kern_proc.c revision 1.209.2.14
1 /* $NetBSD: kern_proc.c,v 1.209.2.14 2019/01/21 06:49:28 pgoyette Exp $ */ 2 3 /*- 4 * Copyright (c) 1999, 2006, 2007, 2008 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 1.209.2.14 2019/01/21 06:49:28 pgoyette 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> 109 110 #include <uvm/uvm_extern.h> 111 #include <uvm/uvm.h> 112 113 /* 114 * Process lists. 115 */ 116 117 struct proclist allproc __cacheline_aligned; 118 struct proclist zombproc __cacheline_aligned; 119 120 kmutex_t * proc_lock __cacheline_aligned; 121 122 /* 123 * pid to proc lookup is done by indexing the pid_table array. 124 * Since pid numbers are only allocated when an empty slot 125 * has been found, there is no need to search any lists ever. 126 * (an orphaned pgrp will lock the slot, a session will lock 127 * the pgrp with the same number.) 128 * If the table is too small it is reallocated with twice the 129 * previous size and the entries 'unzipped' into the two halves. 130 * A linked list of free entries is passed through the pt_proc 131 * field of 'free' items - set odd to be an invalid ptr. 132 */ 133 134 struct pid_table { 135 struct proc *pt_proc; 136 struct pgrp *pt_pgrp; 137 pid_t pt_pid; 138 }; 139 #if 1 /* strongly typed cast - should be a noop */ 140 static inline uint p2u(struct proc *p) { return (uint)(uintptr_t)p; } 141 #else 142 #define p2u(p) ((uint)p) 143 #endif 144 #define P_VALID(p) (!(p2u(p) & 1)) 145 #define P_NEXT(p) (p2u(p) >> 1) 146 #define P_FREE(pid) ((struct proc *)(uintptr_t)((pid) << 1 | 1)) 147 148 /* 149 * Table of process IDs (PIDs). 150 */ 151 static struct pid_table *pid_table __read_mostly; 152 153 #define INITIAL_PID_TABLE_SIZE (1 << 5) 154 155 /* Table mask, threshold for growing and number of allocated PIDs. */ 156 static u_int pid_tbl_mask __read_mostly; 157 static u_int pid_alloc_lim __read_mostly; 158 static u_int pid_alloc_cnt __cacheline_aligned; 159 160 /* Next free, last free and maximum PIDs. */ 161 static u_int next_free_pt __cacheline_aligned; 162 static u_int last_free_pt __cacheline_aligned; 163 static pid_t pid_max __read_mostly; 164 165 /* Components of the first process -- never freed. */ 166 167 extern struct emul emul_netbsd; /* defined in kern_exec.c */ 168 169 struct session session0 = { 170 .s_count = 1, 171 .s_sid = 0, 172 }; 173 struct pgrp pgrp0 = { 174 .pg_members = LIST_HEAD_INITIALIZER(&pgrp0.pg_members), 175 .pg_session = &session0, 176 }; 177 filedesc_t filedesc0; 178 struct cwdinfo cwdi0 = { 179 .cwdi_cmask = CMASK, 180 .cwdi_refcnt = 1, 181 }; 182 struct plimit limit0; 183 struct pstats pstat0; 184 struct vmspace vmspace0; 185 struct sigacts sigacts0; 186 struct proc proc0 = { 187 .p_lwps = LIST_HEAD_INITIALIZER(&proc0.p_lwps), 188 .p_sigwaiters = LIST_HEAD_INITIALIZER(&proc0.p_sigwaiters), 189 .p_nlwps = 1, 190 .p_nrlwps = 1, 191 .p_nlwpid = 1, /* must match lwp0.l_lid */ 192 .p_pgrp = &pgrp0, 193 .p_comm = "system", 194 /* 195 * Set P_NOCLDWAIT so that kernel threads are reparented to init(8) 196 * when they exit. init(8) can easily wait them out for us. 197 */ 198 .p_flag = PK_SYSTEM | PK_NOCLDWAIT, 199 .p_stat = SACTIVE, 200 .p_nice = NZERO, 201 .p_emul = &emul_netbsd, 202 .p_cwdi = &cwdi0, 203 .p_limit = &limit0, 204 .p_fd = &filedesc0, 205 .p_vmspace = &vmspace0, 206 .p_stats = &pstat0, 207 .p_sigacts = &sigacts0, 208 #ifdef PROC0_MD_INITIALIZERS 209 PROC0_MD_INITIALIZERS 210 #endif 211 }; 212 kauth_cred_t cred0; 213 214 static const int nofile = NOFILE; 215 static const int maxuprc = MAXUPRC; 216 217 static int sysctl_doeproc(SYSCTLFN_PROTO); 218 static int sysctl_kern_proc_args(SYSCTLFN_PROTO); 219 static int sysctl_security_expose_address(SYSCTLFN_PROTO); 220 221 #ifdef KASLR 222 static int kern_expose_address = 0; 223 #else 224 static int kern_expose_address = 1; 225 #endif 226 /* 227 * The process list descriptors, used during pid allocation and 228 * by sysctl. No locking on this data structure is needed since 229 * it is completely static. 230 */ 231 const struct proclist_desc proclists[] = { 232 { &allproc }, 233 { &zombproc }, 234 { NULL }, 235 }; 236 237 static struct pgrp * pg_remove(pid_t); 238 static void pg_delete(pid_t); 239 static void orphanpg(struct pgrp *); 240 241 static specificdata_domain_t proc_specificdata_domain; 242 243 static pool_cache_t proc_cache; 244 245 static kauth_listener_t proc_listener; 246 247 static void fill_proc(const struct proc *, struct proc *, bool); 248 static int fill_pathname(struct lwp *, pid_t, void *, size_t *); 249 250 static int 251 proc_listener_cb(kauth_cred_t cred, kauth_action_t action, void *cookie, 252 void *arg0, void *arg1, void *arg2, void *arg3) 253 { 254 struct proc *p; 255 int result; 256 257 result = KAUTH_RESULT_DEFER; 258 p = arg0; 259 260 switch (action) { 261 case KAUTH_PROCESS_CANSEE: { 262 enum kauth_process_req req; 263 264 req = (enum kauth_process_req)arg1; 265 266 switch (req) { 267 case KAUTH_REQ_PROCESS_CANSEE_ARGS: 268 case KAUTH_REQ_PROCESS_CANSEE_ENTRY: 269 case KAUTH_REQ_PROCESS_CANSEE_OPENFILES: 270 case KAUTH_REQ_PROCESS_CANSEE_EPROC: 271 result = KAUTH_RESULT_ALLOW; 272 break; 273 274 case KAUTH_REQ_PROCESS_CANSEE_ENV: 275 if (kauth_cred_getuid(cred) != 276 kauth_cred_getuid(p->p_cred) || 277 kauth_cred_getuid(cred) != 278 kauth_cred_getsvuid(p->p_cred)) 279 break; 280 281 result = KAUTH_RESULT_ALLOW; 282 283 break; 284 285 case KAUTH_REQ_PROCESS_CANSEE_KPTR: 286 if (!kern_expose_address) 287 break; 288 289 if (kern_expose_address == 1 && !(p->p_flag & PK_KMEM)) 290 break; 291 292 result = KAUTH_RESULT_ALLOW; 293 294 break; 295 296 default: 297 break; 298 } 299 300 break; 301 } 302 303 case KAUTH_PROCESS_FORK: { 304 int lnprocs = (int)(unsigned long)arg2; 305 306 /* 307 * Don't allow a nonprivileged user to use the last few 308 * processes. The variable lnprocs is the current number of 309 * processes, maxproc is the limit. 310 */ 311 if (__predict_false((lnprocs >= maxproc - 5))) 312 break; 313 314 result = KAUTH_RESULT_ALLOW; 315 316 break; 317 } 318 319 case KAUTH_PROCESS_CORENAME: 320 case KAUTH_PROCESS_STOPFLAG: 321 if (proc_uidmatch(cred, p->p_cred) == 0) 322 result = KAUTH_RESULT_ALLOW; 323 324 break; 325 326 default: 327 break; 328 } 329 330 return result; 331 } 332 333 static int 334 proc_ctor(void *arg __unused, void *obj, int flags __unused) 335 { 336 memset(obj, 0, sizeof(struct proc)); 337 return 0; 338 } 339 340 /* 341 * Initialize global process hashing structures. 342 */ 343 void 344 procinit(void) 345 { 346 const struct proclist_desc *pd; 347 u_int i; 348 #define LINK_EMPTY ((PID_MAX + INITIAL_PID_TABLE_SIZE) & ~(INITIAL_PID_TABLE_SIZE - 1)) 349 350 for (pd = proclists; pd->pd_list != NULL; pd++) 351 LIST_INIT(pd->pd_list); 352 353 proc_lock = mutex_obj_alloc(MUTEX_DEFAULT, IPL_NONE); 354 pid_table = kmem_alloc(INITIAL_PID_TABLE_SIZE 355 * sizeof(struct pid_table), KM_SLEEP); 356 pid_tbl_mask = INITIAL_PID_TABLE_SIZE - 1; 357 pid_max = PID_MAX; 358 359 /* Set free list running through table... 360 Preset 'use count' above PID_MAX so we allocate pid 1 next. */ 361 for (i = 0; i <= pid_tbl_mask; i++) { 362 pid_table[i].pt_proc = P_FREE(LINK_EMPTY + i + 1); 363 pid_table[i].pt_pgrp = 0; 364 pid_table[i].pt_pid = 0; 365 } 366 /* slot 0 is just grabbed */ 367 next_free_pt = 1; 368 /* Need to fix last entry. */ 369 last_free_pt = pid_tbl_mask; 370 pid_table[last_free_pt].pt_proc = P_FREE(LINK_EMPTY); 371 /* point at which we grow table - to avoid reusing pids too often */ 372 pid_alloc_lim = pid_tbl_mask - 1; 373 #undef LINK_EMPTY 374 375 proc_specificdata_domain = specificdata_domain_create(); 376 KASSERT(proc_specificdata_domain != NULL); 377 378 proc_cache = pool_cache_init(sizeof(struct proc), 0, 0, 0, 379 "procpl", NULL, IPL_NONE, proc_ctor, NULL, NULL); 380 381 proc_listener = kauth_listen_scope(KAUTH_SCOPE_PROCESS, 382 proc_listener_cb, NULL); 383 } 384 385 void 386 procinit_sysctl(void) 387 { 388 static struct sysctllog *clog; 389 390 sysctl_createv(&clog, 0, NULL, NULL, 391 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 392 CTLTYPE_INT, "expose_address", 393 SYSCTL_DESCR("Enable exposing kernel addresses"), 394 sysctl_security_expose_address, 0, 395 &kern_expose_address, 0, CTL_KERN, CTL_CREATE, CTL_EOL); 396 sysctl_createv(&clog, 0, NULL, NULL, 397 CTLFLAG_PERMANENT, 398 CTLTYPE_NODE, "proc", 399 SYSCTL_DESCR("System-wide process information"), 400 sysctl_doeproc, 0, NULL, 0, 401 CTL_KERN, KERN_PROC, CTL_EOL); 402 sysctl_createv(&clog, 0, NULL, NULL, 403 CTLFLAG_PERMANENT, 404 CTLTYPE_NODE, "proc2", 405 SYSCTL_DESCR("Machine-independent process information"), 406 sysctl_doeproc, 0, NULL, 0, 407 CTL_KERN, KERN_PROC2, CTL_EOL); 408 sysctl_createv(&clog, 0, NULL, NULL, 409 CTLFLAG_PERMANENT, 410 CTLTYPE_NODE, "proc_args", 411 SYSCTL_DESCR("Process argument information"), 412 sysctl_kern_proc_args, 0, NULL, 0, 413 CTL_KERN, KERN_PROC_ARGS, CTL_EOL); 414 415 /* 416 "nodes" under these: 417 418 KERN_PROC_ALL 419 KERN_PROC_PID pid 420 KERN_PROC_PGRP pgrp 421 KERN_PROC_SESSION sess 422 KERN_PROC_TTY tty 423 KERN_PROC_UID uid 424 KERN_PROC_RUID uid 425 KERN_PROC_GID gid 426 KERN_PROC_RGID gid 427 428 all in all, probably not worth the effort... 429 */ 430 } 431 432 /* 433 * Initialize process 0. 434 */ 435 void 436 proc0_init(void) 437 { 438 struct proc *p; 439 struct pgrp *pg; 440 struct rlimit *rlim; 441 rlim_t lim; 442 int i; 443 444 p = &proc0; 445 pg = &pgrp0; 446 447 mutex_init(&p->p_stmutex, MUTEX_DEFAULT, IPL_HIGH); 448 mutex_init(&p->p_auxlock, MUTEX_DEFAULT, IPL_NONE); 449 p->p_lock = mutex_obj_alloc(MUTEX_DEFAULT, IPL_NONE); 450 451 rw_init(&p->p_reflock); 452 cv_init(&p->p_waitcv, "wait"); 453 cv_init(&p->p_lwpcv, "lwpwait"); 454 455 LIST_INSERT_HEAD(&p->p_lwps, &lwp0, l_sibling); 456 457 pid_table[0].pt_proc = p; 458 LIST_INSERT_HEAD(&allproc, p, p_list); 459 460 pid_table[0].pt_pgrp = pg; 461 LIST_INSERT_HEAD(&pg->pg_members, p, p_pglist); 462 463 #ifdef __HAVE_SYSCALL_INTERN 464 (*p->p_emul->e_syscall_intern)(p); 465 #endif 466 467 /* Create credentials. */ 468 cred0 = kauth_cred_alloc(); 469 p->p_cred = cred0; 470 471 /* Create the CWD info. */ 472 rw_init(&cwdi0.cwdi_lock); 473 474 /* Create the limits structures. */ 475 mutex_init(&limit0.pl_lock, MUTEX_DEFAULT, IPL_NONE); 476 477 rlim = limit0.pl_rlimit; 478 for (i = 0; i < __arraycount(limit0.pl_rlimit); i++) { 479 rlim[i].rlim_cur = RLIM_INFINITY; 480 rlim[i].rlim_max = RLIM_INFINITY; 481 } 482 483 rlim[RLIMIT_NOFILE].rlim_max = maxfiles; 484 rlim[RLIMIT_NOFILE].rlim_cur = maxfiles < nofile ? maxfiles : nofile; 485 486 rlim[RLIMIT_NPROC].rlim_max = maxproc; 487 rlim[RLIMIT_NPROC].rlim_cur = maxproc < maxuprc ? maxproc : maxuprc; 488 489 lim = MIN(VM_MAXUSER_ADDRESS, ctob((rlim_t)uvmexp.free)); 490 rlim[RLIMIT_RSS].rlim_max = lim; 491 rlim[RLIMIT_MEMLOCK].rlim_max = lim; 492 rlim[RLIMIT_MEMLOCK].rlim_cur = lim / 3; 493 494 rlim[RLIMIT_NTHR].rlim_max = maxlwp; 495 rlim[RLIMIT_NTHR].rlim_cur = maxlwp < maxuprc ? maxlwp : maxuprc; 496 497 /* Note that default core name has zero length. */ 498 limit0.pl_corename = defcorename; 499 limit0.pl_cnlen = 0; 500 limit0.pl_refcnt = 1; 501 limit0.pl_writeable = false; 502 limit0.pl_sv_limit = NULL; 503 504 /* Configure virtual memory system, set vm rlimits. */ 505 uvm_init_limits(p); 506 507 /* Initialize file descriptor table for proc0. */ 508 fd_init(&filedesc0); 509 510 /* 511 * Initialize proc0's vmspace, which uses the kernel pmap. 512 * All kernel processes (which never have user space mappings) 513 * share proc0's vmspace, and thus, the kernel pmap. 514 */ 515 uvmspace_init(&vmspace0, pmap_kernel(), round_page(VM_MIN_ADDRESS), 516 trunc_page(VM_MAXUSER_ADDRESS), 517 #ifdef __USE_TOPDOWN_VM 518 true 519 #else 520 false 521 #endif 522 ); 523 524 /* Initialize signal state for proc0. XXX IPL_SCHED */ 525 mutex_init(&p->p_sigacts->sa_mutex, MUTEX_DEFAULT, IPL_SCHED); 526 siginit(p); 527 528 proc_initspecific(p); 529 kdtrace_proc_ctor(NULL, p); 530 } 531 532 /* 533 * Session reference counting. 534 */ 535 536 void 537 proc_sesshold(struct session *ss) 538 { 539 540 KASSERT(mutex_owned(proc_lock)); 541 ss->s_count++; 542 } 543 544 void 545 proc_sessrele(struct session *ss) 546 { 547 548 KASSERT(mutex_owned(proc_lock)); 549 /* 550 * We keep the pgrp with the same id as the session in order to 551 * stop a process being given the same pid. Since the pgrp holds 552 * a reference to the session, it must be a 'zombie' pgrp by now. 553 */ 554 if (--ss->s_count == 0) { 555 struct pgrp *pg; 556 557 pg = pg_remove(ss->s_sid); 558 mutex_exit(proc_lock); 559 560 kmem_free(pg, sizeof(struct pgrp)); 561 kmem_free(ss, sizeof(struct session)); 562 } else { 563 mutex_exit(proc_lock); 564 } 565 } 566 567 /* 568 * Check that the specified process group is in the session of the 569 * specified process. 570 * Treats -ve ids as process ids. 571 * Used to validate TIOCSPGRP requests. 572 */ 573 int 574 pgid_in_session(struct proc *p, pid_t pg_id) 575 { 576 struct pgrp *pgrp; 577 struct session *session; 578 int error; 579 580 mutex_enter(proc_lock); 581 if (pg_id < 0) { 582 struct proc *p1 = proc_find(-pg_id); 583 if (p1 == NULL) { 584 error = EINVAL; 585 goto fail; 586 } 587 pgrp = p1->p_pgrp; 588 } else { 589 pgrp = pgrp_find(pg_id); 590 if (pgrp == NULL) { 591 error = EINVAL; 592 goto fail; 593 } 594 } 595 session = pgrp->pg_session; 596 error = (session != p->p_pgrp->pg_session) ? EPERM : 0; 597 fail: 598 mutex_exit(proc_lock); 599 return error; 600 } 601 602 /* 603 * p_inferior: is p an inferior of q? 604 */ 605 static inline bool 606 p_inferior(struct proc *p, struct proc *q) 607 { 608 609 KASSERT(mutex_owned(proc_lock)); 610 611 for (; p != q; p = p->p_pptr) 612 if (p->p_pid == 0) 613 return false; 614 return true; 615 } 616 617 /* 618 * proc_find: locate a process by the ID. 619 * 620 * => Must be called with proc_lock held. 621 */ 622 proc_t * 623 proc_find_raw(pid_t pid) 624 { 625 struct pid_table *pt; 626 proc_t *p; 627 628 KASSERT(mutex_owned(proc_lock)); 629 pt = &pid_table[pid & pid_tbl_mask]; 630 p = pt->pt_proc; 631 if (__predict_false(!P_VALID(p) || pt->pt_pid != pid)) { 632 return NULL; 633 } 634 return p; 635 } 636 637 proc_t * 638 proc_find(pid_t pid) 639 { 640 proc_t *p; 641 642 p = proc_find_raw(pid); 643 if (__predict_false(p == NULL)) { 644 return NULL; 645 } 646 647 /* 648 * Only allow live processes to be found by PID. 649 * XXX: p_stat might change, since unlocked. 650 */ 651 if (__predict_true(p->p_stat == SACTIVE || p->p_stat == SSTOP)) { 652 return p; 653 } 654 return NULL; 655 } 656 657 /* 658 * pgrp_find: locate a process group by the ID. 659 * 660 * => Must be called with proc_lock held. 661 */ 662 struct pgrp * 663 pgrp_find(pid_t pgid) 664 { 665 struct pgrp *pg; 666 667 KASSERT(mutex_owned(proc_lock)); 668 669 pg = pid_table[pgid & pid_tbl_mask].pt_pgrp; 670 671 /* 672 * Cannot look up a process group that only exists because the 673 * session has not died yet (traditional). 674 */ 675 if (pg == NULL || pg->pg_id != pgid || LIST_EMPTY(&pg->pg_members)) { 676 return NULL; 677 } 678 return pg; 679 } 680 681 static void 682 expand_pid_table(void) 683 { 684 size_t pt_size, tsz; 685 struct pid_table *n_pt, *new_pt; 686 struct proc *proc; 687 struct pgrp *pgrp; 688 pid_t pid, rpid; 689 u_int i; 690 uint new_pt_mask; 691 692 pt_size = pid_tbl_mask + 1; 693 tsz = pt_size * 2 * sizeof(struct pid_table); 694 new_pt = kmem_alloc(tsz, KM_SLEEP); 695 new_pt_mask = pt_size * 2 - 1; 696 697 mutex_enter(proc_lock); 698 if (pt_size != pid_tbl_mask + 1) { 699 /* Another process beat us to it... */ 700 mutex_exit(proc_lock); 701 kmem_free(new_pt, tsz); 702 return; 703 } 704 705 /* 706 * Copy entries from old table into new one. 707 * If 'pid' is 'odd' we need to place in the upper half, 708 * even pid's to the lower half. 709 * Free items stay in the low half so we don't have to 710 * fixup the reference to them. 711 * We stuff free items on the front of the freelist 712 * because we can't write to unmodified entries. 713 * Processing the table backwards maintains a semblance 714 * of issuing pid numbers that increase with time. 715 */ 716 i = pt_size - 1; 717 n_pt = new_pt + i; 718 for (; ; i--, n_pt--) { 719 proc = pid_table[i].pt_proc; 720 pgrp = pid_table[i].pt_pgrp; 721 if (!P_VALID(proc)) { 722 /* Up 'use count' so that link is valid */ 723 pid = (P_NEXT(proc) + pt_size) & ~pt_size; 724 rpid = 0; 725 proc = P_FREE(pid); 726 if (pgrp) 727 pid = pgrp->pg_id; 728 } else { 729 pid = pid_table[i].pt_pid; 730 rpid = pid; 731 } 732 733 /* Save entry in appropriate half of table */ 734 n_pt[pid & pt_size].pt_proc = proc; 735 n_pt[pid & pt_size].pt_pgrp = pgrp; 736 n_pt[pid & pt_size].pt_pid = rpid; 737 738 /* Put other piece on start of free list */ 739 pid = (pid ^ pt_size) & ~pid_tbl_mask; 740 n_pt[pid & pt_size].pt_proc = 741 P_FREE((pid & ~pt_size) | next_free_pt); 742 n_pt[pid & pt_size].pt_pgrp = 0; 743 n_pt[pid & pt_size].pt_pid = 0; 744 745 next_free_pt = i | (pid & pt_size); 746 if (i == 0) 747 break; 748 } 749 750 /* Save old table size and switch tables */ 751 tsz = pt_size * sizeof(struct pid_table); 752 n_pt = pid_table; 753 pid_table = new_pt; 754 pid_tbl_mask = new_pt_mask; 755 756 /* 757 * pid_max starts as PID_MAX (= 30000), once we have 16384 758 * allocated pids we need it to be larger! 759 */ 760 if (pid_tbl_mask > PID_MAX) { 761 pid_max = pid_tbl_mask * 2 + 1; 762 pid_alloc_lim |= pid_alloc_lim << 1; 763 } else 764 pid_alloc_lim <<= 1; /* doubles number of free slots... */ 765 766 mutex_exit(proc_lock); 767 kmem_free(n_pt, tsz); 768 } 769 770 struct proc * 771 proc_alloc(void) 772 { 773 struct proc *p; 774 775 p = pool_cache_get(proc_cache, PR_WAITOK); 776 p->p_stat = SIDL; /* protect against others */ 777 proc_initspecific(p); 778 kdtrace_proc_ctor(NULL, p); 779 p->p_pid = -1; 780 proc_alloc_pid(p); 781 return p; 782 } 783 784 /* 785 * proc_alloc_pid: allocate PID and record the given proc 'p' so that 786 * proc_find_raw() can find it by the PID. 787 */ 788 789 pid_t 790 proc_alloc_pid(struct proc *p) 791 { 792 struct pid_table *pt; 793 pid_t pid; 794 int nxt; 795 796 for (;;expand_pid_table()) { 797 if (__predict_false(pid_alloc_cnt >= pid_alloc_lim)) 798 /* ensure pids cycle through 2000+ values */ 799 continue; 800 mutex_enter(proc_lock); 801 pt = &pid_table[next_free_pt]; 802 #ifdef DIAGNOSTIC 803 if (__predict_false(P_VALID(pt->pt_proc) || pt->pt_pgrp)) 804 panic("proc_alloc: slot busy"); 805 #endif 806 nxt = P_NEXT(pt->pt_proc); 807 if (nxt & pid_tbl_mask) 808 break; 809 /* Table full - expand (NB last entry not used....) */ 810 mutex_exit(proc_lock); 811 } 812 813 /* pid is 'saved use count' + 'size' + entry */ 814 pid = (nxt & ~pid_tbl_mask) + pid_tbl_mask + 1 + next_free_pt; 815 if ((uint)pid > (uint)pid_max) 816 pid &= pid_tbl_mask; 817 next_free_pt = nxt & pid_tbl_mask; 818 819 /* Grab table slot */ 820 pt->pt_proc = p; 821 822 KASSERT(pt->pt_pid == 0); 823 pt->pt_pid = pid; 824 if (p->p_pid == -1) { 825 p->p_pid = pid; 826 } 827 pid_alloc_cnt++; 828 mutex_exit(proc_lock); 829 830 return pid; 831 } 832 833 /* 834 * Free a process id - called from proc_free (in kern_exit.c) 835 * 836 * Called with the proc_lock held. 837 */ 838 void 839 proc_free_pid(pid_t pid) 840 { 841 struct pid_table *pt; 842 843 KASSERT(mutex_owned(proc_lock)); 844 845 pt = &pid_table[pid & pid_tbl_mask]; 846 847 /* save pid use count in slot */ 848 pt->pt_proc = P_FREE(pid & ~pid_tbl_mask); 849 KASSERT(pt->pt_pid == pid); 850 pt->pt_pid = 0; 851 852 if (pt->pt_pgrp == NULL) { 853 /* link last freed entry onto ours */ 854 pid &= pid_tbl_mask; 855 pt = &pid_table[last_free_pt]; 856 pt->pt_proc = P_FREE(P_NEXT(pt->pt_proc) | pid); 857 pt->pt_pid = 0; 858 last_free_pt = pid; 859 pid_alloc_cnt--; 860 } 861 862 atomic_dec_uint(&nprocs); 863 } 864 865 void 866 proc_free_mem(struct proc *p) 867 { 868 869 kdtrace_proc_dtor(NULL, p); 870 pool_cache_put(proc_cache, p); 871 } 872 873 /* 874 * proc_enterpgrp: move p to a new or existing process group (and session). 875 * 876 * If we are creating a new pgrp, the pgid should equal 877 * the calling process' pid. 878 * If is only valid to enter a process group that is in the session 879 * of the process. 880 * Also mksess should only be set if we are creating a process group 881 * 882 * Only called from sys_setsid, sys_setpgid and posix_spawn/spawn_return. 883 */ 884 int 885 proc_enterpgrp(struct proc *curp, pid_t pid, pid_t pgid, bool mksess) 886 { 887 struct pgrp *new_pgrp, *pgrp; 888 struct session *sess; 889 struct proc *p; 890 int rval; 891 pid_t pg_id = NO_PGID; 892 893 sess = mksess ? kmem_alloc(sizeof(*sess), KM_SLEEP) : NULL; 894 895 /* Allocate data areas we might need before doing any validity checks */ 896 mutex_enter(proc_lock); /* Because pid_table might change */ 897 if (pid_table[pgid & pid_tbl_mask].pt_pgrp == 0) { 898 mutex_exit(proc_lock); 899 new_pgrp = kmem_alloc(sizeof(*new_pgrp), KM_SLEEP); 900 mutex_enter(proc_lock); 901 } else 902 new_pgrp = NULL; 903 rval = EPERM; /* most common error (to save typing) */ 904 905 /* Check pgrp exists or can be created */ 906 pgrp = pid_table[pgid & pid_tbl_mask].pt_pgrp; 907 if (pgrp != NULL && pgrp->pg_id != pgid) 908 goto done; 909 910 /* Can only set another process under restricted circumstances. */ 911 if (pid != curp->p_pid) { 912 /* Must exist and be one of our children... */ 913 p = proc_find(pid); 914 if (p == NULL || !p_inferior(p, curp)) { 915 rval = ESRCH; 916 goto done; 917 } 918 /* ... in the same session... */ 919 if (sess != NULL || p->p_session != curp->p_session) 920 goto done; 921 /* ... existing pgid must be in same session ... */ 922 if (pgrp != NULL && pgrp->pg_session != p->p_session) 923 goto done; 924 /* ... and not done an exec. */ 925 if (p->p_flag & PK_EXEC) { 926 rval = EACCES; 927 goto done; 928 } 929 } else { 930 /* ... setsid() cannot re-enter a pgrp */ 931 if (mksess && (curp->p_pgid == curp->p_pid || 932 pgrp_find(curp->p_pid))) 933 goto done; 934 p = curp; 935 } 936 937 /* Changing the process group/session of a session 938 leader is definitely off limits. */ 939 if (SESS_LEADER(p)) { 940 if (sess == NULL && p->p_pgrp == pgrp) 941 /* unless it's a definite noop */ 942 rval = 0; 943 goto done; 944 } 945 946 /* Can only create a process group with id of process */ 947 if (pgrp == NULL && pgid != pid) 948 goto done; 949 950 /* Can only create a session if creating pgrp */ 951 if (sess != NULL && pgrp != NULL) 952 goto done; 953 954 /* Check we allocated memory for a pgrp... */ 955 if (pgrp == NULL && new_pgrp == NULL) 956 goto done; 957 958 /* Don't attach to 'zombie' pgrp */ 959 if (pgrp != NULL && LIST_EMPTY(&pgrp->pg_members)) 960 goto done; 961 962 /* Expect to succeed now */ 963 rval = 0; 964 965 if (pgrp == p->p_pgrp) 966 /* nothing to do */ 967 goto done; 968 969 /* Ok all setup, link up required structures */ 970 971 if (pgrp == NULL) { 972 pgrp = new_pgrp; 973 new_pgrp = NULL; 974 if (sess != NULL) { 975 sess->s_sid = p->p_pid; 976 sess->s_leader = p; 977 sess->s_count = 1; 978 sess->s_ttyvp = NULL; 979 sess->s_ttyp = NULL; 980 sess->s_flags = p->p_session->s_flags & ~S_LOGIN_SET; 981 memcpy(sess->s_login, p->p_session->s_login, 982 sizeof(sess->s_login)); 983 p->p_lflag &= ~PL_CONTROLT; 984 } else { 985 sess = p->p_pgrp->pg_session; 986 proc_sesshold(sess); 987 } 988 pgrp->pg_session = sess; 989 sess = NULL; 990 991 pgrp->pg_id = pgid; 992 LIST_INIT(&pgrp->pg_members); 993 #ifdef DIAGNOSTIC 994 if (__predict_false(pid_table[pgid & pid_tbl_mask].pt_pgrp)) 995 panic("enterpgrp: pgrp table slot in use"); 996 if (__predict_false(mksess && p != curp)) 997 panic("enterpgrp: mksession and p != curproc"); 998 #endif 999 pid_table[pgid & pid_tbl_mask].pt_pgrp = pgrp; 1000 pgrp->pg_jobc = 0; 1001 } 1002 1003 /* 1004 * Adjust eligibility of affected pgrps to participate in job control. 1005 * Increment eligibility counts before decrementing, otherwise we 1006 * could reach 0 spuriously during the first call. 1007 */ 1008 fixjobc(p, pgrp, 1); 1009 fixjobc(p, p->p_pgrp, 0); 1010 1011 /* Interlock with ttread(). */ 1012 mutex_spin_enter(&tty_lock); 1013 1014 /* Move process to requested group. */ 1015 LIST_REMOVE(p, p_pglist); 1016 if (LIST_EMPTY(&p->p_pgrp->pg_members)) 1017 /* defer delete until we've dumped the lock */ 1018 pg_id = p->p_pgrp->pg_id; 1019 p->p_pgrp = pgrp; 1020 LIST_INSERT_HEAD(&pgrp->pg_members, p, p_pglist); 1021 1022 /* Done with the swap; we can release the tty mutex. */ 1023 mutex_spin_exit(&tty_lock); 1024 1025 done: 1026 if (pg_id != NO_PGID) { 1027 /* Releases proc_lock. */ 1028 pg_delete(pg_id); 1029 } else { 1030 mutex_exit(proc_lock); 1031 } 1032 if (sess != NULL) 1033 kmem_free(sess, sizeof(*sess)); 1034 if (new_pgrp != NULL) 1035 kmem_free(new_pgrp, sizeof(*new_pgrp)); 1036 #ifdef DEBUG_PGRP 1037 if (__predict_false(rval)) 1038 printf("enterpgrp(%d,%d,%d), curproc %d, rval %d\n", 1039 pid, pgid, mksess, curp->p_pid, rval); 1040 #endif 1041 return rval; 1042 } 1043 1044 /* 1045 * proc_leavepgrp: remove a process from its process group. 1046 * => must be called with the proc_lock held, which will be released; 1047 */ 1048 void 1049 proc_leavepgrp(struct proc *p) 1050 { 1051 struct pgrp *pgrp; 1052 1053 KASSERT(mutex_owned(proc_lock)); 1054 1055 /* Interlock with ttread() */ 1056 mutex_spin_enter(&tty_lock); 1057 pgrp = p->p_pgrp; 1058 LIST_REMOVE(p, p_pglist); 1059 p->p_pgrp = NULL; 1060 mutex_spin_exit(&tty_lock); 1061 1062 if (LIST_EMPTY(&pgrp->pg_members)) { 1063 /* Releases proc_lock. */ 1064 pg_delete(pgrp->pg_id); 1065 } else { 1066 mutex_exit(proc_lock); 1067 } 1068 } 1069 1070 /* 1071 * pg_remove: remove a process group from the table. 1072 * => must be called with the proc_lock held; 1073 * => returns process group to free; 1074 */ 1075 static struct pgrp * 1076 pg_remove(pid_t pg_id) 1077 { 1078 struct pgrp *pgrp; 1079 struct pid_table *pt; 1080 1081 KASSERT(mutex_owned(proc_lock)); 1082 1083 pt = &pid_table[pg_id & pid_tbl_mask]; 1084 pgrp = pt->pt_pgrp; 1085 1086 KASSERT(pgrp != NULL); 1087 KASSERT(pgrp->pg_id == pg_id); 1088 KASSERT(LIST_EMPTY(&pgrp->pg_members)); 1089 1090 pt->pt_pgrp = NULL; 1091 1092 if (!P_VALID(pt->pt_proc)) { 1093 /* Orphaned pgrp, put slot onto free list. */ 1094 KASSERT((P_NEXT(pt->pt_proc) & pid_tbl_mask) == 0); 1095 pg_id &= pid_tbl_mask; 1096 pt = &pid_table[last_free_pt]; 1097 pt->pt_proc = P_FREE(P_NEXT(pt->pt_proc) | pg_id); 1098 KASSERT(pt->pt_pid == 0); 1099 last_free_pt = pg_id; 1100 pid_alloc_cnt--; 1101 } 1102 return pgrp; 1103 } 1104 1105 /* 1106 * pg_delete: delete and free a process group. 1107 * => must be called with the proc_lock held, which will be released. 1108 */ 1109 static void 1110 pg_delete(pid_t pg_id) 1111 { 1112 struct pgrp *pg; 1113 struct tty *ttyp; 1114 struct session *ss; 1115 1116 KASSERT(mutex_owned(proc_lock)); 1117 1118 pg = pid_table[pg_id & pid_tbl_mask].pt_pgrp; 1119 if (pg == NULL || pg->pg_id != pg_id || !LIST_EMPTY(&pg->pg_members)) { 1120 mutex_exit(proc_lock); 1121 return; 1122 } 1123 1124 ss = pg->pg_session; 1125 1126 /* Remove reference (if any) from tty to this process group */ 1127 mutex_spin_enter(&tty_lock); 1128 ttyp = ss->s_ttyp; 1129 if (ttyp != NULL && ttyp->t_pgrp == pg) { 1130 ttyp->t_pgrp = NULL; 1131 KASSERT(ttyp->t_session == ss); 1132 } 1133 mutex_spin_exit(&tty_lock); 1134 1135 /* 1136 * The leading process group in a session is freed by proc_sessrele(), 1137 * if last reference. Note: proc_sessrele() releases proc_lock. 1138 */ 1139 pg = (ss->s_sid != pg->pg_id) ? pg_remove(pg_id) : NULL; 1140 proc_sessrele(ss); 1141 1142 if (pg != NULL) { 1143 /* Free it, if was not done by proc_sessrele(). */ 1144 kmem_free(pg, sizeof(struct pgrp)); 1145 } 1146 } 1147 1148 /* 1149 * Adjust pgrp jobc counters when specified process changes process group. 1150 * We count the number of processes in each process group that "qualify" 1151 * the group for terminal job control (those with a parent in a different 1152 * process group of the same session). If that count reaches zero, the 1153 * process group becomes orphaned. Check both the specified process' 1154 * process group and that of its children. 1155 * entering == 0 => p is leaving specified group. 1156 * entering == 1 => p is entering specified group. 1157 * 1158 * Call with proc_lock held. 1159 */ 1160 void 1161 fixjobc(struct proc *p, struct pgrp *pgrp, int entering) 1162 { 1163 struct pgrp *hispgrp; 1164 struct session *mysession = pgrp->pg_session; 1165 struct proc *child; 1166 1167 KASSERT(mutex_owned(proc_lock)); 1168 1169 /* 1170 * Check p's parent to see whether p qualifies its own process 1171 * group; if so, adjust count for p's process group. 1172 */ 1173 hispgrp = p->p_pptr->p_pgrp; 1174 if (hispgrp != pgrp && hispgrp->pg_session == mysession) { 1175 if (entering) { 1176 pgrp->pg_jobc++; 1177 p->p_lflag &= ~PL_ORPHANPG; 1178 } else if (--pgrp->pg_jobc == 0) 1179 orphanpg(pgrp); 1180 } 1181 1182 /* 1183 * Check this process' children to see whether they qualify 1184 * their process groups; if so, adjust counts for children's 1185 * process groups. 1186 */ 1187 LIST_FOREACH(child, &p->p_children, p_sibling) { 1188 hispgrp = child->p_pgrp; 1189 if (hispgrp != pgrp && hispgrp->pg_session == mysession && 1190 !P_ZOMBIE(child)) { 1191 if (entering) { 1192 child->p_lflag &= ~PL_ORPHANPG; 1193 hispgrp->pg_jobc++; 1194 } else if (--hispgrp->pg_jobc == 0) 1195 orphanpg(hispgrp); 1196 } 1197 } 1198 } 1199 1200 /* 1201 * A process group has become orphaned; 1202 * if there are any stopped processes in the group, 1203 * hang-up all process in that group. 1204 * 1205 * Call with proc_lock held. 1206 */ 1207 static void 1208 orphanpg(struct pgrp *pg) 1209 { 1210 struct proc *p; 1211 1212 KASSERT(mutex_owned(proc_lock)); 1213 1214 LIST_FOREACH(p, &pg->pg_members, p_pglist) { 1215 if (p->p_stat == SSTOP) { 1216 p->p_lflag |= PL_ORPHANPG; 1217 psignal(p, SIGHUP); 1218 psignal(p, SIGCONT); 1219 } 1220 } 1221 } 1222 1223 #ifdef DDB 1224 #include <ddb/db_output.h> 1225 void pidtbl_dump(void); 1226 void 1227 pidtbl_dump(void) 1228 { 1229 struct pid_table *pt; 1230 struct proc *p; 1231 struct pgrp *pgrp; 1232 int id; 1233 1234 db_printf("pid table %p size %x, next %x, last %x\n", 1235 pid_table, pid_tbl_mask+1, 1236 next_free_pt, last_free_pt); 1237 for (pt = pid_table, id = 0; id <= pid_tbl_mask; id++, pt++) { 1238 p = pt->pt_proc; 1239 if (!P_VALID(p) && !pt->pt_pgrp) 1240 continue; 1241 db_printf(" id %x: ", id); 1242 if (P_VALID(p)) 1243 db_printf("slotpid %d proc %p id %d (0x%x) %s\n", 1244 pt->pt_pid, p, p->p_pid, p->p_pid, p->p_comm); 1245 else 1246 db_printf("next %x use %x\n", 1247 P_NEXT(p) & pid_tbl_mask, 1248 P_NEXT(p) & ~pid_tbl_mask); 1249 if ((pgrp = pt->pt_pgrp)) { 1250 db_printf("\tsession %p, sid %d, count %d, login %s\n", 1251 pgrp->pg_session, pgrp->pg_session->s_sid, 1252 pgrp->pg_session->s_count, 1253 pgrp->pg_session->s_login); 1254 db_printf("\tpgrp %p, pg_id %d, pg_jobc %d, members %p\n", 1255 pgrp, pgrp->pg_id, pgrp->pg_jobc, 1256 LIST_FIRST(&pgrp->pg_members)); 1257 LIST_FOREACH(p, &pgrp->pg_members, p_pglist) { 1258 db_printf("\t\tpid %d addr %p pgrp %p %s\n", 1259 p->p_pid, p, p->p_pgrp, p->p_comm); 1260 } 1261 } 1262 } 1263 } 1264 #endif /* DDB */ 1265 1266 #ifdef KSTACK_CHECK_MAGIC 1267 1268 #define KSTACK_MAGIC 0xdeadbeaf 1269 1270 /* XXX should be per process basis? */ 1271 static int kstackleftmin = KSTACK_SIZE; 1272 static int kstackleftthres = KSTACK_SIZE / 8; 1273 1274 void 1275 kstack_setup_magic(const struct lwp *l) 1276 { 1277 uint32_t *ip; 1278 uint32_t const *end; 1279 1280 KASSERT(l != NULL); 1281 KASSERT(l != &lwp0); 1282 1283 /* 1284 * fill all the stack with magic number 1285 * so that later modification on it can be detected. 1286 */ 1287 ip = (uint32_t *)KSTACK_LOWEST_ADDR(l); 1288 end = (uint32_t *)((char *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE); 1289 for (; ip < end; ip++) { 1290 *ip = KSTACK_MAGIC; 1291 } 1292 } 1293 1294 void 1295 kstack_check_magic(const struct lwp *l) 1296 { 1297 uint32_t const *ip, *end; 1298 int stackleft; 1299 1300 KASSERT(l != NULL); 1301 1302 /* don't check proc0 */ /*XXX*/ 1303 if (l == &lwp0) 1304 return; 1305 1306 #ifdef __MACHINE_STACK_GROWS_UP 1307 /* stack grows upwards (eg. hppa) */ 1308 ip = (uint32_t *)((void *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE); 1309 end = (uint32_t *)KSTACK_LOWEST_ADDR(l); 1310 for (ip--; ip >= end; ip--) 1311 if (*ip != KSTACK_MAGIC) 1312 break; 1313 1314 stackleft = (void *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE - (void *)ip; 1315 #else /* __MACHINE_STACK_GROWS_UP */ 1316 /* stack grows downwards (eg. i386) */ 1317 ip = (uint32_t *)KSTACK_LOWEST_ADDR(l); 1318 end = (uint32_t *)((char *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE); 1319 for (; ip < end; ip++) 1320 if (*ip != KSTACK_MAGIC) 1321 break; 1322 1323 stackleft = ((const char *)ip) - (const char *)KSTACK_LOWEST_ADDR(l); 1324 #endif /* __MACHINE_STACK_GROWS_UP */ 1325 1326 if (kstackleftmin > stackleft) { 1327 kstackleftmin = stackleft; 1328 if (stackleft < kstackleftthres) 1329 printf("warning: kernel stack left %d bytes" 1330 "(pid %u:lid %u)\n", stackleft, 1331 (u_int)l->l_proc->p_pid, (u_int)l->l_lid); 1332 } 1333 1334 if (stackleft <= 0) { 1335 panic("magic on the top of kernel stack changed for " 1336 "pid %u, lid %u: maybe kernel stack overflow", 1337 (u_int)l->l_proc->p_pid, (u_int)l->l_lid); 1338 } 1339 } 1340 #endif /* KSTACK_CHECK_MAGIC */ 1341 1342 int 1343 proclist_foreach_call(struct proclist *list, 1344 int (*callback)(struct proc *, void *arg), void *arg) 1345 { 1346 struct proc marker; 1347 struct proc *p; 1348 int ret = 0; 1349 1350 marker.p_flag = PK_MARKER; 1351 mutex_enter(proc_lock); 1352 for (p = LIST_FIRST(list); ret == 0 && p != NULL;) { 1353 if (p->p_flag & PK_MARKER) { 1354 p = LIST_NEXT(p, p_list); 1355 continue; 1356 } 1357 LIST_INSERT_AFTER(p, &marker, p_list); 1358 ret = (*callback)(p, arg); 1359 KASSERT(mutex_owned(proc_lock)); 1360 p = LIST_NEXT(&marker, p_list); 1361 LIST_REMOVE(&marker, p_list); 1362 } 1363 mutex_exit(proc_lock); 1364 1365 return ret; 1366 } 1367 1368 int 1369 proc_vmspace_getref(struct proc *p, struct vmspace **vm) 1370 { 1371 1372 /* XXXCDC: how should locking work here? */ 1373 1374 /* curproc exception is for coredump. */ 1375 1376 if ((p != curproc && (p->p_sflag & PS_WEXIT) != 0) || 1377 (p->p_vmspace->vm_refcnt < 1)) { /* XXX */ 1378 return EFAULT; 1379 } 1380 1381 uvmspace_addref(p->p_vmspace); 1382 *vm = p->p_vmspace; 1383 1384 return 0; 1385 } 1386 1387 /* 1388 * Acquire a write lock on the process credential. 1389 */ 1390 void 1391 proc_crmod_enter(void) 1392 { 1393 struct lwp *l = curlwp; 1394 struct proc *p = l->l_proc; 1395 kauth_cred_t oc; 1396 1397 /* Reset what needs to be reset in plimit. */ 1398 if (p->p_limit->pl_corename != defcorename) { 1399 lim_setcorename(p, defcorename, 0); 1400 } 1401 1402 mutex_enter(p->p_lock); 1403 1404 /* Ensure the LWP cached credentials are up to date. */ 1405 if ((oc = l->l_cred) != p->p_cred) { 1406 kauth_cred_hold(p->p_cred); 1407 l->l_cred = p->p_cred; 1408 kauth_cred_free(oc); 1409 } 1410 } 1411 1412 /* 1413 * Set in a new process credential, and drop the write lock. The credential 1414 * must have a reference already. Optionally, free a no-longer required 1415 * credential. The scheduler also needs to inspect p_cred, so we also 1416 * briefly acquire the sched state mutex. 1417 */ 1418 void 1419 proc_crmod_leave(kauth_cred_t scred, kauth_cred_t fcred, bool sugid) 1420 { 1421 struct lwp *l = curlwp, *l2; 1422 struct proc *p = l->l_proc; 1423 kauth_cred_t oc; 1424 1425 KASSERT(mutex_owned(p->p_lock)); 1426 1427 /* Is there a new credential to set in? */ 1428 if (scred != NULL) { 1429 p->p_cred = scred; 1430 LIST_FOREACH(l2, &p->p_lwps, l_sibling) { 1431 if (l2 != l) 1432 l2->l_prflag |= LPR_CRMOD; 1433 } 1434 1435 /* Ensure the LWP cached credentials are up to date. */ 1436 if ((oc = l->l_cred) != scred) { 1437 kauth_cred_hold(scred); 1438 l->l_cred = scred; 1439 } 1440 } else 1441 oc = NULL; /* XXXgcc */ 1442 1443 if (sugid) { 1444 /* 1445 * Mark process as having changed credentials, stops 1446 * tracing etc. 1447 */ 1448 p->p_flag |= PK_SUGID; 1449 } 1450 1451 mutex_exit(p->p_lock); 1452 1453 /* If there is a credential to be released, free it now. */ 1454 if (fcred != NULL) { 1455 KASSERT(scred != NULL); 1456 kauth_cred_free(fcred); 1457 if (oc != scred) 1458 kauth_cred_free(oc); 1459 } 1460 } 1461 1462 /* 1463 * proc_specific_key_create -- 1464 * Create a key for subsystem proc-specific data. 1465 */ 1466 int 1467 proc_specific_key_create(specificdata_key_t *keyp, specificdata_dtor_t dtor) 1468 { 1469 1470 return (specificdata_key_create(proc_specificdata_domain, keyp, dtor)); 1471 } 1472 1473 /* 1474 * proc_specific_key_delete -- 1475 * Delete a key for subsystem proc-specific data. 1476 */ 1477 void 1478 proc_specific_key_delete(specificdata_key_t key) 1479 { 1480 1481 specificdata_key_delete(proc_specificdata_domain, key); 1482 } 1483 1484 /* 1485 * proc_initspecific -- 1486 * Initialize a proc's specificdata container. 1487 */ 1488 void 1489 proc_initspecific(struct proc *p) 1490 { 1491 int error __diagused; 1492 1493 error = specificdata_init(proc_specificdata_domain, &p->p_specdataref); 1494 KASSERT(error == 0); 1495 } 1496 1497 /* 1498 * proc_finispecific -- 1499 * Finalize a proc's specificdata container. 1500 */ 1501 void 1502 proc_finispecific(struct proc *p) 1503 { 1504 1505 specificdata_fini(proc_specificdata_domain, &p->p_specdataref); 1506 } 1507 1508 /* 1509 * proc_getspecific -- 1510 * Return proc-specific data corresponding to the specified key. 1511 */ 1512 void * 1513 proc_getspecific(struct proc *p, specificdata_key_t key) 1514 { 1515 1516 return (specificdata_getspecific(proc_specificdata_domain, 1517 &p->p_specdataref, key)); 1518 } 1519 1520 /* 1521 * proc_setspecific -- 1522 * Set proc-specific data corresponding to the specified key. 1523 */ 1524 void 1525 proc_setspecific(struct proc *p, specificdata_key_t key, void *data) 1526 { 1527 1528 specificdata_setspecific(proc_specificdata_domain, 1529 &p->p_specdataref, key, data); 1530 } 1531 1532 int 1533 proc_uidmatch(kauth_cred_t cred, kauth_cred_t target) 1534 { 1535 int r = 0; 1536 1537 if (kauth_cred_getuid(cred) != kauth_cred_getuid(target) || 1538 kauth_cred_getuid(cred) != kauth_cred_getsvuid(target)) { 1539 /* 1540 * suid proc of ours or proc not ours 1541 */ 1542 r = EPERM; 1543 } else if (kauth_cred_getgid(target) != kauth_cred_getsvgid(target)) { 1544 /* 1545 * sgid proc has sgid back to us temporarily 1546 */ 1547 r = EPERM; 1548 } else { 1549 /* 1550 * our rgid must be in target's group list (ie, 1551 * sub-processes started by a sgid process) 1552 */ 1553 int ismember = 0; 1554 1555 if (kauth_cred_ismember_gid(cred, 1556 kauth_cred_getgid(target), &ismember) != 0 || 1557 !ismember) 1558 r = EPERM; 1559 } 1560 1561 return (r); 1562 } 1563 1564 /* 1565 * sysctl stuff 1566 */ 1567 1568 #define KERN_PROCSLOP (5 * sizeof(struct kinfo_proc)) 1569 1570 static const u_int sysctl_flagmap[] = { 1571 PK_ADVLOCK, P_ADVLOCK, 1572 PK_EXEC, P_EXEC, 1573 PK_NOCLDWAIT, P_NOCLDWAIT, 1574 PK_32, P_32, 1575 PK_CLDSIGIGN, P_CLDSIGIGN, 1576 PK_SUGID, P_SUGID, 1577 0 1578 }; 1579 1580 static const u_int sysctl_sflagmap[] = { 1581 PS_NOCLDSTOP, P_NOCLDSTOP, 1582 PS_WEXIT, P_WEXIT, 1583 PS_STOPFORK, P_STOPFORK, 1584 PS_STOPEXEC, P_STOPEXEC, 1585 PS_STOPEXIT, P_STOPEXIT, 1586 0 1587 }; 1588 1589 static const u_int sysctl_slflagmap[] = { 1590 PSL_TRACED, P_TRACED, 1591 PSL_CHTRACED, P_CHTRACED, 1592 PSL_SYSCALL, P_SYSCALL, 1593 0 1594 }; 1595 1596 static const u_int sysctl_lflagmap[] = { 1597 PL_CONTROLT, P_CONTROLT, 1598 PL_PPWAIT, P_PPWAIT, 1599 0 1600 }; 1601 1602 static const u_int sysctl_stflagmap[] = { 1603 PST_PROFIL, P_PROFIL, 1604 0 1605 1606 }; 1607 1608 /* used by kern_lwp also */ 1609 const u_int sysctl_lwpflagmap[] = { 1610 LW_SINTR, L_SINTR, 1611 LW_SYSTEM, L_SYSTEM, 1612 0 1613 }; 1614 1615 /* 1616 * Find the most ``active'' lwp of a process and return it for ps display 1617 * purposes 1618 */ 1619 static struct lwp * 1620 proc_active_lwp(struct proc *p) 1621 { 1622 static const int ostat[] = { 1623 0, 1624 2, /* LSIDL */ 1625 6, /* LSRUN */ 1626 5, /* LSSLEEP */ 1627 4, /* LSSTOP */ 1628 0, /* LSZOMB */ 1629 1, /* LSDEAD */ 1630 7, /* LSONPROC */ 1631 3 /* LSSUSPENDED */ 1632 }; 1633 1634 struct lwp *l, *lp = NULL; 1635 LIST_FOREACH(l, &p->p_lwps, l_sibling) { 1636 KASSERT(l->l_stat >= 0 && l->l_stat < __arraycount(ostat)); 1637 if (lp == NULL || 1638 ostat[l->l_stat] > ostat[lp->l_stat] || 1639 (ostat[l->l_stat] == ostat[lp->l_stat] && 1640 l->l_cpticks > lp->l_cpticks)) { 1641 lp = l; 1642 continue; 1643 } 1644 } 1645 return lp; 1646 } 1647 1648 static int 1649 sysctl_doeproc(SYSCTLFN_ARGS) 1650 { 1651 union { 1652 struct kinfo_proc kproc; 1653 struct kinfo_proc2 kproc2; 1654 } *kbuf; 1655 struct proc *p, *next, *marker; 1656 char *where, *dp; 1657 int type, op, arg, error; 1658 u_int elem_size, kelem_size, elem_count; 1659 size_t buflen, needed; 1660 bool match, zombie, mmmbrains; 1661 const bool allowaddr = get_expose_address(curproc); 1662 1663 if (namelen == 1 && name[0] == CTL_QUERY) 1664 return (sysctl_query(SYSCTLFN_CALL(rnode))); 1665 1666 dp = where = oldp; 1667 buflen = where != NULL ? *oldlenp : 0; 1668 error = 0; 1669 needed = 0; 1670 type = rnode->sysctl_num; 1671 1672 if (type == KERN_PROC) { 1673 if (namelen == 0) 1674 return EINVAL; 1675 switch (op = name[0]) { 1676 case KERN_PROC_ALL: 1677 if (namelen != 1) 1678 return EINVAL; 1679 arg = 0; 1680 break; 1681 default: 1682 if (namelen != 2) 1683 return EINVAL; 1684 arg = name[1]; 1685 break; 1686 } 1687 elem_count = 0; /* Hush little compiler, don't you cry */ 1688 kelem_size = elem_size = sizeof(kbuf->kproc); 1689 } else { 1690 if (namelen != 4) 1691 return EINVAL; 1692 op = name[0]; 1693 arg = name[1]; 1694 elem_size = name[2]; 1695 elem_count = name[3]; 1696 kelem_size = sizeof(kbuf->kproc2); 1697 } 1698 1699 sysctl_unlock(); 1700 1701 kbuf = kmem_zalloc(sizeof(*kbuf), KM_SLEEP); 1702 marker = kmem_alloc(sizeof(*marker), KM_SLEEP); 1703 marker->p_flag = PK_MARKER; 1704 1705 mutex_enter(proc_lock); 1706 /* 1707 * Start with zombies to prevent reporting processes twice, in case they 1708 * are dying and being moved from the list of alive processes to zombies. 1709 */ 1710 mmmbrains = true; 1711 for (p = LIST_FIRST(&zombproc);; p = next) { 1712 if (p == NULL) { 1713 if (mmmbrains) { 1714 p = LIST_FIRST(&allproc); 1715 mmmbrains = false; 1716 } 1717 if (p == NULL) 1718 break; 1719 } 1720 next = LIST_NEXT(p, p_list); 1721 if ((p->p_flag & PK_MARKER) != 0) 1722 continue; 1723 1724 /* 1725 * Skip embryonic processes. 1726 */ 1727 if (p->p_stat == SIDL) 1728 continue; 1729 1730 mutex_enter(p->p_lock); 1731 error = kauth_authorize_process(l->l_cred, 1732 KAUTH_PROCESS_CANSEE, p, 1733 KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_EPROC), NULL, NULL); 1734 if (error != 0) { 1735 mutex_exit(p->p_lock); 1736 continue; 1737 } 1738 1739 /* 1740 * Hande all the operations in one switch on the cost of 1741 * algorithm complexity is on purpose. The win splitting this 1742 * function into several similar copies makes maintenance burden 1743 * burden, code grow and boost is neglible in practical systems. 1744 */ 1745 switch (op) { 1746 case KERN_PROC_PID: 1747 match = (p->p_pid == (pid_t)arg); 1748 break; 1749 1750 case KERN_PROC_PGRP: 1751 match = (p->p_pgrp->pg_id == (pid_t)arg); 1752 break; 1753 1754 case KERN_PROC_SESSION: 1755 match = (p->p_session->s_sid == (pid_t)arg); 1756 break; 1757 1758 case KERN_PROC_TTY: 1759 match = true; 1760 if (arg == (int) KERN_PROC_TTY_REVOKE) { 1761 if ((p->p_lflag & PL_CONTROLT) == 0 || 1762 p->p_session->s_ttyp == NULL || 1763 p->p_session->s_ttyvp != NULL) { 1764 match = false; 1765 } 1766 } else if ((p->p_lflag & PL_CONTROLT) == 0 || 1767 p->p_session->s_ttyp == NULL) { 1768 if ((dev_t)arg != KERN_PROC_TTY_NODEV) { 1769 match = false; 1770 } 1771 } else if (p->p_session->s_ttyp->t_dev != (dev_t)arg) { 1772 match = false; 1773 } 1774 break; 1775 1776 case KERN_PROC_UID: 1777 match = (kauth_cred_geteuid(p->p_cred) == (uid_t)arg); 1778 break; 1779 1780 case KERN_PROC_RUID: 1781 match = (kauth_cred_getuid(p->p_cred) == (uid_t)arg); 1782 break; 1783 1784 case KERN_PROC_GID: 1785 match = (kauth_cred_getegid(p->p_cred) == (uid_t)arg); 1786 break; 1787 1788 case KERN_PROC_RGID: 1789 match = (kauth_cred_getgid(p->p_cred) == (uid_t)arg); 1790 break; 1791 1792 case KERN_PROC_ALL: 1793 match = true; 1794 /* allow everything */ 1795 break; 1796 1797 default: 1798 error = EINVAL; 1799 mutex_exit(p->p_lock); 1800 goto cleanup; 1801 } 1802 if (!match) { 1803 mutex_exit(p->p_lock); 1804 continue; 1805 } 1806 1807 /* 1808 * Grab a hold on the process. 1809 */ 1810 if (mmmbrains) { 1811 zombie = true; 1812 } else { 1813 zombie = !rw_tryenter(&p->p_reflock, RW_READER); 1814 } 1815 if (zombie) { 1816 LIST_INSERT_AFTER(p, marker, p_list); 1817 } 1818 1819 if (buflen >= elem_size && 1820 (type == KERN_PROC || elem_count > 0)) { 1821 if (type == KERN_PROC) { 1822 fill_proc(p, &kbuf->kproc.kp_proc, allowaddr); 1823 fill_eproc(p, &kbuf->kproc.kp_eproc, zombie, 1824 allowaddr); 1825 } else { 1826 fill_kproc2(p, &kbuf->kproc2, zombie, 1827 allowaddr); 1828 elem_count--; 1829 } 1830 mutex_exit(p->p_lock); 1831 mutex_exit(proc_lock); 1832 /* 1833 * Copy out elem_size, but not larger than kelem_size 1834 */ 1835 error = sysctl_copyout(l, kbuf, dp, 1836 uimin(kelem_size, elem_size)); 1837 mutex_enter(proc_lock); 1838 if (error) { 1839 goto bah; 1840 } 1841 dp += elem_size; 1842 buflen -= elem_size; 1843 } else { 1844 mutex_exit(p->p_lock); 1845 } 1846 needed += elem_size; 1847 1848 /* 1849 * Release reference to process. 1850 */ 1851 if (zombie) { 1852 next = LIST_NEXT(marker, p_list); 1853 LIST_REMOVE(marker, p_list); 1854 } else { 1855 rw_exit(&p->p_reflock); 1856 next = LIST_NEXT(p, p_list); 1857 } 1858 1859 /* 1860 * Short-circuit break quickly! 1861 */ 1862 if (op == KERN_PROC_PID) 1863 break; 1864 } 1865 mutex_exit(proc_lock); 1866 1867 if (where != NULL) { 1868 *oldlenp = dp - where; 1869 if (needed > *oldlenp) { 1870 error = ENOMEM; 1871 goto out; 1872 } 1873 } else { 1874 needed += KERN_PROCSLOP; 1875 *oldlenp = needed; 1876 } 1877 kmem_free(kbuf, sizeof(*kbuf)); 1878 kmem_free(marker, sizeof(*marker)); 1879 sysctl_relock(); 1880 return 0; 1881 bah: 1882 if (zombie) 1883 LIST_REMOVE(marker, p_list); 1884 else 1885 rw_exit(&p->p_reflock); 1886 cleanup: 1887 mutex_exit(proc_lock); 1888 out: 1889 kmem_free(kbuf, sizeof(*kbuf)); 1890 kmem_free(marker, sizeof(*marker)); 1891 sysctl_relock(); 1892 return error; 1893 } 1894 1895 /* 1896 * compat_netbsd32 hooks 1897 */ 1898 #if !defined(_RUMPKERNEL) 1899 MODULE_CALL_HOOK_DECL(kern_proc_32_copyin_hook, int, 1900 (struct proc *p, struct ps_strings *s)); 1901 MODULE_CALL_HOOK(kern_proc_32_copyin_hook, int, 1902 (struct proc *p, struct ps_strings *s), (p, s), enosys()); 1903 1904 MODULE_CALL_HOOK_DECL(kern_proc_32_base_hook, void, 1905 (char **argv, size_t i, vaddr_t *base)); 1906 MODULE_CALL_VOID_HOOK(kern_proc_32_base_hook, 1907 (char **argv, size_t i, vaddr_t *base), (argv, i, base), __nothing); 1908 #endif /* !defined(_RUMPKERNEL) */ 1909 1910 int 1911 copyin_psstrings(struct proc *p, struct ps_strings *arginfo) 1912 { 1913 1914 #if !defined(_RUMPKERNEL) 1915 if (p->p_flag & PK_32) 1916 return kern_proc_32_copyin_hook_call(p, arginfo); 1917 #endif /* !defined(_RUMPKERNEL) */ 1918 1919 return copyin_proc(p, (void *)p->p_psstrp, arginfo, sizeof(*arginfo)); 1920 } 1921 1922 static int 1923 copy_procargs_sysctl_cb(void *cookie_, const void *src, size_t off, size_t len) 1924 { 1925 void **cookie = cookie_; 1926 struct lwp *l = cookie[0]; 1927 char *dst = cookie[1]; 1928 1929 return sysctl_copyout(l, src, dst + off, len); 1930 } 1931 1932 /* 1933 * sysctl helper routine for kern.proc_args pseudo-subtree. 1934 */ 1935 static int 1936 sysctl_kern_proc_args(SYSCTLFN_ARGS) 1937 { 1938 struct ps_strings pss; 1939 struct proc *p; 1940 pid_t pid; 1941 int type, error; 1942 void *cookie[2]; 1943 1944 if (namelen == 1 && name[0] == CTL_QUERY) 1945 return (sysctl_query(SYSCTLFN_CALL(rnode))); 1946 1947 if (newp != NULL || namelen != 2) 1948 return (EINVAL); 1949 pid = name[0]; 1950 type = name[1]; 1951 1952 switch (type) { 1953 case KERN_PROC_PATHNAME: 1954 sysctl_unlock(); 1955 error = fill_pathname(l, pid, oldp, oldlenp); 1956 sysctl_relock(); 1957 return error; 1958 1959 case KERN_PROC_ARGV: 1960 case KERN_PROC_NARGV: 1961 case KERN_PROC_ENV: 1962 case KERN_PROC_NENV: 1963 /* ok */ 1964 break; 1965 default: 1966 return (EINVAL); 1967 } 1968 1969 sysctl_unlock(); 1970 1971 /* check pid */ 1972 mutex_enter(proc_lock); 1973 if ((p = proc_find(pid)) == NULL) { 1974 error = EINVAL; 1975 goto out_locked; 1976 } 1977 mutex_enter(p->p_lock); 1978 1979 /* Check permission. */ 1980 if (type == KERN_PROC_ARGV || type == KERN_PROC_NARGV) 1981 error = kauth_authorize_process(l->l_cred, KAUTH_PROCESS_CANSEE, 1982 p, KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_ARGS), NULL, NULL); 1983 else if (type == KERN_PROC_ENV || type == KERN_PROC_NENV) 1984 error = kauth_authorize_process(l->l_cred, KAUTH_PROCESS_CANSEE, 1985 p, KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_ENV), NULL, NULL); 1986 else 1987 error = EINVAL; /* XXXGCC */ 1988 if (error) { 1989 mutex_exit(p->p_lock); 1990 goto out_locked; 1991 } 1992 1993 if (oldp == NULL) { 1994 if (type == KERN_PROC_NARGV || type == KERN_PROC_NENV) 1995 *oldlenp = sizeof (int); 1996 else 1997 *oldlenp = ARG_MAX; /* XXX XXX XXX */ 1998 error = 0; 1999 mutex_exit(p->p_lock); 2000 goto out_locked; 2001 } 2002 2003 /* 2004 * Zombies don't have a stack, so we can't read their psstrings. 2005 * System processes also don't have a user stack. 2006 */ 2007 if (P_ZOMBIE(p) || (p->p_flag & PK_SYSTEM) != 0) { 2008 error = EINVAL; 2009 mutex_exit(p->p_lock); 2010 goto out_locked; 2011 } 2012 2013 error = rw_tryenter(&p->p_reflock, RW_READER) ? 0 : EBUSY; 2014 mutex_exit(p->p_lock); 2015 if (error) { 2016 goto out_locked; 2017 } 2018 mutex_exit(proc_lock); 2019 2020 if (type == KERN_PROC_NARGV || type == KERN_PROC_NENV) { 2021 int value; 2022 if ((error = copyin_psstrings(p, &pss)) == 0) { 2023 if (type == KERN_PROC_NARGV) 2024 value = pss.ps_nargvstr; 2025 else 2026 value = pss.ps_nenvstr; 2027 error = sysctl_copyout(l, &value, oldp, sizeof(value)); 2028 *oldlenp = sizeof(value); 2029 } 2030 } else { 2031 cookie[0] = l; 2032 cookie[1] = oldp; 2033 error = copy_procargs(p, type, oldlenp, 2034 copy_procargs_sysctl_cb, cookie); 2035 } 2036 rw_exit(&p->p_reflock); 2037 sysctl_relock(); 2038 return error; 2039 2040 out_locked: 2041 mutex_exit(proc_lock); 2042 sysctl_relock(); 2043 return error; 2044 } 2045 2046 int 2047 copy_procargs(struct proc *p, int oid, size_t *limit, 2048 int (*cb)(void *, const void *, size_t, size_t), void *cookie) 2049 { 2050 struct ps_strings pss; 2051 size_t len, i, loaded, entry_len; 2052 struct uio auio; 2053 struct iovec aiov; 2054 int error, argvlen; 2055 char *arg; 2056 char **argv; 2057 vaddr_t user_argv; 2058 struct vmspace *vmspace; 2059 2060 /* 2061 * Allocate a temporary buffer to hold the argument vector and 2062 * the arguments themselve. 2063 */ 2064 arg = kmem_alloc(PAGE_SIZE, KM_SLEEP); 2065 argv = kmem_alloc(PAGE_SIZE, KM_SLEEP); 2066 2067 /* 2068 * Lock the process down in memory. 2069 */ 2070 vmspace = p->p_vmspace; 2071 uvmspace_addref(vmspace); 2072 2073 /* 2074 * Read in the ps_strings structure. 2075 */ 2076 if ((error = copyin_psstrings(p, &pss)) != 0) 2077 goto done; 2078 2079 /* 2080 * Now read the address of the argument vector. 2081 */ 2082 switch (oid) { 2083 case KERN_PROC_ARGV: 2084 user_argv = (uintptr_t)pss.ps_argvstr; 2085 argvlen = pss.ps_nargvstr; 2086 break; 2087 case KERN_PROC_ENV: 2088 user_argv = (uintptr_t)pss.ps_envstr; 2089 argvlen = pss.ps_nenvstr; 2090 break; 2091 default: 2092 error = EINVAL; 2093 goto done; 2094 } 2095 2096 if (argvlen < 0) { 2097 error = EIO; 2098 goto done; 2099 } 2100 2101 2102 /* 2103 * Now copy each string. 2104 */ 2105 len = 0; /* bytes written to user buffer */ 2106 loaded = 0; /* bytes from argv already processed */ 2107 i = 0; /* To make compiler happy */ 2108 entry_len = PROC_PTRSZ(p); 2109 2110 for (; argvlen; --argvlen) { 2111 int finished = 0; 2112 vaddr_t base; 2113 size_t xlen; 2114 int j; 2115 2116 if (loaded == 0) { 2117 size_t rem = entry_len * argvlen; 2118 loaded = MIN(rem, PAGE_SIZE); 2119 error = copyin_vmspace(vmspace, 2120 (const void *)user_argv, argv, loaded); 2121 if (error) 2122 break; 2123 user_argv += loaded; 2124 i = 0; 2125 } 2126 2127 #if !defined(_RUMPKERNEL) 2128 if (p->p_flag & PK_32) 2129 kern_proc_32_base_hook_call(argv, i++, &base); 2130 else 2131 #endif /* !defined(_RUMPKERNEL) */ 2132 base = (vaddr_t)argv[i++]; 2133 loaded -= entry_len; 2134 2135 /* 2136 * The program has messed around with its arguments, 2137 * possibly deleting some, and replacing them with 2138 * NULL's. Treat this as the last argument and not 2139 * a failure. 2140 */ 2141 if (base == 0) 2142 break; 2143 2144 while (!finished) { 2145 xlen = PAGE_SIZE - (base & PAGE_MASK); 2146 2147 aiov.iov_base = arg; 2148 aiov.iov_len = PAGE_SIZE; 2149 auio.uio_iov = &aiov; 2150 auio.uio_iovcnt = 1; 2151 auio.uio_offset = base; 2152 auio.uio_resid = xlen; 2153 auio.uio_rw = UIO_READ; 2154 UIO_SETUP_SYSSPACE(&auio); 2155 error = uvm_io(&vmspace->vm_map, &auio, 0); 2156 if (error) 2157 goto done; 2158 2159 /* Look for the end of the string */ 2160 for (j = 0; j < xlen; j++) { 2161 if (arg[j] == '\0') { 2162 xlen = j + 1; 2163 finished = 1; 2164 break; 2165 } 2166 } 2167 2168 /* Check for user buffer overflow */ 2169 if (len + xlen > *limit) { 2170 finished = 1; 2171 if (len > *limit) 2172 xlen = 0; 2173 else 2174 xlen = *limit - len; 2175 } 2176 2177 /* Copyout the page */ 2178 error = (*cb)(cookie, arg, len, xlen); 2179 if (error) 2180 goto done; 2181 2182 len += xlen; 2183 base += xlen; 2184 } 2185 } 2186 *limit = len; 2187 2188 done: 2189 kmem_free(argv, PAGE_SIZE); 2190 kmem_free(arg, PAGE_SIZE); 2191 uvmspace_free(vmspace); 2192 return error; 2193 } 2194 2195 /* 2196 * Fill in a proc structure for the specified process. 2197 */ 2198 static void 2199 fill_proc(const struct proc *psrc, struct proc *p, bool allowaddr) 2200 { 2201 COND_SET_VALUE(p->p_list, psrc->p_list, allowaddr); 2202 COND_SET_VALUE(p->p_auxlock, psrc->p_auxlock, allowaddr); 2203 COND_SET_VALUE(p->p_lock, psrc->p_lock, allowaddr); 2204 COND_SET_VALUE(p->p_stmutex, psrc->p_stmutex, allowaddr); 2205 COND_SET_VALUE(p->p_reflock, psrc->p_reflock, allowaddr); 2206 COND_SET_VALUE(p->p_waitcv, psrc->p_waitcv, allowaddr); 2207 COND_SET_VALUE(p->p_lwpcv, psrc->p_lwpcv, allowaddr); 2208 COND_SET_VALUE(p->p_cred, psrc->p_cred, allowaddr); 2209 COND_SET_VALUE(p->p_fd, psrc->p_fd, allowaddr); 2210 COND_SET_VALUE(p->p_cwdi, psrc->p_cwdi, allowaddr); 2211 COND_SET_VALUE(p->p_stats, psrc->p_stats, allowaddr); 2212 COND_SET_VALUE(p->p_limit, psrc->p_limit, allowaddr); 2213 COND_SET_VALUE(p->p_vmspace, psrc->p_vmspace, allowaddr); 2214 COND_SET_VALUE(p->p_sigacts, psrc->p_sigacts, allowaddr); 2215 COND_SET_VALUE(p->p_aio, psrc->p_aio, allowaddr); 2216 p->p_mqueue_cnt = psrc->p_mqueue_cnt; 2217 COND_SET_VALUE(p->p_specdataref, psrc->p_specdataref, allowaddr); 2218 p->p_exitsig = psrc->p_exitsig; 2219 p->p_flag = psrc->p_flag; 2220 p->p_sflag = psrc->p_sflag; 2221 p->p_slflag = psrc->p_slflag; 2222 p->p_lflag = psrc->p_lflag; 2223 p->p_stflag = psrc->p_stflag; 2224 p->p_stat = psrc->p_stat; 2225 p->p_trace_enabled = psrc->p_trace_enabled; 2226 p->p_pid = psrc->p_pid; 2227 COND_SET_VALUE(p->p_pglist, psrc->p_pglist, allowaddr); 2228 COND_SET_VALUE(p->p_pptr, psrc->p_pptr, allowaddr); 2229 COND_SET_VALUE(p->p_sibling, psrc->p_sibling, allowaddr); 2230 COND_SET_VALUE(p->p_children, psrc->p_children, allowaddr); 2231 COND_SET_VALUE(p->p_lwps, psrc->p_lwps, allowaddr); 2232 COND_SET_VALUE(p->p_raslist, psrc->p_raslist, allowaddr); 2233 p->p_nlwps = psrc->p_nlwps; 2234 p->p_nzlwps = psrc->p_nzlwps; 2235 p->p_nrlwps = psrc->p_nrlwps; 2236 p->p_nlwpwait = psrc->p_nlwpwait; 2237 p->p_ndlwps = psrc->p_ndlwps; 2238 p->p_nlwpid = psrc->p_nlwpid; 2239 p->p_nstopchild = psrc->p_nstopchild; 2240 p->p_waited = psrc->p_waited; 2241 COND_SET_VALUE(p->p_zomblwp, psrc->p_zomblwp, allowaddr); 2242 COND_SET_VALUE(p->p_vforklwp, psrc->p_vforklwp, allowaddr); 2243 COND_SET_VALUE(p->p_sched_info, psrc->p_sched_info, allowaddr); 2244 p->p_estcpu = psrc->p_estcpu; 2245 p->p_estcpu_inherited = psrc->p_estcpu_inherited; 2246 p->p_forktime = psrc->p_forktime; 2247 p->p_pctcpu = psrc->p_pctcpu; 2248 COND_SET_VALUE(p->p_opptr, psrc->p_opptr, allowaddr); 2249 COND_SET_VALUE(p->p_timers, psrc->p_timers, allowaddr); 2250 p->p_rtime = psrc->p_rtime; 2251 p->p_uticks = psrc->p_uticks; 2252 p->p_sticks = psrc->p_sticks; 2253 p->p_iticks = psrc->p_iticks; 2254 p->p_xutime = psrc->p_xutime; 2255 p->p_xstime = psrc->p_xstime; 2256 p->p_traceflag = psrc->p_traceflag; 2257 COND_SET_VALUE(p->p_tracep, psrc->p_tracep, allowaddr); 2258 COND_SET_VALUE(p->p_textvp, psrc->p_textvp, allowaddr); 2259 COND_SET_VALUE(p->p_emul, psrc->p_emul, allowaddr); 2260 COND_SET_VALUE(p->p_emuldata, psrc->p_emuldata, allowaddr); 2261 COND_SET_VALUE(p->p_execsw, psrc->p_execsw, allowaddr); 2262 COND_SET_VALUE(p->p_klist, psrc->p_klist, allowaddr); 2263 COND_SET_VALUE(p->p_sigwaiters, psrc->p_sigwaiters, allowaddr); 2264 COND_SET_VALUE(p->p_sigpend, psrc->p_sigpend, allowaddr); 2265 COND_SET_VALUE(p->p_lwpctl, psrc->p_lwpctl, allowaddr); 2266 p->p_ppid = psrc->p_ppid; 2267 p->p_fpid = psrc->p_fpid; 2268 p->p_vfpid = psrc->p_vfpid; 2269 p->p_vfpid_done = psrc->p_vfpid_done; 2270 p->p_lwp_created = psrc->p_lwp_created; 2271 p->p_lwp_exited = psrc->p_lwp_exited; 2272 p->p_nsems = psrc->p_nsems; 2273 COND_SET_VALUE(p->p_path, psrc->p_path, allowaddr); 2274 COND_SET_VALUE(p->p_sigctx, psrc->p_sigctx, allowaddr); 2275 p->p_nice = psrc->p_nice; 2276 memcpy(p->p_comm, psrc->p_comm, sizeof(p->p_comm)); 2277 COND_SET_VALUE(p->p_pgrp, psrc->p_pgrp, allowaddr); 2278 COND_SET_VALUE(p->p_psstrp, psrc->p_psstrp, allowaddr); 2279 p->p_pax = psrc->p_pax; 2280 p->p_xexit = psrc->p_xexit; 2281 p->p_xsig = psrc->p_xsig; 2282 p->p_acflag = psrc->p_acflag; 2283 COND_SET_VALUE(p->p_md, psrc->p_md, allowaddr); 2284 p->p_stackbase = psrc->p_stackbase; 2285 COND_SET_VALUE(p->p_dtrace, psrc->p_dtrace, allowaddr); 2286 } 2287 2288 /* 2289 * Fill in an eproc structure for the specified process. 2290 */ 2291 void 2292 fill_eproc(struct proc *p, struct eproc *ep, bool zombie, bool allowaddr) 2293 { 2294 struct tty *tp; 2295 struct lwp *l; 2296 2297 KASSERT(mutex_owned(proc_lock)); 2298 KASSERT(mutex_owned(p->p_lock)); 2299 2300 COND_SET_VALUE(ep->e_paddr, p, allowaddr); 2301 COND_SET_VALUE(ep->e_sess, p->p_session, allowaddr); 2302 if (p->p_cred) { 2303 kauth_cred_topcred(p->p_cred, &ep->e_pcred); 2304 kauth_cred_toucred(p->p_cred, &ep->e_ucred); 2305 } 2306 if (p->p_stat != SIDL && !P_ZOMBIE(p) && !zombie) { 2307 struct vmspace *vm = p->p_vmspace; 2308 2309 ep->e_vm.vm_rssize = vm_resident_count(vm); 2310 ep->e_vm.vm_tsize = vm->vm_tsize; 2311 ep->e_vm.vm_dsize = vm->vm_dsize; 2312 ep->e_vm.vm_ssize = vm->vm_ssize; 2313 ep->e_vm.vm_map.size = vm->vm_map.size; 2314 2315 /* Pick the primary (first) LWP */ 2316 l = proc_active_lwp(p); 2317 KASSERT(l != NULL); 2318 lwp_lock(l); 2319 if (l->l_wchan) 2320 strncpy(ep->e_wmesg, l->l_wmesg, WMESGLEN); 2321 lwp_unlock(l); 2322 } 2323 ep->e_ppid = p->p_ppid; 2324 if (p->p_pgrp && p->p_session) { 2325 ep->e_pgid = p->p_pgrp->pg_id; 2326 ep->e_jobc = p->p_pgrp->pg_jobc; 2327 ep->e_sid = p->p_session->s_sid; 2328 if ((p->p_lflag & PL_CONTROLT) && 2329 (tp = p->p_session->s_ttyp)) { 2330 ep->e_tdev = tp->t_dev; 2331 ep->e_tpgid = tp->t_pgrp ? tp->t_pgrp->pg_id : NO_PGID; 2332 COND_SET_VALUE(ep->e_tsess, tp->t_session, allowaddr); 2333 } else 2334 ep->e_tdev = (uint32_t)NODEV; 2335 ep->e_flag = p->p_session->s_ttyvp ? EPROC_CTTY : 0; 2336 if (SESS_LEADER(p)) 2337 ep->e_flag |= EPROC_SLEADER; 2338 strncpy(ep->e_login, p->p_session->s_login, MAXLOGNAME); 2339 } 2340 ep->e_xsize = ep->e_xrssize = 0; 2341 ep->e_xccount = ep->e_xswrss = 0; 2342 } 2343 2344 /* 2345 * Fill in a kinfo_proc2 structure for the specified process. 2346 */ 2347 void 2348 fill_kproc2(struct proc *p, struct kinfo_proc2 *ki, bool zombie, bool allowaddr) 2349 { 2350 struct tty *tp; 2351 struct lwp *l, *l2; 2352 struct timeval ut, st, rt; 2353 sigset_t ss1, ss2; 2354 struct rusage ru; 2355 struct vmspace *vm; 2356 2357 KASSERT(mutex_owned(proc_lock)); 2358 KASSERT(mutex_owned(p->p_lock)); 2359 2360 sigemptyset(&ss1); 2361 sigemptyset(&ss2); 2362 2363 COND_SET_VALUE(ki->p_paddr, PTRTOUINT64(p), allowaddr); 2364 COND_SET_VALUE(ki->p_fd, PTRTOUINT64(p->p_fd), allowaddr); 2365 COND_SET_VALUE(ki->p_cwdi, PTRTOUINT64(p->p_cwdi), allowaddr); 2366 COND_SET_VALUE(ki->p_stats, PTRTOUINT64(p->p_stats), allowaddr); 2367 COND_SET_VALUE(ki->p_limit, PTRTOUINT64(p->p_limit), allowaddr); 2368 COND_SET_VALUE(ki->p_vmspace, PTRTOUINT64(p->p_vmspace), allowaddr); 2369 COND_SET_VALUE(ki->p_sigacts, PTRTOUINT64(p->p_sigacts), allowaddr); 2370 COND_SET_VALUE(ki->p_sess, PTRTOUINT64(p->p_session), allowaddr); 2371 ki->p_tsess = 0; /* may be changed if controlling tty below */ 2372 COND_SET_VALUE(ki->p_ru, PTRTOUINT64(&p->p_stats->p_ru), allowaddr); 2373 ki->p_eflag = 0; 2374 ki->p_exitsig = p->p_exitsig; 2375 ki->p_flag = L_INMEM; /* Process never swapped out */ 2376 ki->p_flag |= sysctl_map_flags(sysctl_flagmap, p->p_flag); 2377 ki->p_flag |= sysctl_map_flags(sysctl_sflagmap, p->p_sflag); 2378 ki->p_flag |= sysctl_map_flags(sysctl_slflagmap, p->p_slflag); 2379 ki->p_flag |= sysctl_map_flags(sysctl_lflagmap, p->p_lflag); 2380 ki->p_flag |= sysctl_map_flags(sysctl_stflagmap, p->p_stflag); 2381 ki->p_pid = p->p_pid; 2382 ki->p_ppid = p->p_ppid; 2383 ki->p_uid = kauth_cred_geteuid(p->p_cred); 2384 ki->p_ruid = kauth_cred_getuid(p->p_cred); 2385 ki->p_gid = kauth_cred_getegid(p->p_cred); 2386 ki->p_rgid = kauth_cred_getgid(p->p_cred); 2387 ki->p_svuid = kauth_cred_getsvuid(p->p_cred); 2388 ki->p_svgid = kauth_cred_getsvgid(p->p_cred); 2389 ki->p_ngroups = kauth_cred_ngroups(p->p_cred); 2390 kauth_cred_getgroups(p->p_cred, ki->p_groups, 2391 uimin(ki->p_ngroups, sizeof(ki->p_groups) / sizeof(ki->p_groups[0])), 2392 UIO_SYSSPACE); 2393 2394 ki->p_uticks = p->p_uticks; 2395 ki->p_sticks = p->p_sticks; 2396 ki->p_iticks = p->p_iticks; 2397 ki->p_tpgid = NO_PGID; /* may be changed if controlling tty below */ 2398 COND_SET_VALUE(ki->p_tracep, PTRTOUINT64(p->p_tracep), allowaddr); 2399 ki->p_traceflag = p->p_traceflag; 2400 2401 memcpy(&ki->p_sigignore, &p->p_sigctx.ps_sigignore,sizeof(ki_sigset_t)); 2402 memcpy(&ki->p_sigcatch, &p->p_sigctx.ps_sigcatch, sizeof(ki_sigset_t)); 2403 2404 ki->p_cpticks = 0; 2405 ki->p_pctcpu = p->p_pctcpu; 2406 ki->p_estcpu = 0; 2407 ki->p_stat = p->p_stat; /* Will likely be overridden by LWP status */ 2408 ki->p_realstat = p->p_stat; 2409 ki->p_nice = p->p_nice; 2410 ki->p_xstat = P_WAITSTATUS(p); 2411 ki->p_acflag = p->p_acflag; 2412 2413 strncpy(ki->p_comm, p->p_comm, 2414 uimin(sizeof(ki->p_comm), sizeof(p->p_comm))); 2415 strncpy(ki->p_ename, p->p_emul->e_name, sizeof(ki->p_ename)); 2416 2417 ki->p_nlwps = p->p_nlwps; 2418 ki->p_realflag = ki->p_flag; 2419 2420 if (p->p_stat != SIDL && !P_ZOMBIE(p) && !zombie) { 2421 vm = p->p_vmspace; 2422 ki->p_vm_rssize = vm_resident_count(vm); 2423 ki->p_vm_tsize = vm->vm_tsize; 2424 ki->p_vm_dsize = vm->vm_dsize; 2425 ki->p_vm_ssize = vm->vm_ssize; 2426 ki->p_vm_vsize = atop(vm->vm_map.size); 2427 /* 2428 * Since the stack is initially mapped mostly with 2429 * PROT_NONE and grown as needed, adjust the "mapped size" 2430 * to skip the unused stack portion. 2431 */ 2432 ki->p_vm_msize = 2433 atop(vm->vm_map.size) - vm->vm_issize + vm->vm_ssize; 2434 2435 /* Pick the primary (first) LWP */ 2436 l = proc_active_lwp(p); 2437 KASSERT(l != NULL); 2438 lwp_lock(l); 2439 ki->p_nrlwps = p->p_nrlwps; 2440 ki->p_forw = 0; 2441 ki->p_back = 0; 2442 COND_SET_VALUE(ki->p_addr, PTRTOUINT64(l->l_addr), allowaddr); 2443 ki->p_stat = l->l_stat; 2444 ki->p_flag |= sysctl_map_flags(sysctl_lwpflagmap, l->l_flag); 2445 ki->p_swtime = l->l_swtime; 2446 ki->p_slptime = l->l_slptime; 2447 if (l->l_stat == LSONPROC) 2448 ki->p_schedflags = l->l_cpu->ci_schedstate.spc_flags; 2449 else 2450 ki->p_schedflags = 0; 2451 ki->p_priority = lwp_eprio(l); 2452 ki->p_usrpri = l->l_priority; 2453 if (l->l_wchan) 2454 strncpy(ki->p_wmesg, l->l_wmesg, sizeof(ki->p_wmesg)); 2455 COND_SET_VALUE(ki->p_wchan, PTRTOUINT64(l->l_wchan), allowaddr); 2456 ki->p_cpuid = cpu_index(l->l_cpu); 2457 lwp_unlock(l); 2458 LIST_FOREACH(l, &p->p_lwps, l_sibling) { 2459 /* This is hardly correct, but... */ 2460 sigplusset(&l->l_sigpend.sp_set, &ss1); 2461 sigplusset(&l->l_sigmask, &ss2); 2462 ki->p_cpticks += l->l_cpticks; 2463 ki->p_pctcpu += l->l_pctcpu; 2464 ki->p_estcpu += l->l_estcpu; 2465 } 2466 } 2467 sigplusset(&p->p_sigpend.sp_set, &ss2); 2468 memcpy(&ki->p_siglist, &ss1, sizeof(ki_sigset_t)); 2469 memcpy(&ki->p_sigmask, &ss2, sizeof(ki_sigset_t)); 2470 2471 if (p->p_session != NULL) { 2472 ki->p_sid = p->p_session->s_sid; 2473 ki->p__pgid = p->p_pgrp->pg_id; 2474 if (p->p_session->s_ttyvp) 2475 ki->p_eflag |= EPROC_CTTY; 2476 if (SESS_LEADER(p)) 2477 ki->p_eflag |= EPROC_SLEADER; 2478 strncpy(ki->p_login, p->p_session->s_login, 2479 uimin(sizeof ki->p_login - 1, sizeof p->p_session->s_login)); 2480 ki->p_jobc = p->p_pgrp->pg_jobc; 2481 if ((p->p_lflag & PL_CONTROLT) && (tp = p->p_session->s_ttyp)) { 2482 ki->p_tdev = tp->t_dev; 2483 ki->p_tpgid = tp->t_pgrp ? tp->t_pgrp->pg_id : NO_PGID; 2484 COND_SET_VALUE(ki->p_tsess, PTRTOUINT64(tp->t_session), 2485 allowaddr); 2486 } else { 2487 ki->p_tdev = (int32_t)NODEV; 2488 } 2489 } 2490 2491 if (!P_ZOMBIE(p) && !zombie) { 2492 ki->p_uvalid = 1; 2493 ki->p_ustart_sec = p->p_stats->p_start.tv_sec; 2494 ki->p_ustart_usec = p->p_stats->p_start.tv_usec; 2495 2496 calcru(p, &ut, &st, NULL, &rt); 2497 ki->p_rtime_sec = rt.tv_sec; 2498 ki->p_rtime_usec = rt.tv_usec; 2499 ki->p_uutime_sec = ut.tv_sec; 2500 ki->p_uutime_usec = ut.tv_usec; 2501 ki->p_ustime_sec = st.tv_sec; 2502 ki->p_ustime_usec = st.tv_usec; 2503 2504 memcpy(&ru, &p->p_stats->p_ru, sizeof(ru)); 2505 ki->p_uru_nvcsw = 0; 2506 ki->p_uru_nivcsw = 0; 2507 LIST_FOREACH(l2, &p->p_lwps, l_sibling) { 2508 ki->p_uru_nvcsw += (l2->l_ncsw - l2->l_nivcsw); 2509 ki->p_uru_nivcsw += l2->l_nivcsw; 2510 ruadd(&ru, &l2->l_ru); 2511 } 2512 ki->p_uru_maxrss = ru.ru_maxrss; 2513 ki->p_uru_ixrss = ru.ru_ixrss; 2514 ki->p_uru_idrss = ru.ru_idrss; 2515 ki->p_uru_isrss = ru.ru_isrss; 2516 ki->p_uru_minflt = ru.ru_minflt; 2517 ki->p_uru_majflt = ru.ru_majflt; 2518 ki->p_uru_nswap = ru.ru_nswap; 2519 ki->p_uru_inblock = ru.ru_inblock; 2520 ki->p_uru_oublock = ru.ru_oublock; 2521 ki->p_uru_msgsnd = ru.ru_msgsnd; 2522 ki->p_uru_msgrcv = ru.ru_msgrcv; 2523 ki->p_uru_nsignals = ru.ru_nsignals; 2524 2525 timeradd(&p->p_stats->p_cru.ru_utime, 2526 &p->p_stats->p_cru.ru_stime, &ut); 2527 ki->p_uctime_sec = ut.tv_sec; 2528 ki->p_uctime_usec = ut.tv_usec; 2529 } 2530 } 2531 2532 2533 int 2534 proc_find_locked(struct lwp *l, struct proc **p, pid_t pid) 2535 { 2536 int error; 2537 2538 mutex_enter(proc_lock); 2539 if (pid == -1) 2540 *p = l->l_proc; 2541 else 2542 *p = proc_find(pid); 2543 2544 if (*p == NULL) { 2545 if (pid != -1) 2546 mutex_exit(proc_lock); 2547 return ESRCH; 2548 } 2549 if (pid != -1) 2550 mutex_enter((*p)->p_lock); 2551 mutex_exit(proc_lock); 2552 2553 error = kauth_authorize_process(l->l_cred, 2554 KAUTH_PROCESS_CANSEE, *p, 2555 KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_ENTRY), NULL, NULL); 2556 if (error) { 2557 if (pid != -1) 2558 mutex_exit((*p)->p_lock); 2559 } 2560 return error; 2561 } 2562 2563 static int 2564 fill_pathname(struct lwp *l, pid_t pid, void *oldp, size_t *oldlenp) 2565 { 2566 int error; 2567 struct proc *p; 2568 2569 if ((error = proc_find_locked(l, &p, pid)) != 0) 2570 return error; 2571 2572 if (p->p_path == NULL) { 2573 if (pid != -1) 2574 mutex_exit(p->p_lock); 2575 return ENOENT; 2576 } 2577 2578 size_t len = strlen(p->p_path) + 1; 2579 if (oldp != NULL) { 2580 size_t copylen = uimin(len, *oldlenp); 2581 error = sysctl_copyout(l, p->p_path, oldp, copylen); 2582 if (error == 0 && *oldlenp < len) 2583 error = ENOSPC; 2584 } 2585 *oldlenp = len; 2586 if (pid != -1) 2587 mutex_exit(p->p_lock); 2588 return error; 2589 } 2590 2591 int 2592 proc_getauxv(struct proc *p, void **buf, size_t *len) 2593 { 2594 struct ps_strings pss; 2595 int error; 2596 void *uauxv, *kauxv; 2597 size_t size; 2598 2599 if ((error = copyin_psstrings(p, &pss)) != 0) 2600 return error; 2601 if (pss.ps_envstr == NULL) 2602 return EIO; 2603 2604 size = p->p_execsw->es_arglen; 2605 if (size == 0) 2606 return EIO; 2607 2608 size_t ptrsz = PROC_PTRSZ(p); 2609 uauxv = (void *)((char *)pss.ps_envstr + (pss.ps_nenvstr + 1) * ptrsz); 2610 2611 kauxv = kmem_alloc(size, KM_SLEEP); 2612 2613 error = copyin_proc(p, uauxv, kauxv, size); 2614 if (error) { 2615 kmem_free(kauxv, size); 2616 return error; 2617 } 2618 2619 *buf = kauxv; 2620 *len = size; 2621 2622 return 0; 2623 } 2624 2625 2626 static int 2627 sysctl_security_expose_address(SYSCTLFN_ARGS) 2628 { 2629 int expose_address, error; 2630 struct sysctlnode node; 2631 2632 node = *rnode; 2633 node.sysctl_data = &expose_address; 2634 expose_address = *(int *)rnode->sysctl_data; 2635 error = sysctl_lookup(SYSCTLFN_CALL(&node)); 2636 if (error || newp == NULL) 2637 return error; 2638 2639 if (kauth_authorize_system(l->l_cred, KAUTH_SYSTEM_KERNADDR, 2640 0, NULL, NULL, NULL)) 2641 return EPERM; 2642 2643 switch (expose_address) { 2644 case 0: 2645 case 1: 2646 case 2: 2647 break; 2648 default: 2649 return EINVAL; 2650 } 2651 2652 *(int *)rnode->sysctl_data = expose_address; 2653 2654 return 0; 2655 } 2656 2657 bool 2658 get_expose_address(struct proc *p) 2659 { 2660 /* allow only if sysctl variable is set or privileged */ 2661 return kauth_authorize_process(kauth_cred_get(), KAUTH_PROCESS_CANSEE, 2662 p, KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_KPTR), NULL, NULL) == 0; 2663 } 2664
Indexes created Sun Sep 21 16:09:51 GMT 2025