1 /* utimecmp.c -- compare file timestamps
2
3 Copyright (C) 2004-2007, 2009-2021 Free Software Foundation, Inc.
4
5 This program is free software: you can redistribute it and/or modify
6 it under the terms of the GNU General Public License as published by
7 the Free Software Foundation; either version 3 of the License, or
8 (at your option) any later version.
9
10 This program is distributed in the hope that it will be useful,
11 but WITHOUT ANY WARRANTY; without even the implied warranty of
12 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 GNU General Public License for more details.
14
15 You should have received a copy of the GNU General Public License
16 along with this program. If not, see <https://www.gnu.org/licenses/>. */
17
18 /* Written by Paul Eggert. */
19
20 #include <config.h>
21
22 #include "utimecmp.h"
23
24 #include <fcntl.h>
25 #include <limits.h>
26 #include <stdbool.h>
27 #include <stdint.h>
28 #include <stdlib.h>
29 #include <sys/stat.h>
30 #include <time.h>
31 #include <unistd.h>
32
33 #include "dirname.h"
34 #include "hash.h"
35 #include "intprops.h"
36 #include "stat-time.h"
37 #include "verify.h"
38
39 #ifndef MAX
40 # define MAX(a, b) ((a) > (b) ? (a) : (b))
41 #endif
42
43 #define BILLION (1000 * 1000 * 1000)
44
45 /* Best possible resolution that utimens can set and stat can return,
46 due to system-call limitations. It must be a power of 10 that is
47 no greater than 1 billion. */
48 #if HAVE_UTIMENSAT
49 enum { SYSCALL_RESOLUTION = 1 };
50 #elif defined _WIN32 && ! defined __CYGWIN__
51 /* On native Windows, file times have 100 ns resolution. See
52 <https://docs.microsoft.com/en-us/windows/desktop/api/minwinbase/ns-minwinbase-filetime> */
53 enum { SYSCALL_RESOLUTION = 100 };
54 #elif ((HAVE_FUTIMESAT || HAVE_WORKING_UTIMES) \
55 && (defined HAVE_STRUCT_STAT_ST_ATIM_TV_NSEC \
56 || defined HAVE_STRUCT_STAT_ST_ATIMESPEC_TV_NSEC \
57 || defined HAVE_STRUCT_STAT_ST_ATIMENSEC \
58 || defined HAVE_STRUCT_STAT_ST_ATIM_ST__TIM_TV_NSEC \
59 || defined HAVE_STRUCT_STAT_ST_SPARE1))
60 enum { SYSCALL_RESOLUTION = 1000 };
61 #else
62 enum { SYSCALL_RESOLUTION = BILLION };
63 #endif
64
65 /* Describe a file system and its timestamp resolution in nanoseconds. */
66 struct fs_res
67 {
68 /* Device number of file system. */
69 dev_t dev;
70
71 /* An upper bound on the timestamp resolution of this file system,
72 ignoring any resolution that cannot be set via utimens. It is
73 represented by an integer count of nanoseconds. It must be
74 either 2 billion, or a power of 10 that is no greater than a
75 billion and is no less than SYSCALL_RESOLUTION. */
76 int resolution;
77
78 /* True if RESOLUTION is known to be exact, and is not merely an
79 upper bound on the true resolution. */
80 bool exact;
81 };
82
83 /* Hash some device info. */
84 static size_t
85 dev_info_hash (void const *x, size_t table_size)
/* ![[previous]](../icons/n_left.png)
![[next]](../icons/right.png)
![[first]](../icons/n_first.png)
![[last]](../icons/last.png)
![[top]](../icons/top.png)
![[bottom]](../icons/bottom.png)
![[index]](../icons/index.png)
*/
86 {
87 struct fs_res const *p = x;
88
89 /* Beware signed arithmetic gotchas. */
90 if (TYPE_SIGNED (dev_t) && SIZE_MAX < MAX (INT_MAX, TYPE_MAXIMUM (dev_t)))
91 {
92 uintmax_t dev = p->dev;
93 return dev % table_size;
94 }
95
96 return p->dev % table_size;
97 }
98
99 /* Compare two dev_info structs. */
100 static bool
101 dev_info_compare (void const *x, void const *y)
/* ![[previous]](../icons/left.png)
![[next]](../icons/right.png)
![[first]](../icons/first.png)
![[last]](../icons/last.png)
![[top]](../icons/top.png)
![[bottom]](../icons/bottom.png)
![[index]](../icons/index.png)
*/
102 {
103 struct fs_res const *a = x;
104 struct fs_res const *b = y;
105 return a->dev == b->dev;
106 }
107
108 /* Return -1, 0, 1 based on whether the destination file (relative
109 to openat-like directory file descriptor DFD with name
110 DST_NAME and status DST_STAT) is older than SRC_STAT, the same age
111 as SRC_STAT, or newer than SRC_STAT, respectively.
112
113 DST_NAME may be NULL if OPTIONS is 0.
114
115 If OPTIONS & UTIMECMP_TRUNCATE_SOURCE, do the comparison after SRC is
116 converted to the destination's timestamp resolution as filtered through
117 utimens. In this case, return -2 if the exact answer cannot be
118 determined; this can happen only if the timestamps are very close and
119 there is some trouble accessing the file system (e.g., the user does not
120 have permission to futz with the destination's timestamps). */
121
122 int
123 utimecmp (char const *dst_name,
/* ![[previous]](../icons/left.png)
![[next]](../icons/right.png)
![[first]](../icons/first.png)
![[last]](../icons/last.png)
![[top]](../icons/top.png)
![[bottom]](../icons/bottom.png)
![[index]](../icons/index.png)
*/
124 struct stat const *dst_stat,
125 struct stat const *src_stat,
126 int options)
127 {
128 return utimecmpat (AT_FDCWD, dst_name, dst_stat, src_stat, options);
129 }
130
131 int
132 utimecmpat (int dfd, char const *dst_name,
/* ![[previous]](../icons/left.png)
![[next]](../icons/n_right.png)
![[first]](../icons/first.png)
![[last]](../icons/n_last.png)
![[top]](../icons/top.png)
![[bottom]](../icons/bottom.png)
![[index]](../icons/index.png)
*/
133 struct stat const *dst_stat,
134 struct stat const *src_stat,
135 int options)
136 {
137 /* Things to watch out for:
138
139 The code uses a static hash table internally and is not safe in the
140 presence of signals, multiple threads, etc. However, memory pressure
141 that prevents use of the hash table is not fatal - we just fall back
142 to redoing the computations on every call in that case.
143
144 int and long int might be 32 bits. Many of the calculations store
145 numbers up to 2 billion, and multiply by 10; they have to avoid
146 multiplying 2 billion by 10, as this exceeds 32-bit capabilities.
147
148 time_t might be unsigned. */
149
150 verify (TYPE_IS_INTEGER (time_t));
151
152 /* Destination and source timestamps. */
153 time_t dst_s = dst_stat->st_mtime;
154 time_t src_s = src_stat->st_mtime;
155 int dst_ns = get_stat_mtime_ns (dst_stat);
156 int src_ns = get_stat_mtime_ns (src_stat);
157
158 if (options & UTIMECMP_TRUNCATE_SOURCE)
159 {
160 #if defined _AIX
161 /* On AIX 7.2, on a jfs2 file system, the times may differ by up to
162 0.01 seconds in either direction. But it does not seem to come
163 from clock ticks of 0.01 seconds each. */
164 long long difference =
165 ((long long) dst_s - (long long) src_s) * BILLION
166 + ((long long) dst_ns - (long long) src_ns);
167 if (difference < 10000000 && difference > -10000000)
168 return 0;
169 #endif
170
171 /* Look up the timestamp resolution for the destination device. */
172
173 /* Hash table for caching information learned about devices. */
174 static Hash_table *ht;
175
176 /* Information about the destination file system. */
177 static struct fs_res *new_dst_res;
178 struct fs_res *dst_res = NULL;
179 struct fs_res tmp_dst_res;
180
181 /* timestamp resolution in nanoseconds. */
182 int res;
183
184 /* Quick exit, if possible. Since the worst resolution is 2
185 seconds, anything that differs by more than that does not
186 needs source truncation. */
187 if (dst_s == src_s && dst_ns == src_ns)
188 return 0;
189 if (dst_s <= src_s - 2)
190 return -1;
191 if (src_s <= dst_s - 2)
192 return 1;
193
194 /* Try to do a hash lookup, but fall back to stack variables and
195 recomputation on low memory situations. */
196 if (! ht)
197 ht = hash_initialize (16, NULL, dev_info_hash, dev_info_compare, free);
198 if (ht)
199 {
200 if (! new_dst_res)
201 {
202 new_dst_res = malloc (sizeof *new_dst_res);
203 if (!new_dst_res)
204 goto low_memory;
205 new_dst_res->resolution = 2 * BILLION;
206 new_dst_res->exact = false;
207 }
208 new_dst_res->dev = dst_stat->st_dev;
209 dst_res = hash_insert (ht, new_dst_res);
210 if (! dst_res)
211 goto low_memory;
212
213 if (dst_res == new_dst_res)
214 {
215 /* NEW_DST_RES is now in use in the hash table, so allocate a
216 new entry next time. */
217 new_dst_res = NULL;
218 }
219 }
220 else
221 {
222 low_memory:
223 if (ht)
224 {
225 tmp_dst_res.dev = dst_stat->st_dev;
226 dst_res = hash_lookup (ht, &tmp_dst_res);
227 }
228 if (!dst_res)
229 {
230 dst_res = &tmp_dst_res;
231 dst_res->resolution = 2 * BILLION;
232 dst_res->exact = false;
233 }
234 }
235
236 res = dst_res->resolution;
237
238 #ifdef _PC_TIMESTAMP_RESOLUTION
239 /* If the system will tell us the resolution, we're set! */
240 if (! dst_res->exact)
241 {
242 res = -1;
243 if (dfd == AT_FDCWD)
244 res = pathconf (dst_name, _PC_TIMESTAMP_RESOLUTION);
245 else
246 {
247 char *dstdir = mdir_name (dst_name);
248 if (dstdir)
249 {
250 int destdirfd = openat (dfd, dstdir,
251 O_SEARCH | O_CLOEXEC | O_DIRECTORY);
252 if (0 <= destdirfd)
253 {
254 res = fpathconf (destdirfd, _PC_TIMESTAMP_RESOLUTION);
255 close (destdirfd);
256 }
257 free (dstdir);
258 }
259 }
260 if (0 < res)
261 {
262 dst_res->resolution = res;
263 dst_res->exact = true;
264 }
265 }
266 #endif
267
268 if (! dst_res->exact)
269 {
270 /* This file system's resolution is not known exactly.
271 Deduce it, and store the result in the hash table. */
272
273 time_t dst_a_s = dst_stat->st_atime;
274 time_t dst_c_s = dst_stat->st_ctime;
275 time_t dst_m_s = dst_s;
276 int dst_a_ns = get_stat_atime_ns (dst_stat);
277 int dst_c_ns = get_stat_ctime_ns (dst_stat);
278 int dst_m_ns = dst_ns;
279
280 /* Set RES to an upper bound on the file system resolution
281 (after truncation due to SYSCALL_RESOLUTION) by inspecting
282 the atime, ctime and mtime of the existing destination.
283 We don't know of any file system that stores atime or
284 ctime with a higher precision than mtime, so it's valid to
285 look at them too. */
286 {
287 bool odd_second = (dst_a_s | dst_c_s | dst_m_s) & 1;
288
289 if (SYSCALL_RESOLUTION == BILLION)
290 {
291 if (odd_second | dst_a_ns | dst_c_ns | dst_m_ns)
292 res = BILLION;
293 }
294 else
295 {
296 int a = dst_a_ns;
297 int c = dst_c_ns;
298 int m = dst_m_ns;
299
300 /* Write it this way to avoid mistaken GCC warning
301 about integer overflow in constant expression. */
302 int SR10 = SYSCALL_RESOLUTION; SR10 *= 10;
303
304 if ((a % SR10 | c % SR10 | m % SR10) != 0)
305 res = SYSCALL_RESOLUTION;
306 else
307 for (res = SR10, a /= SR10, c /= SR10, m /= SR10;
308 (res < dst_res->resolution
309 && (a % 10 | c % 10 | m % 10) == 0);
310 res *= 10, a /= 10, c /= 10, m /= 10)
311 if (res == BILLION)
312 {
313 if (! odd_second)
314 res *= 2;
315 break;
316 }
317 }
318
319 dst_res->resolution = res;
320 }
321
322 if (SYSCALL_RESOLUTION < res)
323 {
324 struct timespec timespec[2];
325 struct stat dst_status;
326
327 /* Ignore source timestamp information that must necessarily
328 be lost when filtered through utimens. */
329 src_ns -= src_ns % SYSCALL_RESOLUTION;
330
331 /* If the timestamps disagree widely enough, there's no need
332 to interrogate the file system to deduce the exact
333 timestamp resolution; return the answer directly. */
334 {
335 time_t s = src_s & ~ (res == 2 * BILLION ? 1 : 0);
336 if (src_s < dst_s || (src_s == dst_s && src_ns <= dst_ns))
337 return 1;
338 if (dst_s < s
339 || (dst_s == s && dst_ns < src_ns - src_ns % res))
340 return -1;
341 }
342
343 /* Determine the actual timestamp resolution for the
344 destination file system (after truncation due to
345 SYSCALL_RESOLUTION) by setting the access timestamp of the
346 destination to the existing access time, except with
347 trailing nonzero digits. */
348
349 timespec[0].tv_sec = dst_a_s;
350 timespec[0].tv_nsec = dst_a_ns;
351 timespec[1].tv_sec = dst_m_s | (res == 2 * BILLION);
352 timespec[1].tv_nsec = dst_m_ns + res / 9;
353
354 if (utimensat (dfd, dst_name, timespec, AT_SYMLINK_NOFOLLOW))
355 return -2;
356
357 /* Read the modification time that was set. */
358 {
359 int stat_result
360 = fstatat (dfd, dst_name, &dst_status, AT_SYMLINK_NOFOLLOW);
361
362 if (stat_result
363 | (dst_status.st_mtime ^ dst_m_s)
364 | (get_stat_mtime_ns (&dst_status) ^ dst_m_ns))
365 {
366 /* The modification time changed, or we can't tell whether
367 it changed. Change it back as best we can. */
368 timespec[1].tv_sec = dst_m_s;
369 timespec[1].tv_nsec = dst_m_ns;
370 utimensat (dfd, dst_name, timespec, AT_SYMLINK_NOFOLLOW);
371 }
372
373 if (stat_result != 0)
374 return -2;
375 }
376
377 /* Determine the exact resolution from the modification time
378 that was read back. */
379 {
380 int old_res = res;
381 int a = (BILLION * (dst_status.st_mtime & 1)
382 + get_stat_mtime_ns (&dst_status));
383
384 res = SYSCALL_RESOLUTION;
385
386 for (a /= res; a % 10 == 0; a /= 10)
387 {
388 if (res == BILLION)
389 {
390 res *= 2;
391 break;
392 }
393 res *= 10;
394 if (res == old_res)
395 break;
396 }
397 }
398 }
399
400 dst_res->resolution = res;
401 dst_res->exact = true;
402 }
403
404 /* Truncate the source's timestamp according to the resolution. */
405 src_s &= ~ (res == 2 * BILLION ? 1 : 0);
406 src_ns -= src_ns % res;
407 }
408
409 /* Compare the timestamps and return -1, 0, 1 accordingly. */
410 return (_GL_CMP (dst_s, src_s)
411 + ((dst_s == src_s ? ~0 : 0) & _GL_CMP (dst_ns, src_ns)));
412 }