vserver.h (37571B)
1 /* $Id$ 2 3 * Copyright (C) 2003 Enrico Scholz <enrico.scholz@informatik.tu-chemnitz.de> 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 2, or (at your option) 8 * 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, write to the Free Software 17 * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. 18 */ 19 20 /** \file vserver.h 21 * \brief The public interface of the the libvserver library. 22 */ 23 24 #ifndef H_VSERVER_SYSCALL_H 25 #define H_VSERVER_SYSCALL_H 26 27 #include <stdint.h> 28 #include <stdlib.h> 29 #include <stdbool.h> 30 #include <sys/types.h> 31 #include <sched.h> 32 #include <netinet/in.h> 33 34 #ifndef IS_DOXYGEN 35 #if defined(__GNUC__) 36 # define VC_ATTR_UNUSED __attribute__((__unused__)) 37 # define VC_ATTR_NORETURN __attribute__((__noreturn__)) 38 # define VC_ATTR_CONST __attribute__((__const__)) 39 # define VC_ATTR_DEPRECATED __attribute__((__deprecated__)) 40 # if __GNUC__*0x10000 + __GNUC_MINOR__*0x100 + __GNUC_PATCHLEVEL__ >= 0x30300 41 # define VC_ATTR_NONNULL(ARGS) __attribute__((__nonnull__ ARGS)) 42 # define VC_ATTR_ALWAYSINLINE __attribute__((__always_inline__)) 43 # else 44 # define VC_ATTR_NONNULL(ARGS) 45 # define VC_ATTR_ALWAYSINLINE 46 # endif 47 # if __GNUC__*0x10000 + __GNUC_MINOR__*0x100 + __GNUC_PATCHLEVEL__ >= 0x30303 48 # define VC_ATTR_PURE __attribute__((__pure__)) 49 # else 50 # define VC_ATTR_PURE 51 # endif 52 #else 53 # define VC_ATTR_NONNULL(ARGS) 54 # define VC_ATTR_UNUSED 55 # define VC_ATTR_NORETURN 56 # define VC_ATTR_ALWAYSINLINE 57 # define VC_ATTR_DEPRECATED 58 # define VC_ATTR_PURE 59 # define VC_ATTR_CONST 60 #endif 61 #endif // IS_DOXYGEN 62 63 /** the value which is returned in error-case (no ctx found) */ 64 #define VC_NOCTX ((xid_t)(-1)) 65 #define VC_NOXID ((xid_t)(-1)) 66 /** the value which means a random (the next free) ctx */ 67 #define VC_DYNAMIC_XID ((xid_t)(-1)) 68 /** the value which means the current ctx */ 69 #define VC_SAMECTX ((xid_t)(-2)) 70 71 #define VC_NONID ((nid_t)(-1)) 72 #define VC_DYNAMIC_NID ((nid_t)(-1)) 73 74 #define VC_LIM_INFINITY (~0ULL) 75 #define VC_LIM_KEEP (~1ULL) 76 77 #define VC_CDLIM_UNSET (0U) 78 #define VC_CDLIM_INFINITY (~0U) 79 #define VC_CDLIM_KEEP (~1U) 80 81 #ifndef S_CTX_INFO_LOCK 82 # define S_CTX_INFO_LOCK 1 83 #endif 84 85 #ifndef S_CTX_INFO_SCHED 86 # define S_CTX_INFO_SCHED 2 87 #endif 88 89 #ifndef S_CTX_INFO_NPROC 90 # define S_CTX_INFO_NPROC 4 91 #endif 92 93 #ifndef S_CTX_INFO_PRIVATE 94 # define S_CTX_INFO_PRIVATE 8 95 #endif 96 97 #ifndef S_CTX_INFO_INIT 98 # define S_CTX_INFO_INIT 16 99 #endif 100 101 #ifndef S_CTX_INFO_HIDEINFO 102 # define S_CTX_INFO_HIDEINFO 32 103 #endif 104 105 #ifndef S_CTX_INFO_ULIMIT 106 # define S_CTX_INFO_ULIMIT 64 107 #endif 108 109 #ifndef S_CTX_INFO_NAMESPACE 110 # define S_CTX_INFO_NAMESPACE 128 111 #endif 112 113 #define VC_CAP_CHOWN 0 114 #define VC_CAP_DAC_OVERRIDE 1 115 #define VC_CAP_DAC_READ_SEARCH 2 116 #define VC_CAP_FOWNER 3 117 #define VC_CAP_FSETID 4 118 #define VC_CAP_KILL 5 119 #define VC_CAP_SETGID 6 120 #define VC_CAP_SETUID 7 121 #define VC_CAP_SETPCAP 8 122 #define VC_CAP_LINUX_IMMUTABLE 9 123 #define VC_CAP_NET_BIND_SERVICE 10 124 #define VC_CAP_NET_BROADCAST 11 125 #define VC_CAP_NET_ADMIN 12 126 #define VC_CAP_NET_RAW 13 127 #define VC_CAP_IPC_LOCK 14 128 #define VC_CAP_IPC_OWNER 15 129 #define VC_CAP_SYS_MODULE 16 130 #define VC_CAP_SYS_RAWIO 17 131 #define VC_CAP_SYS_CHROOT 18 132 #define VC_CAP_SYS_PTRACE 19 133 #define VC_CAP_SYS_PACCT 20 134 #define VC_CAP_SYS_ADMIN 21 135 #define VC_CAP_SYS_BOOT 22 136 #define VC_CAP_SYS_NICE 23 137 #define VC_CAP_SYS_RESOURCE 24 138 #define VC_CAP_SYS_TIME 25 139 #define VC_CAP_SYS_TTY_CONFIG 26 140 #define VC_CAP_MKNOD 27 141 #define VC_CAP_LEASE 28 142 #define VC_CAP_AUDIT_WRITE 29 143 #define VC_CAP_AUDIT_CONTROL 30 144 #define VC_CAP_SETFCAP 31 145 #define VC_CAP_MAC_OVERRIDE 32 146 #define VC_CAP_MAC_ADMIN 33 147 148 #define VC_IMMUTABLE_FILE_FL 0x0000010lu 149 #define VC_IMMUTABLE_LINK_FL 0x0008000lu 150 #define VC_IMMUTABLE_ALL (VC_IMMUTABLE_LINK_FL|VC_IMMUTABLE_FILE_FL) 151 152 #define VC_IATTR_XID 0x01000000u 153 154 #define VC_IATTR_ADMIN 0x00000001u 155 #define VC_IATTR_WATCH 0x00000002u 156 #define VC_IATTR_HIDE 0x00000004u 157 #define VC_IATTR_WRITE 0x00000008u 158 #define VC_IATTR_FLAGS 0x0000000fu 159 160 #define VC_IATTR_BARRIER 0x00010000u 161 #define VC_IATTR_IUNLINK 0x00020000u 162 #define VC_IATTR_IMMUTABLE 0x00040000u 163 #define VC_IATTR_COW 0x00080000u 164 165 166 // the flags 167 #define VC_VXF_INFO_LOCK 0x00000001ull 168 #define VC_VXF_INFO_NPROC 0x00000004ull 169 #define VC_VXF_INFO_PRIVATE 0x00000008ull 170 #define VC_VXF_INFO_INIT 0x00000010ull 171 172 #define VC_VXF_INFO_HIDEINFO 0x00000020ull 173 #define VC_VXF_INFO_ULIMIT 0x00000040ull 174 #define VC_VXF_INFO_NAMESPACE 0x00000080ull 175 176 #define VC_VXF_SCHED_HARD 0x00000100ull 177 #define VC_VXF_SCHED_PRIO 0x00000200ull 178 #define VC_VXF_SCHED_PAUSE 0x00000400ull 179 180 #define VC_VXF_VIRT_MEM 0x00010000ull 181 #define VC_VXF_VIRT_UPTIME 0x00020000ull 182 #define VC_VXF_VIRT_CPU 0x00040000ull 183 #define VC_VXF_VIRT_LOAD 0x00080000ull 184 #define VC_VXF_VIRT_TIME 0x00100000ull 185 186 #define VC_VXF_HIDE_MOUNT 0x01000000ull 187 #define VC_VXF_HIDE_NETIF 0x02000000ull 188 #define VC_VXF_HIDE_VINFO 0x04000000ull 189 190 #define VC_VXF_STATE_SETUP (1ULL<<32) 191 #define VC_VXF_STATE_INIT (1ULL<<33) 192 #define VC_VXF_STATE_ADMIN (1ULL<<34) 193 194 #define VC_VXF_SC_HELPER (1ULL<<36) 195 #define VC_VXF_REBOOT_KILL (1ULL<<37) 196 #define VC_VXF_PERSISTENT (1ULL<<38) 197 198 #define VC_VXF_FORK_RSS (1ULL<<48) 199 #define VC_VXF_PROLIFIC (1ULL<<49) 200 201 #define VC_VXF_IGNEG_NICE (1ULL<<52) 202 #define VC_VXF_IGNEG_IONICE (1ULL<<53) 203 204 205 // the ccapabilities 206 #define VC_VXC_SET_UTSNAME 0x00000001ull 207 #define VC_VXC_SET_RLIMIT 0x00000002ull 208 #define VC_VXC_FS_SECURITY 0x00000004ull 209 #define VC_VXC_FS_TRUSTED 0x00000008ull 210 #define VC_VXC_TIOCSTI 0x00000010ull 211 212 #define VC_VXC_RAW_ICMP 0x00000100ull 213 #define VC_VXC_SYSLOG 0x00001000ull 214 #define VC_VXC_OOM_ADJUST 0x00002000ull 215 #define VC_VXC_AUDIT_CONTROL 0x00004000ull 216 217 #define VC_VXC_SECURE_MOUNT 0x00010000ull 218 #define VC_VXC_SECURE_REMOUNT 0x00020000ull 219 #define VC_VXC_BINARY_MOUNT 0x00040000ull 220 #define VC_VXC_DEV_MOUNT 0x00080000ull 221 222 #define VC_VXC_QUOTA_CTL 0x00100000ull 223 #define VC_VXC_ADMIN_MAPPER 0x00200000ull 224 #define VC_VXC_ADMIN_CLOOP 0x00400000ull 225 226 #define VC_VXC_KTHREAD 0x01000000ull 227 #define VC_VXC_NAMESPACE 0x02000000ull 228 229 230 // the scheduler flags 231 #define VC_VXSM_FILL_RATE 0x0001 232 #define VC_VXSM_INTERVAL 0x0002 233 #define VC_VXSM_FILL_RATE2 0x0004 234 #define VC_VXSM_INTERVAL2 0x0008 235 #define VC_VXSM_TOKENS 0x0010 236 #define VC_VXSM_TOKENS_MIN 0x0020 237 #define VC_VXSM_TOKENS_MAX 0x0040 238 #define VC_VXSM_PRIO_BIAS 0x0100 239 #define VC_VXSM_CPU_ID 0x1000 240 #define VC_VXSM_BUCKET_ID 0x2000 241 242 #define VC_VXSM_IDLE_TIME 0x0200 243 #define VC_VXSM_FORCE 0x0400 244 #define VC_VXSM_MSEC 0x4000 245 246 #define VC_VXSM_V3_MASK 0x0173 247 248 249 // the network flags 250 #define VC_NXF_INFO_LOCK 0x00000001ull 251 #define VC_NXF_INFO_PRIVATE 0x00000008ull 252 253 #define VC_NXF_SINGLE_IP 0x00000100ull 254 #define VC_NXF_LBACK_REMAP 0x00000200ull 255 #define VC_NXF_LBACK_ALLOW 0x00000400ull 256 257 #define VC_NXF_HIDE_NETIF 0x02000000ull 258 #define VC_NXF_HIDE_LBACK 0x04000000ull 259 260 #define VC_NXF_STATE_SETUP (1ULL<<32) 261 #define VC_NXF_STATE_ADMIN (1ULL<<34) 262 263 #define VC_NXF_SC_HELPER (1ULL<<36) 264 #define VC_NXF_PERSISTENT (1ULL<<38) 265 266 267 // the network capabilities 268 #define VC_NXC_TUN_CREATE 0x00000001ull 269 270 #define VC_NXC_RAW_ICMP 0x00000100ull 271 272 273 // the vserver specific limits 274 #define VC_VLIMIT_NSOCK 16 275 #define VC_VLIMIT_OPENFD 17 276 #define VC_VLIMIT_ANON 18 277 #define VC_VLIMIT_SHMEM 19 278 #define VC_VLIMIT_SEMARY 20 279 #define VC_VLIMIT_NSEMS 21 280 #define VC_VLIMIT_DENTRY 22 281 #define VC_VLIMIT_MAPPED 23 282 283 284 // the VCI bit values 285 #define VC_VCI_NO_DYNAMIC (1 << 0) 286 #define VC_VCI_PROC_SECURE (1 << 4) 287 #define VC_VCI_HARDCPU (1 << 5) 288 #define VC_VCI_IDLELIMIT (1 << 6) 289 #define VC_VCI_IDLETIME (1 << 7) 290 #define VC_VCI_COWBL (1 << 8) 291 #define VC_VCI_FULLCOWBL (1 << 9) 292 #define VC_VCI_SPACES (1 << 10) 293 #define VC_VCI_NETV2 (1 << 11) 294 #define VC_VCI_MEMCG (1 << 12) 295 #define VC_VCI_DEBUG (1 << 16) 296 #define VC_VCI_HISTORY (1 << 20) 297 #define VC_VCI_TAGGED (1 << 24) 298 #define VC_VCI_PPTAG (1 << 28) 299 300 301 // the device mapping flags 302 #define VC_DATTR_CREATE 0x00000001 303 #define VC_DATTR_OPEN 0x00000002 304 305 #define VC_DATTR_REMAP 0x00000010 306 307 308 // the process context migration flags 309 #define VC_VXM_SET_INIT 0x00000001 310 #define VC_VXM_SET_REAPER 0x00000002 311 312 313 // the network address flags 314 #define VC_NXA_TYPE_IPV4 0x0001 315 #define VC_NXA_TYPE_IPV6 0x0002 316 317 #define VC_NXA_TYPE_NONE 0x0000 318 #define VC_NXA_TYPE_ANY 0x00FF 319 320 #define VC_NXA_TYPE_ADDR 0x0010 321 #define VC_NXA_TYPE_MASK 0x0020 322 #define VC_NXA_TYPE_RANGE 0x0040 323 324 #define VC_NXA_MOD_BCAST 0x0100 325 #define VC_NXA_MOD_LBACK 0x0200 326 327 328 #ifndef CLONE_NEWNS 329 # define CLONE_NEWNS 0x00020000 330 #endif 331 #ifndef CLONE_NEWUTS 332 # define CLONE_NEWUTS 0x04000000 333 #endif 334 #ifndef CLONE_NEWIPC 335 # define CLONE_NEWIPC 0x08000000 336 #endif 337 #ifndef CLONE_NEWUSER 338 # define CLONE_NEWUSER 0x10000000 339 #endif 340 #ifndef CLONE_NEWPID 341 # define CLONE_NEWPID 0x20000000 342 #endif 343 #ifndef CLONE_NEWNET 344 # define CLONE_NEWNET 0x40000000 345 #endif 346 347 348 #define VC_BAD_PERSONALITY ((uint_least32_t)(-1)) 349 350 351 /** \defgroup syscalls Syscall wrappers 352 * Functions which are calling the vserver syscall directly. */ 353 354 /** \defgroup helper Helper functions 355 * Functions which are doing general helper tasks like parameter parsing. */ 356 357 /** \typedef an_unsigned_integer_type xid_t 358 * The identifier of a context. */ 359 360 #ifdef IS_DOXYGEN 361 typedef an_unsigned_integer_type xid_t; 362 typedef an_unsigned_integer_type nid_t; 363 typedef an_unsigned_integer_type tag_t; 364 #endif 365 366 #ifdef __cplusplus 367 extern "C" { 368 #endif 369 370 /** \brief The generic vserver syscall 371 * \ingroup syscalls 372 * 373 * This function executes the generic vserver syscall. It uses the 374 * correct syscallnumber (which may differ between the different 375 * architectures). 376 * 377 * \param cmd the command to be executed 378 * \param xid the xid on which the cmd shall be applied 379 * \param data additional arguments; depends on \c cmd 380 * \returns depends on \c cmd; usually, -1 stands for an error 381 */ 382 int vc_syscall(uint32_t cmd, xid_t xid, void *data); 383 384 /** \brief Returns the version of the current kernel API. 385 * \ingroup syscalls 386 * \returns The versionnumber of the kernel API 387 */ 388 int vc_get_version(); 389 390 typedef uint64_t vc_vci_t; 391 /** \brief Returns the kernel configuration bits 392 * \ingroup syscalls 393 * \returns The kernel configuration bits 394 */ 395 vc_vci_t vc_get_vci(); 396 397 int vc_get_kernel(); 398 399 /** \brief Moves current process into a context 400 * \ingroup syscalls 401 * 402 * Puts current process into context \a ctx, removes the capabilities 403 * given in \a remove_cap and sets \a flags. 404 * 405 * \param ctx The new context; special values for are 406 * - VC_SAMECTX which means the current context (just for changing caps and flags) 407 * - VC_DYNAMIC_XID which means the next free context; this value can be used by 408 * ordinary users also 409 * \param remove_cap The linux capabilities which will be \b removed. 410 * \param flags Special flags which will be set. 411 * 412 * \returns The new context-id, or VC_NOCTX on errors; \c errno 413 * will be set appropriately 414 * 415 * See http://vserver.13thfloor.at/Stuff/Logic.txt for details */ 416 xid_t vc_new_s_context(xid_t ctx, unsigned int remove_cap, unsigned int flags); 417 418 struct vc_ip_mask_pair { 419 uint32_t ip; 420 uint32_t mask; 421 }; 422 423 /** \brief Sets the ipv4root information. 424 * \ingroup syscalls 425 * \pre \a nb < NB_IPV4ROOT && \a ips != 0 */ 426 int vc_set_ipv4root(uint32_t bcast, size_t nb, 427 struct vc_ip_mask_pair const *ips) VC_ATTR_NONNULL((3)); 428 429 /** \brief Returns the value of NB_IPV4ROOT. 430 * \ingroup helper 431 * 432 * This function returns the value of NB_IPV4ROOT which was used when the 433 * library was built, but \b not the value which is used by the currently 434 * running kernel. */ 435 size_t vc_get_nb_ipv4root() VC_ATTR_CONST VC_ATTR_PURE; 436 437 /* process context */ 438 /** \brief Flags of process-contexts 439 */ 440 struct vc_ctx_flags { 441 /** \brief Mask of set context flags */ 442 uint_least64_t flagword; 443 /** \brief Mask of set and unset context flags when used by set 444 * operations, or modifiable flags when used by get 445 * operations */ 446 uint_least64_t mask; 447 }; 448 449 /** \brief Creates a context without starting it. 450 * \ingroup syscalls 451 * 452 * This functions initializes a new context. When already in a freshly 453 * created context, this old context will be discarded. 454 * 455 * \param xid The new context; special values are: 456 * - VC_DYNAMIC_XID which means to create a dynamic context 457 * 458 * \returns the xid of the created context, or VC_NOCTX on errors. \c errno 459 * will be set appropriately. */ 460 xid_t vc_ctx_create(xid_t xid, struct vc_ctx_flags *flags); 461 462 /** \brief Moves the current process into the specified context. 463 * \ingroup syscalls 464 * 465 * \param xid The new context 466 * \param flags The flags, see VC_VXM_* 467 * \returns 0 on success, -1 on errors */ 468 int vc_ctx_migrate(xid_t xid, uint_least64_t flags); 469 470 /** \brief Statistics about a context */ 471 struct vc_ctx_stat { 472 uint_least32_t usecnt; ///< number of uses 473 uint_least32_t tasks; ///< number of tasks 474 }; 475 476 /** \brief Get some statistics about a context. 477 * \ingroup syscalls 478 * 479 * \param xid The context to get stats about 480 * \param stat Where to store the result 481 * 482 * \returns 0 on success, -1 on errors. */ 483 int vc_ctx_stat(xid_t xid, struct vc_ctx_stat /*@out@*/ *stat) VC_ATTR_NONNULL((2)); 484 485 /** \brief Contains further statistics about a context. */ 486 struct vc_virt_stat { 487 uint_least64_t offset; 488 uint_least64_t uptime; 489 uint_least32_t nr_threads; 490 uint_least32_t nr_running; 491 uint_least32_t nr_uninterruptible; 492 uint_least32_t nr_onhold; 493 uint_least32_t nr_forks; 494 uint_least32_t load[3]; 495 }; 496 497 /** \brief Get more statistics about a context. 498 * \ingroup syscalls 499 * 500 * \param xid The context to get stats about 501 * \param stat Where to store the result 502 * 503 * \returns 0 on success, -1 on errors. */ 504 int vc_virt_stat(xid_t xid, struct vc_virt_stat /*@out@*/ *stat) VC_ATTR_NONNULL((2)); 505 506 /** \brief Sends a signal to a context/pid 507 * \ingroup syscalls 508 * 509 * Special values for \a pid are: 510 * - -1 which means every process in ctx except the init-process 511 * - 0 which means every process in ctx inclusive the init-process */ 512 int vc_ctx_kill(xid_t ctx, pid_t pid, int sig); 513 514 int vc_get_cflags(xid_t xid, struct vc_ctx_flags /*@out@*/ *) VC_ATTR_NONNULL((2)); 515 int vc_set_cflags(xid_t xid, struct vc_ctx_flags /*@in@*/ const *) VC_ATTR_NONNULL((2)); 516 517 /** \brief Capabilities of process-contexts */ 518 struct vc_ctx_caps { 519 /** \brief Mask of set common system capabilities */ 520 uint_least64_t bcaps; 521 /** \brief Mask of set and unset common system capabilities when used by 522 * set operations, or the modifiable capabilities when used by 523 * get operations */ 524 uint_least64_t bmask; 525 /** \brief Mask of set process context capabilities */ 526 uint_least64_t ccaps; 527 /** \brief Mask of set and unset process context capabilities when used 528 * by set operations, or the modifiable capabilities when used 529 * by get operations */ 530 uint_least64_t cmask; 531 }; 532 533 int vc_get_ccaps(xid_t xid, struct vc_ctx_caps *); 534 int vc_set_ccaps(xid_t xid, struct vc_ctx_caps const *); 535 536 struct vc_vx_info { 537 xid_t xid; 538 pid_t initpid; 539 }; 540 541 int vc_get_vx_info(xid_t xid, struct vc_vx_info *info) VC_ATTR_NONNULL((2)); 542 543 /** \brief Returns the context of the given process. 544 * \ingroup syscalls 545 * 546 * \param pid the process-id whose xid shall be determined; 547 * pid==0 means the current process. 548 * \returns the xid of process \c pid or -1 on errors 549 */ 550 xid_t vc_get_task_xid(pid_t pid); 551 552 /** \brief Waits for the end of a context 553 * \ingroup syscalls 554 */ 555 int vc_wait_exit(xid_t xid); 556 557 /* rlimit related functions */ 558 559 /** \brief The type which is used for a single limit value. 560 * 561 * Special values are 562 * - VC_LIM_INFINITY ... which is the infinite value 563 * - VC_LIM_KEEP ... which is used to mark values which shall not be 564 * modified by the vc_set_rlimit() operation. 565 * 566 * Else, the interpretation of the value depends on the corresponding 567 * resource; it might be bytes, pages, seconds or litres of beer. */ 568 typedef uint_least64_t vc_limit_t; 569 570 /** \brief Masks describing the supported limits. */ 571 struct vc_rlimit_mask { 572 uint_least32_t min; ///< masks the resources supporting a minimum limit 573 uint_least32_t soft; ///< masks the resources supporting a soft limit 574 uint_least32_t hard; ///< masks the resources supporting a hard limit 575 }; 576 577 /** \brief Returns the limits supported by the kernel */ 578 int vc_get_rlimit_mask(xid_t xid, 579 struct vc_rlimit_mask /*@out@*/ *lim) VC_ATTR_NONNULL((2)); 580 581 /** \brief The limits of a resources. 582 * 583 * This is a triple consisting of a minimum, soft and hardlimit. */ 584 struct vc_rlimit { 585 vc_limit_t min; ///< the guaranted minimum of a resources 586 vc_limit_t soft; ///< the softlimit of a resource 587 vc_limit_t hard; ///< the absolute hardlimit of a resource 588 }; 589 590 /** \brief Returns the limits of \a resource. 591 * \ingroup syscalls 592 * 593 * \param xid The id of the context 594 * \param resource The resource which will be queried 595 * \param lim The result which will be filled with the limits 596 * 597 * \returns 0 on success, and -1 on errors. */ 598 int vc_get_rlimit(xid_t xid, int resource, 599 struct vc_rlimit /*@out@*/ *lim) VC_ATTR_NONNULL((3)); 600 /** \brief Sets the limits of \a resource. 601 * \ingroup syscalls 602 * 603 * \param xid The id of the context 604 * \param resource The resource which will be queried 605 * \param lim The new limits 606 * 607 * \returns 0 on success, and -1 on errors. */ 608 int vc_set_rlimit(xid_t xid, int resource, 609 struct vc_rlimit const /*@in@*/ *lim) VC_ATTR_NONNULL((3)); 610 611 /** \brief Statistics for a resource limit. */ 612 struct vc_rlimit_stat { 613 uint_least32_t hits; ///< number of hits on the limit 614 vc_limit_t value; ///< current value 615 vc_limit_t minimum; ///< minimum value observed 616 vc_limit_t maximum; ///< maximum value observed 617 }; 618 619 /** \brief Returns the current stats of \a resource. 620 * \ingroup syscalls 621 * 622 * \param xid The id of the context 623 * \param resource The resource which will be queried 624 * \param stat The result which will be filled with the stats 625 * 626 * \returns 0 on success, and -1 on errors. */ 627 int vc_rlimit_stat(xid_t xid, int resource, 628 struct vc_rlimit_stat /*@out@*/ *stat) VC_ATTR_NONNULL((3)); 629 630 /** \brief Resets the minimum and maximum observed values of all resources. 631 * \ingroup syscalls 632 * 633 * \param xid The id of the context 634 * 635 * \returns 0 on success, and -1 on errors. */ 636 int vc_reset_minmax(xid_t xid); 637 638 /** \brief Parses a string describing a limit 639 * \ingroup helper 640 * 641 * This function parses \a str and interprets special words like \p "inf" 642 * or suffixes. Valid suffixes are 643 * - \p k ... 1000 644 * - \p m ... 1000000 645 * - \p K ... 1024 646 * - \p M ... 1048576 647 * 648 * \param str The string which shall be parsed 649 * \param res Will be filled with the interpreted value; in errorcase, 650 * this value is undefined. 651 * 652 * \returns \a true, iff the string \a str could be parsed. \a res will 653 * be filled with the interpreted value in this case. 654 * 655 * \pre \a str!=0 && \a res!=0 656 */ 657 bool vc_parseLimit(char const /*@in@*/ *str, vc_limit_t /*@out@*/ *res) VC_ATTR_NONNULL((1,2)); 658 659 660 /* network context */ 661 struct vc_nx_info { 662 nid_t nid; 663 }; 664 665 nid_t vc_get_task_nid(pid_t pid); 666 int vc_get_nx_info(nid_t nid, struct vc_nx_info *) VC_ATTR_NONNULL((2)); 667 668 struct vc_net_addr { 669 uint16_t vna_type; 670 uint16_t vna_flags; 671 uint16_t vna_prefix; 672 uint16_t vna_parent; 673 struct { 674 union { 675 struct in_addr v4; 676 struct in6_addr v6; 677 } ip; 678 union { 679 struct in_addr v4; 680 struct in6_addr v6; 681 } ip2; 682 union { 683 struct in_addr v4; 684 struct in6_addr v6; 685 } mask; 686 } s; 687 #define vna_v4_ip s.ip.v4 688 #define vna_v4_ip2 s.ip2.v4 689 #define vna_v4_mask s.mask.v4 690 #define vna_v6_ip s.ip.v6 691 #define vna_v6_ip2 s.ip2.v6 692 #define vna_v6_mask s.mask.v6 693 }; 694 695 struct vc_net_flags { 696 uint_least64_t flagword; 697 uint_least64_t mask; 698 }; 699 700 nid_t vc_net_create(nid_t nid); 701 int vc_net_migrate(nid_t nid); 702 703 int vc_net_add(nid_t nid, struct vc_net_addr const *info); 704 int vc_net_remove(nid_t nid, struct vc_net_addr const *info); 705 706 int vc_get_nflags(nid_t, struct vc_net_flags *); 707 int vc_set_nflags(nid_t, struct vc_net_flags const *); 708 709 struct vc_net_caps { 710 uint_least64_t ncaps; 711 uint_least64_t cmask; 712 }; 713 714 int vc_get_ncaps(nid_t, struct vc_net_caps *); 715 int vc_set_ncaps(nid_t, struct vc_net_caps const *); 716 717 718 /* iattr related functions */ 719 720 int vc_set_iattr(char const *filename, xid_t xid, 721 uint_least32_t flags, uint_least32_t mask) VC_ATTR_NONNULL((1)); 722 723 int vc_fset_iattr(int fd, xid_t xid, 724 uint_least32_t flags, uint_least32_t mask); 725 726 /** \brief Returns information about attributes and assigned context of a file. 727 * \ingroup syscalls 728 * 729 * This function returns the VC_IATTR_XXX flags and about the assigned 730 * context of a file. To request an information, the appropriate bit in 731 * \c mask must be set and the corresponding parameter (\a xid or \a 732 * flags) must not be NULL. 733 * 734 * E.g. to receive the assigned context, the \c VC_IATTR_XID bit must be 735 * set in \a mask, and \a xid must point to valid memory. 736 * 737 * Possible flags are \c VC_IATTR_ADMIN, \c VC_IATTR_WATCH , \c VC_IATTR_HIDE, 738 * \c VC_IATTR_BARRIER, \c VC_IATTR_IUNLINK and \c VC_IATTR_IMMUTABLE. 739 * 740 * \param filename The name of the file whose attributes shall be determined. 741 742 * \param xid When non-zero and the VC_IATTR_XID bit is set in \a mask, 743 * the assigned context of \a filename will be stored there. 744 * \param flags When non-zero, a bitmask of current attributes will be 745 * stored there. These attributes must be requested explicitly 746 * by setting the appropriate bit in \a mask 747 * \param mask Points to a bitmask which tells which attributes shall be 748 * determined. On return, it will masquerade the attributes 749 * which were determined. 750 * 751 * \pre mask!=0 && !((*mask&VC_IATTR_XID) && xid==0) && !((*mask&~VC_IATTR_XID) && flags==0) */ 752 int vc_get_iattr(char const *filename, xid_t * /*@null@*/ xid, 753 uint_least32_t * /*@null@*/ flags, 754 uint_least32_t * /*@null@*/ mask) VC_ATTR_NONNULL((1)); 755 756 int vc_fget_iattr(int fd, xid_t * /*@null@*/ xid, 757 uint_least32_t * /*@null@*/ flags, 758 uint_least32_t * /*@null@*/ mask) VC_ATTR_NONNULL((4)); 759 760 /** \brief Returns the context of \c filename 761 * \ingroup syscalls 762 * 763 * This function calls vc_get_iattr() with appropriate arguments to 764 * determine the context of \c filename. In error-case or when no context 765 * is assigned, \c VC_NOCTX will be returned. To differ between both cases, 766 * \c errno must be examined. 767 * 768 * \b WARNING: this function can modify \c errno although no error happened. 769 * 770 * \param filename The file to check 771 * \returns The assigned context, or VC_NOCTX when an error occured or no 772 * such assignment exists. \c errno will be 0 in the latter case */ 773 xid_t vc_getfilecontext(char const *filename) VC_ATTR_NONNULL((1)); 774 775 776 /* vhi related functions */ 777 typedef enum { vcVHI_CONTEXT, vcVHI_SYSNAME, vcVHI_NODENAME, 778 vcVHI_RELEASE, vcVHI_VERSION, vcVHI_MACHINE, 779 vcVHI_DOMAINNAME } vc_uts_type; 780 781 int vc_set_vhi_name(xid_t xid, vc_uts_type type, 782 char const *val, size_t len) VC_ATTR_NONNULL((3)); 783 int vc_get_vhi_name(xid_t xid, vc_uts_type type, 784 char *val, size_t len) VC_ATTR_NONNULL((3)); 785 786 /* namespace related functions */ 787 int vc_enter_namespace(xid_t xid, uint_least64_t mask, uint32_t index); 788 int vc_set_namespace(xid_t xid, uint_least64_t mask, uint32_t index); 789 int vc_cleanup_namespace(void); 790 uint_least64_t vc_get_space_mask(void); 791 uint_least64_t vc_get_space_default(void); 792 793 794 /* disk limit related things */ 795 struct vc_ctx_dlimit { 796 uint_least32_t space_used; 797 uint_least32_t space_total; 798 uint_least32_t inodes_used; 799 uint_least32_t inodes_total; 800 uint_least32_t reserved; 801 }; 802 803 804 /** Add a disk limit to a file system. */ 805 int vc_add_dlimit(char const *filename, xid_t xid, 806 uint_least32_t flags) VC_ATTR_NONNULL((1)); 807 /** Remove a disk limit from a file system. */ 808 int vc_rem_dlimit(char const *filename, xid_t xid, 809 uint_least32_t flags) VC_ATTR_NONNULL((1)); 810 811 /** Set a disk limit. */ 812 int vc_set_dlimit(char const *filename, xid_t xid, 813 uint_least32_t flags, 814 struct vc_ctx_dlimit const *limits) VC_ATTR_NONNULL((1,4)); 815 /** Get a disk limit. */ 816 int vc_get_dlimit(char const *filename, xid_t xid, 817 uint_least32_t flags, 818 struct vc_ctx_dlimit *limits) VC_ATTR_NONNULL((1)); 819 820 /** Get the filesystem tag for a process. */ 821 tag_t vc_get_task_tag(pid_t pid); 822 823 /** Create a new filesystem tag space. */ 824 int vc_tag_create(tag_t tag); 825 826 /** Migrate to an existing filesystem tag space. */ 827 int vc_tag_migrate(tag_t tag); 828 829 /* scheduler related syscalls */ 830 struct vc_set_sched { 831 uint_least32_t set_mask; 832 int_least32_t fill_rate; 833 int_least32_t interval; 834 int_least32_t fill_rate2; 835 int_least32_t interval2; 836 int_least32_t tokens; 837 int_least32_t tokens_min; 838 int_least32_t tokens_max; 839 int_least32_t priority_bias; 840 int_least32_t cpu_id; 841 int_least32_t bucket_id; 842 }; 843 844 int vc_set_sched(xid_t xid, struct vc_set_sched const *) VC_ATTR_NONNULL((2)); 845 int vc_get_sched(xid_t xid, struct vc_set_sched *) VC_ATTR_NONNULL((2)); 846 847 struct vc_sched_info { 848 int_least32_t cpu_id; 849 int_least32_t bucket_id; 850 uint_least64_t user_msec; 851 uint_least64_t sys_msec; 852 uint_least64_t hold_msec; 853 uint_least32_t token_usec; 854 int_least32_t vavavoom; 855 }; 856 857 int vc_sched_info(xid_t xid, struct vc_sched_info *info) VC_ATTR_NONNULL((2)); 858 859 /* misc. syscalls */ 860 int vc_set_mapping(xid_t xid, const char *device, const char *target, uint32_t flags); 861 int vc_unset_mapping(xid_t xid, const char *device, const char *target, uint32_t flags); 862 863 int vc_get_badness(xid_t xid, int64_t *badness); 864 int vc_set_badness(xid_t xid, int64_t badness); 865 866 867 /** \brief Namespaces allowed to unshare */ 868 struct vc_umask { 869 uint_least64_t umask; 870 uint_least64_t mask; 871 }; 872 873 int vc_get_umask(xid_t xid, struct vc_umask *umask); 874 int vc_set_umask(xid_t xid, struct vc_umask const *umask); 875 876 /** \brief Information about parsing errors 877 * \ingroup helper 878 */ 879 struct vc_err_listparser { 880 char const *ptr; ///< Pointer to the first character of an erroneous string 881 size_t len; ///< Length of the erroneous string 882 }; 883 884 /** \brief Converts a single string into bcapability 885 * \ingroup helper 886 * 887 * \param str The string to be parsed; 888 * both "CAP_xxx" and "xxx" will be accepted 889 * \param len The length of the string, or \c 0 for automatic detection 890 * 891 * \returns 0 on error; a bitmask on success 892 * \pre \a str != 0 893 */ 894 uint_least64_t vc_text2bcap(char const *str, size_t len); 895 896 /** \brief Converts the lowest bit of a bcapability or the entire value 897 * (when possible) to a textual representation 898 * \ingroup helper 899 * 900 * \param val The string to be converted; on success, the detected bit(s) 901 * will be unset, in errorcase only the lowest set bit 902 * 903 * \returns A textual representation of \a val resp. of its lowest set bit; 904 * or \c NULL in errorcase. 905 * \pre \a val!=0 906 * \post \a *val<sub>old</sub> \c != 0 \c <--> 907 * \a *val<sub>old</sub> > \a *val<sub>new</sub> 908 * \post \a *val<sub>old</sub> \c == 0 \c ---> \a result == 0 909 */ 910 char const * vc_lobcap2text(uint_least64_t *val) VC_ATTR_NONNULL((1)); 911 912 /** \brief Converts a string into a bcapability-bitmask 913 * \ingroup helper 914 * 915 * Syntax of \a str: \verbinclude list2xxx.syntax 916 * 917 * When the \c `~' prefix is used, the bits will be unset and a `~' after 918 * another `~' will cancel both ones. The \c `^' prefix specifies a 919 * bitnumber instead of a bitmask. 920 * 921 * "literal name" is everything which will be accepted by the 922 * vc_text2bcap() function. The special values for \c NAME will be 923 * recognized case insensitively 924 * 925 * \param str The string to be parsed 926 * \param len The length of the string, or \c 0 for automatic detection 927 * \param err Pointer to a structure for error-information, or \c NULL. 928 * \param cap Pointer to a vc_ctx_caps structure holding the results; 929 * only the \a bcaps and \a bmask fields will be changed and 930 * already set values will not be honored. When an error 931 * occured, \a cap will have the value of all processed valid 932 * \c BCAP parts. 933 * 934 * \returns 0 on success, -1 on error. In error case, \a err will hold 935 * position and length of the first not understood BCAP part 936 * \pre \a str != 0 && \a cap != 0; 937 * \a cap->bcaps and \a cap->bmask must be initialized 938 */ 939 int vc_list2bcap(char const *str, size_t len, 940 struct vc_err_listparser *err, 941 struct vc_ctx_caps *cap) VC_ATTR_NONNULL((1,4)); 942 943 uint_least64_t vc_text2ccap(char const *, size_t len); 944 char const * vc_loccap2text(uint_least64_t *); 945 int vc_list2ccap(char const *, size_t len, 946 struct vc_err_listparser *err, 947 struct vc_ctx_caps *); 948 char const * vc_loumask2text(uint_least64_t *); 949 int vc_list2umask(char const *, size_t len, 950 struct vc_err_listparser *err, 951 struct vc_umask *); 952 int vc_list2cflag(char const *, size_t len, 953 struct vc_err_listparser *err, 954 struct vc_ctx_flags *flags); 955 uint_least64_t vc_text2umask(char const *str, size_t len); 956 uint_least64_t vc_text2cflag(char const *, size_t len); 957 char const * vc_locflag2text(uint_least64_t *); 958 959 uint_least32_t vc_list2cflag_compat(char const *, size_t len, 960 struct vc_err_listparser *err); 961 uint_least32_t vc_text2cflag_compat(char const *, size_t len); 962 char const * vc_hicflag2text_compat(uint_least32_t); 963 964 int vc_text2cap(char const *); 965 char const * vc_cap2text(unsigned int); 966 967 968 int vc_list2nflag(char const *, size_t len, 969 struct vc_err_listparser *err, 970 struct vc_net_flags *flags); 971 uint_least64_t vc_text2nflag(char const *, size_t len); 972 char const * vc_lonflag2text(uint_least64_t *); 973 974 uint_least64_t vc_text2ncap(char const *, size_t len); 975 char const * vc_loncap2text(uint_least64_t *); 976 int vc_list2ncap(char const *, size_t len, 977 struct vc_err_listparser *err, 978 struct vc_net_caps *); 979 980 uint_least64_t vc_get_insecurebcaps() VC_ATTR_CONST; 981 inline static uint_least64_t vc_get_insecureccaps() { 982 return ~(VC_VXC_SET_UTSNAME|VC_VXC_RAW_ICMP|VC_VXC_AUDIT_CONTROL); 983 } 984 985 inline static int vc_setfilecontext(char const *filename, xid_t xid) { 986 return vc_set_iattr(filename, xid, 0, VC_IATTR_XID); 987 } 988 989 990 uint_least32_t vc_text2personalityflag(char const *str, 991 size_t len) VC_ATTR_NONNULL((1)); 992 993 char const * vc_lopersonality2text(uint_least32_t *) VC_ATTR_NONNULL((1)); 994 995 int vc_list2personalityflag(char const /*@in@*/ *, 996 size_t len, 997 uint_least32_t /*@out@*/ *personality, 998 struct vc_err_listparser /*@out@*/ *err) VC_ATTR_NONNULL((1,3)); 999 1000 uint_least32_t vc_str2personalitytype(char const /*@in@*/*, 1001 size_t len) VC_ATTR_NONNULL((1)); 1002 1003 1004 typedef enum { vcFEATURE_VKILL, vcFEATURE_IATTR, vcFEATURE_RLIMIT, 1005 vcFEATURE_COMPAT, vcFEATURE_MIGRATE, vcFEATURE_NAMESPACE, 1006 vcFEATURE_SCHED, vcFEATURE_VINFO, vcFEATURE_VHI, 1007 vcFEATURE_VSHELPER0, vcFEATURE_VSHELPER, vcFEATURE_VWAIT, 1008 vcFEATURE_VNET, vcFEATURE_VSTAT, vcFEATURE_PPTAG, 1009 vcFEATURE_PIDSPACE, vcFEATURE_SPACES, vcFEATURE_PERSISTENT, 1010 vcFEATURE_PIVOT_ROOT, vcFEATURE_MEMCG, vcFEATURE_DYNAMIC, 1011 vcFEATURE_BME, 1012 } 1013 vcFeatureSet; 1014 1015 bool vc_isSupported(vcFeatureSet) VC_ATTR_CONST; 1016 bool vc_isSupportedString(char const *); 1017 1018 1019 typedef enum { vcTYPE_INVALID, vcTYPE_MAIN, vcTYPE_WATCH, 1020 vcTYPE_STATIC, vcTYPE_DYNAMIC } 1021 vcXidType; 1022 1023 vcXidType vc_getXIDType(xid_t xid) VC_ATTR_CONST; 1024 1025 /** Returns true iff \a xid is a dynamic xid */ 1026 bool vc_is_dynamic_xid(xid_t xid); 1027 1028 1029 /* The management part */ 1030 1031 #define VC_LIMIT_VSERVER_NAME_LEN 1024 1032 1033 typedef enum { vcCFG_NONE, vcCFG_AUTO, 1034 vcCFG_LEGACY, 1035 vcCFG_RECENT_SHORT, 1036 vcCFG_RECENT_FULL } vcCfgStyle; 1037 1038 1039 /** Maps an xid given at '--xid' options to an xid_t */ 1040 xid_t vc_xidopt2xid(char const *, bool honor_static, char const **err_info); 1041 /** Maps a nid given at '--nid' options to a nid_t */ 1042 nid_t vc_nidopt2nid(char const *, bool honor_static, char const **err_info); 1043 /** Maps a tag given at '--tag' options to a tag_t */ 1044 tag_t vc_tagopt2tag(char const *, bool honor_static, char const **err_info); 1045 1046 vcCfgStyle vc_getVserverCfgStyle(char const *id); 1047 1048 /** Resolves the name of the vserver. The result will be allocated and must 1049 be freed by the caller. */ 1050 char * vc_getVserverName(char const *id, vcCfgStyle style); 1051 1052 /** Returns the path of the vserver configuration directory. When the given 1053 * vserver does not exist, or when it does not have such a directory, NULL 1054 * will be returned. Else, the result will be allocated and must be freed 1055 * by the caller. */ 1056 char * vc_getVserverCfgDir(char const *id, vcCfgStyle style); 1057 1058 /** Returns the path of the configuration directory for the given 1059 * application. The result will be allocated and must be freed by the 1060 * caller. */ 1061 char * vc_getVserverAppDir(char const *id, vcCfgStyle style, char const *app); 1062 1063 /** Returns the path to the vserver root-directory. The result will be 1064 * allocated and must be freed by the caller. */ 1065 char * vc_getVserverVdir(char const *id, vcCfgStyle style, bool physical); 1066 1067 typedef enum { vcCTX_XID = 1, 1068 vcCTX_NID, 1069 vcCTX_TAG, 1070 } vcCtxType; 1071 1072 /** Returns the ctx of the given vserver. When vserver is not running and 1073 * 'honor_static' is false, VC_NOCTX will be returned. Else, when 1074 * 'honor_static' is true and a static assignment exists, those value will 1075 * be returned. Else, the result will be VC_NOCTX. 1076 * 1077 * When 'is_running' is not null, the status of the vserver will be 1078 * assigned to this variable. */ 1079 xid_t vc_getVserverCtx(char const *id, vcCfgStyle style, 1080 bool honor_static, bool /*@null@*/ *is_running, 1081 vcCtxType type); 1082 1083 /** Resolves the cfg-path of the vserver owning the given ctx. 'revdir' will 1084 be used as the directory holding the mapping-links; when NULL, the 1085 default value will be assumed. The result will be allocated and must be 1086 freed by the caller. */ 1087 char * vc_getVserverByCtx(xid_t ctx, /*@null@*/vcCfgStyle *style, 1088 /*@null@*/char const *revdir); 1089 1090 int vc_compareVserverById(char const *lhs, vcCfgStyle lhs_style, 1091 char const *rhs, vcCfgStyle rhs_style); 1092 1093 void vc_exitLikeProcess(int pid, int ret) VC_ATTR_NORETURN; 1094 1095 #define vcSKEL_INTERFACES 1u 1096 #define vcSKEL_PKGMGMT 2u 1097 #define vcSKEL_FILESYSTEM 4u 1098 1099 /** Create a basic configuration skeleton for a vserver plus toplevel 1100 * directories for pkgmanagemt and filesystem (when requested). */ 1101 int vc_createSkeleton(char const *id, vcCfgStyle style, int flags); 1102 1103 1104 #ifdef __cplusplus 1105 } 1106 #endif 1107 1108 #undef VC_ATTR_PURE 1109 #undef VC_ATTR_ALWAYSINLINE 1110 #undef VC_ATTR_NORETURN 1111 #undef VC_ATTR_UNUSED 1112 #undef VC_ATTR_NONNULL 1113 1114 #endif