kern_proc.c revision 1.212.2.2
1 /* $NetBSD: kern_proc.c,v 1.212.2.2 2020/04/13 08:05:03 martin Exp $ */ 2 3 /*- 4 * Copyright (c) 1999, 2006, 2007, 2008, 2020 The NetBSD Foundation, Inc. 5 * All rights reserved. 6 * 7 * This code is derived from software contributed to The NetBSD Foundation 8 * by Jason R. Thorpe of the Numerical Aerospace Simulation Facility, 9 * NASA Ames Research Center, and by Andrew Doran. 10 * 11 * Redistribution and use in source and binary forms, with or without 12 * modification, are permitted provided that the following conditions 13 * are met: 14 * 1. Redistributions of source code must retain the above copyright 15 * notice, this list of conditions and the following disclaimer. 16 * 2. Redistributions in binary form must reproduce the above copyright 17 * notice, this list of conditions and the following disclaimer in the 18 * documentation and/or other materials provided with the distribution. 19 * 20 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 21 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 22 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 23 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 24 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 25 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 26 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 27 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 28 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 29 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 30 * POSSIBILITY OF SUCH DAMAGE. 31 */ 32 33 /* 34 * Copyright (c) 1982, 1986, 1989, 1991, 1993 35 * The Regents of the University of California. All rights reserved. 36 * 37 * Redistribution and use in source and binary forms, with or without 38 * modification, are permitted provided that the following conditions 39 * are met: 40 * 1. Redistributions of source code must retain the above copyright 41 * notice, this list of conditions and the following disclaimer. 42 * 2. Redistributions in binary form must reproduce the above copyright 43 * notice, this list of conditions and the following disclaimer in the 44 * documentation and/or other materials provided with the distribution. 45 * 3. Neither the name of the University nor the names of its contributors 46 * may be used to endorse or promote products derived from this software 47 * without specific prior written permission. 48 * 49 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 50 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 51 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 52 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 53 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 54 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 55 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 56 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 57 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 58 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 59 * SUCH DAMAGE. 60 * 61 * @(#)kern_proc.c 8.7 (Berkeley) 2/14/95 62 */ 63 64 #include <sys/cdefs.h> 65 __KERNEL_RCSID(0, "$NetBSD: kern_proc.c,v 1.212.2.2 2020/04/13 08:05:03 martin Exp $"); 66 67 #ifdef _KERNEL_OPT 68 #include "opt_kstack.h" 69 #include "opt_maxuprc.h" 70 #include "opt_dtrace.h" 71 #include "opt_compat_netbsd32.h" 72 #include "opt_kaslr.h" 73 #endif 74 75 #if defined(__HAVE_COMPAT_NETBSD32) && !defined(COMPAT_NETBSD32) \ 76 && !defined(_RUMPKERNEL) 77 #define COMPAT_NETBSD32 78 #endif 79 80 #include <sys/param.h> 81 #include <sys/systm.h> 82 #include <sys/kernel.h> 83 #include <sys/proc.h> 84 #include <sys/resourcevar.h> 85 #include <sys/buf.h> 86 #include <sys/acct.h> 87 #include <sys/wait.h> 88 #include <sys/file.h> 89 #include <ufs/ufs/quota.h> 90 #include <sys/uio.h> 91 #include <sys/pool.h> 92 #include <sys/pset.h> 93 #include <sys/ioctl.h> 94 #include <sys/tty.h> 95 #include <sys/signalvar.h> 96 #include <sys/ras.h> 97 #include <sys/filedesc.h> 98 #include <sys/syscall_stats.h> 99 #include <sys/kauth.h> 100 #include <sys/sleepq.h> 101 #include <sys/atomic.h> 102 #include <sys/kmem.h> 103 #include <sys/namei.h> 104 #include <sys/dtrace_bsd.h> 105 #include <sys/sysctl.h> 106 #include <sys/exec.h> 107 #include <sys/cpu.h> 108 #include <sys/compat_stub.h> 109 #include <sys/vnode.h> 110 111 #include <uvm/uvm_extern.h> 112 #include <uvm/uvm.h> 113 114 /* 115 * Process lists. 116 */ 117 118 struct proclist allproc __cacheline_aligned; 119 struct proclist zombproc __cacheline_aligned; 120 121 kmutex_t * proc_lock __cacheline_aligned; 122 123 /* 124 * pid to proc lookup is done by indexing the pid_table array. 125 * Since pid numbers are only allocated when an empty slot 126 * has been found, there is no need to search any lists ever. 127 * (an orphaned pgrp will lock the slot, a session will lock 128 * the pgrp with the same number.) 129 * If the table is too small it is reallocated with twice the 130 * previous size and the entries 'unzipped' into the two halves. 131 * A linked list of free entries is passed through the pt_proc 132 * field of 'free' items - set odd to be an invalid ptr. 133 */ 134 135 struct pid_table { 136 struct proc *pt_proc; 137 struct pgrp *pt_pgrp; 138 pid_t pt_pid; 139 }; 140 #if 1 /* strongly typed cast - should be a noop */ 141 static inline uint p2u(struct proc *p) { return (uint)(uintptr_t)p; } 142 #else 143 #define p2u(p) ((uint)p) 144 #endif 145 #define P_VALID(p) (!(p2u(p) & 1)) 146 #define P_NEXT(p) (p2u(p) >> 1) 147 #define P_FREE(pid) ((struct proc *)(uintptr_t)((pid) << 1 | 1)) 148 149 /* 150 * Table of process IDs (PIDs). 151 */ 152 static struct pid_table *pid_table __read_mostly; 153 154 #define INITIAL_PID_TABLE_SIZE (1 << 5) 155 156 /* Table mask, threshold for growing and number of allocated PIDs. */ 157 static u_int pid_tbl_mask __read_mostly; 158 static u_int pid_alloc_lim __read_mostly; 159 static u_int pid_alloc_cnt __cacheline_aligned; 160 161 /* Next free, last free and maximum PIDs. */ 162 static u_int next_free_pt __cacheline_aligned; 163 static u_int last_free_pt __cacheline_aligned; 164 static pid_t pid_max __read_mostly; 165 166 /* Components of the first process -- never freed. */ 167 168 extern struct emul emul_netbsd; /* defined in kern_exec.c */ 169 170 struct session session0 = { 171 .s_count = 1, 172 .s_sid = 0, 173 }; 174 struct pgrp pgrp0 = { 175 .pg_members = LIST_HEAD_INITIALIZER(&pgrp0.pg_members), 176 .pg_session = &session0, 177 }; 178 filedesc_t filedesc0; 179 struct cwdinfo cwdi0 = { 180 .cwdi_cmask = CMASK, 181 .cwdi_refcnt = 1, 182 }; 183 struct plimit limit0; 184 struct pstats pstat0; 185 struct vmspace vmspace0; 186 struct sigacts sigacts0; 187 struct proc proc0 = { 188 .p_lwps = LIST_HEAD_INITIALIZER(&proc0.p_lwps), 189 .p_sigwaiters = LIST_HEAD_INITIALIZER(&proc0.p_sigwaiters), 190 .p_nlwps = 1, 191 .p_nrlwps = 1, 192 .p_nlwpid = 1, /* must match lwp0.l_lid */ 193 .p_pgrp = &pgrp0, 194 .p_comm = "system", 195 /* 196 * Set P_NOCLDWAIT so that kernel threads are reparented to init(8) 197 * when they exit. init(8) can easily wait them out for us. 198 */ 199 .p_flag = PK_SYSTEM | PK_NOCLDWAIT, 200 .p_stat = SACTIVE, 201 .p_nice = NZERO, 202 .p_emul = &emul_netbsd, 203 .p_cwdi = &cwdi0, 204 .p_limit = &limit0, 205 .p_fd = &filedesc0, 206 .p_vmspace = &vmspace0, 207 .p_stats = &pstat0, 208 .p_sigacts = &sigacts0, 209 #ifdef PROC0_MD_INITIALIZERS 210 PROC0_MD_INITIALIZERS 211 #endif 212 }; 213 kauth_cred_t cred0; 214 215 static const int nofile = NOFILE; 216 static const int maxuprc = MAXUPRC; 217 218 static int sysctl_doeproc(SYSCTLFN_PROTO); 219 static int sysctl_kern_proc_args(SYSCTLFN_PROTO); 220 static int sysctl_security_expose_address(SYSCTLFN_PROTO); 221 222 #ifdef KASLR 223 static int kern_expose_address = 0; 224 #else 225 static int kern_expose_address = 1; 226 #endif 227 /* 228 * The process list descriptors, used during pid allocation and 229 * by sysctl. No locking on this data structure is needed since 230 * it is completely static. 231 */ 232 const struct proclist_desc proclists[] = { 233 { &allproc }, 234 { &zombproc }, 235 { NULL }, 236 }; 237 238 static struct pgrp * pg_remove(pid_t); 239 static void pg_delete(pid_t); 240 static void orphanpg(struct pgrp *); 241 242 static specificdata_domain_t proc_specificdata_domain; 243 244 static pool_cache_t proc_cache; 245 246 static kauth_listener_t proc_listener; 247 248 static void fill_proc(const struct proc *, struct proc *, bool); 249 static int fill_pathname(struct lwp *, pid_t, void *, size_t *); 250 static int fill_cwd(struct lwp *, pid_t, void *, size_t *); 251 252 static int 253 proc_listener_cb(kauth_cred_t cred, kauth_action_t action, void *cookie, 254 void *arg0, void *arg1, void *arg2, void *arg3) 255 { 256 struct proc *p; 257 int result; 258 259 result = KAUTH_RESULT_DEFER; 260 p = arg0; 261 262 switch (action) { 263 case KAUTH_PROCESS_CANSEE: { 264 enum kauth_process_req req; 265 266 req = (enum kauth_process_req)(uintptr_t)arg1; 267 268 switch (req) { 269 case KAUTH_REQ_PROCESS_CANSEE_ARGS: 270 case KAUTH_REQ_PROCESS_CANSEE_ENTRY: 271 case KAUTH_REQ_PROCESS_CANSEE_OPENFILES: 272 case KAUTH_REQ_PROCESS_CANSEE_EPROC: 273 result = KAUTH_RESULT_ALLOW; 274 break; 275 276 case KAUTH_REQ_PROCESS_CANSEE_ENV: 277 if (kauth_cred_getuid(cred) != 278 kauth_cred_getuid(p->p_cred) || 279 kauth_cred_getuid(cred) != 280 kauth_cred_getsvuid(p->p_cred)) 281 break; 282 283 result = KAUTH_RESULT_ALLOW; 284 285 break; 286 287 case KAUTH_REQ_PROCESS_CANSEE_KPTR: 288 if (!kern_expose_address) 289 break; 290 291 if (kern_expose_address == 1 && !(p->p_flag & PK_KMEM)) 292 break; 293 294 result = KAUTH_RESULT_ALLOW; 295 296 break; 297 298 default: 299 break; 300 } 301 302 break; 303 } 304 305 case KAUTH_PROCESS_FORK: { 306 int lnprocs = (int)(unsigned long)arg2; 307 308 /* 309 * Don't allow a nonprivileged user to use the last few 310 * processes. The variable lnprocs is the current number of 311 * processes, maxproc is the limit. 312 */ 313 if (__predict_false((lnprocs >= maxproc - 5))) 314 break; 315 316 result = KAUTH_RESULT_ALLOW; 317 318 break; 319 } 320 321 case KAUTH_PROCESS_CORENAME: 322 case KAUTH_PROCESS_STOPFLAG: 323 if (proc_uidmatch(cred, p->p_cred) == 0) 324 result = KAUTH_RESULT_ALLOW; 325 326 break; 327 328 default: 329 break; 330 } 331 332 return result; 333 } 334 335 static int 336 proc_ctor(void *arg __unused, void *obj, int flags __unused) 337 { 338 memset(obj, 0, sizeof(struct proc)); 339 return 0; 340 } 341 342 /* 343 * Initialize global process hashing structures. 344 */ 345 void 346 procinit(void) 347 { 348 const struct proclist_desc *pd; 349 u_int i; 350 #define LINK_EMPTY ((PID_MAX + INITIAL_PID_TABLE_SIZE) & ~(INITIAL_PID_TABLE_SIZE - 1)) 351 352 for (pd = proclists; pd->pd_list != NULL; pd++) 353 LIST_INIT(pd->pd_list); 354 355 proc_lock = mutex_obj_alloc(MUTEX_DEFAULT, IPL_NONE); 356 pid_table = kmem_alloc(INITIAL_PID_TABLE_SIZE 357 * sizeof(struct pid_table), KM_SLEEP); 358 pid_tbl_mask = INITIAL_PID_TABLE_SIZE - 1; 359 pid_max = PID_MAX; 360 361 /* Set free list running through table... 362 Preset 'use count' above PID_MAX so we allocate pid 1 next. */ 363 for (i = 0; i <= pid_tbl_mask; i++) { 364 pid_table[i].pt_proc = P_FREE(LINK_EMPTY + i + 1); 365 pid_table[i].pt_pgrp = 0; 366 pid_table[i].pt_pid = 0; 367 } 368 /* slot 0 is just grabbed */ 369 next_free_pt = 1; 370 /* Need to fix last entry. */ 371 last_free_pt = pid_tbl_mask; 372 pid_table[last_free_pt].pt_proc = P_FREE(LINK_EMPTY); 373 /* point at which we grow table - to avoid reusing pids too often */ 374 pid_alloc_lim = pid_tbl_mask - 1; 375 #undef LINK_EMPTY 376 377 proc_specificdata_domain = specificdata_domain_create(); 378 KASSERT(proc_specificdata_domain != NULL); 379 380 proc_cache = pool_cache_init(sizeof(struct proc), coherency_unit, 0, 0, 381 "procpl", NULL, IPL_NONE, proc_ctor, NULL, NULL); 382 383 proc_listener = kauth_listen_scope(KAUTH_SCOPE_PROCESS, 384 proc_listener_cb, NULL); 385 } 386 387 void 388 procinit_sysctl(void) 389 { 390 static struct sysctllog *clog; 391 392 sysctl_createv(&clog, 0, NULL, NULL, 393 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 394 CTLTYPE_INT, "expose_address", 395 SYSCTL_DESCR("Enable exposing kernel addresses"), 396 sysctl_security_expose_address, 0, 397 &kern_expose_address, 0, CTL_KERN, CTL_CREATE, CTL_EOL); 398 sysctl_createv(&clog, 0, NULL, NULL, 399 CTLFLAG_PERMANENT, 400 CTLTYPE_NODE, "proc", 401 SYSCTL_DESCR("System-wide process information"), 402 sysctl_doeproc, 0, NULL, 0, 403 CTL_KERN, KERN_PROC, CTL_EOL); 404 sysctl_createv(&clog, 0, NULL, NULL, 405 CTLFLAG_PERMANENT, 406 CTLTYPE_NODE, "proc2", 407 SYSCTL_DESCR("Machine-independent process information"), 408 sysctl_doeproc, 0, NULL, 0, 409 CTL_KERN, KERN_PROC2, CTL_EOL); 410 sysctl_createv(&clog, 0, NULL, NULL, 411 CTLFLAG_PERMANENT, 412 CTLTYPE_NODE, "proc_args", 413 SYSCTL_DESCR("Process argument information"), 414 sysctl_kern_proc_args, 0, NULL, 0, 415 CTL_KERN, KERN_PROC_ARGS, CTL_EOL); 416 417 /* 418 "nodes" under these: 419 420 KERN_PROC_ALL 421 KERN_PROC_PID pid 422 KERN_PROC_PGRP pgrp 423 KERN_PROC_SESSION sess 424 KERN_PROC_TTY tty 425 KERN_PROC_UID uid 426 KERN_PROC_RUID uid 427 KERN_PROC_GID gid 428 KERN_PROC_RGID gid 429 430 all in all, probably not worth the effort... 431 */ 432 } 433 434 /* 435 * Initialize process 0. 436 */ 437 void 438 proc0_init(void) 439 { 440 struct proc *p; 441 struct pgrp *pg; 442 struct rlimit *rlim; 443 rlim_t lim; 444 int error __diagused; 445 int i; 446 447 p = &proc0; 448 pg = &pgrp0; 449 450 mutex_init(&p->p_stmutex, MUTEX_DEFAULT, IPL_HIGH); 451 mutex_init(&p->p_auxlock, MUTEX_DEFAULT, IPL_NONE); 452 p->p_lock = mutex_obj_alloc(MUTEX_DEFAULT, IPL_NONE); 453 454 rw_init(&p->p_reflock); 455 rw_init(&p->p_treelock); 456 cv_init(&p->p_waitcv, "wait"); 457 cv_init(&p->p_lwpcv, "lwpwait"); 458 459 LIST_INSERT_HEAD(&p->p_lwps, &lwp0, l_sibling); 460 radix_tree_init_tree(&p->p_lwptree); 461 error = radix_tree_insert_node(&p->p_lwptree, 462 (uint64_t)(lwp0.l_lid - 1), &lwp0); 463 KASSERT(error == 0); 464 465 pid_table[0].pt_proc = p; 466 LIST_INSERT_HEAD(&allproc, p, p_list); 467 468 pid_table[0].pt_pgrp = pg; 469 LIST_INSERT_HEAD(&pg->pg_members, p, p_pglist); 470 471 #ifdef __HAVE_SYSCALL_INTERN 472 (*p->p_emul->e_syscall_intern)(p); 473 #endif 474 475 /* Create credentials. */ 476 cred0 = kauth_cred_alloc(); 477 p->p_cred = cred0; 478 479 /* Create the CWD info. */ 480 mutex_init(&cwdi0.cwdi_lock, MUTEX_DEFAULT, IPL_NONE); 481 482 /* Create the limits structures. */ 483 mutex_init(&limit0.pl_lock, MUTEX_DEFAULT, IPL_NONE); 484 485 rlim = limit0.pl_rlimit; 486 for (i = 0; i < __arraycount(limit0.pl_rlimit); i++) { 487 rlim[i].rlim_cur = RLIM_INFINITY; 488 rlim[i].rlim_max = RLIM_INFINITY; 489 } 490 491 rlim[RLIMIT_NOFILE].rlim_max = maxfiles; 492 rlim[RLIMIT_NOFILE].rlim_cur = maxfiles < nofile ? maxfiles : nofile; 493 494 rlim[RLIMIT_NPROC].rlim_max = maxproc; 495 rlim[RLIMIT_NPROC].rlim_cur = maxproc < maxuprc ? maxproc : maxuprc; 496 497 lim = MIN(VM_MAXUSER_ADDRESS, ctob((rlim_t)uvm_availmem())); 498 rlim[RLIMIT_RSS].rlim_max = lim; 499 rlim[RLIMIT_MEMLOCK].rlim_max = lim; 500 rlim[RLIMIT_MEMLOCK].rlim_cur = lim / 3; 501 502 rlim[RLIMIT_NTHR].rlim_max = maxlwp; 503 rlim[RLIMIT_NTHR].rlim_cur = maxlwp < maxuprc ? maxlwp : maxuprc; 504 505 /* Note that default core name has zero length. */ 506 limit0.pl_corename = defcorename; 507 limit0.pl_cnlen = 0; 508 limit0.pl_refcnt = 1; 509 limit0.pl_writeable = false; 510 limit0.pl_sv_limit = NULL; 511 512 /* Configure virtual memory system, set vm rlimits. */ 513 uvm_init_limits(p); 514 515 /* Initialize file descriptor table for proc0. */ 516 fd_init(&filedesc0); 517 518 /* 519 * Initialize proc0's vmspace, which uses the kernel pmap. 520 * All kernel processes (which never have user space mappings) 521 * share proc0's vmspace, and thus, the kernel pmap. 522 */ 523 uvmspace_init(&vmspace0, pmap_kernel(), round_page(VM_MIN_ADDRESS), 524 trunc_page(VM_MAXUSER_ADDRESS), 525 #ifdef __USE_TOPDOWN_VM 526 true 527 #else 528 false 529 #endif 530 ); 531 532 /* Initialize signal state for proc0. XXX IPL_SCHED */ 533 mutex_init(&p->p_sigacts->sa_mutex, MUTEX_DEFAULT, IPL_SCHED); 534 siginit(p); 535 536 proc_initspecific(p); 537 kdtrace_proc_ctor(NULL, p); 538 } 539 540 /* 541 * Session reference counting. 542 */ 543 544 void 545 proc_sesshold(struct session *ss) 546 { 547 548 KASSERT(mutex_owned(proc_lock)); 549 ss->s_count++; 550 } 551 552 void 553 proc_sessrele(struct session *ss) 554 { 555 556 KASSERT(mutex_owned(proc_lock)); 557 /* 558 * We keep the pgrp with the same id as the session in order to 559 * stop a process being given the same pid. Since the pgrp holds 560 * a reference to the session, it must be a 'zombie' pgrp by now. 561 */ 562 if (--ss->s_count == 0) { 563 struct pgrp *pg; 564 565 pg = pg_remove(ss->s_sid); 566 mutex_exit(proc_lock); 567 568 kmem_free(pg, sizeof(struct pgrp)); 569 kmem_free(ss, sizeof(struct session)); 570 } else { 571 mutex_exit(proc_lock); 572 } 573 } 574 575 /* 576 * Check that the specified process group is in the session of the 577 * specified process. 578 * Treats -ve ids as process ids. 579 * Used to validate TIOCSPGRP requests. 580 */ 581 int 582 pgid_in_session(struct proc *p, pid_t pg_id) 583 { 584 struct pgrp *pgrp; 585 struct session *session; 586 int error; 587 588 mutex_enter(proc_lock); 589 if (pg_id < 0) { 590 struct proc *p1 = proc_find(-pg_id); 591 if (p1 == NULL) { 592 error = EINVAL; 593 goto fail; 594 } 595 pgrp = p1->p_pgrp; 596 } else { 597 pgrp = pgrp_find(pg_id); 598 if (pgrp == NULL) { 599 error = EINVAL; 600 goto fail; 601 } 602 } 603 session = pgrp->pg_session; 604 error = (session != p->p_pgrp->pg_session) ? EPERM : 0; 605 fail: 606 mutex_exit(proc_lock); 607 return error; 608 } 609 610 /* 611 * p_inferior: is p an inferior of q? 612 */ 613 static inline bool 614 p_inferior(struct proc *p, struct proc *q) 615 { 616 617 KASSERT(mutex_owned(proc_lock)); 618 619 for (; p != q; p = p->p_pptr) 620 if (p->p_pid == 0) 621 return false; 622 return true; 623 } 624 625 /* 626 * proc_find: locate a process by the ID. 627 * 628 * => Must be called with proc_lock held. 629 */ 630 proc_t * 631 proc_find_raw(pid_t pid) 632 { 633 struct pid_table *pt; 634 proc_t *p; 635 636 KASSERT(mutex_owned(proc_lock)); 637 pt = &pid_table[pid & pid_tbl_mask]; 638 p = pt->pt_proc; 639 if (__predict_false(!P_VALID(p) || pt->pt_pid != pid)) { 640 return NULL; 641 } 642 return p; 643 } 644 645 proc_t * 646 proc_find(pid_t pid) 647 { 648 proc_t *p; 649 650 p = proc_find_raw(pid); 651 if (__predict_false(p == NULL)) { 652 return NULL; 653 } 654 655 /* 656 * Only allow live processes to be found by PID. 657 * XXX: p_stat might change, since unlocked. 658 */ 659 if (__predict_true(p->p_stat == SACTIVE || p->p_stat == SSTOP)) { 660 return p; 661 } 662 return NULL; 663 } 664 665 /* 666 * pgrp_find: locate a process group by the ID. 667 * 668 * => Must be called with proc_lock held. 669 */ 670 struct pgrp * 671 pgrp_find(pid_t pgid) 672 { 673 struct pgrp *pg; 674 675 KASSERT(mutex_owned(proc_lock)); 676 677 pg = pid_table[pgid & pid_tbl_mask].pt_pgrp; 678 679 /* 680 * Cannot look up a process group that only exists because the 681 * session has not died yet (traditional). 682 */ 683 if (pg == NULL || pg->pg_id != pgid || LIST_EMPTY(&pg->pg_members)) { 684 return NULL; 685 } 686 return pg; 687 } 688 689 static void 690 expand_pid_table(void) 691 { 692 size_t pt_size, tsz; 693 struct pid_table *n_pt, *new_pt; 694 struct proc *proc; 695 struct pgrp *pgrp; 696 pid_t pid, rpid; 697 u_int i; 698 uint new_pt_mask; 699 700 pt_size = pid_tbl_mask + 1; 701 tsz = pt_size * 2 * sizeof(struct pid_table); 702 new_pt = kmem_alloc(tsz, KM_SLEEP); 703 new_pt_mask = pt_size * 2 - 1; 704 705 mutex_enter(proc_lock); 706 if (pt_size != pid_tbl_mask + 1) { 707 /* Another process beat us to it... */ 708 mutex_exit(proc_lock); 709 kmem_free(new_pt, tsz); 710 return; 711 } 712 713 /* 714 * Copy entries from old table into new one. 715 * If 'pid' is 'odd' we need to place in the upper half, 716 * even pid's to the lower half. 717 * Free items stay in the low half so we don't have to 718 * fixup the reference to them. 719 * We stuff free items on the front of the freelist 720 * because we can't write to unmodified entries. 721 * Processing the table backwards maintains a semblance 722 * of issuing pid numbers that increase with time. 723 */ 724 i = pt_size - 1; 725 n_pt = new_pt + i; 726 for (; ; i--, n_pt--) { 727 proc = pid_table[i].pt_proc; 728 pgrp = pid_table[i].pt_pgrp; 729 if (!P_VALID(proc)) { 730 /* Up 'use count' so that link is valid */ 731 pid = (P_NEXT(proc) + pt_size) & ~pt_size; 732 rpid = 0; 733 proc = P_FREE(pid); 734 if (pgrp) 735 pid = pgrp->pg_id; 736 } else { 737 pid = pid_table[i].pt_pid; 738 rpid = pid; 739 } 740 741 /* Save entry in appropriate half of table */ 742 n_pt[pid & pt_size].pt_proc = proc; 743 n_pt[pid & pt_size].pt_pgrp = pgrp; 744 n_pt[pid & pt_size].pt_pid = rpid; 745 746 /* Put other piece on start of free list */ 747 pid = (pid ^ pt_size) & ~pid_tbl_mask; 748 n_pt[pid & pt_size].pt_proc = 749 P_FREE((pid & ~pt_size) | next_free_pt); 750 n_pt[pid & pt_size].pt_pgrp = 0; 751 n_pt[pid & pt_size].pt_pid = 0; 752 753 next_free_pt = i | (pid & pt_size); 754 if (i == 0) 755 break; 756 } 757 758 /* Save old table size and switch tables */ 759 tsz = pt_size * sizeof(struct pid_table); 760 n_pt = pid_table; 761 pid_table = new_pt; 762 pid_tbl_mask = new_pt_mask; 763 764 /* 765 * pid_max starts as PID_MAX (= 30000), once we have 16384 766 * allocated pids we need it to be larger! 767 */ 768 if (pid_tbl_mask > PID_MAX) { 769 pid_max = pid_tbl_mask * 2 + 1; 770 pid_alloc_lim |= pid_alloc_lim << 1; 771 } else 772 pid_alloc_lim <<= 1; /* doubles number of free slots... */ 773 774 mutex_exit(proc_lock); 775 kmem_free(n_pt, tsz); 776 } 777 778 struct proc * 779 proc_alloc(void) 780 { 781 struct proc *p; 782 783 p = pool_cache_get(proc_cache, PR_WAITOK); 784 p->p_stat = SIDL; /* protect against others */ 785 proc_initspecific(p); 786 kdtrace_proc_ctor(NULL, p); 787 p->p_pid = -1; 788 proc_alloc_pid(p); 789 return p; 790 } 791 792 /* 793 * proc_alloc_pid: allocate PID and record the given proc 'p' so that 794 * proc_find_raw() can find it by the PID. 795 */ 796 797 pid_t 798 proc_alloc_pid(struct proc *p) 799 { 800 struct pid_table *pt; 801 pid_t pid; 802 int nxt; 803 804 for (;;expand_pid_table()) { 805 if (__predict_false(pid_alloc_cnt >= pid_alloc_lim)) 806 /* ensure pids cycle through 2000+ values */ 807 continue; 808 mutex_enter(proc_lock); 809 pt = &pid_table[next_free_pt]; 810 #ifdef DIAGNOSTIC 811 if (__predict_false(P_VALID(pt->pt_proc) || pt->pt_pgrp)) 812 panic("proc_alloc: slot busy"); 813 #endif 814 nxt = P_NEXT(pt->pt_proc); 815 if (nxt & pid_tbl_mask) 816 break; 817 /* Table full - expand (NB last entry not used....) */ 818 mutex_exit(proc_lock); 819 } 820 821 /* pid is 'saved use count' + 'size' + entry */ 822 pid = (nxt & ~pid_tbl_mask) + pid_tbl_mask + 1 + next_free_pt; 823 if ((uint)pid > (uint)pid_max) 824 pid &= pid_tbl_mask; 825 next_free_pt = nxt & pid_tbl_mask; 826 827 /* Grab table slot */ 828 pt->pt_proc = p; 829 830 KASSERT(pt->pt_pid == 0); 831 pt->pt_pid = pid; 832 if (p->p_pid == -1) { 833 p->p_pid = pid; 834 } 835 pid_alloc_cnt++; 836 mutex_exit(proc_lock); 837 838 return pid; 839 } 840 841 /* 842 * Free a process id - called from proc_free (in kern_exit.c) 843 * 844 * Called with the proc_lock held. 845 */ 846 void 847 proc_free_pid(pid_t pid) 848 { 849 struct pid_table *pt; 850 851 KASSERT(mutex_owned(proc_lock)); 852 853 pt = &pid_table[pid & pid_tbl_mask]; 854 855 /* save pid use count in slot */ 856 pt->pt_proc = P_FREE(pid & ~pid_tbl_mask); 857 KASSERT(pt->pt_pid == pid); 858 pt->pt_pid = 0; 859 860 if (pt->pt_pgrp == NULL) { 861 /* link last freed entry onto ours */ 862 pid &= pid_tbl_mask; 863 pt = &pid_table[last_free_pt]; 864 pt->pt_proc = P_FREE(P_NEXT(pt->pt_proc) | pid); 865 pt->pt_pid = 0; 866 last_free_pt = pid; 867 pid_alloc_cnt--; 868 } 869 870 atomic_dec_uint(&nprocs); 871 } 872 873 void 874 proc_free_mem(struct proc *p) 875 { 876 877 kdtrace_proc_dtor(NULL, p); 878 pool_cache_put(proc_cache, p); 879 } 880 881 /* 882 * proc_enterpgrp: move p to a new or existing process group (and session). 883 * 884 * If we are creating a new pgrp, the pgid should equal 885 * the calling process' pid. 886 * If is only valid to enter a process group that is in the session 887 * of the process. 888 * Also mksess should only be set if we are creating a process group 889 * 890 * Only called from sys_setsid, sys_setpgid and posix_spawn/spawn_return. 891 */ 892 int 893 proc_enterpgrp(struct proc *curp, pid_t pid, pid_t pgid, bool mksess) 894 { 895 struct pgrp *new_pgrp, *pgrp; 896 struct session *sess; 897 struct proc *p; 898 int rval; 899 pid_t pg_id = NO_PGID; 900 901 sess = mksess ? kmem_alloc(sizeof(*sess), KM_SLEEP) : NULL; 902 903 /* Allocate data areas we might need before doing any validity checks */ 904 mutex_enter(proc_lock); /* Because pid_table might change */ 905 if (pid_table[pgid & pid_tbl_mask].pt_pgrp == 0) { 906 mutex_exit(proc_lock); 907 new_pgrp = kmem_alloc(sizeof(*new_pgrp), KM_SLEEP); 908 mutex_enter(proc_lock); 909 } else 910 new_pgrp = NULL; 911 rval = EPERM; /* most common error (to save typing) */ 912 913 /* Check pgrp exists or can be created */ 914 pgrp = pid_table[pgid & pid_tbl_mask].pt_pgrp; 915 if (pgrp != NULL && pgrp->pg_id != pgid) 916 goto done; 917 918 /* Can only set another process under restricted circumstances. */ 919 if (pid != curp->p_pid) { 920 /* Must exist and be one of our children... */ 921 p = proc_find(pid); 922 if (p == NULL || !p_inferior(p, curp)) { 923 rval = ESRCH; 924 goto done; 925 } 926 /* ... in the same session... */ 927 if (sess != NULL || p->p_session != curp->p_session) 928 goto done; 929 /* ... existing pgid must be in same session ... */ 930 if (pgrp != NULL && pgrp->pg_session != p->p_session) 931 goto done; 932 /* ... and not done an exec. */ 933 if (p->p_flag & PK_EXEC) { 934 rval = EACCES; 935 goto done; 936 } 937 } else { 938 /* ... setsid() cannot re-enter a pgrp */ 939 if (mksess && (curp->p_pgid == curp->p_pid || 940 pgrp_find(curp->p_pid))) 941 goto done; 942 p = curp; 943 } 944 945 /* Changing the process group/session of a session 946 leader is definitely off limits. */ 947 if (SESS_LEADER(p)) { 948 if (sess == NULL && p->p_pgrp == pgrp) 949 /* unless it's a definite noop */ 950 rval = 0; 951 goto done; 952 } 953 954 /* Can only create a process group with id of process */ 955 if (pgrp == NULL && pgid != pid) 956 goto done; 957 958 /* Can only create a session if creating pgrp */ 959 if (sess != NULL && pgrp != NULL) 960 goto done; 961 962 /* Check we allocated memory for a pgrp... */ 963 if (pgrp == NULL && new_pgrp == NULL) 964 goto done; 965 966 /* Don't attach to 'zombie' pgrp */ 967 if (pgrp != NULL && LIST_EMPTY(&pgrp->pg_members)) 968 goto done; 969 970 /* Expect to succeed now */ 971 rval = 0; 972 973 if (pgrp == p->p_pgrp) 974 /* nothing to do */ 975 goto done; 976 977 /* Ok all setup, link up required structures */ 978 979 if (pgrp == NULL) { 980 pgrp = new_pgrp; 981 new_pgrp = NULL; 982 if (sess != NULL) { 983 sess->s_sid = p->p_pid; 984 sess->s_leader = p; 985 sess->s_count = 1; 986 sess->s_ttyvp = NULL; 987 sess->s_ttyp = NULL; 988 sess->s_flags = p->p_session->s_flags & ~S_LOGIN_SET; 989 memcpy(sess->s_login, p->p_session->s_login, 990 sizeof(sess->s_login)); 991 p->p_lflag &= ~PL_CONTROLT; 992 } else { 993 sess = p->p_pgrp->pg_session; 994 proc_sesshold(sess); 995 } 996 pgrp->pg_session = sess; 997 sess = NULL; 998 999 pgrp->pg_id = pgid; 1000 LIST_INIT(&pgrp->pg_members); 1001 #ifdef DIAGNOSTIC 1002 if (__predict_false(pid_table[pgid & pid_tbl_mask].pt_pgrp)) 1003 panic("enterpgrp: pgrp table slot in use"); 1004 if (__predict_false(mksess && p != curp)) 1005 panic("enterpgrp: mksession and p != curproc"); 1006 #endif 1007 pid_table[pgid & pid_tbl_mask].pt_pgrp = pgrp; 1008 pgrp->pg_jobc = 0; 1009 } 1010 1011 /* 1012 * Adjust eligibility of affected pgrps to participate in job control. 1013 * Increment eligibility counts before decrementing, otherwise we 1014 * could reach 0 spuriously during the first call. 1015 */ 1016 fixjobc(p, pgrp, 1); 1017 fixjobc(p, p->p_pgrp, 0); 1018 1019 /* Interlock with ttread(). */ 1020 mutex_spin_enter(&tty_lock); 1021 1022 /* Move process to requested group. */ 1023 LIST_REMOVE(p, p_pglist); 1024 if (LIST_EMPTY(&p->p_pgrp->pg_members)) 1025 /* defer delete until we've dumped the lock */ 1026 pg_id = p->p_pgrp->pg_id; 1027 p->p_pgrp = pgrp; 1028 LIST_INSERT_HEAD(&pgrp->pg_members, p, p_pglist); 1029 1030 /* Done with the swap; we can release the tty mutex. */ 1031 mutex_spin_exit(&tty_lock); 1032 1033 done: 1034 if (pg_id != NO_PGID) { 1035 /* Releases proc_lock. */ 1036 pg_delete(pg_id); 1037 } else { 1038 mutex_exit(proc_lock); 1039 } 1040 if (sess != NULL) 1041 kmem_free(sess, sizeof(*sess)); 1042 if (new_pgrp != NULL) 1043 kmem_free(new_pgrp, sizeof(*new_pgrp)); 1044 #ifdef DEBUG_PGRP 1045 if (__predict_false(rval)) 1046 printf("enterpgrp(%d,%d,%d), curproc %d, rval %d\n", 1047 pid, pgid, mksess, curp->p_pid, rval); 1048 #endif 1049 return rval; 1050 } 1051 1052 /* 1053 * proc_leavepgrp: remove a process from its process group. 1054 * => must be called with the proc_lock held, which will be released; 1055 */ 1056 void 1057 proc_leavepgrp(struct proc *p) 1058 { 1059 struct pgrp *pgrp; 1060 1061 KASSERT(mutex_owned(proc_lock)); 1062 1063 /* Interlock with ttread() */ 1064 mutex_spin_enter(&tty_lock); 1065 pgrp = p->p_pgrp; 1066 LIST_REMOVE(p, p_pglist); 1067 p->p_pgrp = NULL; 1068 mutex_spin_exit(&tty_lock); 1069 1070 if (LIST_EMPTY(&pgrp->pg_members)) { 1071 /* Releases proc_lock. */ 1072 pg_delete(pgrp->pg_id); 1073 } else { 1074 mutex_exit(proc_lock); 1075 } 1076 } 1077 1078 /* 1079 * pg_remove: remove a process group from the table. 1080 * => must be called with the proc_lock held; 1081 * => returns process group to free; 1082 */ 1083 static struct pgrp * 1084 pg_remove(pid_t pg_id) 1085 { 1086 struct pgrp *pgrp; 1087 struct pid_table *pt; 1088 1089 KASSERT(mutex_owned(proc_lock)); 1090 1091 pt = &pid_table[pg_id & pid_tbl_mask]; 1092 pgrp = pt->pt_pgrp; 1093 1094 KASSERT(pgrp != NULL); 1095 KASSERT(pgrp->pg_id == pg_id); 1096 KASSERT(LIST_EMPTY(&pgrp->pg_members)); 1097 1098 pt->pt_pgrp = NULL; 1099 1100 if (!P_VALID(pt->pt_proc)) { 1101 /* Orphaned pgrp, put slot onto free list. */ 1102 KASSERT((P_NEXT(pt->pt_proc) & pid_tbl_mask) == 0); 1103 pg_id &= pid_tbl_mask; 1104 pt = &pid_table[last_free_pt]; 1105 pt->pt_proc = P_FREE(P_NEXT(pt->pt_proc) | pg_id); 1106 KASSERT(pt->pt_pid == 0); 1107 last_free_pt = pg_id; 1108 pid_alloc_cnt--; 1109 } 1110 return pgrp; 1111 } 1112 1113 /* 1114 * pg_delete: delete and free a process group. 1115 * => must be called with the proc_lock held, which will be released. 1116 */ 1117 static void 1118 pg_delete(pid_t pg_id) 1119 { 1120 struct pgrp *pg; 1121 struct tty *ttyp; 1122 struct session *ss; 1123 1124 KASSERT(mutex_owned(proc_lock)); 1125 1126 pg = pid_table[pg_id & pid_tbl_mask].pt_pgrp; 1127 if (pg == NULL || pg->pg_id != pg_id || !LIST_EMPTY(&pg->pg_members)) { 1128 mutex_exit(proc_lock); 1129 return; 1130 } 1131 1132 ss = pg->pg_session; 1133 1134 /* Remove reference (if any) from tty to this process group */ 1135 mutex_spin_enter(&tty_lock); 1136 ttyp = ss->s_ttyp; 1137 if (ttyp != NULL && ttyp->t_pgrp == pg) { 1138 ttyp->t_pgrp = NULL; 1139 KASSERT(ttyp->t_session == ss); 1140 } 1141 mutex_spin_exit(&tty_lock); 1142 1143 /* 1144 * The leading process group in a session is freed by proc_sessrele(), 1145 * if last reference. Note: proc_sessrele() releases proc_lock. 1146 */ 1147 pg = (ss->s_sid != pg->pg_id) ? pg_remove(pg_id) : NULL; 1148 proc_sessrele(ss); 1149 1150 if (pg != NULL) { 1151 /* Free it, if was not done by proc_sessrele(). */ 1152 kmem_free(pg, sizeof(struct pgrp)); 1153 } 1154 } 1155 1156 /* 1157 * Adjust pgrp jobc counters when specified process changes process group. 1158 * We count the number of processes in each process group that "qualify" 1159 * the group for terminal job control (those with a parent in a different 1160 * process group of the same session). If that count reaches zero, the 1161 * process group becomes orphaned. Check both the specified process' 1162 * process group and that of its children. 1163 * entering == 0 => p is leaving specified group. 1164 * entering == 1 => p is entering specified group. 1165 * 1166 * Call with proc_lock held. 1167 */ 1168 void 1169 fixjobc(struct proc *p, struct pgrp *pgrp, int entering) 1170 { 1171 struct pgrp *hispgrp; 1172 struct session *mysession = pgrp->pg_session; 1173 struct proc *child; 1174 1175 KASSERT(mutex_owned(proc_lock)); 1176 1177 /* 1178 * Check p's parent to see whether p qualifies its own process 1179 * group; if so, adjust count for p's process group. 1180 */ 1181 hispgrp = p->p_pptr->p_pgrp; 1182 if (hispgrp != pgrp && hispgrp->pg_session == mysession) { 1183 if (entering) { 1184 pgrp->pg_jobc++; 1185 p->p_lflag &= ~PL_ORPHANPG; 1186 } else if (--pgrp->pg_jobc == 0) 1187 orphanpg(pgrp); 1188 } 1189 1190 /* 1191 * Check this process' children to see whether they qualify 1192 * their process groups; if so, adjust counts for children's 1193 * process groups. 1194 */ 1195 LIST_FOREACH(child, &p->p_children, p_sibling) { 1196 hispgrp = child->p_pgrp; 1197 if (hispgrp != pgrp && hispgrp->pg_session == mysession && 1198 !P_ZOMBIE(child)) { 1199 if (entering) { 1200 child->p_lflag &= ~PL_ORPHANPG; 1201 hispgrp->pg_jobc++; 1202 } else if (--hispgrp->pg_jobc == 0) 1203 orphanpg(hispgrp); 1204 } 1205 } 1206 } 1207 1208 /* 1209 * A process group has become orphaned; 1210 * if there are any stopped processes in the group, 1211 * hang-up all process in that group. 1212 * 1213 * Call with proc_lock held. 1214 */ 1215 static void 1216 orphanpg(struct pgrp *pg) 1217 { 1218 struct proc *p; 1219 1220 KASSERT(mutex_owned(proc_lock)); 1221 1222 LIST_FOREACH(p, &pg->pg_members, p_pglist) { 1223 if (p->p_stat == SSTOP) { 1224 p->p_lflag |= PL_ORPHANPG; 1225 psignal(p, SIGHUP); 1226 psignal(p, SIGCONT); 1227 } 1228 } 1229 } 1230 1231 #ifdef DDB 1232 #include <ddb/db_output.h> 1233 void pidtbl_dump(void); 1234 void 1235 pidtbl_dump(void) 1236 { 1237 struct pid_table *pt; 1238 struct proc *p; 1239 struct pgrp *pgrp; 1240 int id; 1241 1242 db_printf("pid table %p size %x, next %x, last %x\n", 1243 pid_table, pid_tbl_mask+1, 1244 next_free_pt, last_free_pt); 1245 for (pt = pid_table, id = 0; id <= pid_tbl_mask; id++, pt++) { 1246 p = pt->pt_proc; 1247 if (!P_VALID(p) && !pt->pt_pgrp) 1248 continue; 1249 db_printf(" id %x: ", id); 1250 if (P_VALID(p)) 1251 db_printf("slotpid %d proc %p id %d (0x%x) %s\n", 1252 pt->pt_pid, p, p->p_pid, p->p_pid, p->p_comm); 1253 else 1254 db_printf("next %x use %x\n", 1255 P_NEXT(p) & pid_tbl_mask, 1256 P_NEXT(p) & ~pid_tbl_mask); 1257 if ((pgrp = pt->pt_pgrp)) { 1258 db_printf("\tsession %p, sid %d, count %d, login %s\n", 1259 pgrp->pg_session, pgrp->pg_session->s_sid, 1260 pgrp->pg_session->s_count, 1261 pgrp->pg_session->s_login); 1262 db_printf("\tpgrp %p, pg_id %d, pg_jobc %d, members %p\n", 1263 pgrp, pgrp->pg_id, pgrp->pg_jobc, 1264 LIST_FIRST(&pgrp->pg_members)); 1265 LIST_FOREACH(p, &pgrp->pg_members, p_pglist) { 1266 db_printf("\t\tpid %d addr %p pgrp %p %s\n", 1267 p->p_pid, p, p->p_pgrp, p->p_comm); 1268 } 1269 } 1270 } 1271 } 1272 #endif /* DDB */ 1273 1274 #ifdef KSTACK_CHECK_MAGIC 1275 1276 #define KSTACK_MAGIC 0xdeadbeaf 1277 1278 /* XXX should be per process basis? */ 1279 static int kstackleftmin = KSTACK_SIZE; 1280 static int kstackleftthres = KSTACK_SIZE / 8; 1281 1282 void 1283 kstack_setup_magic(const struct lwp *l) 1284 { 1285 uint32_t *ip; 1286 uint32_t const *end; 1287 1288 KASSERT(l != NULL); 1289 KASSERT(l != &lwp0); 1290 1291 /* 1292 * fill all the stack with magic number 1293 * so that later modification on it can be detected. 1294 */ 1295 ip = (uint32_t *)KSTACK_LOWEST_ADDR(l); 1296 end = (uint32_t *)((char *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE); 1297 for (; ip < end; ip++) { 1298 *ip = KSTACK_MAGIC; 1299 } 1300 } 1301 1302 void 1303 kstack_check_magic(const struct lwp *l) 1304 { 1305 uint32_t const *ip, *end; 1306 int stackleft; 1307 1308 KASSERT(l != NULL); 1309 1310 /* don't check proc0 */ /*XXX*/ 1311 if (l == &lwp0) 1312 return; 1313 1314 #ifdef __MACHINE_STACK_GROWS_UP 1315 /* stack grows upwards (eg. hppa) */ 1316 ip = (uint32_t *)((void *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE); 1317 end = (uint32_t *)KSTACK_LOWEST_ADDR(l); 1318 for (ip--; ip >= end; ip--) 1319 if (*ip != KSTACK_MAGIC) 1320 break; 1321 1322 stackleft = (void *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE - (void *)ip; 1323 #else /* __MACHINE_STACK_GROWS_UP */ 1324 /* stack grows downwards (eg. i386) */ 1325 ip = (uint32_t *)KSTACK_LOWEST_ADDR(l); 1326 end = (uint32_t *)((char *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE); 1327 for (; ip < end; ip++) 1328 if (*ip != KSTACK_MAGIC) 1329 break; 1330 1331 stackleft = ((const char *)ip) - (const char *)KSTACK_LOWEST_ADDR(l); 1332 #endif /* __MACHINE_STACK_GROWS_UP */ 1333 1334 if (kstackleftmin > stackleft) { 1335 kstackleftmin = stackleft; 1336 if (stackleft < kstackleftthres) 1337 printf("warning: kernel stack left %d bytes" 1338 "(pid %u:lid %u)\n", stackleft, 1339 (u_int)l->l_proc->p_pid, (u_int)l->l_lid); 1340 } 1341 1342 if (stackleft <= 0) { 1343 panic("magic on the top of kernel stack changed for " 1344 "pid %u, lid %u: maybe kernel stack overflow", 1345 (u_int)l->l_proc->p_pid, (u_int)l->l_lid); 1346 } 1347 } 1348 #endif /* KSTACK_CHECK_MAGIC */ 1349 1350 int 1351 proclist_foreach_call(struct proclist *list, 1352 int (*callback)(struct proc *, void *arg), void *arg) 1353 { 1354 struct proc marker; 1355 struct proc *p; 1356 int ret = 0; 1357 1358 marker.p_flag = PK_MARKER; 1359 mutex_enter(proc_lock); 1360 for (p = LIST_FIRST(list); ret == 0 && p != NULL;) { 1361 if (p->p_flag & PK_MARKER) { 1362 p = LIST_NEXT(p, p_list); 1363 continue; 1364 } 1365 LIST_INSERT_AFTER(p, &marker, p_list); 1366 ret = (*callback)(p, arg); 1367 KASSERT(mutex_owned(proc_lock)); 1368 p = LIST_NEXT(&marker, p_list); 1369 LIST_REMOVE(&marker, p_list); 1370 } 1371 mutex_exit(proc_lock); 1372 1373 return ret; 1374 } 1375 1376 int 1377 proc_vmspace_getref(struct proc *p, struct vmspace **vm) 1378 { 1379 1380 /* XXXCDC: how should locking work here? */ 1381 1382 /* curproc exception is for coredump. */ 1383 1384 if ((p != curproc && (p->p_sflag & PS_WEXIT) != 0) || 1385 (p->p_vmspace->vm_refcnt < 1)) { /* XXX */ 1386 return EFAULT; 1387 } 1388 1389 uvmspace_addref(p->p_vmspace); 1390 *vm = p->p_vmspace; 1391 1392 return 0; 1393 } 1394 1395 /* 1396 * Acquire a write lock on the process credential. 1397 */ 1398 void 1399 proc_crmod_enter(void) 1400 { 1401 struct lwp *l = curlwp; 1402 struct proc *p = l->l_proc; 1403 kauth_cred_t oc; 1404 1405 /* Reset what needs to be reset in plimit. */ 1406 if (p->p_limit->pl_corename != defcorename) { 1407 lim_setcorename(p, defcorename, 0); 1408 } 1409 1410 mutex_enter(p->p_lock); 1411 1412 /* Ensure the LWP cached credentials are up to date. */ 1413 if ((oc = l->l_cred) != p->p_cred) { 1414 kauth_cred_hold(p->p_cred); 1415 l->l_cred = p->p_cred; 1416 kauth_cred_free(oc); 1417 } 1418 } 1419 1420 /* 1421 * Set in a new process credential, and drop the write lock. The credential 1422 * must have a reference already. Optionally, free a no-longer required 1423 * credential. The scheduler also needs to inspect p_cred, so we also 1424 * briefly acquire the sched state mutex. 1425 */ 1426 void 1427 proc_crmod_leave(kauth_cred_t scred, kauth_cred_t fcred, bool sugid) 1428 { 1429 struct lwp *l = curlwp, *l2; 1430 struct proc *p = l->l_proc; 1431 kauth_cred_t oc; 1432 1433 KASSERT(mutex_owned(p->p_lock)); 1434 1435 /* Is there a new credential to set in? */ 1436 if (scred != NULL) { 1437 p->p_cred = scred; 1438 LIST_FOREACH(l2, &p->p_lwps, l_sibling) { 1439 if (l2 != l) 1440 l2->l_prflag |= LPR_CRMOD; 1441 } 1442 1443 /* Ensure the LWP cached credentials are up to date. */ 1444 if ((oc = l->l_cred) != scred) { 1445 kauth_cred_hold(scred); 1446 l->l_cred = scred; 1447 } 1448 } else 1449 oc = NULL; /* XXXgcc */ 1450 1451 if (sugid) { 1452 /* 1453 * Mark process as having changed credentials, stops 1454 * tracing etc. 1455 */ 1456 p->p_flag |= PK_SUGID; 1457 } 1458 1459 mutex_exit(p->p_lock); 1460 1461 /* If there is a credential to be released, free it now. */ 1462 if (fcred != NULL) { 1463 KASSERT(scred != NULL); 1464 kauth_cred_free(fcred); 1465 if (oc != scred) 1466 kauth_cred_free(oc); 1467 } 1468 } 1469 1470 /* 1471 * proc_specific_key_create -- 1472 * Create a key for subsystem proc-specific data. 1473 */ 1474 int 1475 proc_specific_key_create(specificdata_key_t *keyp, specificdata_dtor_t dtor) 1476 { 1477 1478 return (specificdata_key_create(proc_specificdata_domain, keyp, dtor)); 1479 } 1480 1481 /* 1482 * proc_specific_key_delete -- 1483 * Delete a key for subsystem proc-specific data. 1484 */ 1485 void 1486 proc_specific_key_delete(specificdata_key_t key) 1487 { 1488 1489 specificdata_key_delete(proc_specificdata_domain, key); 1490 } 1491 1492 /* 1493 * proc_initspecific -- 1494 * Initialize a proc's specificdata container. 1495 */ 1496 void 1497 proc_initspecific(struct proc *p) 1498 { 1499 int error __diagused; 1500 1501 error = specificdata_init(proc_specificdata_domain, &p->p_specdataref); 1502 KASSERT(error == 0); 1503 } 1504 1505 /* 1506 * proc_finispecific -- 1507 * Finalize a proc's specificdata container. 1508 */ 1509 void 1510 proc_finispecific(struct proc *p) 1511 { 1512 1513 specificdata_fini(proc_specificdata_domain, &p->p_specdataref); 1514 } 1515 1516 /* 1517 * proc_getspecific -- 1518 * Return proc-specific data corresponding to the specified key. 1519 */ 1520 void * 1521 proc_getspecific(struct proc *p, specificdata_key_t key) 1522 { 1523 1524 return (specificdata_getspecific(proc_specificdata_domain, 1525 &p->p_specdataref, key)); 1526 } 1527 1528 /* 1529 * proc_setspecific -- 1530 * Set proc-specific data corresponding to the specified key. 1531 */ 1532 void 1533 proc_setspecific(struct proc *p, specificdata_key_t key, void *data) 1534 { 1535 1536 specificdata_setspecific(proc_specificdata_domain, 1537 &p->p_specdataref, key, data); 1538 } 1539 1540 int 1541 proc_uidmatch(kauth_cred_t cred, kauth_cred_t target) 1542 { 1543 int r = 0; 1544 1545 if (kauth_cred_getuid(cred) != kauth_cred_getuid(target) || 1546 kauth_cred_getuid(cred) != kauth_cred_getsvuid(target)) { 1547 /* 1548 * suid proc of ours or proc not ours 1549 */ 1550 r = EPERM; 1551 } else if (kauth_cred_getgid(target) != kauth_cred_getsvgid(target)) { 1552 /* 1553 * sgid proc has sgid back to us temporarily 1554 */ 1555 r = EPERM; 1556 } else { 1557 /* 1558 * our rgid must be in target's group list (ie, 1559 * sub-processes started by a sgid process) 1560 */ 1561 int ismember = 0; 1562 1563 if (kauth_cred_ismember_gid(cred, 1564 kauth_cred_getgid(target), &ismember) != 0 || 1565 !ismember) 1566 r = EPERM; 1567 } 1568 1569 return (r); 1570 } 1571 1572 /* 1573 * sysctl stuff 1574 */ 1575 1576 #define KERN_PROCSLOP (5 * sizeof(struct kinfo_proc)) 1577 1578 static const u_int sysctl_flagmap[] = { 1579 PK_ADVLOCK, P_ADVLOCK, 1580 PK_EXEC, P_EXEC, 1581 PK_NOCLDWAIT, P_NOCLDWAIT, 1582 PK_32, P_32, 1583 PK_CLDSIGIGN, P_CLDSIGIGN, 1584 PK_SUGID, P_SUGID, 1585 0 1586 }; 1587 1588 static const u_int sysctl_sflagmap[] = { 1589 PS_NOCLDSTOP, P_NOCLDSTOP, 1590 PS_WEXIT, P_WEXIT, 1591 PS_STOPFORK, P_STOPFORK, 1592 PS_STOPEXEC, P_STOPEXEC, 1593 PS_STOPEXIT, P_STOPEXIT, 1594 0 1595 }; 1596 1597 static const u_int sysctl_slflagmap[] = { 1598 PSL_TRACED, P_TRACED, 1599 PSL_CHTRACED, P_CHTRACED, 1600 PSL_SYSCALL, P_SYSCALL, 1601 0 1602 }; 1603 1604 static const u_int sysctl_lflagmap[] = { 1605 PL_CONTROLT, P_CONTROLT, 1606 PL_PPWAIT, P_PPWAIT, 1607 0 1608 }; 1609 1610 static const u_int sysctl_stflagmap[] = { 1611 PST_PROFIL, P_PROFIL, 1612 0 1613 1614 }; 1615 1616 /* used by kern_lwp also */ 1617 const u_int sysctl_lwpflagmap[] = { 1618 LW_SINTR, L_SINTR, 1619 LW_SYSTEM, L_SYSTEM, 1620 0 1621 }; 1622 1623 /* 1624 * Find the most ``active'' lwp of a process and return it for ps display 1625 * purposes 1626 */ 1627 static struct lwp * 1628 proc_active_lwp(struct proc *p) 1629 { 1630 static const int ostat[] = { 1631 0, 1632 2, /* LSIDL */ 1633 6, /* LSRUN */ 1634 5, /* LSSLEEP */ 1635 4, /* LSSTOP */ 1636 0, /* LSZOMB */ 1637 1, /* LSDEAD */ 1638 7, /* LSONPROC */ 1639 3 /* LSSUSPENDED */ 1640 }; 1641 1642 struct lwp *l, *lp = NULL; 1643 LIST_FOREACH(l, &p->p_lwps, l_sibling) { 1644 KASSERT(l->l_stat >= 0 && l->l_stat < __arraycount(ostat)); 1645 if (lp == NULL || 1646 ostat[l->l_stat] > ostat[lp->l_stat] || 1647 (ostat[l->l_stat] == ostat[lp->l_stat] && 1648 l->l_cpticks > lp->l_cpticks)) { 1649 lp = l; 1650 continue; 1651 } 1652 } 1653 return lp; 1654 } 1655 1656 static int 1657 sysctl_doeproc(SYSCTLFN_ARGS) 1658 { 1659 union { 1660 struct kinfo_proc kproc; 1661 struct kinfo_proc2 kproc2; 1662 } *kbuf; 1663 struct proc *p, *next, *marker; 1664 char *where, *dp; 1665 int type, op, arg, error; 1666 u_int elem_size, kelem_size, elem_count; 1667 size_t buflen, needed; 1668 bool match, zombie, mmmbrains; 1669 const bool allowaddr = get_expose_address(curproc); 1670 1671 if (namelen == 1 && name[0] == CTL_QUERY) 1672 return (sysctl_query(SYSCTLFN_CALL(rnode))); 1673 1674 dp = where = oldp; 1675 buflen = where != NULL ? *oldlenp : 0; 1676 error = 0; 1677 needed = 0; 1678 type = rnode->sysctl_num; 1679 1680 if (type == KERN_PROC) { 1681 if (namelen == 0) 1682 return EINVAL; 1683 switch (op = name[0]) { 1684 case KERN_PROC_ALL: 1685 if (namelen != 1) 1686 return EINVAL; 1687 arg = 0; 1688 break; 1689 default: 1690 if (namelen != 2) 1691 return EINVAL; 1692 arg = name[1]; 1693 break; 1694 } 1695 elem_count = 0; /* Hush little compiler, don't you cry */ 1696 kelem_size = elem_size = sizeof(kbuf->kproc); 1697 } else { 1698 if (namelen != 4) 1699 return EINVAL; 1700 op = name[0]; 1701 arg = name[1]; 1702 elem_size = name[2]; 1703 elem_count = name[3]; 1704 kelem_size = sizeof(kbuf->kproc2); 1705 } 1706 1707 sysctl_unlock(); 1708 1709 kbuf = kmem_zalloc(sizeof(*kbuf), KM_SLEEP); 1710 marker = kmem_alloc(sizeof(*marker), KM_SLEEP); 1711 marker->p_flag = PK_MARKER; 1712 1713 mutex_enter(proc_lock); 1714 /* 1715 * Start with zombies to prevent reporting processes twice, in case they 1716 * are dying and being moved from the list of alive processes to zombies. 1717 */ 1718 mmmbrains = true; 1719 for (p = LIST_FIRST(&zombproc);; p = next) { 1720 if (p == NULL) { 1721 if (mmmbrains) { 1722 p = LIST_FIRST(&allproc); 1723 mmmbrains = false; 1724 } 1725 if (p == NULL) 1726 break; 1727 } 1728 next = LIST_NEXT(p, p_list); 1729 if ((p->p_flag & PK_MARKER) != 0) 1730 continue; 1731 1732 /* 1733 * Skip embryonic processes. 1734 */ 1735 if (p->p_stat == SIDL) 1736 continue; 1737 1738 mutex_enter(p->p_lock); 1739 error = kauth_authorize_process(l->l_cred, 1740 KAUTH_PROCESS_CANSEE, p, 1741 KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_EPROC), NULL, NULL); 1742 if (error != 0) { 1743 mutex_exit(p->p_lock); 1744 continue; 1745 } 1746 1747 /* 1748 * Hande all the operations in one switch on the cost of 1749 * algorithm complexity is on purpose. The win splitting this 1750 * function into several similar copies makes maintenance burden 1751 * burden, code grow and boost is neglible in practical systems. 1752 */ 1753 switch (op) { 1754 case KERN_PROC_PID: 1755 match = (p->p_pid == (pid_t)arg); 1756 break; 1757 1758 case KERN_PROC_PGRP: 1759 match = (p->p_pgrp->pg_id == (pid_t)arg); 1760 break; 1761 1762 case KERN_PROC_SESSION: 1763 match = (p->p_session->s_sid == (pid_t)arg); 1764 break; 1765 1766 case KERN_PROC_TTY: 1767 match = true; 1768 if (arg == (int) KERN_PROC_TTY_REVOKE) { 1769 if ((p->p_lflag & PL_CONTROLT) == 0 || 1770 p->p_session->s_ttyp == NULL || 1771 p->p_session->s_ttyvp != NULL) { 1772 match = false; 1773 } 1774 } else if ((p->p_lflag & PL_CONTROLT) == 0 || 1775 p->p_session->s_ttyp == NULL) { 1776 if ((dev_t)arg != KERN_PROC_TTY_NODEV) { 1777 match = false; 1778 } 1779 } else if (p->p_session->s_ttyp->t_dev != (dev_t)arg) { 1780 match = false; 1781 } 1782 break; 1783 1784 case KERN_PROC_UID: 1785 match = (kauth_cred_geteuid(p->p_cred) == (uid_t)arg); 1786 break; 1787 1788 case KERN_PROC_RUID: 1789 match = (kauth_cred_getuid(p->p_cred) == (uid_t)arg); 1790 break; 1791 1792 case KERN_PROC_GID: 1793 match = (kauth_cred_getegid(p->p_cred) == (uid_t)arg); 1794 break; 1795 1796 case KERN_PROC_RGID: 1797 match = (kauth_cred_getgid(p->p_cred) == (uid_t)arg); 1798 break; 1799 1800 case KERN_PROC_ALL: 1801 match = true; 1802 /* allow everything */ 1803 break; 1804 1805 default: 1806 error = EINVAL; 1807 mutex_exit(p->p_lock); 1808 goto cleanup; 1809 } 1810 if (!match) { 1811 mutex_exit(p->p_lock); 1812 continue; 1813 } 1814 1815 /* 1816 * Grab a hold on the process. 1817 */ 1818 if (mmmbrains) { 1819 zombie = true; 1820 } else { 1821 zombie = !rw_tryenter(&p->p_reflock, RW_READER); 1822 } 1823 if (zombie) { 1824 LIST_INSERT_AFTER(p, marker, p_list); 1825 } 1826 1827 if (buflen >= elem_size && 1828 (type == KERN_PROC || elem_count > 0)) { 1829 ruspace(p); /* Update process vm resource use */ 1830 1831 if (type == KERN_PROC) { 1832 fill_proc(p, &kbuf->kproc.kp_proc, allowaddr); 1833 fill_eproc(p, &kbuf->kproc.kp_eproc, zombie, 1834 allowaddr); 1835 } else { 1836 fill_kproc2(p, &kbuf->kproc2, zombie, 1837 allowaddr); 1838 elem_count--; 1839 } 1840 mutex_exit(p->p_lock); 1841 mutex_exit(proc_lock); 1842 /* 1843 * Copy out elem_size, but not larger than kelem_size 1844 */ 1845 error = sysctl_copyout(l, kbuf, dp, 1846 uimin(kelem_size, elem_size)); 1847 mutex_enter(proc_lock); 1848 if (error) { 1849 goto bah; 1850 } 1851 dp += elem_size; 1852 buflen -= elem_size; 1853 } else { 1854 mutex_exit(p->p_lock); 1855 } 1856 needed += elem_size; 1857 1858 /* 1859 * Release reference to process. 1860 */ 1861 if (zombie) { 1862 next = LIST_NEXT(marker, p_list); 1863 LIST_REMOVE(marker, p_list); 1864 } else { 1865 rw_exit(&p->p_reflock); 1866 next = LIST_NEXT(p, p_list); 1867 } 1868 1869 /* 1870 * Short-circuit break quickly! 1871 */ 1872 if (op == KERN_PROC_PID) 1873 break; 1874 } 1875 mutex_exit(proc_lock); 1876 1877 if (where != NULL) { 1878 *oldlenp = dp - where; 1879 if (needed > *oldlenp) { 1880 error = ENOMEM; 1881 goto out; 1882 } 1883 } else { 1884 needed += KERN_PROCSLOP; 1885 *oldlenp = needed; 1886 } 1887 kmem_free(kbuf, sizeof(*kbuf)); 1888 kmem_free(marker, sizeof(*marker)); 1889 sysctl_relock(); 1890 return 0; 1891 bah: 1892 if (zombie) 1893 LIST_REMOVE(marker, p_list); 1894 else 1895 rw_exit(&p->p_reflock); 1896 cleanup: 1897 mutex_exit(proc_lock); 1898 out: 1899 kmem_free(kbuf, sizeof(*kbuf)); 1900 kmem_free(marker, sizeof(*marker)); 1901 sysctl_relock(); 1902 return error; 1903 } 1904 1905 int 1906 copyin_psstrings(struct proc *p, struct ps_strings *arginfo) 1907 { 1908 #if !defined(_RUMPKERNEL) 1909 int retval; 1910 1911 if (p->p_flag & PK_32) { 1912 MODULE_HOOK_CALL(kern_proc32_copyin_hook, (p, arginfo), 1913 enosys(), retval); 1914 return retval; 1915 } 1916 #endif /* !defined(_RUMPKERNEL) */ 1917 1918 return copyin_proc(p, (void *)p->p_psstrp, arginfo, sizeof(*arginfo)); 1919 } 1920 1921 static int 1922 copy_procargs_sysctl_cb(void *cookie_, const void *src, size_t off, size_t len) 1923 { 1924 void **cookie = cookie_; 1925 struct lwp *l = cookie[0]; 1926 char *dst = cookie[1]; 1927 1928 return sysctl_copyout(l, src, dst + off, len); 1929 } 1930 1931 /* 1932 * sysctl helper routine for kern.proc_args pseudo-subtree. 1933 */ 1934 static int 1935 sysctl_kern_proc_args(SYSCTLFN_ARGS) 1936 { 1937 struct ps_strings pss; 1938 struct proc *p; 1939 pid_t pid; 1940 int type, error; 1941 void *cookie[2]; 1942 1943 if (namelen == 1 && name[0] == CTL_QUERY) 1944 return (sysctl_query(SYSCTLFN_CALL(rnode))); 1945 1946 if (newp != NULL || namelen != 2) 1947 return (EINVAL); 1948 pid = name[0]; 1949 type = name[1]; 1950 1951 switch (type) { 1952 case KERN_PROC_PATHNAME: 1953 sysctl_unlock(); 1954 error = fill_pathname(l, pid, oldp, oldlenp); 1955 sysctl_relock(); 1956 return error; 1957 1958 case KERN_PROC_CWD: 1959 sysctl_unlock(); 1960 error = fill_cwd(l, pid, oldp, oldlenp); 1961 sysctl_relock(); 1962 return error; 1963 1964 case KERN_PROC_ARGV: 1965 case KERN_PROC_NARGV: 1966 case KERN_PROC_ENV: 1967 case KERN_PROC_NENV: 1968 /* ok */ 1969 break; 1970 default: 1971 return (EINVAL); 1972 } 1973 1974 sysctl_unlock(); 1975 1976 /* check pid */ 1977 mutex_enter(proc_lock); 1978 if ((p = proc_find(pid)) == NULL) { 1979 error = EINVAL; 1980 goto out_locked; 1981 } 1982 mutex_enter(p->p_lock); 1983 1984 /* Check permission. */ 1985 if (type == KERN_PROC_ARGV || type == KERN_PROC_NARGV) 1986 error = kauth_authorize_process(l->l_cred, KAUTH_PROCESS_CANSEE, 1987 p, KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_ARGS), NULL, NULL); 1988 else if (type == KERN_PROC_ENV || type == KERN_PROC_NENV) 1989 error = kauth_authorize_process(l->l_cred, KAUTH_PROCESS_CANSEE, 1990 p, KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_ENV), NULL, NULL); 1991 else 1992 error = EINVAL; /* XXXGCC */ 1993 if (error) { 1994 mutex_exit(p->p_lock); 1995 goto out_locked; 1996 } 1997 1998 if (oldp == NULL) { 1999 if (type == KERN_PROC_NARGV || type == KERN_PROC_NENV) 2000 *oldlenp = sizeof (int); 2001 else 2002 *oldlenp = ARG_MAX; /* XXX XXX XXX */ 2003 error = 0; 2004 mutex_exit(p->p_lock); 2005 goto out_locked; 2006 } 2007 2008 /* 2009 * Zombies don't have a stack, so we can't read their psstrings. 2010 * System processes also don't have a user stack. 2011 */ 2012 if (P_ZOMBIE(p) || (p->p_flag & PK_SYSTEM) != 0) { 2013 error = EINVAL; 2014 mutex_exit(p->p_lock); 2015 goto out_locked; 2016 } 2017 2018 error = rw_tryenter(&p->p_reflock, RW_READER) ? 0 : EBUSY; 2019 mutex_exit(p->p_lock); 2020 if (error) { 2021 goto out_locked; 2022 } 2023 mutex_exit(proc_lock); 2024 2025 if (type == KERN_PROC_NARGV || type == KERN_PROC_NENV) { 2026 int value; 2027 if ((error = copyin_psstrings(p, &pss)) == 0) { 2028 if (type == KERN_PROC_NARGV) 2029 value = pss.ps_nargvstr; 2030 else 2031 value = pss.ps_nenvstr; 2032 error = sysctl_copyout(l, &value, oldp, sizeof(value)); 2033 *oldlenp = sizeof(value); 2034 } 2035 } else { 2036 cookie[0] = l; 2037 cookie[1] = oldp; 2038 error = copy_procargs(p, type, oldlenp, 2039 copy_procargs_sysctl_cb, cookie); 2040 } 2041 rw_exit(&p->p_reflock); 2042 sysctl_relock(); 2043 return error; 2044 2045 out_locked: 2046 mutex_exit(proc_lock); 2047 sysctl_relock(); 2048 return error; 2049 } 2050 2051 int 2052 copy_procargs(struct proc *p, int oid, size_t *limit, 2053 int (*cb)(void *, const void *, size_t, size_t), void *cookie) 2054 { 2055 struct ps_strings pss; 2056 size_t len, i, loaded, entry_len; 2057 struct uio auio; 2058 struct iovec aiov; 2059 int error, argvlen; 2060 char *arg; 2061 char **argv; 2062 vaddr_t user_argv; 2063 struct vmspace *vmspace; 2064 2065 /* 2066 * Allocate a temporary buffer to hold the argument vector and 2067 * the arguments themselve. 2068 */ 2069 arg = kmem_alloc(PAGE_SIZE, KM_SLEEP); 2070 argv = kmem_alloc(PAGE_SIZE, KM_SLEEP); 2071 2072 /* 2073 * Lock the process down in memory. 2074 */ 2075 vmspace = p->p_vmspace; 2076 uvmspace_addref(vmspace); 2077 2078 /* 2079 * Read in the ps_strings structure. 2080 */ 2081 if ((error = copyin_psstrings(p, &pss)) != 0) 2082 goto done; 2083 2084 /* 2085 * Now read the address of the argument vector. 2086 */ 2087 switch (oid) { 2088 case KERN_PROC_ARGV: 2089 user_argv = (uintptr_t)pss.ps_argvstr; 2090 argvlen = pss.ps_nargvstr; 2091 break; 2092 case KERN_PROC_ENV: 2093 user_argv = (uintptr_t)pss.ps_envstr; 2094 argvlen = pss.ps_nenvstr; 2095 break; 2096 default: 2097 error = EINVAL; 2098 goto done; 2099 } 2100 2101 if (argvlen < 0) { 2102 error = EIO; 2103 goto done; 2104 } 2105 2106 2107 /* 2108 * Now copy each string. 2109 */ 2110 len = 0; /* bytes written to user buffer */ 2111 loaded = 0; /* bytes from argv already processed */ 2112 i = 0; /* To make compiler happy */ 2113 entry_len = PROC_PTRSZ(p); 2114 2115 for (; argvlen; --argvlen) { 2116 int finished = 0; 2117 vaddr_t base; 2118 size_t xlen; 2119 int j; 2120 2121 if (loaded == 0) { 2122 size_t rem = entry_len * argvlen; 2123 loaded = MIN(rem, PAGE_SIZE); 2124 error = copyin_vmspace(vmspace, 2125 (const void *)user_argv, argv, loaded); 2126 if (error) 2127 break; 2128 user_argv += loaded; 2129 i = 0; 2130 } 2131 2132 #if !defined(_RUMPKERNEL) 2133 if (p->p_flag & PK_32) 2134 MODULE_HOOK_CALL(kern_proc32_base_hook, 2135 (argv, i++), 0, base); 2136 else 2137 #endif /* !defined(_RUMPKERNEL) */ 2138 base = (vaddr_t)argv[i++]; 2139 loaded -= entry_len; 2140 2141 /* 2142 * The program has messed around with its arguments, 2143 * possibly deleting some, and replacing them with 2144 * NULL's. Treat this as the last argument and not 2145 * a failure. 2146 */ 2147 if (base == 0) 2148 break; 2149 2150 while (!finished) { 2151 xlen = PAGE_SIZE - (base & PAGE_MASK); 2152 2153 aiov.iov_base = arg; 2154 aiov.iov_len = PAGE_SIZE; 2155 auio.uio_iov = &aiov; 2156 auio.uio_iovcnt = 1; 2157 auio.uio_offset = base; 2158 auio.uio_resid = xlen; 2159 auio.uio_rw = UIO_READ; 2160 UIO_SETUP_SYSSPACE(&auio); 2161 error = uvm_io(&vmspace->vm_map, &auio, 0); 2162 if (error) 2163 goto done; 2164 2165 /* Look for the end of the string */ 2166 for (j = 0; j < xlen; j++) { 2167 if (arg[j] == '\0') { 2168 xlen = j + 1; 2169 finished = 1; 2170 break; 2171 } 2172 } 2173 2174 /* Check for user buffer overflow */ 2175 if (len + xlen > *limit) { 2176 finished = 1; 2177 if (len > *limit) 2178 xlen = 0; 2179 else 2180 xlen = *limit - len; 2181 } 2182 2183 /* Copyout the page */ 2184 error = (*cb)(cookie, arg, len, xlen); 2185 if (error) 2186 goto done; 2187 2188 len += xlen; 2189 base += xlen; 2190 } 2191 } 2192 *limit = len; 2193 2194 done: 2195 kmem_free(argv, PAGE_SIZE); 2196 kmem_free(arg, PAGE_SIZE); 2197 uvmspace_free(vmspace); 2198 return error; 2199 } 2200 2201 /* 2202 * Fill in a proc structure for the specified process. 2203 */ 2204 static void 2205 fill_proc(const struct proc *psrc, struct proc *p, bool allowaddr) 2206 { 2207 COND_SET_VALUE(p->p_list, psrc->p_list, allowaddr); 2208 COND_SET_VALUE(p->p_auxlock, psrc->p_auxlock, allowaddr); 2209 COND_SET_VALUE(p->p_lock, psrc->p_lock, allowaddr); 2210 COND_SET_VALUE(p->p_stmutex, psrc->p_stmutex, allowaddr); 2211 COND_SET_VALUE(p->p_reflock, psrc->p_reflock, allowaddr); 2212 COND_SET_VALUE(p->p_waitcv, psrc->p_waitcv, allowaddr); 2213 COND_SET_VALUE(p->p_lwpcv, psrc->p_lwpcv, allowaddr); 2214 COND_SET_VALUE(p->p_cred, psrc->p_cred, allowaddr); 2215 COND_SET_VALUE(p->p_fd, psrc->p_fd, allowaddr); 2216 COND_SET_VALUE(p->p_cwdi, psrc->p_cwdi, allowaddr); 2217 COND_SET_VALUE(p->p_stats, psrc->p_stats, allowaddr); 2218 COND_SET_VALUE(p->p_limit, psrc->p_limit, allowaddr); 2219 COND_SET_VALUE(p->p_vmspace, psrc->p_vmspace, allowaddr); 2220 COND_SET_VALUE(p->p_sigacts, psrc->p_sigacts, allowaddr); 2221 COND_SET_VALUE(p->p_aio, psrc->p_aio, allowaddr); 2222 p->p_mqueue_cnt = psrc->p_mqueue_cnt; 2223 COND_SET_VALUE(p->p_specdataref, psrc->p_specdataref, allowaddr); 2224 p->p_exitsig = psrc->p_exitsig; 2225 p->p_flag = psrc->p_flag; 2226 p->p_sflag = psrc->p_sflag; 2227 p->p_slflag = psrc->p_slflag; 2228 p->p_lflag = psrc->p_lflag; 2229 p->p_stflag = psrc->p_stflag; 2230 p->p_stat = psrc->p_stat; 2231 p->p_trace_enabled = psrc->p_trace_enabled; 2232 p->p_pid = psrc->p_pid; 2233 COND_SET_VALUE(p->p_pglist, psrc->p_pglist, allowaddr); 2234 COND_SET_VALUE(p->p_pptr, psrc->p_pptr, allowaddr); 2235 COND_SET_VALUE(p->p_sibling, psrc->p_sibling, allowaddr); 2236 COND_SET_VALUE(p->p_children, psrc->p_children, allowaddr); 2237 COND_SET_VALUE(p->p_lwps, psrc->p_lwps, allowaddr); 2238 COND_SET_VALUE(p->p_raslist, psrc->p_raslist, allowaddr); 2239 p->p_nlwps = psrc->p_nlwps; 2240 p->p_nzlwps = psrc->p_nzlwps; 2241 p->p_nrlwps = psrc->p_nrlwps; 2242 p->p_nlwpwait = psrc->p_nlwpwait; 2243 p->p_ndlwps = psrc->p_ndlwps; 2244 p->p_nlwpid = psrc->p_nlwpid; 2245 p->p_nstopchild = psrc->p_nstopchild; 2246 p->p_waited = psrc->p_waited; 2247 COND_SET_VALUE(p->p_zomblwp, psrc->p_zomblwp, allowaddr); 2248 COND_SET_VALUE(p->p_vforklwp, psrc->p_vforklwp, allowaddr); 2249 COND_SET_VALUE(p->p_sched_info, psrc->p_sched_info, allowaddr); 2250 p->p_estcpu = psrc->p_estcpu; 2251 p->p_estcpu_inherited = psrc->p_estcpu_inherited; 2252 p->p_forktime = psrc->p_forktime; 2253 p->p_pctcpu = psrc->p_pctcpu; 2254 COND_SET_VALUE(p->p_opptr, psrc->p_opptr, allowaddr); 2255 COND_SET_VALUE(p->p_timers, psrc->p_timers, allowaddr); 2256 p->p_rtime = psrc->p_rtime; 2257 p->p_uticks = psrc->p_uticks; 2258 p->p_sticks = psrc->p_sticks; 2259 p->p_iticks = psrc->p_iticks; 2260 p->p_xutime = psrc->p_xutime; 2261 p->p_xstime = psrc->p_xstime; 2262 p->p_traceflag = psrc->p_traceflag; 2263 COND_SET_VALUE(p->p_tracep, psrc->p_tracep, allowaddr); 2264 COND_SET_VALUE(p->p_textvp, psrc->p_textvp, allowaddr); 2265 COND_SET_VALUE(p->p_emul, psrc->p_emul, allowaddr); 2266 COND_SET_VALUE(p->p_emuldata, psrc->p_emuldata, allowaddr); 2267 COND_SET_VALUE(p->p_execsw, psrc->p_execsw, allowaddr); 2268 COND_SET_VALUE(p->p_klist, psrc->p_klist, allowaddr); 2269 COND_SET_VALUE(p->p_sigwaiters, psrc->p_sigwaiters, allowaddr); 2270 COND_SET_VALUE(p->p_sigpend, psrc->p_sigpend, allowaddr); 2271 COND_SET_VALUE(p->p_lwpctl, psrc->p_lwpctl, allowaddr); 2272 p->p_ppid = psrc->p_ppid; 2273 p->p_oppid = psrc->p_oppid; 2274 COND_SET_VALUE(p->p_path, psrc->p_path, allowaddr); 2275 COND_SET_VALUE(p->p_sigctx, psrc->p_sigctx, allowaddr); 2276 p->p_nice = psrc->p_nice; 2277 memcpy(p->p_comm, psrc->p_comm, sizeof(p->p_comm)); 2278 COND_SET_VALUE(p->p_pgrp, psrc->p_pgrp, allowaddr); 2279 COND_SET_VALUE(p->p_psstrp, psrc->p_psstrp, allowaddr); 2280 p->p_pax = psrc->p_pax; 2281 p->p_xexit = psrc->p_xexit; 2282 p->p_xsig = psrc->p_xsig; 2283 p->p_acflag = psrc->p_acflag; 2284 COND_SET_VALUE(p->p_md, psrc->p_md, allowaddr); 2285 p->p_stackbase = psrc->p_stackbase; 2286 COND_SET_VALUE(p->p_dtrace, psrc->p_dtrace, allowaddr); 2287 } 2288 2289 /* 2290 * Fill in an eproc structure for the specified process. 2291 */ 2292 void 2293 fill_eproc(struct proc *p, struct eproc *ep, bool zombie, bool allowaddr) 2294 { 2295 struct tty *tp; 2296 struct lwp *l; 2297 2298 KASSERT(mutex_owned(proc_lock)); 2299 KASSERT(mutex_owned(p->p_lock)); 2300 2301 COND_SET_VALUE(ep->e_paddr, p, allowaddr); 2302 COND_SET_VALUE(ep->e_sess, p->p_session, allowaddr); 2303 if (p->p_cred) { 2304 kauth_cred_topcred(p->p_cred, &ep->e_pcred); 2305 kauth_cred_toucred(p->p_cred, &ep->e_ucred); 2306 } 2307 if (p->p_stat != SIDL && !P_ZOMBIE(p) && !zombie) { 2308 struct vmspace *vm = p->p_vmspace; 2309 2310 ep->e_vm.vm_rssize = vm_resident_count(vm); 2311 ep->e_vm.vm_tsize = vm->vm_tsize; 2312 ep->e_vm.vm_dsize = vm->vm_dsize; 2313 ep->e_vm.vm_ssize = vm->vm_ssize; 2314 ep->e_vm.vm_map.size = vm->vm_map.size; 2315 2316 /* Pick the primary (first) LWP */ 2317 l = proc_active_lwp(p); 2318 KASSERT(l != NULL); 2319 lwp_lock(l); 2320 if (l->l_wchan) 2321 strncpy(ep->e_wmesg, l->l_wmesg, WMESGLEN); 2322 lwp_unlock(l); 2323 } 2324 ep->e_ppid = p->p_ppid; 2325 if (p->p_pgrp && p->p_session) { 2326 ep->e_pgid = p->p_pgrp->pg_id; 2327 ep->e_jobc = p->p_pgrp->pg_jobc; 2328 ep->e_sid = p->p_session->s_sid; 2329 if ((p->p_lflag & PL_CONTROLT) && 2330 (tp = p->p_session->s_ttyp)) { 2331 ep->e_tdev = tp->t_dev; 2332 ep->e_tpgid = tp->t_pgrp ? tp->t_pgrp->pg_id : NO_PGID; 2333 COND_SET_VALUE(ep->e_tsess, tp->t_session, allowaddr); 2334 } else 2335 ep->e_tdev = (uint32_t)NODEV; 2336 ep->e_flag = p->p_session->s_ttyvp ? EPROC_CTTY : 0; 2337 if (SESS_LEADER(p)) 2338 ep->e_flag |= EPROC_SLEADER; 2339 strncpy(ep->e_login, p->p_session->s_login, MAXLOGNAME); 2340 } 2341 ep->e_xsize = ep->e_xrssize = 0; 2342 ep->e_xccount = ep->e_xswrss = 0; 2343 } 2344 2345 /* 2346 * Fill in a kinfo_proc2 structure for the specified process. 2347 */ 2348 void 2349 fill_kproc2(struct proc *p, struct kinfo_proc2 *ki, bool zombie, bool allowaddr) 2350 { 2351 struct tty *tp; 2352 struct lwp *l, *l2; 2353 struct timeval ut, st, rt; 2354 sigset_t ss1, ss2; 2355 struct rusage ru; 2356 struct vmspace *vm; 2357 2358 KASSERT(mutex_owned(proc_lock)); 2359 KASSERT(mutex_owned(p->p_lock)); 2360 2361 sigemptyset(&ss1); 2362 sigemptyset(&ss2); 2363 2364 COND_SET_VALUE(ki->p_paddr, PTRTOUINT64(p), allowaddr); 2365 COND_SET_VALUE(ki->p_fd, PTRTOUINT64(p->p_fd), allowaddr); 2366 COND_SET_VALUE(ki->p_cwdi, PTRTOUINT64(p->p_cwdi), allowaddr); 2367 COND_SET_VALUE(ki->p_stats, PTRTOUINT64(p->p_stats), allowaddr); 2368 COND_SET_VALUE(ki->p_limit, PTRTOUINT64(p->p_limit), allowaddr); 2369 COND_SET_VALUE(ki->p_vmspace, PTRTOUINT64(p->p_vmspace), allowaddr); 2370 COND_SET_VALUE(ki->p_sigacts, PTRTOUINT64(p->p_sigacts), allowaddr); 2371 COND_SET_VALUE(ki->p_sess, PTRTOUINT64(p->p_session), allowaddr); 2372 ki->p_tsess = 0; /* may be changed if controlling tty below */ 2373 COND_SET_VALUE(ki->p_ru, PTRTOUINT64(&p->p_stats->p_ru), allowaddr); 2374 ki->p_eflag = 0; 2375 ki->p_exitsig = p->p_exitsig; 2376 ki->p_flag = L_INMEM; /* Process never swapped out */ 2377 ki->p_flag |= sysctl_map_flags(sysctl_flagmap, p->p_flag); 2378 ki->p_flag |= sysctl_map_flags(sysctl_sflagmap, p->p_sflag); 2379 ki->p_flag |= sysctl_map_flags(sysctl_slflagmap, p->p_slflag); 2380 ki->p_flag |= sysctl_map_flags(sysctl_lflagmap, p->p_lflag); 2381 ki->p_flag |= sysctl_map_flags(sysctl_stflagmap, p->p_stflag); 2382 ki->p_pid = p->p_pid; 2383 ki->p_ppid = p->p_ppid; 2384 ki->p_uid = kauth_cred_geteuid(p->p_cred); 2385 ki->p_ruid = kauth_cred_getuid(p->p_cred); 2386 ki->p_gid = kauth_cred_getegid(p->p_cred); 2387 ki->p_rgid = kauth_cred_getgid(p->p_cred); 2388 ki->p_svuid = kauth_cred_getsvuid(p->p_cred); 2389 ki->p_svgid = kauth_cred_getsvgid(p->p_cred); 2390 ki->p_ngroups = kauth_cred_ngroups(p->p_cred); 2391 kauth_cred_getgroups(p->p_cred, ki->p_groups, 2392 uimin(ki->p_ngroups, sizeof(ki->p_groups) / sizeof(ki->p_groups[0])), 2393 UIO_SYSSPACE); 2394 2395 ki->p_uticks = p->p_uticks; 2396 ki->p_sticks = p->p_sticks; 2397 ki->p_iticks = p->p_iticks; 2398 ki->p_tpgid = NO_PGID; /* may be changed if controlling tty below */ 2399 COND_SET_VALUE(ki->p_tracep, PTRTOUINT64(p->p_tracep), allowaddr); 2400 ki->p_traceflag = p->p_traceflag; 2401 2402 memcpy(&ki->p_sigignore, &p->p_sigctx.ps_sigignore,sizeof(ki_sigset_t)); 2403 memcpy(&ki->p_sigcatch, &p->p_sigctx.ps_sigcatch, sizeof(ki_sigset_t)); 2404 2405 ki->p_cpticks = 0; 2406 ki->p_pctcpu = p->p_pctcpu; 2407 ki->p_estcpu = 0; 2408 ki->p_stat = p->p_stat; /* Will likely be overridden by LWP status */ 2409 ki->p_realstat = p->p_stat; 2410 ki->p_nice = p->p_nice; 2411 ki->p_xstat = P_WAITSTATUS(p); 2412 ki->p_acflag = p->p_acflag; 2413 2414 strncpy(ki->p_comm, p->p_comm, 2415 uimin(sizeof(ki->p_comm), sizeof(p->p_comm))); 2416 strncpy(ki->p_ename, p->p_emul->e_name, sizeof(ki->p_ename)); 2417 2418 ki->p_nlwps = p->p_nlwps; 2419 ki->p_realflag = ki->p_flag; 2420 2421 if (p->p_stat != SIDL && !P_ZOMBIE(p) && !zombie) { 2422 vm = p->p_vmspace; 2423 ki->p_vm_rssize = vm_resident_count(vm); 2424 ki->p_vm_tsize = vm->vm_tsize; 2425 ki->p_vm_dsize = vm->vm_dsize; 2426 ki->p_vm_ssize = vm->vm_ssize; 2427 ki->p_vm_vsize = atop(vm->vm_map.size); 2428 /* 2429 * Since the stack is initially mapped mostly with 2430 * PROT_NONE and grown as needed, adjust the "mapped size" 2431 * to skip the unused stack portion. 2432 */ 2433 ki->p_vm_msize = 2434 atop(vm->vm_map.size) - vm->vm_issize + vm->vm_ssize; 2435 2436 /* Pick the primary (first) LWP */ 2437 l = proc_active_lwp(p); 2438 KASSERT(l != NULL); 2439 lwp_lock(l); 2440 ki->p_nrlwps = p->p_nrlwps; 2441 ki->p_forw = 0; 2442 ki->p_back = 0; 2443 COND_SET_VALUE(ki->p_addr, PTRTOUINT64(l->l_addr), allowaddr); 2444 ki->p_stat = l->l_stat; 2445 ki->p_flag |= sysctl_map_flags(sysctl_lwpflagmap, l->l_flag); 2446 ki->p_swtime = l->l_swtime; 2447 ki->p_slptime = l->l_slptime; 2448 if (l->l_stat == LSONPROC) 2449 ki->p_schedflags = l->l_cpu->ci_schedstate.spc_flags; 2450 else 2451 ki->p_schedflags = 0; 2452 ki->p_priority = lwp_eprio(l); 2453 ki->p_usrpri = l->l_priority; 2454 if (l->l_wchan) 2455 strncpy(ki->p_wmesg, l->l_wmesg, sizeof(ki->p_wmesg)); 2456 COND_SET_VALUE(ki->p_wchan, PTRTOUINT64(l->l_wchan), allowaddr); 2457 ki->p_cpuid = cpu_index(l->l_cpu); 2458 lwp_unlock(l); 2459 LIST_FOREACH(l, &p->p_lwps, l_sibling) { 2460 /* This is hardly correct, but... */ 2461 sigplusset(&l->l_sigpend.sp_set, &ss1); 2462 sigplusset(&l->l_sigmask, &ss2); 2463 ki->p_cpticks += l->l_cpticks; 2464 ki->p_pctcpu += l->l_pctcpu; 2465 ki->p_estcpu += l->l_estcpu; 2466 } 2467 } 2468 sigplusset(&p->p_sigpend.sp_set, &ss1); 2469 memcpy(&ki->p_siglist, &ss1, sizeof(ki_sigset_t)); 2470 memcpy(&ki->p_sigmask, &ss2, sizeof(ki_sigset_t)); 2471 2472 if (p->p_session != NULL) { 2473 ki->p_sid = p->p_session->s_sid; 2474 ki->p__pgid = p->p_pgrp->pg_id; 2475 if (p->p_session->s_ttyvp) 2476 ki->p_eflag |= EPROC_CTTY; 2477 if (SESS_LEADER(p)) 2478 ki->p_eflag |= EPROC_SLEADER; 2479 strncpy(ki->p_login, p->p_session->s_login, 2480 uimin(sizeof ki->p_login - 1, sizeof p->p_session->s_login)); 2481 ki->p_jobc = p->p_pgrp->pg_jobc; 2482 if ((p->p_lflag & PL_CONTROLT) && (tp = p->p_session->s_ttyp)) { 2483 ki->p_tdev = tp->t_dev; 2484 ki->p_tpgid = tp->t_pgrp ? tp->t_pgrp->pg_id : NO_PGID; 2485 COND_SET_VALUE(ki->p_tsess, PTRTOUINT64(tp->t_session), 2486 allowaddr); 2487 } else { 2488 ki->p_tdev = (int32_t)NODEV; 2489 } 2490 } 2491 2492 if (!P_ZOMBIE(p) && !zombie) { 2493 ki->p_uvalid = 1; 2494 ki->p_ustart_sec = p->p_stats->p_start.tv_sec; 2495 ki->p_ustart_usec = p->p_stats->p_start.tv_usec; 2496 2497 calcru(p, &ut, &st, NULL, &rt); 2498 ki->p_rtime_sec = rt.tv_sec; 2499 ki->p_rtime_usec = rt.tv_usec; 2500 ki->p_uutime_sec = ut.tv_sec; 2501 ki->p_uutime_usec = ut.tv_usec; 2502 ki->p_ustime_sec = st.tv_sec; 2503 ki->p_ustime_usec = st.tv_usec; 2504 2505 memcpy(&ru, &p->p_stats->p_ru, sizeof(ru)); 2506 ki->p_uru_nvcsw = 0; 2507 ki->p_uru_nivcsw = 0; 2508 LIST_FOREACH(l2, &p->p_lwps, l_sibling) { 2509 ki->p_uru_nvcsw += (l2->l_ncsw - l2->l_nivcsw); 2510 ki->p_uru_nivcsw += l2->l_nivcsw; 2511 ruadd(&ru, &l2->l_ru); 2512 } 2513 ki->p_uru_maxrss = ru.ru_maxrss; 2514 ki->p_uru_ixrss = ru.ru_ixrss; 2515 ki->p_uru_idrss = ru.ru_idrss; 2516 ki->p_uru_isrss = ru.ru_isrss; 2517 ki->p_uru_minflt = ru.ru_minflt; 2518 ki->p_uru_majflt = ru.ru_majflt; 2519 ki->p_uru_nswap = ru.ru_nswap; 2520 ki->p_uru_inblock = ru.ru_inblock; 2521 ki->p_uru_oublock = ru.ru_oublock; 2522 ki->p_uru_msgsnd = ru.ru_msgsnd; 2523 ki->p_uru_msgrcv = ru.ru_msgrcv; 2524 ki->p_uru_nsignals = ru.ru_nsignals; 2525 2526 timeradd(&p->p_stats->p_cru.ru_utime, 2527 &p->p_stats->p_cru.ru_stime, &ut); 2528 ki->p_uctime_sec = ut.tv_sec; 2529 ki->p_uctime_usec = ut.tv_usec; 2530 } 2531 } 2532 2533 2534 int 2535 proc_find_locked(struct lwp *l, struct proc **p, pid_t pid) 2536 { 2537 int error; 2538 2539 mutex_enter(proc_lock); 2540 if (pid == -1) 2541 *p = l->l_proc; 2542 else 2543 *p = proc_find(pid); 2544 2545 if (*p == NULL) { 2546 if (pid != -1) 2547 mutex_exit(proc_lock); 2548 return ESRCH; 2549 } 2550 if (pid != -1) 2551 mutex_enter((*p)->p_lock); 2552 mutex_exit(proc_lock); 2553 2554 error = kauth_authorize_process(l->l_cred, 2555 KAUTH_PROCESS_CANSEE, *p, 2556 KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_ENTRY), NULL, NULL); 2557 if (error) { 2558 if (pid != -1) 2559 mutex_exit((*p)->p_lock); 2560 } 2561 return error; 2562 } 2563 2564 static int 2565 fill_pathname(struct lwp *l, pid_t pid, void *oldp, size_t *oldlenp) 2566 { 2567 int error; 2568 struct proc *p; 2569 2570 if ((error = proc_find_locked(l, &p, pid)) != 0) 2571 return error; 2572 2573 if (p->p_path == NULL) { 2574 if (pid != -1) 2575 mutex_exit(p->p_lock); 2576 return ENOENT; 2577 } 2578 2579 size_t len = strlen(p->p_path) + 1; 2580 if (oldp != NULL) { 2581 size_t copylen = uimin(len, *oldlenp); 2582 error = sysctl_copyout(l, p->p_path, oldp, copylen); 2583 if (error == 0 && *oldlenp < len) 2584 error = ENOSPC; 2585 } 2586 *oldlenp = len; 2587 if (pid != -1) 2588 mutex_exit(p->p_lock); 2589 return error; 2590 } 2591 2592 static int 2593 fill_cwd(struct lwp *l, pid_t pid, void *oldp, size_t *oldlenp) 2594 { 2595 int error; 2596 struct proc *p; 2597 char *path; 2598 char *bp, *bend; 2599 const struct cwdinfo *cwdi; 2600 struct vnode *vp; 2601 size_t len, lenused; 2602 2603 if ((error = proc_find_locked(l, &p, pid)) != 0) 2604 return error; 2605 2606 len = MAXPATHLEN * 4; 2607 2608 path = kmem_alloc(len, KM_SLEEP); 2609 2610 bp = &path[len]; 2611 bend = bp; 2612 *(--bp) = '\0'; 2613 2614 cwdi = cwdlock(p); 2615 vp = cwdi->cwdi_cdir; 2616 vref(vp); 2617 cwdunlock(p); 2618 error = getcwd_common(vp, NULL, &bp, path, len/2, 0, l); 2619 vrele(vp); 2620 2621 if (error) 2622 goto out; 2623 2624 lenused = bend - bp; 2625 2626 if (oldp != NULL) { 2627 size_t copylen = uimin(lenused, *oldlenp); 2628 error = sysctl_copyout(l, bp, oldp, copylen); 2629 if (error == 0 && *oldlenp < lenused) 2630 error = ENOSPC; 2631 } 2632 *oldlenp = lenused; 2633 out: 2634 if (pid != -1) 2635 mutex_exit(p->p_lock); 2636 kmem_free(path, len); 2637 return error; 2638 } 2639 2640 int 2641 proc_getauxv(struct proc *p, void **buf, size_t *len) 2642 { 2643 struct ps_strings pss; 2644 int error; 2645 void *uauxv, *kauxv; 2646 size_t size; 2647 2648 if ((error = copyin_psstrings(p, &pss)) != 0) 2649 return error; 2650 if (pss.ps_envstr == NULL) 2651 return EIO; 2652 2653 size = p->p_execsw->es_arglen; 2654 if (size == 0) 2655 return EIO; 2656 2657 size_t ptrsz = PROC_PTRSZ(p); 2658 uauxv = (void *)((char *)pss.ps_envstr + (pss.ps_nenvstr + 1) * ptrsz); 2659 2660 kauxv = kmem_alloc(size, KM_SLEEP); 2661 2662 error = copyin_proc(p, uauxv, kauxv, size); 2663 if (error) { 2664 kmem_free(kauxv, size); 2665 return error; 2666 } 2667 2668 *buf = kauxv; 2669 *len = size; 2670 2671 return 0; 2672 } 2673 2674 2675 static int 2676 sysctl_security_expose_address(SYSCTLFN_ARGS) 2677 { 2678 int expose_address, error; 2679 struct sysctlnode node; 2680 2681 node = *rnode; 2682 node.sysctl_data = &expose_address; 2683 expose_address = *(int *)rnode->sysctl_data; 2684 error = sysctl_lookup(SYSCTLFN_CALL(&node)); 2685 if (error || newp == NULL) 2686 return error; 2687 2688 if (kauth_authorize_system(l->l_cred, KAUTH_SYSTEM_KERNADDR, 2689 0, NULL, NULL, NULL)) 2690 return EPERM; 2691 2692 switch (expose_address) { 2693 case 0: 2694 case 1: 2695 case 2: 2696 break; 2697 default: 2698 return EINVAL; 2699 } 2700 2701 *(int *)rnode->sysctl_data = expose_address; 2702 2703 return 0; 2704 } 2705 2706 bool 2707 get_expose_address(struct proc *p) 2708 { 2709 /* allow only if sysctl variable is set or privileged */ 2710 return kauth_authorize_process(kauth_cred_get(), KAUTH_PROCESS_CANSEE, 2711 p, KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_KPTR), NULL, NULL) == 0; 2712 } 2713
Indexes created Sat Sep 20 22:09:52 GMT 2025