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
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
|
// SPDX-License-Identifier: GPL-2.0-only
/*
* Copyright (C) 2024, SUSE LLC
*
* Authors: Enzo Matsumiya <ematsumiya@suse.de>
*
* Implementation of the LZ77 "plain" compression algorithm, as per MS-XCA spec.
*/
#include <linux/slab.h>
#include <linux/sizes.h>
#include <linux/count_zeros.h>
#include <linux/unaligned.h>
#include "lz77.h"
/*
* Compression parameters.
*/
#define LZ77_MATCH_MIN_LEN 3
#define LZ77_MATCH_MIN_DIST 1
#define LZ77_MATCH_MAX_DIST SZ_1K
#define LZ77_HASH_LOG 15
#define LZ77_HASH_SIZE (1 << LZ77_HASH_LOG)
#define LZ77_STEP_SIZE sizeof(u64)
static __always_inline u8 lz77_read8(const u8 *ptr)
{
return get_unaligned(ptr);
}
static __always_inline u16 lz77_read16(const u16 *ptr)
{
return get_unaligned(ptr);
}
static __always_inline u32 lz77_read32(const u32 *ptr)
{
return get_unaligned(ptr);
}
static __always_inline u64 lz77_read64(const u64 *ptr)
{
return get_unaligned(ptr);
}
static __always_inline void lz77_write8(u8 *ptr, u8 v)
{
put_unaligned(v, ptr);
}
static __always_inline void lz77_write16(u16 *ptr, u16 v)
{
put_unaligned_le16(v, ptr);
}
static __always_inline void lz77_write32(u32 *ptr, u32 v)
{
put_unaligned_le32(v, ptr);
}
static __always_inline u32 lz77_match_len(const void *wnd, const void *cur, const void *end)
{
const void *start = cur;
/*
* Safe for a do/while (i.e. no bounds check the first iteration) because otherwise we
* wouldn't reach here from the main loop.
*/
do {
const u64 diff = lz77_read64(cur) ^ lz77_read64(wnd);
if (!diff) {
cur += LZ77_STEP_SIZE;
wnd += LZ77_STEP_SIZE;
continue;
}
/* This computes the number of common bytes in @diff. */
cur += count_trailing_zeros(diff) >> 3;
return (cur - start);
} while (likely(cur + LZ77_STEP_SIZE < end));
while (cur < end && lz77_read8(cur) == lz77_read8(wnd)) {
cur++;
wnd++;
}
return (cur - start);
}
static __always_inline void *lz77_write_match(void *dst, void **nib, u32 dist, u32 len)
{
len -= 3;
dist--;
dist <<= 3;
if (len < 7) {
lz77_write16(dst, dist + len);
return dst + 2;
}
dist |= 7;
lz77_write16(dst, dist);
dst += 2;
len -= 7;
if (!*nib) {
lz77_write8(dst, umin(len, 15));
*nib = dst;
dst++;
} else {
u8 *b = *nib;
lz77_write8(b, *b | umin(len, 15) << 4);
*nib = NULL;
}
if (len < 15)
return dst;
len -= 15;
if (len < 255) {
lz77_write8(dst, len);
return dst + 1;
}
lz77_write8(dst, 0xff);
dst++;
len += 7 + 15;
if (len <= 0xffff) {
lz77_write16(dst, len);
return dst + 2;
}
lz77_write16(dst, 0);
dst += 2;
lz77_write32(dst, len);
return dst + 4;
}
noinline int lz77_compress(const void *src, u32 slen, void *dst, u32 *dlen)
{
const void *srcp, *end, *anchor;
void *dstp, *nib, *flag_pos, *dend;
u32 flag_count = 0;
long flag = 0;
u64 *htable;
srcp = src;
anchor = srcp;
end = src + slen;
dstp = dst;
nib = NULL;
flag_pos = dstp;
dstp += 4;
dend = dst + (*dlen - (*dlen >> 3));
htable = kvcalloc(LZ77_HASH_SIZE, sizeof(*htable), GFP_KERNEL);
if (!htable)
return -ENOMEM;
#if 0
d = 4;
/* warm up loop */
do {
u32 dist, len = 0;
const void *wnd;
u64 hash;
hash = ((lz77_read64(srcp) << 24) * 889523592379ULL) >> (64 - LZ77_HASH_LOG);
wnd = src + htable[hash];
htable[hash] = srcp - src;
dist = srcp - wnd;
if (dist && dist < LZ77_MATCH_MAX_DIST)
len = lz77_match_len(wnd, srcp, end);
if (len < LZ77_MATCH_MIN_LEN) {
srcp++;
d++;
continue;
}
d += 4;
srcp += len;
} while (likely(srcp + LZ77_STEP_SIZE < end));
if (srcp < end)
d += (end - srcp);
if (d > (slen - (slen >> 2))) {
*dlen = slen;
goto out;
}
srcp = src;
anchor = srcp;
#endif
/* Main loop. */
do {
u32 dist, len = 0;
const void *wnd;
u64 hash;
#if 0
/*
* Bail out if @dstp reached >= 7/8 of @slen -- already compressed badly, not worth
* going further.
*/
if (unlikely(dstp - dst >= slen - (slen >> 3))) {
*dlen = slen;
goto out;
}
#endif
hash = ((lz77_read64(srcp) << 24) * 889523592379ULL) >> (64 - LZ77_HASH_LOG);
wnd = src + htable[hash];
htable[hash] = srcp - src;
dist = srcp - wnd;
if (dist && dist < LZ77_MATCH_MAX_DIST)
len = lz77_match_len(wnd, srcp, end);
#if 1
if (len < LZ77_MATCH_MIN_LEN) {
srcp++;
continue;
}
if (dstp + (srcp - anchor) >= dend)
goto out;
while (anchor < srcp) {
u32 c = umin(srcp - anchor, 32 - flag_count);
memcpy(dstp, anchor, c);
anchor += c;
dstp += c;
flag <<= c;
flag_count += c;
if (flag_count == 32) {
lz77_write32(flag_pos, flag);
flag_count = 0;
flag_pos = dstp;
dstp += 4;
}
}
#else
if (len < LZ77_MATCH_MIN_LEN) {
lz77_write8(dstp, lz77_read8(srcp));
dstp++;
srcp++;
if (dstp >= dend)
goto out;
flag <<= 1;
flag_count++;
if (flag_count == 32) {
lz77_write32(flag_pos, flag);
flag_count = 0;
flag_pos = dstp;
dstp += 4;
}
continue;
}
#endif
if (dstp + len >= dend)
goto out;
dstp = lz77_write_match(dstp, &nib, dist, len);
srcp += len;
anchor = srcp;
flag = (flag << 1) | 1;
flag_count++;
if (flag_count == 32) {
lz77_write32(flag_pos, flag);
flag_count = 0;
flag_pos = dstp;
dstp += 4;
}
} while (likely(srcp + LZ77_STEP_SIZE < end));
while (anchor < srcp) {
u32 c = umin(srcp - anchor, 32 - flag_count);
if (dstp + c >= dend)
goto out;
memcpy(dstp, anchor, c);
anchor += c;
dstp += c;
flag <<= c;
flag_count += c;
if (flag_count == 32) {
lz77_write32(flag_pos, flag);
flag_count = 0;
flag_pos = dstp;
dstp += 4;
}
}
while (srcp < end) {
u32 c = umin(end - srcp, 32 - flag_count);
if (dstp + c >= dend)
goto out;
memcpy(dstp, srcp, c);
dstp += c;
srcp += c;
flag <<= c;
flag_count += c;
if (flag_count == 32) {
lz77_write32(flag_pos, flag);
flag_count = 0;
flag_pos = dstp;
dstp += 4;
}
}
flag <<= (32 - flag_count);
flag |= (1UL << (32 - flag_count)) - 1;
lz77_write32(flag_pos, flag);
*dlen = dstp - dst;
out:
kvfree(htable);
if (*dlen < slen)
return 0;
return -EMSGSIZE;
}
static __always_inline const void *lz77_read_match_len(const void *src, const void **nib, u32 *dist,
u32 *len)
{
u32 mlen;
u16 sym;
sym = lz77_read16(src);
src += 2;
*dist = (sym >> 3) + 1;
mlen = sym & 7;
if (mlen == 7) {
if (!*nib) {
*nib = src;
mlen = lz77_read8(src) % 16;
src++;
} else {
mlen = lz77_read8(*nib) >> 4;
*nib = NULL;
}
if (mlen == 15) {
mlen = lz77_read8(src);
src++;
if (mlen == 255) {
mlen = lz77_read16(src);
src += 2;
if (mlen == 0) {
mlen = lz77_read32(src);
src += 4;
}
if (mlen < 15 + 7) {
pr_warn("CIFS: unexpected match length %u\n", mlen);
return NULL;
}
mlen -= (15 + 7)
|