1 /*- 2 * SPDX-License-Identifier: BSD-2-Clause-FreeBSD 3 * 4 * Copyright (c) 1999 Marcel Moolenaar 5 * All rights reserved. 6 * 7 * Redistribution and use in source and binary forms, with or without 8 * modification, are permitted provided that the following conditions 9 * are met: 10 * 1. Redistributions of source code must retain the above copyright 11 * notice, this list of conditions and the following disclaimer. 12 * 2. Redistributions in binary form must reproduce the above copyright 13 * notice, this list of conditions and the following disclaimer in the 14 * documentation and/or other materials provided with the distribution. 15 * 16 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 17 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 18 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 19 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 20 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 21 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 22 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 23 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 24 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 25 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 26 * SUCH DAMAGE. 27 */ 28 29 #include <sys/cdefs.h> 30 __FBSDID("$FreeBSD$"); 31 32 #include <sys/param.h> 33 #include <sys/kernel.h> 34 #include <sys/sdt.h> 35 #include <sys/systm.h> 36 #include <sys/sysctl.h> 37 #include <sys/proc.h> 38 #include <sys/malloc.h> 39 #include <sys/mount.h> 40 #include <sys/jail.h> 41 #include <sys/lock.h> 42 #include <sys/sx.h> 43 44 #include <compat/linux/linux_mib.h> 45 #include <compat/linux/linux_misc.h> 46 47 struct linux_prison { 48 char pr_osname[LINUX_MAX_UTSNAME]; 49 char pr_osrelease[LINUX_MAX_UTSNAME]; 50 int pr_oss_version; 51 int pr_osrel; 52 }; 53 54 static struct linux_prison lprison0 = { 55 .pr_osname = "Linux", 56 .pr_osrelease = LINUX_VERSION_STR, 57 .pr_oss_version = 0x030600, 58 .pr_osrel = LINUX_VERSION_CODE 59 }; 60 61 static unsigned linux_osd_jail_slot; 62 63 SYSCTL_NODE(_compat, OID_AUTO, linux, CTLFLAG_RW, 0, "Linux mode"); 64 65 int linux_ignore_ip_recverr = 1; 66 SYSCTL_INT(_compat_linux, OID_AUTO, ignore_ip_recverr, CTLFLAG_RWTUN, 67 &linux_ignore_ip_recverr, 0, "Ignore enabling IP_RECVERR"); 68 69 int linux_preserve_vstatus = 0; 70 SYSCTL_INT(_compat_linux, OID_AUTO, preserve_vstatus, CTLFLAG_RWTUN, 71 &linux_preserve_vstatus, 0, "Preserve VSTATUS termios(4) flag"); 72 73 static int linux_set_osname(struct thread *td, char *osname); 74 static int linux_set_osrelease(struct thread *td, char *osrelease); 75 static int linux_set_oss_version(struct thread *td, int oss_version); 76 77 static int 78 linux_sysctl_osname(SYSCTL_HANDLER_ARGS) 79 { 80 char osname[LINUX_MAX_UTSNAME]; 81 int error; 82 83 linux_get_osname(req->td, osname); 84 error = sysctl_handle_string(oidp, osname, LINUX_MAX_UTSNAME, req); 85 if (error != 0 || req->newptr == NULL) 86 return (error); 87 error = linux_set_osname(req->td, osname); 88 89 return (error); 90 } 91 92 SYSCTL_PROC(_compat_linux, OID_AUTO, osname, 93 CTLTYPE_STRING | CTLFLAG_RW | CTLFLAG_PRISON | CTLFLAG_MPSAFE, 94 0, 0, linux_sysctl_osname, "A", 95 "Linux kernel OS name"); 96 97 static int 98 linux_sysctl_osrelease(SYSCTL_HANDLER_ARGS) 99 { 100 char osrelease[LINUX_MAX_UTSNAME]; 101 int error; 102 103 linux_get_osrelease(req->td, osrelease); 104 error = sysctl_handle_string(oidp, osrelease, LINUX_MAX_UTSNAME, req); 105 if (error != 0 || req->newptr == NULL) 106 return (error); 107 error = linux_set_osrelease(req->td, osrelease); 108 109 return (error); 110 } 111 112 SYSCTL_PROC(_compat_linux, OID_AUTO, osrelease, 113 CTLTYPE_STRING | CTLFLAG_RW | CTLFLAG_PRISON | CTLFLAG_MPSAFE, 114 0, 0, linux_sysctl_osrelease, "A", 115 "Linux kernel OS release"); 116 117 static int 118 linux_sysctl_oss_version(SYSCTL_HANDLER_ARGS) 119 { 120 int oss_version; 121 int error; 122 123 oss_version = linux_get_oss_version(req->td); 124 error = sysctl_handle_int(oidp, &oss_version, 0, req); 125 if (error != 0 || req->newptr == NULL) 126 return (error); 127 error = linux_set_oss_version(req->td, oss_version); 128 129 return (error); 130 } 131 132 SYSCTL_PROC(_compat_linux, OID_AUTO, oss_version, 133 CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_PRISON | CTLFLAG_MPSAFE, 134 0, 0, linux_sysctl_oss_version, "I", 135 "Linux OSS version"); 136 137 /* 138 * Map the osrelease into integer 139 */ 140 static int 141 linux_map_osrel(char *osrelease, int *osrel) 142 { 143 char *sep, *eosrelease; 144 int len, v0, v1, v2, v; 145 146 len = strlen(osrelease); 147 eosrelease = osrelease + len; 148 v0 = strtol(osrelease, &sep, 10); 149 if (osrelease == sep || sep + 1 >= eosrelease || *sep != '.') 150 return (EINVAL); 151 osrelease = sep + 1; 152 v1 = strtol(osrelease, &sep, 10); 153 if (osrelease == sep || sep + 1 >= eosrelease || *sep != '.') 154 return (EINVAL); 155 osrelease = sep + 1; 156 v2 = strtol(osrelease, &sep, 10); 157 if (osrelease == sep || 158 (sep != eosrelease && (sep + 1 >= eosrelease || *sep != '-'))) 159 return (EINVAL); 160 161 v = LINUX_KERNVER(v0, v1, v2); 162 if (v < LINUX_KERNVER(1, 0, 0)) 163 return (EINVAL); 164 165 if (osrel != NULL) 166 *osrel = v; 167 168 return (0); 169 } 170 171 /* 172 * Find a prison with Linux info. 173 * Return the Linux info and the (locked) prison. 174 */ 175 static struct linux_prison * 176 linux_find_prison(struct prison *spr, struct prison **prp) 177 { 178 struct prison *pr; 179 struct linux_prison *lpr; 180 181 for (pr = spr;; pr = pr->pr_parent) { 182 mtx_lock(&pr->pr_mtx); 183 lpr = (pr == &prison0) 184 ? &lprison0 185 : osd_jail_get(pr, linux_osd_jail_slot); 186 if (lpr != NULL) 187 break; 188 mtx_unlock(&pr->pr_mtx); 189 } 190 *prp = pr; 191 192 return (lpr); 193 } 194 195 /* 196 * Ensure a prison has its own Linux info. If lprp is non-null, point it to 197 * the Linux info and lock the prison. 198 */ 199 static void 200 linux_alloc_prison(struct prison *pr, struct linux_prison **lprp) 201 { 202 struct prison *ppr; 203 struct linux_prison *lpr, *nlpr; 204 void **rsv; 205 206 /* If this prison already has Linux info, return that. */ 207 lpr = linux_find_prison(pr, &ppr); 208 if (ppr == pr) 209 goto done; 210 /* 211 * Allocate a new info record. Then check again, in case something 212 * changed during the allocation. 213 */ 214 mtx_unlock(&ppr->pr_mtx); 215 nlpr = malloc(sizeof(struct linux_prison), M_PRISON, M_WAITOK); 216 rsv = osd_reserve(linux_osd_jail_slot); 217 lpr = linux_find_prison(pr, &ppr); 218 if (ppr == pr) { 219 free(nlpr, M_PRISON); 220 osd_free_reserved(rsv); 221 goto done; 222 } 223 /* Inherit the initial values from the ancestor. */ 224 mtx_lock(&pr->pr_mtx); 225 (void)osd_jail_set_reserved(pr, linux_osd_jail_slot, rsv, nlpr); 226 bcopy(lpr, nlpr, sizeof(*lpr)); 227 lpr = nlpr; 228 mtx_unlock(&ppr->pr_mtx); 229 done: 230 if (lprp != NULL) 231 *lprp = lpr; 232 else 233 mtx_unlock(&pr->pr_mtx); 234 } 235 236 /* 237 * Jail OSD methods for Linux prison data. 238 */ 239 static int 240 linux_prison_create(void *obj, void *data) 241 { 242 struct prison *pr = obj; 243 struct vfsoptlist *opts = data; 244 int jsys; 245 246 if (vfs_copyopt(opts, "linux", &jsys, sizeof(jsys)) == 0 && 247 jsys == JAIL_SYS_INHERIT) 248 return (0); 249 /* 250 * Inherit a prison's initial values from its parent 251 * (different from JAIL_SYS_INHERIT which also inherits changes). 252 */ 253 linux_alloc_prison(pr, NULL); 254 return (0); 255 } 256 257 static int 258 linux_prison_check(void *obj __unused, void *data) 259 { 260 struct vfsoptlist *opts = data; 261 char *osname, *osrelease; 262 int error, jsys, len, oss_version; 263 264 /* Check that the parameters are correct. */ 265 error = vfs_copyopt(opts, "linux", &jsys, sizeof(jsys)); 266 if (error != ENOENT) { 267 if (error != 0) 268 return (error); 269 if (jsys != JAIL_SYS_NEW && jsys != JAIL_SYS_INHERIT) 270 return (EINVAL); 271 } 272 error = vfs_getopt(opts, "linux.osname", (void **)&osname, &len); 273 if (error != ENOENT) { 274 if (error != 0) 275 return (error); 276 if (len == 0 || osname[len - 1] != '\0') 277 return (EINVAL); 278 if (len > LINUX_MAX_UTSNAME) { 279 vfs_opterror(opts, "linux.osname too long"); 280 return (ENAMETOOLONG); 281 } 282 } 283 error = vfs_getopt(opts, "linux.osrelease", (void **)&osrelease, &len); 284 if (error != ENOENT) { 285 if (error != 0) 286 return (error); 287 if (len == 0 || osrelease[len - 1] != '\0') 288 return (EINVAL); 289 if (len > LINUX_MAX_UTSNAME) { 290 vfs_opterror(opts, "linux.osrelease too long"); 291 return (ENAMETOOLONG); 292 } 293 error = linux_map_osrel(osrelease, NULL); 294 if (error != 0) { 295 vfs_opterror(opts, "linux.osrelease format error"); 296 return (error); 297 } 298 } 299 error = vfs_copyopt(opts, "linux.oss_version", &oss_version, 300 sizeof(oss_version)); 301 302 if (error == ENOENT) 303 error = 0; 304 return (error); 305 } 306 307 static int 308 linux_prison_set(void *obj, void *data) 309 { 310 struct linux_prison *lpr; 311 struct prison *pr = obj; 312 struct vfsoptlist *opts = data; 313 char *osname, *osrelease; 314 int error, gotversion, jsys, len, oss_version; 315 316 /* Set the parameters, which should be correct. */ 317 error = vfs_copyopt(opts, "linux", &jsys, sizeof(jsys)); 318 if (error == ENOENT) 319 jsys = -1; 320 error = vfs_getopt(opts, "linux.osname", (void **)&osname, &len); 321 if (error == ENOENT) 322 osname = NULL; 323 else 324 jsys = JAIL_SYS_NEW; 325 error = vfs_getopt(opts, "linux.osrelease", (void **)&osrelease, &len); 326 if (error == ENOENT) 327 osrelease = NULL; 328 else 329 jsys = JAIL_SYS_NEW; 330 error = vfs_copyopt(opts, "linux.oss_version", &oss_version, 331 sizeof(oss_version)); 332 if (error == ENOENT) 333 gotversion = 0; 334 else { 335 gotversion = 1; 336 jsys = JAIL_SYS_NEW; 337 } 338 switch (jsys) { 339 case JAIL_SYS_INHERIT: 340 /* "linux=inherit": inherit the parent's Linux info. */ 341 mtx_lock(&pr->pr_mtx); 342 osd_jail_del(pr, linux_osd_jail_slot); 343 mtx_unlock(&pr->pr_mtx); 344 break; 345 case JAIL_SYS_NEW: 346 /* 347 * "linux=new" or "linux.*": 348 * the prison gets its own Linux info. 349 */ 350 linux_alloc_prison(pr, &lpr); 351 if (osrelease) { 352 (void)linux_map_osrel(osrelease, &lpr->pr_osrel); 353 strlcpy(lpr->pr_osrelease, osrelease, 354 LINUX_MAX_UTSNAME); 355 } 356 if (osname) 357 strlcpy(lpr->pr_osname, osname, LINUX_MAX_UTSNAME); 358 if (gotversion) 359 lpr->pr_oss_version = oss_version; 360 mtx_unlock(&pr->pr_mtx); 361 } 362 363 return (0); 364 } 365 366 SYSCTL_JAIL_PARAM_SYS_NODE(linux, CTLFLAG_RW, "Jail Linux parameters"); 367 SYSCTL_JAIL_PARAM_STRING(_linux, osname, CTLFLAG_RW, LINUX_MAX_UTSNAME, 368 "Jail Linux kernel OS name"); 369 SYSCTL_JAIL_PARAM_STRING(_linux, osrelease, CTLFLAG_RW, LINUX_MAX_UTSNAME, 370 "Jail Linux kernel OS release"); 371 SYSCTL_JAIL_PARAM(_linux, oss_version, CTLTYPE_INT | CTLFLAG_RW, 372 "I", "Jail Linux OSS version"); 373 374 static int 375 linux_prison_get(void *obj, void *data) 376 { 377 struct linux_prison *lpr; 378 struct prison *ppr; 379 struct prison *pr = obj; 380 struct vfsoptlist *opts = data; 381 int error, i; 382 383 static int version0; 384 385 /* See if this prison is the one with the Linux info. */ 386 lpr = linux_find_prison(pr, &ppr); 387 i = (ppr == pr) ? JAIL_SYS_NEW : JAIL_SYS_INHERIT; 388 error = vfs_setopt(opts, "linux", &i, sizeof(i)); 389 if (error != 0 && error != ENOENT) 390 goto done; 391 if (i) { 392 error = vfs_setopts(opts, "linux.osname", lpr->pr_osname); 393 if (error != 0 && error != ENOENT) 394 goto done; 395 error = vfs_setopts(opts, "linux.osrelease", lpr->pr_osrelease); 396 if (error != 0 && error != ENOENT) 397 goto done; 398 error = vfs_setopt(opts, "linux.oss_version", 399 &lpr->pr_oss_version, sizeof(lpr->pr_oss_version)); 400 if (error != 0 && error != ENOENT) 401 goto done; 402 } else { 403 /* 404 * If this prison is inheriting its Linux info, report 405 * empty/zero parameters. 406 */ 407 error = vfs_setopts(opts, "linux.osname", ""); 408 if (error != 0 && error != ENOENT) 409 goto done; 410 error = vfs_setopts(opts, "linux.osrelease", ""); 411 if (error != 0 && error != ENOENT) 412 goto done; 413 error = vfs_setopt(opts, "linux.oss_version", &version0, 414 sizeof(lpr->pr_oss_version)); 415 if (error != 0 && error != ENOENT) 416 goto done; 417 } 418 error = 0; 419 420 done: 421 mtx_unlock(&ppr->pr_mtx); 422 423 return (error); 424 } 425 426 static void 427 linux_prison_destructor(void *data) 428 { 429 430 free(data, M_PRISON); 431 } 432 433 void 434 linux_osd_jail_register(void) 435 { 436 struct prison *pr; 437 osd_method_t methods[PR_MAXMETHOD] = { 438 [PR_METHOD_CREATE] = linux_prison_create, 439 [PR_METHOD_GET] = linux_prison_get, 440 [PR_METHOD_SET] = linux_prison_set, 441 [PR_METHOD_CHECK] = linux_prison_check 442 }; 443 444 linux_osd_jail_slot = 445 osd_jail_register(linux_prison_destructor, methods); 446 /* Copy the system Linux info to any current prisons. */ 447 sx_slock(&allprison_lock); 448 TAILQ_FOREACH(pr, &allprison, pr_list) 449 linux_alloc_prison(pr, NULL); 450 sx_sunlock(&allprison_lock); 451 } 452 453 void 454 linux_osd_jail_deregister(void) 455 { 456 457 osd_jail_deregister(linux_osd_jail_slot); 458 } 459 460 void 461 linux_get_osname(struct thread *td, char *dst) 462 { 463 struct prison *pr; 464 struct linux_prison *lpr; 465 466 lpr = linux_find_prison(td->td_ucred->cr_prison, &pr); 467 bcopy(lpr->pr_osname, dst, LINUX_MAX_UTSNAME); 468 mtx_unlock(&pr->pr_mtx); 469 } 470 471 static int 472 linux_set_osname(struct thread *td, char *osname) 473 { 474 struct prison *pr; 475 struct linux_prison *lpr; 476 477 lpr = linux_find_prison(td->td_ucred->cr_prison, &pr); 478 strlcpy(lpr->pr_osname, osname, LINUX_MAX_UTSNAME); 479 mtx_unlock(&pr->pr_mtx); 480 481 return (0); 482 } 483 484 void 485 linux_get_osrelease(struct thread *td, char *dst) 486 { 487 struct prison *pr; 488 struct linux_prison *lpr; 489 490 lpr = linux_find_prison(td->td_ucred->cr_prison, &pr); 491 bcopy(lpr->pr_osrelease, dst, LINUX_MAX_UTSNAME); 492 mtx_unlock(&pr->pr_mtx); 493 } 494 495 int 496 linux_kernver(struct thread *td) 497 { 498 struct prison *pr; 499 struct linux_prison *lpr; 500 int osrel; 501 502 lpr = linux_find_prison(td->td_ucred->cr_prison, &pr); 503 osrel = lpr->pr_osrel; 504 mtx_unlock(&pr->pr_mtx); 505 506 return (osrel); 507 } 508 509 static int 510 linux_set_osrelease(struct thread *td, char *osrelease) 511 { 512 struct prison *pr; 513 struct linux_prison *lpr; 514 int error; 515 516 lpr = linux_find_prison(td->td_ucred->cr_prison, &pr); 517 error = linux_map_osrel(osrelease, &lpr->pr_osrel); 518 if (error == 0) 519 strlcpy(lpr->pr_osrelease, osrelease, LINUX_MAX_UTSNAME); 520 mtx_unlock(&pr->pr_mtx); 521 522 return (error); 523 } 524 525 int 526 linux_get_oss_version(struct thread *td) 527 { 528 struct prison *pr; 529 struct linux_prison *lpr; 530 int version; 531 532 lpr = linux_find_prison(td->td_ucred->cr_prison, &pr); 533 version = lpr->pr_oss_version; 534 mtx_unlock(&pr->pr_mtx); 535 536 return (version); 537 } 538 539 static int 540 linux_set_oss_version(struct thread *td, int oss_version) 541 { 542 struct prison *pr; 543 struct linux_prison *lpr; 544 545 lpr = linux_find_prison(td->td_ucred->cr_prison, &pr); 546 lpr->pr_oss_version = oss_version; 547 mtx_unlock(&pr->pr_mtx); 548 549 return (0); 550 } 551