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
366
367
368
369
370
371
372
|
/* Copyright (C) 2001-2023 Artifex Software, Inc.
All Rights Reserved.
This software is provided AS-IS with no warranty, either express or
implied.
This software is distributed under license and may not be copied,
modified or distributed except as expressly authorized under the terms
of the license contained in the file LICENSE in this distribution.
Refer to licensing information at http://www.artifex.com or contact
Artifex Software, Inc., 39 Mesa Street, Suite 108A, San Francisco,
CA 94129, USA, for further information.
*/
/* This file is repeatedly included by gsroprun.c to 'autogenerate' many
* different versions of roprun code. DO NOT USE THIS FILE EXCEPT FROM
* gsroprun.c.
*/
/* Set the following defines as appropriate on entry:
* TEMPLATE_NAME (Compulsory) The name of the function to generate
* SPECIFIC_ROP (Optional) If set, the function will base its decision
* about whether to provide S and T upon
* this value.
* SPECIFIC_CODE (Optional) If set, this should expand out to code to
* perform the rop. Will be invoked as:
* SPECIFIC_ROP(OUT,D,S,T)
* S_CONST (Optional) If set, S will be taken to be constant, else
* S will be read from a pointer.
* T_CONST (Optional) If set, T will be taken to be constant, else
* T will be read from a pointer.
*/
#if defined(TEMPLATE_NAME)
#ifdef SPECIFIC_ROP
#if rop3_uses_S(SPECIFIC_ROP)
#define S_USED
#endif
#if rop3_uses_T(SPECIFIC_ROP)
#define T_USED
#endif
#else /* !SPECIFIC_ROP */
#define S_USED
#define T_USED
#endif /* SPECIFIC_ROP */
/* We work in 'chunks' here; for bigendian machines, we can safely use
* chunks of 'int' size. For little endian machines where we have a cheap
* endian swap, we can do likewise. For others, we'll work at the byte
* level. */
#if !ARCH_IS_BIG_ENDIAN && !defined(ENDIAN_SWAP_INT)
#define CHUNKSIZE 8
#define CHUNK byte
#define CHUNKONES 255
#define ADJUST_TO_CHUNK(d,dpos) do {} while (0)
#else /* ARCH_IS_BIG_ENDIAN || defined(ENDIAN_SWAP_INT) */
#if ARCH_LOG2_SIZEOF_INT == 2
#define CHUNKSIZE 32
#define CHUNK unsigned int
#define CHUNKONES 0xFFFFFFFFU
#if ARCH_SIZEOF_PTR == (1<<ARCH_LOG2_SIZEOF_INT)
#define ROP_PTRDIFF_T int
#else
#define ROP_PTRDIFF_T int64_t
#endif
#define ADJUST_TO_CHUNK(d, dpos) \
do { int offset = ((ROP_PTRDIFF_T)d) & ((CHUNKSIZE>>3)-1); \
d = (CHUNK *)(void *)(((byte *)(void *)d)-offset); \
dpos += offset<<3; \
} while (0)
#else
/* FIXME: Write more code in here when we find an example. */
#endif
#endif /* ARCH_IS_BIG_ENDIAN || defined(ENDIAN_SWAP_INT) */
/* We define an 'RE' macro that reverses the endianness of a chunk, if we
* need it, and does nothing otherwise. */
#if !ARCH_IS_BIG_ENDIAN && defined(ENDIAN_SWAP_INT) && (CHUNKSIZE != 8)
#define RE(I) ((CHUNK)ENDIAN_SWAP_INT(I))
#else /* ARCH_IS_BIG_ENDIAN || !defined(ENDIAN_SWAP_INT) || (CHUNKSIZE == 8) */
#define RE(I) (I)
#endif /* ARCH_IS_BIG_ENDIAN || !defined(ENDIAN_SWAP_INT) || (CHUNKSIZE == 8) */
/* In some cases we will need to fetch values from a pointer, and 'skew'
* them. We need 2 variants of this macro. One that is 'SAFE' to use when
* SKEW might be 0, and one that can be faster, because we know that SKEW
* is non zero. */
#define SKEW_FETCH(S,s,SKEW) \
do { S = RE((RE(s[0])<<SKEW) | (RE(s[1])>>(CHUNKSIZE-SKEW))); s++; } while (0)
#define SAFE_SKEW_FETCH(S,s,SKEW,L,R) \
do { S = RE(((L) ? 0 : (RE(s[0])<<SKEW)) | ((R) ? 0 : (RE(s[1])>>(CHUNKSIZE-SKEW)))); s++; } while (0)
#if defined(S_USED) && !defined(S_CONST)
#define S_SKEW
#define FETCH_S SKEW_FETCH(S,s,s_skew)
#define SAFE_FETCH_S(L,R) SAFE_SKEW_FETCH(S,s,s_skew,L,R)
#else /* !defined(S_USED) || defined(S_CONST) */
#define FETCH_S
#define SAFE_FETCH_S(L,R)
#endif /* !defined(S_USED) || defined(S_CONST) */
#if defined(T_USED) && !defined(T_CONST)
#define T_SKEW
#define FETCH_T SKEW_FETCH(T,t,t_skew)
#define SAFE_FETCH_T(L,R) SAFE_SKEW_FETCH(T,t,t_skew,L,R)
#else /* !defined(T_USED) || defined(T_CONST) */
#define FETCH_T
#define SAFE_FETCH_T(L,R)
#endif /* !defined(T_USED) || defined(T_CONST) */
static void TEMPLATE_NAME(rop_run_op *op, byte *d_, int len)
{
#ifndef SPECIFIC_CODE
rop_proc proc = rop_proc_table[op->rop];
#define SPECIFIC_CODE(OUT_, D_,S_,T_) OUT_ = proc(D_,S_,T_)
#endif /* !defined(SPECIFIC_CODE) */
CHUNK lmask, rmask;
#ifdef S_USED
#ifdef S_CONST
CHUNK S = (CHUNK)op->s.c;
#else /* !defined(S_CONST) */
const CHUNK *s = (CHUNK *)(void *)op->s.b.ptr;
CHUNK S;
int s_skew;
#endif /* !defined(S_CONST) */
#else /* !defined(S_USED) */
#define S 0
#undef S_CONST
#endif /* !defined(S_USED) */
#ifdef T_USED
#ifdef T_CONST
CHUNK T = (CHUNK)op->t.c;
#else /* !defined(T_CONST) */
const CHUNK *t = (CHUNK *)(void *)op->t.b.ptr;
CHUNK T;
int t_skew;
#endif /* !defined(T_CONST) */
#else /* !defined(T_USED) */
#define T 0
#undef T_CONST
#endif /* !defined(T_USED) */
#if defined(S_SKEW) || defined(T_SKEW)
int skewflags = 0;
#endif
CHUNK D;
int dpos = op->dpos;
CHUNK *d = (CHUNK *)(void *)d_;
/* Align d to CHUNKSIZE */
ADJUST_TO_CHUNK(d,dpos);
/* On entry len = length in 'depth' chunks. Change it to be the length
* in bits, and add on the number of bits we skip at the start of the
* run. */
len = len * op->depth + dpos;
/* lmask = the set of bits to alter in the output bitmap on the left
* hand edge of the run. rmask = the set of bits NOT to alter in the
* output bitmap on the right hand edge of the run. */
lmask = RE((CHUNKONES>>((CHUNKSIZE-1) & dpos)));
rmask = RE((CHUNKONES>>((CHUNKSIZE-1) & len)));
if (rmask == CHUNKONES) rmask = 0;
#if defined(S_CONST) || defined(T_CONST)
/* S and T should be supplied as 'depth' bits. Duplicate them up to be
* byte size (if they are supplied byte sized, that's fine too). */
if (op->depth & 1) {
#ifdef S_CONST
S |= S<<1;
#endif /* !defined(S_CONST) */
#ifdef T_CONST
T |= T<<1;
#endif /* !defined(T_CONST) */
}
if (op->depth & 3) {
#ifdef S_CONST
S |= S<<2;
#endif /* !defined(S_CONST) */
#ifdef T_CONST
T |= T<<2;
#endif /* !defined(T_CONST) */
}
if (op->depth & 7) {
#ifdef S_CONST
S |= S<<4;
#endif /* !defined(S_CONST) */
#ifdef T_CONST
T |= T<<4;
#endif /* !defined(T_CONST) */
}
#if CHUNKSIZE > 8
if (op->depth & 15) {
#ifdef S_CONST
S |= S<<8;
#endif /* !defined(S_CONST) */
#ifdef T_CONST
T |= T<<8;
#endif /* !defined(T_CONST) */
}
#endif /* CHUNKSIZE > 8 */
#if CHUNKSIZE > 16
if (op->depth & 31) {
#ifdef S_CONST
S |= S<<16;
#endif /* !defined(S_CONST) */
#ifdef T_CONST
T |= T<<16;
#endif /* !defined(T_CONST) */
}
#endif /* CHUNKSIZE > 16 */
#endif /* defined(S_CONST) || defined(T_CONST) */
/* Note #1: This mirrors what the original code did, but I think it has
* the risk of moving s and t back beyond officially allocated space. We
* may be saved by the fact that all blocks have a word or two in front
* of them due to the allocator. If we ever get valgrind properly marking
* allocated blocks as readable etc, then this may throw some spurious
* errors. RJW. */
#ifdef S_SKEW
{
int slen, slen2;
int spos = op->s.b.pos;
ADJUST_TO_CHUNK(s, spos);
s_skew = spos - dpos;
if (s_skew < 0) {
s_skew += CHUNKSIZE;
s--;
skewflags |= 1; /* Suppress reading off left edge */
}
/* We are allowed to read all the data bits, so: len - dpos + tpos
* We're allowed to read in CHUNKS, so: CHUNKUP(len-dpos+tpos).
* This code will actually read CHUNKUP(len)+CHUNKSIZE bits. If
* This is larger, then suppress. */
slen = (len + s_skew + CHUNKSIZE-1) & ~(CHUNKSIZE-1);
slen2 = (len + CHUNKSIZE + CHUNKSIZE-1) & ~(CHUNKSIZE-1);
if ((s_skew == 0) || (slen < slen2)) {
skewflags |= 4; /* Suppress reading off the right edge */
}
}
#endif /* !defined(S_SKEW) */
#ifdef T_SKEW
{
int tlen, tlen2;
int tpos = op->t.b.pos;
ADJUST_TO_CHUNK(t, tpos);
t_skew = tpos - dpos;
if (t_skew < 0) {
t_skew += CHUNKSIZE;
t--;
skewflags |= 2; /* Suppress reading off left edge */
}
/* We are allowed to read all the data bits, so: len - dpos + tpos
* We're allowed to read in CHUNKS, so: CHUNKUP(len-dpos+tpos).
* This code will actually read CHUNKUP(len)+CHUNKSIZE bits. If
* This is larger, then suppress. */
tlen = (len + t_skew + CHUNKSIZE-1) & ~(CHUNKSIZE-1);
tlen2 = (len + CHUNKSIZE + CHUNKSIZE-1) & ~(CHUNKSIZE-1);
if ((t_skew == 0) || (tlen < tlen2)) {
skewflags |= 8; /* Suppress reading off the right edge */
}
}
#endif /* !defined(T_SKEW) */
len -= CHUNKSIZE; /* len = bytes to do - CHUNKSIZE */
/* len <= 0 means 1 word or less to do */
if (len <= 0) {
/* Short case - starts and ends in the same chunk */
lmask &= ~rmask; /* Combined mask = bits to alter */
SAFE_FETCH_S(skewflags & 1,skewflags & 4);
SAFE_FETCH_T(skewflags & 2,skewflags & 8);
SPECIFIC_CODE(D, *d, S, T);
*d = (*d & ~lmask) | (D & lmask);
return;
}
if ((lmask != CHUNKONES)
#if defined(S_SKEW) || defined(T_SKEW)
|| (skewflags & 3)
#endif
) {
/* Unaligned left hand case */
SAFE_FETCH_S(skewflags & 1,s_skew == 0);
SAFE_FETCH_T(skewflags & 2,t_skew == 0);
SPECIFIC_CODE(D, *d, S, T);
*d = (*d & ~lmask) | (D & lmask);
d++;
len -= CHUNKSIZE;
}
if (len > 0) {
/* Simple middle case (complete destination chunks). */
#ifdef S_SKEW
if (s_skew == 0) {
#ifdef T_SKEW
if (t_skew == 0) {
do {
SPECIFIC_CODE(*d, *d, *s++, *t++);
d++;
len -= CHUNKSIZE;
} while (len > 0);
} else
#endif /* !defined(T_SKEW) */
{
do {
FETCH_T;
SPECIFIC_CODE(*d, *d, *s++, T);
d++;
len -= CHUNKSIZE;
} while (len > 0);
}
} else
#endif /* !defined(S_SKEW) */
{
#ifdef T_SKEW
if (t_skew == 0) {
do {
FETCH_S;
SPECIFIC_CODE(*d, *d, S, *t++);
d++;
len -= CHUNKSIZE;
} while (len > 0);
} else
#endif /* !defined(T_SKEW) */
{
do {
FETCH_S;
FETCH_T;
SPECIFIC_CODE(*d, *d, S, T);
d++;
len -= CHUNKSIZE;
} while (len > 0);
}
}
}
/* Unaligned right hand case */
SAFE_FETCH_S(0,skewflags & 4);
SAFE_FETCH_T(0,skewflags & 8);
SPECIFIC_CODE(D, *d, S, T);
*d = (*d & rmask) | (D & ~rmask);
}
#undef ADJUST_TO_CHUNK
#undef CHUNKSIZE
#undef CHUNK
#undef CHUNKONES
#undef FETCH_S
#undef FETCH_T
#undef SAFE_FETCH_S
#undef SAFE_FETCH_T
#undef RE
#undef S
#undef S_USED
#undef S_CONST
#undef S_SKEW
#undef SKEW_FETCH
#undef SAFE_SKEW_FETCH
#undef SPECIFIC_CODE
#undef SPECIFIC_ROP
#undef T
#undef T_USED
#undef T_CONST
#undef T_SKEW
#undef TEMPLATE_NAME
#undef ROP_PTRDIFF_T
#else
int dummy;
#endif
|
