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
|
/* Framebuffer Graphics Libary for Linux, Copyright 1993 Harm Hanemaayer */
/* scale.c Scaling routine */
#include <stdlib.h>
#include <vga.h>
#include "inlstring.h" /* include inline string operations */
#include "vgagl.h"
#include "def.h"
#if defined (NO_ASSEMBLY)
#undef USE_ASM
#else
/*
* XXX: The asm code needs fixing
#define USE_ASM
*/
#undef USE_ASM
#endif
#ifdef NO_ASSEMBLY
static inline int muldiv64(int m1, int m2, int d)
{
return (long) m1 *(long) m2 / (long) d;
}
#else
/* We use the 32-bit to 64-bit multiply and 64-bit to 32-bit divide of the */
/* 386 (which gcc doesn't know well enough) to efficiently perform integer */
/* scaling without having to worry about overflows. */
static inline int muldiv64(int m1, int m2, int d)
{
/* int32 * int32 -> int64 / int32 -> int32 */
int result;
int dummy;
__asm__(
"imull %%edx\n\t"
"idivl %4\n\t"
: "=a"(result), "=d"(dummy) /* out */
: "0"(m1), "1"(m2), "g"(d) /* in */
/***rjr***: "ax", "dx"***/ /* mod */
);
return result;
}
#endif
/* This is a DDA-based algorithm. */
/* Iteration over target bitmap. */
void gl_scalebox(int w1, int h1, void *_dp1, int w2, int h2, void *_dp2)
{
uchar *dp1 = _dp1;
uchar *dp2 = _dp2;
int xfactor;
int yfactor;
if (w2 == 0 || h2 == 0)
return;
xfactor = muldiv64(w1, 65536, w2); /* scaled by 65536 */
yfactor = muldiv64(h1, 65536, h2); /* scaled by 65536 */
switch (BYTESPERPIXEL) {
case 1:
{
int y, sy;
sy = 0;
for (y = 0; y < h2;) {
int sx = 0;
uchar *dp2old = dp2;
int x;
x = 0;
while (x < w2 - 8) {
#ifdef USE_ASM
/* This saves just a couple of cycles per */
/* pixel on a 486, but I couldn't resist. */
__asm__ __volatile__("movl %4, %%eax\n\t"
"shrl $16, %%eax\n\t"
"addl %5, %4\n\t"
"movb (%3, %%eax), %%al\n\t"
"movb %%al, (%1, %2)\n\t"
"movl %4, %%eax\n\t"
"shrl $16, %%eax\n\t"
"addl %5, %4\n\t"
"movb (%3, %%eax), %%al\n\t"
"movb %%al, 1 (%1, %2)\n\t"
"movl %4, %%eax\n\t"
"shrl $16, %%eax\n\t"
"addl %5, %4\n\t"
"movb (%3, %%eax), %%al\n\t"
"movb %%al, 2 (%1, %2)\n\t"
"movl %4, %%eax\n\t"
"shrl $16, %%eax\n\t"
"addl %5, %4\n\t"
"movb (%3, %%eax), %%al\n\t"
"movb %%al, 3 (%1, %2)\n\t"
"movl %4, %%eax\n\t"
"shrl $16, %%eax\n\t"
"addl %5, %4\n\t"
"movb (%3, %%eax), %%al\n\t"
"movb %%al, 4 (%1, %2)\n\t"
"movl %4, %%eax\n\t"
"shrl $16, %%eax\n\t"
"addl %5, %4\n\t"
"movb (%3, %%eax), %%al\n\t"
"movb %%al, 5 (%1, %2)\n\t"
"movl %4, %%eax\n\t"
"shrl $16, %%eax\n\t"
"addl %5, %4\n\t"
"movb (%3, %%eax), %%al\n\t"
"movb %%al, 6 (%1, %2)\n\t"
"movl %4, %%eax\n\t"
"shrl $16, %%eax\n\t"
"addl %5, %4\n\t"
"movb (%3, %%eax), %%al\n\t"
"movb %%al, 7 (%1, %2)\n\t"
: /* output */
: /* input */
"ax"(0), "r"(dp2), "r"(x), "r"(dp1),
"r"(sx), "r"(xfactor)
:"ax", "4"
);
#else
*(dp2 + x) = *(dp1 + (sx >> 16));
sx += xfactor;
*(dp2 + x + 1) = *(dp1 + (sx >> 16));
sx += xfactor;
*(dp2 + x + 2) = *(dp1 + (sx >> 16));
sx += xfactor;
*(dp2 + x + 3) = *(dp1 + (sx >> 16));
sx += xfactor;
*(dp2 + x + 4) = *(dp1 + (sx >> 16));
sx += xfactor;
*(dp2 + x + 5) = *(dp1 + (sx >> 16));
sx += xfactor;
*(dp2 + x + 6) = *(dp1 + (sx >> 16));
sx += xfactor;
*(dp2 + x + 7) = *(dp1 + (sx >> 16));
sx += xfactor;
#endif
x += 8;
}
while (x < w2) {
*(dp2 + x) = *(dp1 + (sx >> 16));
sx += xfactor;
x++;
}
dp2 += w2;
y++;
while (y < h2) {
int l;
int syint = sy >> 16;
sy += yfactor;
if ((sy >> 16) != syint)
break;
/* Copy identical lines. */
l = dp2 - dp2old;
__memcpy(dp2, dp2old, l);
dp2old = dp2;
dp2 += l;
y++;
}
dp1 = _dp1 + (sy >> 16) * w1;
}
}
break;
case 2:
{
int y, sy;
sy = 0;
for (y = 0; y < h2;) {
int sx = 0;
uchar *dp2old = dp2;
int x;
x = 0;
/* This can be greatly optimized with loop */
/* unrolling; omitted to save space. */
while (x < w2) {
*(unsigned short *) (dp2 + x * 2) =
*(unsigned short *) (dp1 + (sx >> 16) * 2);
sx += xfactor;
x++;
}
dp2 += w2 * 2;
y++;
while (y < h2) {
int l;
int syint = sy >> 16;
sy += yfactor;
if ((sy >> 16) != syint)
break;
/* Copy identical lines. */
l = dp2 - dp2old;
__memcpy(dp2, dp2old, l);
dp2old = dp2;
dp2 += l;
y++;
}
dp1 = _dp1 + (sy >> 16) * w1 * 2;
}
}
break;
case 3:
{
int y, sy;
sy = 0;
for (y = 0; y < h2;) {
int sx = 0;
uchar *dp2old = dp2;
int x;
x = 0;
/* This can be greatly optimized with loop */
/* unrolling; omitted to save space. */
while (x < w2) {
*(unsigned short *) (dp2 + x * 3) =
*(unsigned short *) (dp1 + (sx >> 16) * 3);
*(unsigned char *) (dp2 + x * 3 + 2) =
*(unsigned char *) (dp1 + (sx >> 16) * 3 + 2);
sx += xfactor;
x++;
}
dp2 += w2 * 3;
y++;
while (y < h2) {
int l;
int syint = sy >> 16;
sy += yfactor;
if ((sy >> 16) != syint)
break;
/* Copy identical lines. */
l = dp2 - dp2old;
__memcpy(dp2, dp2old, l);
dp2old = dp2;
dp2 += l;
y++;
}
dp1 = _dp1 + (sy >> 16) * w1 * 3;
}
}
break;
case 4:
{
int y, sy;
sy = 0;
for (y = 0; y < h2;) {
int sx = 0;
uchar *dp2old = dp2;
int x;
x = 0;
/* This can be greatly optimized with loop */
/* unrolling; omitted to save space. */
while (x < w2) {
*(unsigned *) (dp2 + x * 4) =
*(unsigned *) (dp1 + (sx >> 16) * 4);
sx += xfactor;
x++;
}
dp2 += w2 * 4;
y++;
while (y < h2) {
int l;
int syint = sy >> 16;
sy += yfactor;
if ((sy >> 16) != syint)
break;
/* Copy identical lines. */
l = dp2 - dp2old;
__memcpy(dp2, dp2old, l);
dp2old = dp2;
dp2 += l;
y++;
}
dp1 = _dp1 + (sy >> 16) * w1 * 4;
}
}
break;
}
}
|
