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
|
#include "mupdf/fitz.h"
#include "draw-imp.h"
#include <string.h>
/* Unpack image samples and optionally pad pixels with opaque alpha */
#define get1(buf,x) ((buf[x >> 3] >> ( 7 - (x & 7) ) ) & 1 )
#define get2(buf,x) ((buf[x >> 2] >> ( ( 3 - (x & 3) ) << 1 ) ) & 3 )
#define get4(buf,x) ((buf[x >> 1] >> ( ( 1 - (x & 1) ) << 2 ) ) & 15 )
#define get8(buf,x) (buf[x])
#define get16(buf,x) (buf[x << 1])
#define get24(buf,x) (buf[(x << 1) + x])
#define get32(buf,x) (buf[x << 2])
static unsigned char get1_tab_1[256][8];
static unsigned char get1_tab_1p[256][16];
static unsigned char get1_tab_255[256][8];
static unsigned char get1_tab_255p[256][16];
/*
Bug 697012 shows that the unpacking code can confuse valgrind due
to the use of undefined bits in the padding at the end of lines.
We unpack from bits to bytes by copying from a lookup table.
Valgrind is not capable of understanding that it doesn't matter
what the undefined bits are, as the bytes we copy that correspond
to the defined bits will always agree regardless of these
undefined bits by construction of the table.
We therefore have a VGMASK macro that explicitly masks off these
bits in PACIFY_VALGRIND builds.
*/
#ifdef PACIFY_VALGRIND
static const unsigned char mask[9] = { 0x00, 0x80, 0xc0, 0xe0, 0xf0, 0xf8, 0xfc, 0xfe, 0xff };
#define VGMASK(v,m) (v & mask[(m)])
#else
#define VGMASK(v,m) (v)
#endif
static void
init_get1_tables(void)
{
static int once = 0;
unsigned char bits[1];
int i, k, x;
/* TODO: mutex lock here */
if (once)
return;
for (i = 0; i < 256; i++)
{
bits[0] = i;
for (k = 0; k < 8; k++)
{
x = get1(bits, k);
get1_tab_1[i][k] = x;
get1_tab_1p[i][k * 2] = x;
get1_tab_1p[i][k * 2 + 1] = 255;
get1_tab_255[i][k] = x * 255;
get1_tab_255p[i][k * 2] = x * 255;
get1_tab_255p[i][k * 2 + 1] = 255;
}
}
once = 1;
}
void
fz_unpack_tile(fz_context *ctx, fz_pixmap *dst, unsigned char * restrict src, int n, int depth, size_t stride, int scale)
{
int pad, x, y, k, skip;
int w = dst->w;
pad = 0;
skip = 0;
if (dst->n > n)
pad = 255;
if (dst->n < n)
{
skip = n - dst->n;
n = dst->n;
}
if (depth == 1)
init_get1_tables();
if (scale == 0)
{
switch (depth)
{
case 1: scale = 255; break;
case 2: scale = 85; break;
case 4: scale = 17; break;
}
}
for (y = 0; y < dst->h; y++)
{
unsigned char *sp = src + (y * stride);
unsigned char *dp = dst->samples + (y * dst->stride);
/* Specialized loops */
if (n == 1 && depth == 1 && scale == 1 && !pad && !skip)
{
int w3 = w >> 3;
for (x = 0; x < w3; x++)
{
memcpy(dp, get1_tab_1[*sp++], 8);
dp += 8;
}
x = x << 3;
if (x < w)
memcpy(dp, get1_tab_1[VGMASK(*sp, w - x)], w - x);
}
else if (n == 1 && depth == 1 && scale == 255 && !pad && !skip)
{
int w3 = w >> 3;
for (x = 0; x < w3; x++)
{
memcpy(dp, get1_tab_255[*sp++], 8);
dp += 8;
}
x = x << 3;
if (x < w)
memcpy(dp, get1_tab_255[VGMASK(*sp, w - x)], w - x);
}
else if (n == 1 && depth == 1 && scale == 1 && pad && !skip)
{
int w3 = w >> 3;
for (x = 0; x < w3; x++)
{
memcpy(dp, get1_tab_1p[*sp++], 16);
dp += 16;
}
x = x << 3;
if (x < w)
memcpy(dp, get1_tab_1p[VGMASK(*sp, w - x)], (w - x) << 1);
}
else if (n == 1 && depth == 1 && scale == 255 && pad && !skip)
{
int w3 = w >> 3;
for (x = 0; x < w3; x++)
{
memcpy(dp, get1_tab_255p[*sp++], 16);
dp += 16;
}
x = x << 3;
if (x < w)
memcpy(dp, get1_tab_255p[VGMASK(*sp, w - x)], (w - x) << 1);
}
else if (depth == 8 && !pad && !skip)
{
int len = w * n;
while (len--)
*dp++ = *sp++;
}
else if (depth == 8 && pad && !skip)
{
for (x = 0; x < w; x++)
{
for (k = 0; k < n; k++)
*dp++ = *sp++;
*dp++ = 255;
}
}
else
{
int b = 0;
for (x = 0; x < w; x++)
{
for (k = 0; k < n; k++)
{
switch (depth)
{
case 1: *dp++ = get1(sp, b) * scale; break;
case 2: *dp++ = get2(sp, b) * scale; break;
case 4: *dp++ = get4(sp, b) * scale; break;
case 8: *dp++ = get8(sp, b); break;
case 16: *dp++ = get16(sp, b); break;
case 24: *dp++ = get24(sp, b); break;
case 32: *dp++ = get32(sp, b); break;
default:
fz_throw(ctx, FZ_ERROR_GENERIC, "cannot unpack tile with %d bits per component", depth);
}
b++;
}
b += skip;
if (pad)
*dp++ = 255;
}
}
}
}
/* Apply decode array */
void
fz_decode_indexed_tile(fz_context *ctx, fz_pixmap *pix, const float *decode, int maxval)
{
int add[FZ_MAX_COLORS];
int mul[FZ_MAX_COLORS];
unsigned char *p = pix->samples;
int stride = pix->stride - pix->w * pix->n;
int len;
int pn = pix->n;
int n = pn - pix->alpha;
int needed;
int k;
int h;
needed = 0;
for (k = 0; k < n; k++)
{
int min = decode[k * 2] * 256;
int max = decode[k * 2 + 1] * 256;
add[k] = min;
mul[k] = (max - min) / maxval;
needed |= min != 0 || max != maxval * 256;
}
if (!needed)
return;
h = pix->h;
while (h--)
{
len = pix->w;
while (len--)
{
for (k = 0; k < n; k++)
{
int value = (add[k] + (((p[k] << 8) * mul[k]) >> 8)) >> 8;
p[k] = fz_clampi(value, 0, 255);
}
p += pn;
}
p += stride;
}
}
void
fz_decode_tile(fz_context *ctx, fz_pixmap *pix, const float *decode)
{
int add[FZ_MAX_COLORS];
int mul[FZ_MAX_COLORS];
unsigned char *p = pix->samples;
int stride = pix->stride - pix->w * pix->n;
int len;
int n = fz_maxi(1, pix->n - pix->alpha);
int k;
int h;
for (k = 0; k < n; k++)
{
int min = decode[k * 2] * 255;
int max = decode[k * 2 + 1] * 255;
add[k] = min;
mul[k] = max - min;
}
h = pix->h;
while (h--)
{
len = pix->w;
while (len--)
{
for (k = 0; k < n; k++)
{
int value = add[k] + fz_mul255(p[k], mul[k]);
p[k] = fz_clampi(value, 0, 255);
}
p += pix->n;
}
p += stride;
}
}
|