fix TZ parsing logic for identifying POSIX-form strings
[musl.git] / src / time / __tz.c
blob3e2fcdcbb6a48d000497e2bd36234a2e9c88f0cf
1 #include "time_impl.h"
2 #include <stdint.h>
3 #include <limits.h>
4 #include <stdlib.h>
5 #include <string.h>
6 #include <sys/mman.h>
7 #include <ctype.h>
8 #include "libc.h"
9 #include "lock.h"
10 #include "fork_impl.h"
12 #define malloc __libc_malloc
13 #define calloc undef
14 #define realloc undef
15 #define free undef
17 long __timezone = 0;
18 int __daylight = 0;
19 char *__tzname[2] = { 0, 0 };
21 weak_alias(__timezone, timezone);
22 weak_alias(__daylight, daylight);
23 weak_alias(__tzname, tzname);
25 static char std_name[TZNAME_MAX+1];
26 static char dst_name[TZNAME_MAX+1];
27 const char __utc[] = "UTC";
29 static int dst_off;
30 static int r0[5], r1[5];
32 static const unsigned char *zi, *trans, *index, *types, *abbrevs, *abbrevs_end;
33 static size_t map_size;
35 static char old_tz_buf[32];
36 static char *old_tz = old_tz_buf;
37 static size_t old_tz_size = sizeof old_tz_buf;
39 static volatile int lock[1];
40 volatile int *const __timezone_lockptr = lock;
42 static int getint(const char **p)
44 unsigned x;
45 for (x=0; **p-'0'<10U; (*p)++) x = **p-'0' + 10*x;
46 return x;
49 static int getoff(const char **p)
51 int neg = 0;
52 if (**p == '-') {
53 ++*p;
54 neg = 1;
55 } else if (**p == '+') {
56 ++*p;
58 int off = 3600*getint(p);
59 if (**p == ':') {
60 ++*p;
61 off += 60*getint(p);
62 if (**p == ':') {
63 ++*p;
64 off += getint(p);
67 return neg ? -off : off;
70 static void getrule(const char **p, int rule[5])
72 int r = rule[0] = **p;
74 if (r!='M') {
75 if (r=='J') ++*p;
76 else rule[0] = 0;
77 rule[1] = getint(p);
78 } else {
79 ++*p; rule[1] = getint(p);
80 ++*p; rule[2] = getint(p);
81 ++*p; rule[3] = getint(p);
84 if (**p=='/') {
85 ++*p;
86 rule[4] = getoff(p);
87 } else {
88 rule[4] = 7200;
92 static void getname(char *d, const char **p)
94 int i;
95 if (**p == '<') {
96 ++*p;
97 for (i=0; (*p)[i] && (*p)[i]!='>'; i++)
98 if (i<TZNAME_MAX) d[i] = (*p)[i];
99 if ((*p)[i]) ++*p;
100 } else {
101 for (i=0; ((*p)[i]|32)-'a'<26U; i++)
102 if (i<TZNAME_MAX) d[i] = (*p)[i];
104 *p += i;
105 d[i<TZNAME_MAX?i:TZNAME_MAX] = 0;
108 #define VEC(...) ((const unsigned char[]){__VA_ARGS__})
110 static uint32_t zi_read32(const unsigned char *z)
112 return (unsigned)z[0]<<24 | z[1]<<16 | z[2]<<8 | z[3];
115 static size_t zi_dotprod(const unsigned char *z, const unsigned char *v, size_t n)
117 size_t y;
118 uint32_t x;
119 for (y=0; n; n--, z+=4, v++) {
120 x = zi_read32(z);
121 y += x * *v;
123 return y;
126 static void do_tzset()
128 char buf[NAME_MAX+25], *pathname=buf+24;
129 const char *try, *s, *p;
130 const unsigned char *map = 0;
131 size_t i;
132 static const char search[] =
133 "/usr/share/zoneinfo/\0/share/zoneinfo/\0/etc/zoneinfo/\0";
135 s = getenv("TZ");
136 if (!s) s = "/etc/localtime";
137 if (!*s) s = __utc;
139 if (old_tz && !strcmp(s, old_tz)) return;
141 for (i=0; i<5; i++) r0[i] = r1[i] = 0;
143 if (zi) __munmap((void *)zi, map_size);
145 /* Cache the old value of TZ to check if it has changed. Avoid
146 * free so as not to pull it into static programs. Growth
147 * strategy makes it so free would have minimal benefit anyway. */
148 i = strlen(s);
149 if (i > PATH_MAX+1) s = __utc, i = 3;
150 if (i >= old_tz_size) {
151 old_tz_size *= 2;
152 if (i >= old_tz_size) old_tz_size = i+1;
153 if (old_tz_size > PATH_MAX+2) old_tz_size = PATH_MAX+2;
154 old_tz = malloc(old_tz_size);
156 if (old_tz) memcpy(old_tz, s, i+1);
158 int posix_form = 0;
159 if (*s != ':') {
160 p = s;
161 char dummy_name[TZNAME_MAX+1];
162 getname(dummy_name, &p);
163 if (p!=s && (*p == '+' || *p == '-' || isdigit(*p)
164 || !strcmp(dummy_name, "UTC")
165 || !strcmp(dummy_name, "GMT")))
166 posix_form = 1;
169 /* Non-suid can use an absolute tzfile pathname or a relative
170 * pathame beginning with "."; in secure mode, only the
171 * standard path will be searched. */
172 if (!posix_form) {
173 if (*s == ':') s++;
174 if (*s == '/' || *s == '.') {
175 if (!libc.secure || !strcmp(s, "/etc/localtime"))
176 map = __map_file(s, &map_size);
177 } else {
178 size_t l = strlen(s);
179 if (l <= NAME_MAX && !strchr(s, '.')) {
180 memcpy(pathname, s, l+1);
181 pathname[l] = 0;
182 for (try=search; !map && *try; try+=l+1) {
183 l = strlen(try);
184 memcpy(pathname-l, try, l);
185 map = __map_file(pathname-l, &map_size);
189 if (!map) s = __utc;
191 if (map && (map_size < 44 || memcmp(map, "TZif", 4))) {
192 __munmap((void *)map, map_size);
193 map = 0;
194 s = __utc;
197 zi = map;
198 if (map) {
199 int scale = 2;
200 if (map[4]!='1') {
201 size_t skip = zi_dotprod(zi+20, VEC(1,1,8,5,6,1), 6);
202 trans = zi+skip+44+44;
203 scale++;
204 } else {
205 trans = zi+44;
207 index = trans + (zi_read32(trans-12) << scale);
208 types = index + zi_read32(trans-12);
209 abbrevs = types + 6*zi_read32(trans-8);
210 abbrevs_end = abbrevs + zi_read32(trans-4);
211 if (zi[map_size-1] == '\n') {
212 for (s = (const char *)zi+map_size-2; *s!='\n'; s--);
213 s++;
214 } else {
215 const unsigned char *p;
216 __tzname[0] = __tzname[1] = 0;
217 __daylight = __timezone = dst_off = 0;
218 for (p=types; p<abbrevs; p+=6) {
219 if (!p[4] && !__tzname[0]) {
220 __tzname[0] = (char *)abbrevs + p[5];
221 __timezone = -zi_read32(p);
223 if (p[4] && !__tzname[1]) {
224 __tzname[1] = (char *)abbrevs + p[5];
225 dst_off = -zi_read32(p);
226 __daylight = 1;
229 if (!__tzname[0]) __tzname[0] = __tzname[1];
230 if (!__tzname[0]) __tzname[0] = (char *)__utc;
231 if (!__daylight) {
232 __tzname[1] = __tzname[0];
233 dst_off = __timezone;
235 return;
239 if (!s) s = __utc;
240 getname(std_name, &s);
241 __tzname[0] = std_name;
242 __timezone = getoff(&s);
243 getname(dst_name, &s);
244 __tzname[1] = dst_name;
245 if (dst_name[0]) {
246 __daylight = 1;
247 if (*s == '+' || *s=='-' || *s-'0'<10U)
248 dst_off = getoff(&s);
249 else
250 dst_off = __timezone - 3600;
251 } else {
252 __daylight = 0;
253 dst_off = __timezone;
256 if (*s == ',') s++, getrule(&s, r0);
257 if (*s == ',') s++, getrule(&s, r1);
260 /* Search zoneinfo rules to find the one that applies to the given time,
261 * and determine alternate opposite-DST-status rule that may be needed. */
263 static size_t scan_trans(long long t, int local, size_t *alt)
265 int scale = 3 - (trans == zi+44);
266 uint64_t x;
267 int off = 0;
269 size_t a = 0, n = (index-trans)>>scale, m;
271 if (!n) {
272 if (alt) *alt = 0;
273 return 0;
276 /* Binary search for 'most-recent rule before t'. */
277 while (n > 1) {
278 m = a + n/2;
279 x = zi_read32(trans + (m<<scale));
280 if (scale == 3) x = x<<32 | zi_read32(trans + (m<<scale) + 4);
281 else x = (int32_t)x;
282 if (local) off = (int32_t)zi_read32(types + 6 * index[m-1]);
283 if (t - off < (int64_t)x) {
284 n /= 2;
285 } else {
286 a = m;
287 n -= n/2;
291 /* First and last entry are special. First means to use lowest-index
292 * non-DST type. Last means to apply POSIX-style rule if available. */
293 n = (index-trans)>>scale;
294 if (a == n-1) return -1;
295 if (a == 0) {
296 x = zi_read32(trans + (a<<scale));
297 if (scale == 3) x = x<<32 | zi_read32(trans + (a<<scale) + 4);
298 else x = (int32_t)x;
299 if (local) off = (int32_t)zi_read32(types + 6 * index[a-1]);
300 if (t - off < (int64_t)x) {
301 for (a=0; a<(abbrevs-types)/6; a++) {
302 if (types[6*a+4] != types[4]) break;
304 if (a == (abbrevs-types)/6) a = 0;
305 if (types[6*a+4]) {
306 *alt = a;
307 return 0;
308 } else {
309 *alt = 0;
310 return a;
315 /* Try to find a neighboring opposite-DST-status rule. */
316 if (alt) {
317 if (a && types[6*index[a-1]+4] != types[6*index[a]+4])
318 *alt = index[a-1];
319 else if (a+1<n && types[6*index[a+1]+4] != types[6*index[a]+4])
320 *alt = index[a+1];
321 else
322 *alt = index[a];
325 return index[a];
328 static int days_in_month(int m, int is_leap)
330 if (m==2) return 28+is_leap;
331 else return 30+((0xad5>>(m-1))&1);
334 /* Convert a POSIX DST rule plus year to seconds since epoch. */
336 static long long rule_to_secs(const int *rule, int year)
338 int is_leap;
339 long long t = __year_to_secs(year, &is_leap);
340 int x, m, n, d;
341 if (rule[0]!='M') {
342 x = rule[1];
343 if (rule[0]=='J' && (x < 60 || !is_leap)) x--;
344 t += 86400 * x;
345 } else {
346 m = rule[1];
347 n = rule[2];
348 d = rule[3];
349 t += __month_to_secs(m-1, is_leap);
350 int wday = (int)((t + 4*86400) % (7*86400)) / 86400;
351 int days = d - wday;
352 if (days < 0) days += 7;
353 if (n == 5 && days+28 >= days_in_month(m, is_leap)) n = 4;
354 t += 86400 * (days + 7*(n-1));
356 t += rule[4];
357 return t;
360 /* Determine the time zone in effect for a given time in seconds since the
361 * epoch. It can be given in local or universal time. The results will
362 * indicate whether DST is in effect at the queried time, and will give both
363 * the GMT offset for the active zone/DST rule and the opposite DST. This
364 * enables a caller to efficiently adjust for the case where an explicit
365 * DST specification mismatches what would be in effect at the time. */
367 void __secs_to_zone(long long t, int local, int *isdst, long *offset, long *oppoff, const char **zonename)
369 LOCK(lock);
371 do_tzset();
373 if (zi) {
374 size_t alt, i = scan_trans(t, local, &alt);
375 if (i != -1) {
376 *isdst = types[6*i+4];
377 *offset = (int32_t)zi_read32(types+6*i);
378 *zonename = (const char *)abbrevs + types[6*i+5];
379 if (oppoff) *oppoff = (int32_t)zi_read32(types+6*alt);
380 UNLOCK(lock);
381 return;
385 if (!__daylight) goto std;
387 /* FIXME: may be broken if DST changes right at year boundary?
388 * Also, this could be more efficient.*/
389 long long y = t / 31556952 + 70;
390 while (__year_to_secs(y, 0) > t) y--;
391 while (__year_to_secs(y+1, 0) < t) y++;
393 long long t0 = rule_to_secs(r0, y);
394 long long t1 = rule_to_secs(r1, y);
396 if (!local) {
397 t0 += __timezone;
398 t1 += dst_off;
400 if (t0 < t1) {
401 if (t >= t0 && t < t1) goto dst;
402 goto std;
403 } else {
404 if (t >= t1 && t < t0) goto std;
405 goto dst;
407 std:
408 *isdst = 0;
409 *offset = -__timezone;
410 if (oppoff) *oppoff = -dst_off;
411 *zonename = __tzname[0];
412 UNLOCK(lock);
413 return;
414 dst:
415 *isdst = 1;
416 *offset = -dst_off;
417 if (oppoff) *oppoff = -__timezone;
418 *zonename = __tzname[1];
419 UNLOCK(lock);
422 static void __tzset()
424 LOCK(lock);
425 do_tzset();
426 UNLOCK(lock);
429 weak_alias(__tzset, tzset);
431 const char *__tm_to_tzname(const struct tm *tm)
433 const void *p = tm->__tm_zone;
434 LOCK(lock);
435 do_tzset();
436 if (p != __utc && p != __tzname[0] && p != __tzname[1] &&
437 (!zi || (uintptr_t)p-(uintptr_t)abbrevs >= abbrevs_end - abbrevs))
438 p = "";
439 UNLOCK(lock);
440 return p;