psllq $2, %%mm3 \n\t"\
429			"psllq $2, %%mm4 \n\t"\
430			\
431			"por %%mm3, %%mm2 \n\t"\
432			"por %%mm4, %%mm1 \n\t"\
433			\
434			MOVNTQ(%%mm2, (dst, index, 2))\
435			MOVNTQ(%%mm1, 8(dst, index, 2))\
436			\
437			"add $8, "#index" \n\t"\
438			"cmp "dstw", "#index" \n\t"\
439			" jb 1b \n\t"
440			#define WRITERGB15(dst, dstw, index) REAL_WRITERGB15(dst, dstw, index)
441
442		✗	static void RENAME(yuv2rgb555_X_ar)(SwsInternal *c, const int16_t *lumFilter,
443			const int16_t **lumSrc, int lumFilterSize,
444			const int16_t *chrFilter, const int16_t **chrUSrc,
445			const int16_t **chrVSrc,
446			int chrFilterSize, const int16_t **alpSrc,
447			uint8_t *dest, int dstW, int dstY)
448			{
449		✗	x86_reg dummy=0;
450		✗	x86_reg dstW_reg = dstW;
451		✗	x86_reg uv_off = c->uv_offx2;
452
453		✗	YSCALEYUV2PACKEDX_ACCURATE
454			YSCALEYUV2RGBX
455			"pxor %%mm7, %%mm7 \n\t"
456			/* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */
457			"paddusb "BLUE_DITHER"(%0), %%mm2\n\t"
458			"paddusb "GREEN_DITHER"(%0), %%mm4\n\t"
459			"paddusb "RED_DITHER"(%0), %%mm5\n\t"
460			WRITERGB15(%4, "%5", %%FF_REGa)
461		✗	YSCALEYUV2PACKEDX_END
462		✗	}
463
464		✗	static void RENAME(yuv2rgb555_X)(SwsInternal *c, const int16_t *lumFilter,
465			const int16_t **lumSrc, int lumFilterSize,
466			const int16_t *chrFilter, const int16_t **chrUSrc,
467			const int16_t **chrVSrc,
468			int chrFilterSize, const int16_t **alpSrc,
469			uint8_t *dest, int dstW, int dstY)
470			{
471		✗	x86_reg dummy=0;
472		✗	x86_reg dstW_reg = dstW;
473		✗	x86_reg uv_off = c->uv_offx2;
474
475		✗	YSCALEYUV2PACKEDX
476			YSCALEYUV2RGBX
477			"pxor %%mm7, %%mm7 \n\t"
478			/* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */
479			"paddusb "BLUE_DITHER"(%0), %%mm2 \n\t"
480			"paddusb "GREEN_DITHER"(%0), %%mm4 \n\t"
481			"paddusb "RED_DITHER"(%0), %%mm5 \n\t"
482			WRITERGB15(%4, "%5", %%FF_REGa)
483		✗	YSCALEYUV2PACKEDX_END
484		✗	}
485
486			#define WRITEBGR24MMXEXT(dst, dstw, index) \
487			/* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */\
488			"movq "MANGLE(M24A)", %%mm0 \n\t"\
489			"movq "MANGLE(M24C)", %%mm7 \n\t"\
490			"pshufw $0x50, %%mm2, %%mm1 \n\t" /* B3 B2 B3 B2 B1 B0 B1 B0 */\
491			"pshufw $0x50, %%mm4, %%mm3 \n\t" /* G3 G2 G3 G2 G1 G0 G1 G0 */\
492			"pshufw $0x00, %%mm5, %%mm6 \n\t" /* R1 R0 R1 R0 R1 R0 R1 R0 */\
493			\
494			"pand %%mm0, %%mm1 \n\t" /* B2 B1 B0 */\
495			"pand %%mm0, %%mm3 \n\t" /* G2 G1 G0 */\
496			"pand %%mm7, %%mm6 \n\t" /* R1 R0 */\
497			\
498			"psllq $8, %%mm3 \n\t" /* G2 G1 G0 */\
499			"por %%mm1, %%mm6 \n\t"\
500			"por %%mm3, %%mm6 \n\t"\
501			MOVNTQ(%%mm6, (dst))\
502			\
503			"psrlq $8, %%mm4 \n\t" /* 00 G7 G6 G5 G4 G3 G2 G1 */\
504			"pshufw $0xA5, %%mm2, %%mm1 \n\t" /* B5 B4 B5 B4 B3 B2 B3 B2 */\
505			"pshufw $0x55, %%mm4, %%mm3 \n\t" /* G4 G3 G4 G3 G4 G3 G4 G3 */\
506			"pshufw $0xA5, %%mm5, %%mm6 \n\t" /* R5 R4 R5 R4 R3 R2 R3 R2 */\
507			\
508			"pand "MANGLE(M24B)", %%mm1 \n\t" /* B5 B4 B3 */\
509			"pand %%mm7, %%mm3 \n\t" /* G4 G3 */\
510			"pand %%mm0, %%mm6 \n\t" /* R4 R3 R2 */\
511			\
512			"por %%mm1, %%mm3 \n\t" /* B5 G4 B4 G3 B3 */\
513			"por %%mm3, %%mm6 \n\t"\
514			MOVNTQ(%%mm6, 8(dst))\
515			\
516			"pshufw $0xFF, %%mm2, %%mm1 \n\t" /* B7 B6 B7 B6 B7 B6 B6 B7 */\
517			"pshufw $0xFA, %%mm4, %%mm3 \n\t" /* 00 G7 00 G7 G6 G5 G6 G5 */\
518			"pshufw $0xFA, %%mm5, %%mm6 \n\t" /* R7 R6 R7 R6 R5 R4 R5 R4 */\
519			\
520			"pand %%mm7, %%mm1 \n\t" /* B7 B6 */\
521			"pand %%mm0, %%mm3 \n\t" /* G7 G6 G5 */\
522			"pand "MANGLE(M24B)", %%mm6 \n\t" /* R7 R6 R5 */\
523			\
524			"por %%mm1, %%mm3 \n\t"\
525			"por %%mm3, %%mm6 \n\t"\
526			MOVNTQ(%%mm6, 16(dst))\
527			\
528			"add $24, "#dst" \n\t"\
529			\
530			"add $8, "#index" \n\t"\
531			"cmp "dstw", "#index" \n\t"\
532			" jb 1b \n\t"
533
534			#undef WRITEBGR24
535			#define WRITEBGR24(dst, dstw, index) WRITEBGR24MMXEXT(dst, dstw, index)
536
537			#if HAVE_6REGS
538		✗	static void RENAME(yuv2bgr24_X_ar)(SwsInternal *c, const int16_t *lumFilter,
539			const int16_t **lumSrc, int lumFilterSize,
540			const int16_t *chrFilter, const int16_t **chrUSrc,
541			const int16_t **chrVSrc,
542			int chrFilterSize, const int16_t **alpSrc,
543			uint8_t *dest, int dstW, int dstY)
544			{
545		✗	x86_reg dummy=0;
546		✗	x86_reg dstW_reg = dstW;
547		✗	x86_reg uv_off = c->uv_offx2;
548
549		✗	YSCALEYUV2PACKEDX_ACCURATE
550			YSCALEYUV2RGBX
551			"pxor %%mm7, %%mm7 \n\t"
552			"lea (%%"FF_REG_a", %%"FF_REG_a", 2), %%"FF_REG_c"\n\t" //FIXME optimize
553			"add %4, %%"FF_REG_c" \n\t"
554			WRITEBGR24(%%FF_REGc, "%5", %%FF_REGa)
555		✗	:: "r" (&c->redDither),
556			"m" (dummy), "m" (dummy), "m" (dummy),
557			"r" (dest), "m" (dstW_reg), "m"(uv_off)
558			NAMED_CONSTRAINTS_ADD(M24A,M24C,M24B)
559			: "%"FF_REG_a, "%"FF_REG_c, "%"FF_REG_d, "%"FF_REG_S
560			);
561		✗	}
562
563		✗	static void RENAME(yuv2bgr24_X)(SwsInternal *c, const int16_t *lumFilter,
564			const int16_t **lumSrc, int lumFilterSize,
565			const int16_t *chrFilter, const int16_t **chrUSrc,
566			const int16_t **chrVSrc,
567			int chrFilterSize, const int16_t **alpSrc,
568			uint8_t *dest, int dstW, int dstY)
569			{
570		✗	x86_reg dummy=0;
571		✗	x86_reg dstW_reg = dstW;
572		✗	x86_reg uv_off = c->uv_offx2;
573
574		✗	YSCALEYUV2PACKEDX
575			YSCALEYUV2RGBX
576			"pxor %%mm7, %%mm7 \n\t"
577			"lea (%%"FF_REG_a", %%"FF_REG_a", 2), %%"FF_REG_c" \n\t" //FIXME optimize
578			"add %4, %%"FF_REG_c" \n\t"
579			WRITEBGR24(%%FF_REGc, "%5", %%FF_REGa)
580		✗	:: "r" (&c->redDither),
581			"m" (dummy), "m" (dummy), "m" (dummy),
582			"r" (dest), "m" (dstW_reg), "m"(uv_off)
583			NAMED_CONSTRAINTS_ADD(M24A,M24C,M24B)
584			: "%"FF_REG_a, "%"FF_REG_c, "%"FF_REG_d, "%"FF_REG_S
585			);
586		✗	}
587			#endif /* HAVE_6REGS */
588
589			#define REAL_WRITEYUY2(dst, dstw, index) \
590			"packuswb %%mm3, %%mm3 \n\t"\
591			"packuswb %%mm4, %%mm4 \n\t"\
592			"packuswb %%mm7, %%mm1 \n\t"\
593			"punpcklbw %%mm4, %%mm3 \n\t"\
594			"movq %%mm1, %%mm7 \n\t"\
595			"punpcklbw %%mm3, %%mm1 \n\t"\
596			"punpckhbw %%mm3, %%mm7 \n\t"\
597			\
598			MOVNTQ(%%mm1, (dst, index, 2))\
599			MOVNTQ(%%mm7, 8(dst, index, 2))\
600			\
601			"add $8, "#index" \n\t"\
602			"cmp "dstw", "#index" \n\t"\
603			" jb 1b \n\t"
604			#define WRITEYUY2(dst, dstw, index) REAL_WRITEYUY2(dst, dstw, index)
605
606		✗	static void RENAME(yuv2yuyv422_X_ar)(SwsInternal *c, const int16_t *lumFilter,
607			const int16_t **lumSrc, int lumFilterSize,
608			const int16_t *chrFilter, const int16_t **chrUSrc,
609			const int16_t **chrVSrc,
610			int chrFilterSize, const int16_t **alpSrc,
611			uint8_t *dest, int dstW, int dstY)
612			{
613		✗	x86_reg dummy=0;
614		✗	x86_reg dstW_reg = dstW;
615		✗	x86_reg uv_off = c->uv_offx2;
616
617		✗	YSCALEYUV2PACKEDX_ACCURATE
618			/* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */
619			"psraw $3, %%mm3 \n\t"
620			"psraw $3, %%mm4 \n\t"
621			"psraw $3, %%mm1 \n\t"
622			"psraw $3, %%mm7 \n\t"
623			WRITEYUY2(%4, "%5", %%FF_REGa)
624		✗	YSCALEYUV2PACKEDX_END
625		✗	}
626
627		✗	static void RENAME(yuv2yuyv422_X)(SwsInternal *c, const int16_t *lumFilter,
628			const int16_t **lumSrc, int lumFilterSize,
629			const int16_t *chrFilter, const int16_t **chrUSrc,
630			const int16_t **chrVSrc,
631			int chrFilterSize, const int16_t **alpSrc,
632			uint8_t *dest, int dstW, int dstY)
633			{
634		✗	x86_reg dummy=0;
635		✗	x86_reg dstW_reg = dstW;
636		✗	x86_reg uv_off = c->uv_offx2;
637
638		✗	YSCALEYUV2PACKEDX
639			/* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */
640			"psraw $3, %%mm3 \n\t"
641			"psraw $3, %%mm4 \n\t"
642			"psraw $3, %%mm1 \n\t"
643			"psraw $3, %%mm7 \n\t"
644			WRITEYUY2(%4, "%5", %%FF_REGa)
645		✗	YSCALEYUV2PACKEDX_END
646		✗	}
647
648			#define REAL_YSCALEYUV2RGB_UV(index, c) \
649			"xor "#index", "#index" \n\t"\
650			".p2align 4 \n\t"\
651			"1: \n\t"\
652			"movq (%2, "#index"), %%mm2 \n\t" /* uvbuf0[eax]*/\
653			"movq (%3, "#index"), %%mm3 \n\t" /* uvbuf1[eax]*/\
654			"add "UV_OFF_BYTE"("#c"), "#index" \n\t" \
655			"movq (%2, "#index"), %%mm5 \n\t" /* uvbuf0[eax+2048]*/\
656			"movq (%3, "#index"), %%mm4 \n\t" /* uvbuf1[eax+2048]*/\
657			"sub "UV_OFF_BYTE"("#c"), "#index" \n\t" \
658			"psubw %%mm3, %%mm2 \n\t" /* uvbuf0[eax] - uvbuf1[eax]*/\
659			"psubw %%mm4, %%mm5 \n\t" /* uvbuf0[eax+2048] - uvbuf1[eax+2048]*/\
660			"movq "CHR_MMX_FILTER_OFFSET"+8("#c"), %%mm0 \n\t"\
661			"pmulhw %%mm0, %%mm2 \n\t" /* (uvbuf0[eax] - uvbuf1[eax])uvalpha1>>16*/\
662			"pmulhw %%mm0, %%mm5 \n\t" /* (uvbuf0[eax+2048] - uvbuf1[eax+2048])uvalpha1>>16*/\
663			"psraw $4, %%mm3 \n\t" /* uvbuf0[eax] - uvbuf1[eax] >>4*/\
664			"psraw $4, %%mm4 \n\t" /* uvbuf0[eax+2048] - uvbuf1[eax+2048] >>4*/\
665			"paddw %%mm2, %%mm3 \n\t" /* uvbuf0[eax]uvalpha1 - uvbuf1[eax](1-uvalpha1)*/\
666			"paddw %%mm5, %%mm4 \n\t" /* uvbuf0[eax+2048]uvalpha1 - uvbuf1[eax+2048](1-uvalpha1)*/\
667			"psubw "U_OFFSET"("#c"), %%mm3 \n\t" /* (U-128)8*/\
668			"psubw "V_OFFSET"("#c"), %%mm4 \n\t" /* (V-128)8*/\
669			"movq %%mm3, %%mm2 \n\t" /* (U-128)8*/\
670			"movq %%mm4, %%mm5 \n\t" /* (V-128)8*/\
671			"pmulhw "UG_COEFF"("#c"), %%mm3 \n\t"\
672			"pmulhw "VG_COEFF"("#c"), %%mm4 \n\t"\
673			/* mm2=(U-128)8, mm3=ug, mm4=vg mm5=(V-128)8 */\
674
675			#define REAL_YSCALEYUV2RGB_YA(index, c, b1, b2) \
676			"movq ("#b1", "#index", 2), %%mm0 \n\t" /*buf0[eax]*/\
677			"movq ("#b2", "#index", 2), %%mm1 \n\t" /*buf1[eax]*/\
678			"movq 8("#b1", "#index", 2), %%mm6 \n\t" /*buf0[eax]*/\
679			"movq 8("#b2", "#index", 2), %%mm7 \n\t" /*buf1[eax]*/\
680			"psubw %%mm1, %%mm0 \n\t" /* buf0[eax] - buf1[eax]*/\
681			"psubw %%mm7, %%mm6 \n\t" /* buf0[eax] - buf1[eax]*/\
682			"pmulhw "LUM_MMX_FILTER_OFFSET"+8("#c"), %%mm0 \n\t" /* (buf0[eax] - buf1[eax])yalpha1>>16*/\
683			"pmulhw "LUM_MMX_FILTER_OFFSET"+8("#c"), %%mm6 \n\t" /* (buf0[eax] - buf1[eax])yalpha1>>16*/\
684			"psraw $4, %%mm1 \n\t" /* buf0[eax] - buf1[eax] >>4*/\
685			"psraw $4, %%mm7 \n\t" /* buf0[eax] - buf1[eax] >>4*/\
686			"paddw %%mm0, %%mm1 \n\t" /* buf0[eax]yalpha1 + buf1[eax](1-yalpha1) >>16*/\
687			"paddw %%mm6, %%mm7 \n\t" /* buf0[eax]yalpha1 + buf1[eax](1-yalpha1) >>16*/\
688
689			#define REAL_YSCALEYUV2RGB_COEFF(c) \
690			"pmulhw "UB_COEFF"("#c"), %%mm2 \n\t"\
691			"pmulhw "VR_COEFF"("#c"), %%mm5 \n\t"\
692			"psubw "Y_OFFSET"("#c"), %%mm1 \n\t" /* 8(Y-16)*/\
693			"psubw "Y_OFFSET"("#c"), %%mm7 \n\t" /* 8(Y-16)*/\
694			"pmulhw "Y_COEFF"("#c"), %%mm1 \n\t"\
695			"pmulhw "Y_COEFF"("#c"), %%mm7 \n\t"\
696			/* mm1= Y1, mm2=ub, mm3=ug, mm4=vg mm5=vr, mm7=Y2 */\
697			"paddw %%mm3, %%mm4 \n\t"\
698			"movq %%mm2, %%mm0 \n\t"\
699			"movq %%mm5, %%mm6 \n\t"\
700			"movq %%mm4, %%mm3 \n\t"\
701			"punpcklwd %%mm2, %%mm2 \n\t"\
702			"punpcklwd %%mm5, %%mm5 \n\t"\
703			"punpcklwd %%mm4, %%mm4 \n\t"\
704			"paddw %%mm1, %%mm2 \n\t"\
705			"paddw %%mm1, %%mm5 \n\t"\
706			"paddw %%mm1, %%mm4 \n\t"\
707			"punpckhwd %%mm0, %%mm0 \n\t"\
708			"punpckhwd %%mm6, %%mm6 \n\t"\
709			"punpckhwd %%mm3, %%mm3 \n\t"\
710			"paddw %%mm7, %%mm0 \n\t"\
711			"paddw %%mm7, %%mm6 \n\t"\
712			"paddw %%mm7, %%mm3 \n\t"\
713			/* mm0=B1, mm2=B2, mm3=G2, mm4=G1, mm5=R1, mm6=R2 */\
714			"packuswb %%mm0, %%mm2 \n\t"\
715			"packuswb %%mm6, %%mm5 \n\t"\
716			"packuswb %%mm3, %%mm4 \n\t"\
717
718			#define YSCALEYUV2RGB_YA(index, c, b1, b2) REAL_YSCALEYUV2RGB_YA(index, c, b1, b2)
719
720			#define YSCALEYUV2RGB(index, c) \
721			REAL_YSCALEYUV2RGB_UV(index, c) \
722			REAL_YSCALEYUV2RGB_YA(index, c, %0, %1) \
723			REAL_YSCALEYUV2RGB_COEFF(c)
724
725			/**
726			* vertical bilinear scale YV12 to RGB
727			*/
728		✗	static void RENAME(yuv2rgb32_2)(SwsInternal *c, const int16_t *buf[2],
729			const int16_t *ubuf[2], const int16_t *vbuf[2],
730			const int16_t *abuf[2], uint8_t *dest,
731			int dstW, int yalpha, int uvalpha, int y)
732			{
733		✗	const int16_t *buf0 = buf[0], *buf1 = buf[1],
734		✗	*ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
735
736		✗	if (CONFIG_SWSCALE_ALPHA && c->needAlpha) {
737		✗	const int16_t *abuf0 = abuf[0], *abuf1 = abuf[1];
738			#if ARCH_X86_64
739		✗	__asm__ volatile(
740			YSCALEYUV2RGB(%%r8, %5)
741			YSCALEYUV2RGB_YA(%%r8, %5, %6, %7)
742			"psraw $3, %%mm1 \n\t" /* abuf0[eax] - abuf1[eax] >>7*/
743			"psraw $3, %%mm7 \n\t" /* abuf0[eax] - abuf1[eax] >>7*/
744			"packuswb %%mm7, %%mm1 \n\t"
745			WRITEBGR32(%4, DSTW_OFFSET"(%5)", %%r8, %%mm2, %%mm4, %%mm5, %%mm1, %%mm0, %%mm7, %%mm3, %%mm6)
746			:: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "r" (dest),
747		✗	"a" (&c->redDither),
748			"r" (abuf0), "r" (abuf1)
749			: "%r8"
750			);
751			#else
752			c->u_temp=(intptr_t)abuf0;
753			c->v_temp=(intptr_t)abuf1;
754			__asm__ volatile(
755			"mov %%"FF_REG_b", "ESP_OFFSET"(%5) \n\t"
756			"mov %4, %%"FF_REG_b" \n\t"
757			"push %%"FF_REG_BP" \n\t"
758			YSCALEYUV2RGB(%%FF_REGBP, %5)
759			"push %0 \n\t"
760			"push %1 \n\t"
761			"mov "U_TEMP"(%5), %0 \n\t"
762			"mov "V_TEMP"(%5), %1 \n\t"
763			YSCALEYUV2RGB_YA(%%FF_REGBP, %5, %0, %1)
764			"psraw $3, %%mm1 \n\t" /* abuf0[eax] - abuf1[eax] >>7*/
765			"psraw $3, %%mm7 \n\t" /* abuf0[eax] - abuf1[eax] >>7*/
766			"packuswb %%mm7, %%mm1 \n\t"
767			"pop %1 \n\t"
768			"pop %0 \n\t"
769			WRITEBGR32(%%FF_REGb, DSTW_OFFSET"(%5)", %%FF_REGBP, %%mm2, %%mm4, %%mm5, %%mm1, %%mm0, %%mm7, %%mm3, %%mm6)
770			"pop %%"FF_REG_BP" \n\t"
771			"mov "ESP_OFFSET"(%5), %%"FF_REG_b" \n\t"
772			:: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),
773			"a" (&c->redDither)
774			);
775			#endif
776			} else {
777		✗	__asm__ volatile(
778			"mov %%"FF_REG_b", "ESP_OFFSET"(%5) \n\t"
779			"mov %4, %%"FF_REG_b" \n\t"
780			"push %%"FF_REG_BP" \n\t"
781			YSCALEYUV2RGB(%%FF_REGBP, %5)
782			"pcmpeqd %%mm7, %%mm7 \n\t"
783			WRITEBGR32(%%FF_REGb, DSTW_OFFSET"(%5)", %%FF_REGBP, %%mm2, %%mm4, %%mm5, %%mm7, %%mm0, %%mm1, %%mm3, %%mm6)
784			"pop %%"FF_REG_BP" \n\t"
785			"mov "ESP_OFFSET"(%5), %%"FF_REG_b" \n\t"
786			:: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),
787		✗	"a" (&c->redDither)
788			);
789			}
790		✗	}
791
792		✗	static void RENAME(yuv2bgr24_2)(SwsInternal *c, const int16_t *buf[2],
793			const int16_t *ubuf[2], const int16_t *vbuf[2],
794			const int16_t *abuf[2], uint8_t *dest,
795			int dstW, int yalpha, int uvalpha, int y)
796			{
797		✗	const int16_t *buf0 = buf[0], *buf1 = buf[1],
798		✗	*ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
799
800		✗	__asm__ volatile(
801			"mov %%"FF_REG_b", "ESP_OFFSET"(%5) \n\t"
802			"mov %4, %%"FF_REG_b" \n\t"
803			"push %%"FF_REG_BP" \n\t"
804			YSCALEYUV2RGB(%%FF_REGBP, %5)
805			"pxor %%mm7, %%mm7 \n\t"
806			WRITEBGR24(%%FF_REGb, DSTW_OFFSET"(%5)", %%FF_REGBP)
807			"pop %%"FF_REG_BP" \n\t"
808			"mov "ESP_OFFSET"(%5), %%"FF_REG_b" \n\t"
809			:: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),
810		✗	"a" (&c->redDither)
811			NAMED_CONSTRAINTS_ADD(M24A,M24C,M24B)
812			);
813		✗	}
814
815		✗	static void RENAME(yuv2rgb555_2)(SwsInternal *c, const int16_t *buf[2],
816			const int16_t *ubuf[2], const int16_t *vbuf[2],
817			const int16_t *abuf[2], uint8_t *dest,
818			int dstW, int yalpha, int uvalpha, int y)
819			{
820		✗	const int16_t *buf0 = buf[0], *buf1 = buf[1],
821		✗	*ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
822
823		✗	__asm__ volatile(
824			"mov %%"FF_REG_b", "ESP_OFFSET"(%5) \n\t"
825			"mov %4, %%"FF_REG_b" \n\t"
826			"push %%"FF_REG_BP" \n\t"
827			YSCALEYUV2RGB(%%FF_REGBP, %5)
828			"pxor %%mm7, %%mm7 \n\t"
829			/* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */
830			"paddusb "BLUE_DITHER"(%5), %%mm2 \n\t"
831			"paddusb "GREEN_DITHER"(%5), %%mm4 \n\t"
832			"paddusb "RED_DITHER"(%5), %%mm5 \n\t"
833			WRITERGB15(%%FF_REGb, DSTW_OFFSET"(%5)", %%FF_REGBP)
834			"pop %%"FF_REG_BP" \n\t"
835			"mov "ESP_OFFSET"(%5), %%"FF_REG_b" \n\t"
836			:: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),
837		✗	"a" (&c->redDither)
838			NAMED_CONSTRAINTS_ADD(bF8)
839			);
840		✗	}
841
842		✗	static void RENAME(yuv2rgb565_2)(SwsInternal *c, const int16_t *buf[2],
843			const int16_t *ubuf[2], const int16_t *vbuf[2],
844			const int16_t *abuf[2], uint8_t *dest,
845			int dstW, int yalpha, int uvalpha, int y)
846			{
847		✗	const int16_t *buf0 = buf[0], *buf1 = buf[1],
848		✗	*ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
849
850		✗	__asm__ volatile(
851			"mov %%"FF_REG_b", "ESP_OFFSET"(%5) \n\t"
852			"mov %4, %%"FF_REG_b" \n\t"
853			"push %%"FF_REG_BP" \n\t"
854			YSCALEYUV2RGB(%%FF_REGBP, %5)
855			"pxor %%mm7, %%mm7 \n\t"
856			/* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */
857			"paddusb "BLUE_DITHER"(%5), %%mm2 \n\t"
858			"paddusb "GREEN_DITHER"(%5), %%mm4 \n\t"
859			"paddusb "RED_DITHER"(%5), %%mm5 \n\t"
860			WRITERGB16(%%FF_REGb, DSTW_OFFSET"(%5)", %%FF_REGBP)
861			"pop %%"FF_REG_BP" \n\t"
862			"mov "ESP_OFFSET"(%5), %%"FF_REG_b" \n\t"
863			:: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),
864		✗	"a" (&c->redDither)
865			NAMED_CONSTRAINTS_ADD(bF8,bFC)
866			);
867		✗	}
868
869			#define REAL_YSCALEYUV2PACKED(index, c) \
870			"movq "CHR_MMX_FILTER_OFFSET"+8("#c"), %%mm0 \n\t"\
871			"movq "LUM_MMX_FILTER_OFFSET"+8("#c"), %%mm1 \n\t"\
872			"psraw $3, %%mm0 \n\t"\
873			"psraw $3, %%mm1 \n\t"\
874			"movq %%mm0, "CHR_MMX_FILTER_OFFSET"+8("#c") \n\t"\
875			"movq %%mm1, "LUM_MMX_FILTER_OFFSET"+8("#c") \n\t"\
876			"xor "#index", "#index" \n\t"\
877			".p2align 4 \n\t"\
878			"1: \n\t"\
879			"movq (%2, "#index"), %%mm2 \n\t" /* uvbuf0[eax]*/\
880			"movq (%3, "#index"), %%mm3 \n\t" /* uvbuf1[eax]*/\
881			"add "UV_OFF_BYTE"("#c"), "#index" \n\t" \
882			"movq (%2, "#index"), %%mm5 \n\t" /* uvbuf0[eax+2048]*/\
883			"movq (%3, "#index"), %%mm4 \n\t" /* uvbuf1[eax+2048]*/\
884			"sub "UV_OFF_BYTE"("#c"), "#index" \n\t" \
885			"psubw %%mm3, %%mm2 \n\t" /* uvbuf0[eax] - uvbuf1[eax]*/\
886			"psubw %%mm4, %%mm5 \n\t" /* uvbuf0[eax+2048] - uvbuf1[eax+2048]*/\
887			"movq "CHR_MMX_FILTER_OFFSET"+8("#c"), %%mm0 \n\t"\
888			"pmulhw %%mm0, %%mm2 \n\t" /* (uvbuf0[eax] - uvbuf1[eax])uvalpha1>>16*/\
889			"pmulhw %%mm0, %%mm5 \n\t" /* (uvbuf0[eax+2048] - uvbuf1[eax+2048])uvalpha1>>16*/\
890			"psraw $7, %%mm3 \n\t" /* uvbuf0[eax] - uvbuf1[eax] >>4*/\
891			"psraw $7, %%mm4 \n\t" /* uvbuf0[eax+2048] - uvbuf1[eax+2048] >>4*/\
892			"paddw %%mm2, %%mm3 \n\t" /* uvbuf0[eax]uvalpha1 - uvbuf1[eax](1-uvalpha1)*/\
893			"paddw %%mm5, %%mm4 \n\t" /* uvbuf0[eax+2048]uvalpha1 - uvbuf1[eax+2048](1-uvalpha1)*/\
894			"movq (%0, "#index", 2), %%mm0 \n\t" /*buf0[eax]*/\
895			"movq (%1, "#index", 2), %%mm1 \n\t" /*buf1[eax]*/\
896			"movq 8(%0, "#index", 2), %%mm6 \n\t" /*buf0[eax]*/\
897			"movq 8(%1, "#index", 2), %%mm7 \n\t" /*buf1[eax]*/\
898			"psubw %%mm1, %%mm0 \n\t" /* buf0[eax] - buf1[eax]*/\
899			"psubw %%mm7, %%mm6 \n\t" /* buf0[eax] - buf1[eax]*/\
900			"pmulhw "LUM_MMX_FILTER_OFFSET"+8("#c"), %%mm0 \n\t" /* (buf0[eax] - buf1[eax])yalpha1>>16*/\
901			"pmulhw "LUM_MMX_FILTER_OFFSET"+8("#c"), %%mm6 \n\t" /* (buf0[eax] - buf1[eax])yalpha1>>16*/\
902			"psraw $7, %%mm1 \n\t" /* buf0[eax] - buf1[eax] >>4*/\
903			"psraw $7, %%mm7 \n\t" /* buf0[eax] - buf1[eax] >>4*/\
904			"paddw %%mm0, %%mm1 \n\t" /* buf0[eax]yalpha1 + buf1[eax](1-yalpha1) >>16*/\
905			"paddw %%mm6, %%mm7 \n\t" /* buf0[eax]yalpha1 + buf1[eax](1-yalpha1) >>16*/\
906
907			#define YSCALEYUV2PACKED(index, c) REAL_YSCALEYUV2PACKED(index, c)
908
909		✗	static void RENAME(yuv2yuyv422_2)(SwsInternal *c, const int16_t *buf[2],
910			const int16_t *ubuf[2], const int16_t *vbuf[2],
911			const int16_t *abuf[2], uint8_t *dest,
912			int dstW, int yalpha, int uvalpha, int y)
913			{
914		✗	const int16_t *buf0 = buf[0], *buf1 = buf[1],
915		✗	*ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
916
917		✗	__asm__ volatile(
918			"mov %%"FF_REG_b", "ESP_OFFSET"(%5) \n\t"
919			"mov %4, %%"FF_REG_b" \n\t"
920			"push %%"FF_REG_BP" \n\t"
921			YSCALEYUV2PACKED(%%FF_REGBP, %5)
922			WRITEYUY2(%%FF_REGb, DSTW_OFFSET"(%5)", %%FF_REGBP)
923			"pop %%"FF_REG_BP" \n\t"
924			"mov "ESP_OFFSET"(%5), %%"FF_REG_b" \n\t"
925			:: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),
926		✗	"a" (&c->redDither)
927			);
928		✗	}
929
930			#define REAL_YSCALEYUV2RGB1(index, c) \
931			"xor "#index", "#index" \n\t"\
932			".p2align 4 \n\t"\
933			"1: \n\t"\
934			"movq (%2, "#index"), %%mm3 \n\t" /* uvbuf0[eax]*/\
935			"add "UV_OFF_BYTE"("#c"), "#index" \n\t" \
936			"movq (%2, "#index"), %%mm4 \n\t" /* uvbuf0[eax+2048]*/\
937			"sub "UV_OFF_BYTE"("#c"), "#index" \n\t" \
938			"psraw $4, %%mm3 \n\t" /* uvbuf0[eax] - uvbuf1[eax] >>4*/\
939			"psraw $4, %%mm4 \n\t" /* uvbuf0[eax+2048] - uvbuf1[eax+2048] >>4*/\
940			"psubw "U_OFFSET"("#c"), %%mm3 \n\t" /* (U-128)8*/\
941			"psubw "V_OFFSET"("#c"), %%mm4 \n\t" /* (V-128)8*/\
942			"movq %%mm3, %%mm2 \n\t" /* (U-128)8*/\
943			"movq %%mm4, %%mm5 \n\t" /* (V-128)8*/\
944			"pmulhw "UG_COEFF"("#c"), %%mm3 \n\t"\
945			"pmulhw "VG_COEFF"("#c"), %%mm4 \n\t"\
946			/* mm2=(U-128)8, mm3=ug, mm4=vg mm5=(V-128)8 */\
947			"movq (%0, "#index", 2), %%mm1 \n\t" /*buf0[eax]*/\
948			"movq 8(%0, "#index", 2), %%mm7 \n\t" /*buf0[eax]*/\
949			"psraw $4, %%mm1 \n\t" /* buf0[eax] - buf1[eax] >>4*/\
950			"psraw $4, %%mm7 \n\t" /* buf0[eax] - buf1[eax] >>4*/\
951			"pmulhw "UB_COEFF"("#c"), %%mm2 \n\t"\
952			"pmulhw "VR_COEFF"("#c"), %%mm5 \n\t"\
953			"psubw "Y_OFFSET"("#c"), %%mm1 \n\t" /* 8(Y-16)*/\
954			"psubw "Y_OFFSET"("#c"), %%mm7 \n\t" /* 8(Y-16)*/\
955			"pmulhw "Y_COEFF"("#c"), %%mm1 \n\t"\
956			"pmulhw "Y_COEFF"("#c"), %%mm7 \n\t"\
957			/* mm1= Y1, mm2=ub, mm3=ug, mm4=vg mm5=vr, mm7=Y2 */\
958			"paddw %%mm3, %%mm4 \n\t"\
959			"movq %%mm2, %%mm0 \n\t"\
960			"movq %%mm5, %%mm6 \n\t"\
961			"movq %%mm4, %%mm3 \n\t"\
962			"punpcklwd %%mm2, %%mm2 \n\t"\
963			"punpcklwd %%mm5, %%mm5 \n\t"\
964			"punpcklwd %%mm4, %%mm4 \n\t"\
965			"paddw %%mm1, %%mm2 \n\t"\
966			"paddw %%mm1, %%mm5 \n\t"\
967			"paddw %%mm1, %%mm4 \n\t"\
968			"punpckhwd %%mm0, %%mm0 \n\t"\
969			"punpckhwd %%mm6, %%mm6 \n\t"\
970			"punpckhwd %%mm3, %%mm3 \n\t"\
971			"paddw %%mm7, %%mm0 \n\t"\
972			"paddw %%mm7, %%mm6 \n\t"\
973			"paddw %%mm7, %%mm3 \n\t"\
974			/* mm0=B1, mm2=B2, mm3=G2, mm4=G1, mm5=R1, mm6=R2 */\
975			"packuswb %%mm0, %%mm2 \n\t"\
976			"packuswb %%mm6, %%mm5 \n\t"\
977			"packuswb %%mm3, %%mm4 \n\t"\
978
979			#define YSCALEYUV2RGB1(index, c) REAL_YSCALEYUV2RGB1(index, c)
980
981			// do vertical chrominance interpolation
982			#define REAL_YSCALEYUV2RGB1b(index, c) \
983			"xor "#index", "#index" \n\t"\
984			".p2align 4 \n\t"\
985			"1: \n\t"\
986			"movq (%2, "#index"), %%mm2 \n\t" /* uvbuf0[eax]*/\
987			"movq (%3, "#index"), %%mm3 \n\t" /* uvbuf1[eax]*/\
988			"add "UV_OFF_BYTE"("#c"), "#index" \n\t" \
989			"movq (%2, "#index"), %%mm5 \n\t" /* uvbuf0[eax+2048]*/\
990			"movq (%3, "#index"), %%mm4 \n\t" /* uvbuf1[eax+2048]*/\
991			"sub "UV_OFF_BYTE"("#c"), "#index" \n\t" \
992			"paddw %%mm2, %%mm3 \n\t" /* uvbuf0[eax] + uvbuf1[eax]*/\
993			"paddw %%mm5, %%mm4 \n\t" /* uvbuf0[eax+2048] + uvbuf1[eax+2048]*/\
994			"psrlw $5, %%mm3 \n\t" /*FIXME might overflow*/\
995			"psrlw $5, %%mm4 \n\t" /*FIXME might overflow*/\
996			"psubw "U_OFFSET"("#c"), %%mm3 \n\t" /* (U-128)8*/\
997			"psubw "V_OFFSET"("#c"), %%mm4 \n\t" /* (V-128)8*/\
998			"movq %%mm3, %%mm2 \n\t" /* (U-128)8*/\
999			"movq %%mm4, %%mm5 \n\t" /* (V-128)8*/\
1000			"pmulhw "UG_COEFF"("#c"), %%mm3 \n\t"\
1001			"pmulhw "VG_COEFF"("#c"), %%mm4 \n\t"\
1002			/* mm2=(U-128)8, mm3=ug, mm4=vg mm5=(V-128)8 */\
1003			"movq (%0, "#index", 2), %%mm1 \n\t" /*buf0[eax]*/\
1004			"movq 8(%0, "#index", 2), %%mm7 \n\t" /*buf0[eax]*/\
1005			"psraw $4, %%mm1 \n\t" /* buf0[eax] - buf1[eax] >>4*/\
1006			"psraw $4, %%mm7 \n\t" /* buf0[eax] - buf1[eax] >>4*/\
1007			"pmulhw "UB_COEFF"("#c"), %%mm2 \n\t"\
1008			"pmulhw "VR_COEFF"("#c"), %%mm5 \n\t"\
1009			"psubw "Y_OFFSET"("#c"), %%mm1 \n\t" /* 8(Y-16)*/\
1010			"psubw "Y_OFFSET"("#c"), %%mm7 \n\t" /* 8(Y-16)*/\
1011			"pmulhw "Y_COEFF"("#c"), %%mm1 \n\t"\
1012			"pmulhw "Y_COEFF"("#c"), %%mm7 \n\t"\
1013			/* mm1= Y1, mm2=ub, mm3=ug, mm4=vg mm5=vr, mm7=Y2 */\
1014			"paddw %%mm3, %%mm4 \n\t"\
1015			"movq %%mm2, %%mm0 \n\t"\
1016			"movq %%mm5, %%mm6 \n\t"\
1017			"movq %%mm4, %%mm3 \n\t"\
1018			"punpcklwd %%mm2, %%mm2 \n\t"\
1019			"punpcklwd %%mm5, %%mm5 \n\t"\
1020			"punpcklwd %%mm4, %%mm4 \n\t"\
1021			"paddw %%mm1, %%mm2 \n\t"\
1022			"paddw %%mm1, %%mm5 \n\t"\
1023			"paddw %%mm1, %%mm4 \n\t"\
1024			"punpckhwd %%mm0, %%mm0 \n\t"\
1025			"punpckhwd %%mm6, %%mm6 \n\t"\
1026			"punpckhwd %%mm3, %%mm3 \n\t"\
1027			"paddw %%mm7, %%mm0 \n\t"\
1028			"paddw %%mm7, %%mm6 \n\t"\
1029			"paddw %%mm7, %%mm3 \n\t"\
1030			/* mm0=B1, mm2=B2, mm3=G2, mm4=G1, mm5=R1, mm6=R2 */\
1031			"packuswb %%mm0, %%mm2 \n\t"\
1032			"packuswb %%mm6, %%mm5 \n\t"\
1033			"packuswb %%mm3, %%mm4 \n\t"\
1034
1035			#define YSCALEYUV2RGB1b(index, c) REAL_YSCALEYUV2RGB1b(index, c)
1036
1037			#define REAL_YSCALEYUV2RGB1_ALPHA(index) \
1038			"movq (%1, "#index", 2), %%mm7 \n\t" /* abuf0[index ] */\
1039			"movq 8(%1, "#index", 2), %%mm1 \n\t" /* abuf0[index+4] */\
1040			"psraw $7, %%mm7 \n\t" /* abuf0[index ] >>7 */\
1041			"psraw $7, %%mm1 \n\t" /* abuf0[index+4] >>7 */\
1042			"packuswb %%mm1, %%mm7 \n\t"
1043			#define YSCALEYUV2RGB1_ALPHA(index) REAL_YSCALEYUV2RGB1_ALPHA(index)
1044
1045			/**
1046			* YV12 to RGB without scaling or interpolating
1047			*/
1048		✗	static void RENAME(yuv2rgb32_1)(SwsInternal *c, const int16_t *buf0,
1049			const int16_t *ubuf[2], const int16_t *vbuf[2],
1050			const int16_t *abuf0, uint8_t *dest,
1051			int dstW, int uvalpha, int y)
1052			{
1053		✗	const int16_t *ubuf0 = ubuf[0];
1054		✗	const int16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
1055
1056		✗	if (uvalpha < 2048) { // note this is not correct (shifts chrominance by 0.5 pixels) but it is a bit faster
1057		✗	const int16_t *ubuf1 = ubuf[0];
1058		✗	if (CONFIG_SWSCALE_ALPHA && c->needAlpha) {
1059		✗	__asm__ volatile(
1060			"mov %%"FF_REG_b", "ESP_OFFSET"(%5) \n\t"
1061			"mov %4, %%"FF_REG_b" \n\t"
1062			"push %%"FF_REG_BP" \n\t"
1063			YSCALEYUV2RGB1(%%FF_REGBP, %5)
1064			YSCALEYUV2RGB1_ALPHA(%%FF_REGBP)
1065			WRITEBGR32(%%FF_REGb, DSTW_OFFSET"(%5)", %%FF_REGBP, %%mm2, %%mm4, %%mm5, %%mm7, %%mm0, %%mm1, %%mm3, %%mm6)
1066			"pop %%"FF_REG_BP" \n\t"
1067			"mov "ESP_OFFSET"(%5), %%"FF_REG_b" \n\t"
1068			:: "c" (buf0), "d" (abuf0), "S" (ubuf0), "D" (ubuf1), "m" (dest),
1069		✗	"a" (&c->redDither)
1070			);
1071			} else {
1072		✗	__asm__ volatile(
1073			"mov %%"FF_REG_b", "ESP_OFFSET"(%5) \n\t"
1074			"mov %4, %%"FF_REG_b" \n\t"
1075			"push %%"FF_REG_BP" \n\t"
1076			YSCALEYUV2RGB1(%%FF_REGBP, %5)
1077			"pcmpeqd %%mm7, %%mm7 \n\t"
1078			WRITEBGR32(%%FF_REGb, DSTW_OFFSET"(%5)", %%FF_REGBP, %%mm2, %%mm4, %%mm5, %%mm7, %%mm0, %%mm1, %%mm3, %%mm6)
1079			"pop %%"FF_REG_BP" \n\t"
1080			"mov "ESP_OFFSET"(%5), %%"FF_REG_b" \n\t"
1081			:: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),
1082		✗	"a" (&c->redDither)
1083			);
1084			}
1085			} else {
1086		✗	const int16_t *ubuf1 = ubuf[1];
1087		✗	if (CONFIG_SWSCALE_ALPHA && c->needAlpha) {
1088		✗	__asm__ volatile(
1089			"mov %%"FF_REG_b", "ESP_OFFSET"(%5) \n\t"
1090			"mov %4, %%"FF_REG_b" \n\t"
1091			"push %%"FF_REG_BP" \n\t"
1092			YSCALEYUV2RGB1b(%%FF_REGBP, %5)
1093			YSCALEYUV2RGB1_ALPHA(%%FF_REGBP)
1094			WRITEBGR32(%%FF_REGb, DSTW_OFFSET"(%5)", %%FF_REGBP, %%mm2, %%mm4, %%mm5, %%mm7, %%mm0, %%mm1, %%mm3, %%mm6)
1095			"pop %%"FF_REG_BP" \n\t"
1096			"mov "ESP_OFFSET"(%5), %%"FF_REG_b" \n\t"
1097			:: "c" (buf0), "d" (abuf0), "S" (ubuf0), "D" (ubuf1), "m" (dest),
1098		✗	"a" (&c->redDither)
1099			);
1100			} else {
1101		✗	__asm__ volatile(
1102			"mov %%"FF_REG_b", "ESP_OFFSET"(%5) \n\t"
1103			"mov %4, %%"FF_REG_b" \n\t"
1104			"push %%"FF_REG_BP" \n\t"
1105			YSCALEYUV2RGB1b(%%FF_REGBP, %5)
1106			"pcmpeqd %%mm7, %%mm7 \n\t"
1107			WRITEBGR32(%%FF_REGb, DSTW_OFFSET"(%5)", %%FF_REGBP, %%mm2, %%mm4, %%mm5, %%mm7, %%mm0, %%mm1, %%mm3, %%mm6)
1108			"pop %%"FF_REG_BP" \n\t"
1109			"mov "ESP_OFFSET"(%5), %%"FF_REG_b" \n\t"
1110			:: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),
1111		✗	"a" (&c->redDither)
1112			);
1113			}
1114			}
1115		✗	}
1116
1117		✗	static void RENAME(yuv2bgr24_1)(SwsInternal *c, const int16_t *buf0,
1118			const int16_t *ubuf[2], const int16_t *vbuf[2],
1119			const int16_t *abuf0, uint8_t *dest,
1120			int dstW, int uvalpha, int y)
1121			{
1122		✗	const int16_t *ubuf0 = ubuf[0];
1123		✗	const int16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
1124
1125		✗	if (uvalpha < 2048) { // note this is not correct (shifts chrominance by 0.5 pixels) but it is a bit faster
1126		✗	const int16_t *ubuf1 = ubuf[0];
1127		✗	__asm__ volatile(
1128			"mov %%"FF_REG_b", "ESP_OFFSET"(%5) \n\t"
1129			"mov %4, %%"FF_REG_b" \n\t"
1130			"push %%"FF_REG_BP" \n\t"
1131			YSCALEYUV2RGB1(%%FF_REGBP, %5)
1132			"pxor %%mm7, %%mm7 \n\t"
1133			WRITEBGR24(%%FF_REGb, DSTW_OFFSET"(%5)", %%FF_REGBP)
1134			"pop %%"FF_REG_BP" \n\t"
1135			"mov "ESP_OFFSET"(%5), %%"FF_REG_b" \n\t"
1136			:: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),
1137		✗	"a" (&c->redDither)
1138			NAMED_CONSTRAINTS_ADD(M24A,M24C,M24B)
1139			);
1140			} else {
1141		✗	const int16_t *ubuf1 = ubuf[1];
1142		✗	__asm__ volatile(
1143			"mov %%"FF_REG_b", "ESP_OFFSET"(%5) \n\t"
1144			"mov %4, %%"FF_REG_b" \n\t"
1145			"push %%"FF_REG_BP" \n\t"
1146			YSCALEYUV2RGB1b(%%FF_REGBP, %5)
1147			"pxor %%mm7, %%mm7 \n\t"
1148			WRITEBGR24(%%FF_REGb, DSTW_OFFSET"(%5)", %%FF_REGBP)
1149			"pop %%"FF_REG_BP" \n\t"
1150			"mov "ESP_OFFSET"(%5), %%"FF_REG_b" \n\t"
1151			:: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),
1152		✗	"a" (&c->redDither)
1153			NAMED_CONSTRAINTS_ADD(M24A,M24C,M24B)
1154			);
1155			}
1156		✗	}
1157
1158		940	static void RENAME(yuv2rgb555_1)(SwsInternal *c, const int16_t *buf0,
1159			const int16_t *ubuf[2], const int16_t *vbuf[2],
1160			const int16_t *abuf0, uint8_t *dest,
1161			int dstW, int uvalpha, int y)
1162			{
1163		940	const int16_t *ubuf0 = ubuf[0];
1164		940	const int16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
1165
1166	1/2 ✓ Branch 0 taken 940 times. ✗ Branch 1 not taken.	940	if (uvalpha < 2048) { // note this is not correct (shifts chrominance by 0.5 pixels) but it is a bit faster
1167		940	const int16_t *ubuf1 = ubuf[0];
1168		940	__asm__ volatile(
1169			"mov %%"FF_REG_b", "ESP_OFFSET"(%5) \n\t"
1170			"mov %4, %%"FF_REG_b" \n\t"
1171			"push %%"FF_REG_BP" \n\t"
1172			YSCALEYUV2RGB1(%%FF_REGBP, %5)
1173			"pxor %%mm7, %%mm7 \n\t"
1174			/* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */
1175			"paddusb "BLUE_DITHER"(%5), %%mm2 \n\t"
1176			"paddusb "GREEN_DITHER"(%5), %%mm4 \n\t"
1177			"paddusb "RED_DITHER"(%5), %%mm5 \n\t"
1178			WRITERGB15(%%FF_REGb, DSTW_OFFSET"(%5)", %%FF_REGBP)
1179			"pop %%"FF_REG_BP" \n\t"
1180			"mov "ESP_OFFSET"(%5), %%"FF_REG_b" \n\t"
1181			:: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),
1182		940	"a" (&c->redDither)
1183			NAMED_CONSTRAINTS_ADD(bF8)
1184			);
1185			} else {
1186		✗	const int16_t *ubuf1 = ubuf[1];
1187		✗	__asm__ volatile(
1188			"mov %%"FF_REG_b", "ESP_OFFSET"(%5) \n\t"
1189			"mov %4, %%"FF_REG_b" \n\t"
1190			"push %%"FF_REG_BP" \n\t"
1191			YSCALEYUV2RGB1b(%%FF_REGBP, %5)
1192			"pxor %%mm7, %%mm7 \n\t"
1193			/* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */
1194			"paddusb "BLUE_DITHER"(%5), %%mm2 \n\t"
1195			"paddusb "GREEN_DITHER"(%5), %%mm4 \n\t"
1196			"paddusb "RED_DITHER"(%5), %%mm5 \n\t"
1197			WRITERGB15(%%FF_REGb, DSTW_OFFSET"(%5)", %%FF_REGBP)
1198			"pop %%"FF_REG_BP" \n\t"
1199			"mov "ESP_OFFSET"(%5), %%"FF_REG_b" \n\t"
1200			:: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),
1201		✗	"a" (&c->redDither)
1202			NAMED_CONSTRAINTS_ADD(bF8)
1203			);
1204			}
1205		940	}
1206
1207		940	static void RENAME(yuv2rgb565_1)(SwsInternal *c, const int16_t *buf0,
1208			const int16_t *ubuf[2], const int16_t *vbuf[2],
1209			const int16_t *abuf0, uint8_t *dest,
1210			int dstW, int uvalpha, int y)
1211			{
1212		940	const int16_t *ubuf0 = ubuf[0];
1213		940	const int16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
1214
1215	1/2 ✓ Branch 0 taken 940 times. ✗ Branch 1 not taken.	940	if (uvalpha < 2048) { // note this is not correct (shifts chrominance by 0.5 pixels) but it is a bit faster
1216		940	const int16_t *ubuf1 = ubuf[0];
1217		940	__asm__ volatile(
1218			"mov %%"FF_REG_b", "ESP_OFFSET"(%5) \n\t"
1219			"mov %4, %%"FF_REG_b" \n\t"
1220			"push %%"FF_REG_BP" \n\t"
1221			YSCALEYUV2RGB1(%%FF_REGBP, %5)
1222			"pxor %%mm7, %%mm7 \n\t"
1223			/* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */
1224			"paddusb "BLUE_DITHER"(%5), %%mm2 \n\t"
1225			"paddusb "GREEN_DITHER"(%5), %%mm4 \n\t"
1226			"paddusb "RED_DITHER"(%5), %%mm5 \n\t"
1227			WRITERGB16(%%FF_REGb, DSTW_OFFSET"(%5)", %%FF_REGBP)
1228			"pop %%"FF_REG_BP" \n\t"
1229			"mov "ESP_OFFSET"(%5), %%"FF_REG_b" \n\t"
1230			:: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),
1231		940	"a" (&c->redDither)
1232			NAMED_CONSTRAINTS_ADD(bF8,bFC)
1233			);
1234			} else {
1235		✗	const int16_t *ubuf1 = ubuf[1];
1236		✗	__asm__ volatile(
1237			"mov %%"FF_REG_b", "ESP_OFFSET"(%5) \n\t"
1238			"mov %4, %%"FF_REG_b" \n\t"
1239			"push %%"FF_REG_BP" \n\t"
1240			YSCALEYUV2RGB1b(%%FF_REGBP, %5)
1241			"pxor %%mm7, %%mm7 \n\t"
1242			/* mm2=B, %%mm4=G, %%mm5=R, %%mm7=0 */
1243			"paddusb "BLUE_DITHER"(%5), %%mm2 \n\t"
1244			"paddusb "GREEN_DITHER"(%5), %%mm4 \n\t"
1245			"paddusb "RED_DITHER"(%5), %%mm5 \n\t"
1246			WRITERGB16(%%FF_REGb, DSTW_OFFSET"(%5)", %%FF_REGBP)
1247			"pop %%"FF_REG_BP" \n\t"
1248			"mov "ESP_OFFSET"(%5), %%"FF_REG_b" \n\t"
1249			:: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),
1250		✗	"a" (&c->redDither)
1251			NAMED_CONSTRAINTS_ADD(bF8,bFC)
1252			);
1253			}
1254		940	}
1255
1256			#define REAL_YSCALEYUV2PACKED1(index, c) \
1257			"xor "#index", "#index" \n\t"\
1258			".p2align 4 \n\t"\
1259			"1: \n\t"\
1260			"movq (%2, "#index"), %%mm3 \n\t" /* uvbuf0[eax]*/\
1261			"add "UV_OFF_BYTE"("#c"), "#index" \n\t" \
1262			"movq (%2, "#index"), %%mm4 \n\t" /* uvbuf0[eax+2048]*/\
1263			"sub "UV_OFF_BYTE"("#c"), "#index" \n\t" \
1264			"psraw $7, %%mm3 \n\t" \
1265			"psraw $7, %%mm4 \n\t" \
1266			"movq (%0, "#index", 2), %%mm1 \n\t" /*buf0[eax]*/\
1267			"movq 8(%0, "#index", 2), %%mm7 \n\t" /*buf0[eax]*/\
1268			"psraw $7, %%mm1 \n\t" \
1269			"psraw $7, %%mm7 \n\t" \
1270
1271			#define YSCALEYUV2PACKED1(index, c) REAL_YSCALEYUV2PACKED1(index, c)
1272
1273			#define REAL_YSCALEYUV2PACKED1b(index, c) \
1274			"xor "#index", "#index" \n\t"\
1275			".p2align 4 \n\t"\
1276			"1: \n\t"\
1277			"movq (%2, "#index"), %%mm2 \n\t" /* uvbuf0[eax]*/\
1278			"movq (%3, "#index"), %%mm3 \n\t" /* uvbuf1[eax]*/\
1279			"add "UV_OFF_BYTE"("#c"), "#index" \n\t" \
1280			"movq (%2, "#index"), %%mm5 \n\t" /* uvbuf0[eax+2048]*/\
1281			"movq (%3, "#index"), %%mm4 \n\t" /* uvbuf1[eax+2048]*/\
1282			"sub "UV_OFF_BYTE"("#c"), "#index" \n\t" \
1283			"paddw %%mm2, %%mm3 \n\t" /* uvbuf0[eax] + uvbuf1[eax]*/\
1284			"paddw %%mm5, %%mm4 \n\t" /* uvbuf0[eax+2048] + uvbuf1[eax+2048]*/\
1285			"psrlw $8, %%mm3 \n\t" \
1286			"psrlw $8, %%mm4 \n\t" \
1287			"movq (%0, "#index", 2), %%mm1 \n\t" /*buf0[eax]*/\
1288			"movq 8(%0, "#index", 2), %%mm7 \n\t" /*buf0[eax]*/\
1289			"psraw $7, %%mm1 \n\t" \
1290			"psraw $7, %%mm7 \n\t"
1291			#define YSCALEYUV2PACKED1b(index, c) REAL_YSCALEYUV2PACKED1b(index, c)
1292
1293		✗	static void RENAME(yuv2yuyv422_1)(SwsInternal *c, const int16_t *buf0,
1294			const int16_t *ubuf[2], const int16_t *vbuf[2],
1295			const int16_t *abuf0, uint8_t *dest,
1296			int dstW, int uvalpha, int y)
1297			{
1298		✗	const int16_t *ubuf0 = ubuf[0];
1299		✗	const int16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
1300
1301		✗	if (uvalpha < 2048) { // note this is not correct (shifts chrominance by 0.5 pixels) but it is a bit faster
1302		✗	const int16_t *ubuf1 = ubuf[0];
1303		✗	__asm__ volatile(
1304			"mov %%"FF_REG_b", "ESP_OFFSET"(%5) \n\t"
1305			"mov %4, %%"FF_REG_b" \n\t"
1306			"push %%"FF_REG_BP" \n\t"
1307			YSCALEYUV2PACKED1(%%FF_REGBP, %5)
1308			WRITEYUY2(%%FF_REGb, DSTW_OFFSET"(%5)", %%FF_REGBP)
1309			"pop %%"FF_REG_BP" \n\t"
1310			"mov "ESP_OFFSET"(%5), %%"FF_REG_b" \n\t"
1311			:: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),
1312		✗	"a" (&c->redDither)
1313			);
1314			} else {
1315		✗	const int16_t *ubuf1 = ubuf[1];
1316		✗	__asm__ volatile(
1317			"mov %%"FF_REG_b", "ESP_OFFSET"(%5) \n\t"
1318			"mov %4, %%"FF_REG_b" \n\t"
1319			"push %%"FF_REG_BP" \n\t"
1320			YSCALEYUV2PACKED1b(%%FF_REGBP, %5)
1321			WRITEYUY2(%%FF_REGb, DSTW_OFFSET"(%5)", %%FF_REGBP)
1322			"pop %%"FF_REG_BP" \n\t"
1323			"mov "ESP_OFFSET"(%5), %%"FF_REG_b" \n\t"
1324			:: "c" (buf0), "d" (buf1), "S" (ubuf0), "D" (ubuf1), "m" (dest),
1325		✗	"a" (&c->redDither)
1326			);
1327			}
1328		✗	}
1329