kern_proc.c revision 1.235
1 /* $NetBSD: kern_proc.c,v 1.235 2019/09/30 21:13:33 kamil 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 1.235 2019/09/30 21:13:33 kamil 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 static int fill_cwd(struct lwp *, pid_t, void *, size_t *); 250 251 static int 252 proc_listener_cb(kauth_cred_t cred, kauth_action_t action, void *cookie, 253 void *arg0, void *arg1, void *arg2, void *arg3) 254 { 255 struct proc *p; 256 int result; 257 258 result = KAUTH_RESULT_DEFER; 259 p = arg0; 260 261 switch (action) { 262 case KAUTH_PROCESS_CANSEE: { 263 enum kauth_process_req req; 264 265 req = (enum kauth_process_req)arg1; 266 267 switch (req) { 268 case KAUTH_REQ_PROCESS_CANSEE_ARGS: 269 case KAUTH_REQ_PROCESS_CANSEE_ENTRY: 270 case KAUTH_REQ_PROCESS_CANSEE_OPENFILES: 271 case KAUTH_REQ_PROCESS_CANSEE_EPROC: 272 result = KAUTH_RESULT_ALLOW; 273 break; 274 275 case KAUTH_REQ_PROCESS_CANSEE_ENV: 276 if (kauth_cred_getuid(cred) != 277 kauth_cred_getuid(p->p_cred) || 278 kauth_cred_getuid(cred) != 279 kauth_cred_getsvuid(p->p_cred)) 280 break; 281 282 result = KAUTH_RESULT_ALLOW; 283 284 break; 285 286 case KAUTH_REQ_PROCESS_CANSEE_KPTR: 287 if (!kern_expose_address) 288 break; 289 290 if (kern_expose_address == 1 && !(p->p_flag & PK_KMEM)) 291 break; 292 293 result = KAUTH_RESULT_ALLOW; 294 295 break; 296 297 default: 298 break; 299 } 300 301 break; 302 } 303 304 case KAUTH_PROCESS_FORK: { 305 int lnprocs = (int)(unsigned long)arg2; 306 307 /* 308 * Don't allow a nonprivileged user to use the last few 309 * processes. The variable lnprocs is the current number of 310 * processes, maxproc is the limit. 311 */ 312 if (__predict_false((lnprocs >= maxproc - 5))) 313 break; 314 315 result = KAUTH_RESULT_ALLOW; 316 317 break; 318 } 319 320 case KAUTH_PROCESS_CORENAME: 321 case KAUTH_PROCESS_STOPFLAG: 322 if (proc_uidmatch(cred, p->p_cred) == 0) 323 result = KAUTH_RESULT_ALLOW; 324 325 break; 326 327 default: 328 break; 329 } 330 331 return result; 332 } 333 334 static int 335 proc_ctor(void *arg __unused, void *obj, int flags __unused) 336 { 337 memset(obj, 0, sizeof(struct proc)); 338 return 0; 339 } 340 341 /* 342 * Initialize global process hashing structures. 343 */ 344 void 345 procinit(void) 346 { 347 const struct proclist_desc *pd; 348 u_int i; 349 #define LINK_EMPTY ((PID_MAX + INITIAL_PID_TABLE_SIZE) & ~(INITIAL_PID_TABLE_SIZE - 1)) 350 351 for (pd = proclists; pd->pd_list != NULL; pd++) 352 LIST_INIT(pd->pd_list); 353 354 proc_lock = mutex_obj_alloc(MUTEX_DEFAULT, IPL_NONE); 355 pid_table = kmem_alloc(INITIAL_PID_TABLE_SIZE 356 * sizeof(struct pid_table), KM_SLEEP); 357 pid_tbl_mask = INITIAL_PID_TABLE_SIZE - 1; 358 pid_max = PID_MAX; 359 360 /* Set free list running through table... 361 Preset 'use count' above PID_MAX so we allocate pid 1 next. */ 362 for (i = 0; i <= pid_tbl_mask; i++) { 363 pid_table[i].pt_proc = P_FREE(LINK_EMPTY + i + 1); 364 pid_table[i].pt_pgrp = 0; 365 pid_table[i].pt_pid = 0; 366 } 367 /* slot 0 is just grabbed */ 368 next_free_pt = 1; 369 /* Need to fix last entry. */ 370 last_free_pt = pid_tbl_mask; 371 pid_table[last_free_pt].pt_proc = P_FREE(LINK_EMPTY); 372 /* point at which we grow table - to avoid reusing pids too often */ 373 pid_alloc_lim = pid_tbl_mask - 1; 374 #undef LINK_EMPTY 375 376 proc_specificdata_domain = specificdata_domain_create(); 377 KASSERT(proc_specificdata_domain != NULL); 378 379 proc_cache = pool_cache_init(sizeof(struct proc), 0, 0, 0, 380 "procpl", NULL, IPL_NONE, proc_ctor, NULL, NULL); 381 382 proc_listener = kauth_listen_scope(KAUTH_SCOPE_PROCESS, 383 proc_listener_cb, NULL); 384 } 385 386 void 387 procinit_sysctl(void) 388 { 389 static struct sysctllog *clog; 390 391 sysctl_createv(&clog, 0, NULL, NULL, 392 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 393 CTLTYPE_INT, "expose_address", 394 SYSCTL_DESCR("Enable exposing kernel addresses"), 395 sysctl_security_expose_address, 0, 396 &kern_expose_address, 0, CTL_KERN, CTL_CREATE, CTL_EOL); 397 sysctl_createv(&clog, 0, NULL, NULL, 398 CTLFLAG_PERMANENT, 399 CTLTYPE_NODE, "proc", 400 SYSCTL_DESCR("System-wide process information"), 401 sysctl_doeproc, 0, NULL, 0, 402 CTL_KERN, KERN_PROC, CTL_EOL); 403 sysctl_createv(&clog, 0, NULL, NULL, 404 CTLFLAG_PERMANENT, 405 CTLTYPE_NODE, "proc2", 406 SYSCTL_DESCR("Machine-independent process information"), 407 sysctl_doeproc, 0, NULL, 0, 408 CTL_KERN, KERN_PROC2, CTL_EOL); 409 sysctl_createv(&clog, 0, NULL, NULL, 410 CTLFLAG_PERMANENT, 411 CTLTYPE_NODE, "proc_args", 412 SYSCTL_DESCR("Process argument information"), 413 sysctl_kern_proc_args, 0, NULL, 0, 414 CTL_KERN, KERN_PROC_ARGS, CTL_EOL); 415 416 /* 417 "nodes" under these: 418 419 KERN_PROC_ALL 420 KERN_PROC_PID pid 421 KERN_PROC_PGRP pgrp 422 KERN_PROC_SESSION sess 423 KERN_PROC_TTY tty 424 KERN_PROC_UID uid 425 KERN_PROC_RUID uid 426 KERN_PROC_GID gid 427 KERN_PROC_RGID gid 428 429 all in all, probably not worth the effort... 430 */ 431 } 432 433 /* 434 * Initialize process 0. 435 */ 436 void 437 proc0_init(void) 438 { 439 struct proc *p; 440 struct pgrp *pg; 441 struct rlimit *rlim; 442 rlim_t lim; 443 int i; 444 445 p = &proc0; 446 pg = &pgrp0; 447 448 mutex_init(&p->p_stmutex, MUTEX_DEFAULT, IPL_HIGH); 449 mutex_init(&p->p_auxlock, MUTEX_DEFAULT, IPL_NONE); 450 p->p_lock = mutex_obj_alloc(MUTEX_DEFAULT, IPL_NONE); 451 452 rw_init(&p->p_reflock); 453 cv_init(&p->p_waitcv, "wait"); 454 cv_init(&p->p_lwpcv, "lwpwait"); 455 456 LIST_INSERT_HEAD(&p->p_lwps, &lwp0, l_sibling); 457 458 pid_table[0].pt_proc = p; 459 LIST_INSERT_HEAD(&allproc, p, p_list); 460 461 pid_table[0].pt_pgrp = pg; 462 LIST_INSERT_HEAD(&pg->pg_members, p, p_pglist); 463 464 #ifdef __HAVE_SYSCALL_INTERN 465 (*p->p_emul->e_syscall_intern)(p); 466 #endif 467 468 /* Create credentials. */ 469 cred0 = kauth_cred_alloc(); 470 p->p_cred = cred0; 471 472 /* Create the CWD info. */ 473 rw_init(&cwdi0.cwdi_lock); 474 475 /* Create the limits structures. */ 476 mutex_init(&limit0.pl_lock, MUTEX_DEFAULT, IPL_NONE); 477 478 rlim = limit0.pl_rlimit; 479 for (i = 0; i < __arraycount(limit0.pl_rlimit); i++) { 480 rlim[i].rlim_cur = RLIM_INFINITY; 481 rlim[i].rlim_max = RLIM_INFINITY; 482 } 483 484 rlim[RLIMIT_NOFILE].rlim_max = maxfiles; 485 rlim[RLIMIT_NOFILE].rlim_cur = maxfiles < nofile ? maxfiles : nofile; 486 487 rlim[RLIMIT_NPROC].rlim_max = maxproc; 488 rlim[RLIMIT_NPROC].rlim_cur = maxproc < maxuprc ? maxproc : maxuprc; 489 490 lim = MIN(VM_MAXUSER_ADDRESS, ctob((rlim_t)uvmexp.free)); 491 rlim[RLIMIT_RSS].rlim_max = lim; 492 rlim[RLIMIT_MEMLOCK].rlim_max = lim; 493 rlim[RLIMIT_MEMLOCK].rlim_cur = lim / 3; 494 495 rlim[RLIMIT_NTHR].rlim_max = maxlwp; 496 rlim[RLIMIT_NTHR].rlim_cur = maxlwp < maxuprc ? maxlwp : maxuprc; 497 498 /* Note that default core name has zero length. */ 499 limit0.pl_corename = defcorename; 500 limit0.pl_cnlen = 0; 501 limit0.pl_refcnt = 1; 502 limit0.pl_writeable = false; 503 limit0.pl_sv_limit = NULL; 504 505 /* Configure virtual memory system, set vm rlimits. */ 506 uvm_init_limits(p); 507 508 /* Initialize file descriptor table for proc0. */ 509 fd_init(&filedesc0); 510 511 /* 512 * Initialize proc0's vmspace, which uses the kernel pmap. 513 * All kernel processes (which never have user space mappings) 514 * share proc0's vmspace, and thus, the kernel pmap. 515 */ 516 uvmspace_init(&vmspace0, pmap_kernel(), round_page(VM_MIN_ADDRESS), 517 trunc_page(VM_MAXUSER_ADDRESS), 518 #ifdef __USE_TOPDOWN_VM 519 true 520 #else 521 false 522 #endif 523 ); 524 525 /* Initialize signal state for proc0. XXX IPL_SCHED */ 526 mutex_init(&p->p_sigacts->sa_mutex, MUTEX_DEFAULT, IPL_SCHED); 527 siginit(p); 528 529 proc_initspecific(p); 530 kdtrace_proc_ctor(NULL, p); 531 } 532 533 /* 534 * Session reference counting. 535 */ 536 537 void 538 proc_sesshold(struct session *ss) 539 { 540 541 KASSERT(mutex_owned(proc_lock)); 542 ss->s_count++; 543 } 544 545 void 546 proc_sessrele(struct session *ss) 547 { 548 549 KASSERT(mutex_owned(proc_lock)); 550 /* 551 * We keep the pgrp with the same id as the session in order to 552 * stop a process being given the same pid. Since the pgrp holds 553 * a reference to the session, it must be a 'zombie' pgrp by now. 554 */ 555 if (--ss->s_count == 0) { 556 struct pgrp *pg; 557 558 pg = pg_remove(ss->s_sid); 559 mutex_exit(proc_lock); 560 561 kmem_free(pg, sizeof(struct pgrp)); 562 kmem_free(ss, sizeof(struct session)); 563 } else { 564 mutex_exit(proc_lock); 565 } 566 } 567 568 /* 569 * Check that the specified process group is in the session of the 570 * specified process. 571 * Treats -ve ids as process ids. 572 * Used to validate TIOCSPGRP requests. 573 */ 574 int 575 pgid_in_session(struct proc *p, pid_t pg_id) 576 { 577 struct pgrp *pgrp; 578 struct session *session; 579 int error; 580 581 mutex_enter(proc_lock); 582 if (pg_id < 0) { 583 struct proc *p1 = proc_find(-pg_id); 584 if (p1 == NULL) { 585 error = EINVAL; 586 goto fail; 587 } 588 pgrp = p1->p_pgrp; 589 } else { 590 pgrp = pgrp_find(pg_id); 591 if (pgrp == NULL) { 592 error = EINVAL; 593 goto fail; 594 } 595 } 596 session = pgrp->pg_session; 597 error = (session != p->p_pgrp->pg_session) ? EPERM : 0; 598 fail: 599 mutex_exit(proc_lock); 600 return error; 601 } 602 603 /* 604 * p_inferior: is p an inferior of q? 605 */ 606 static inline bool 607 p_inferior(struct proc *p, struct proc *q) 608 { 609 610 KASSERT(mutex_owned(proc_lock)); 611 612 for (; p != q; p = p->p_pptr) 613 if (p->p_pid == 0) 614 return false; 615 return true; 616 } 617 618 /* 619 * proc_find: locate a process by the ID. 620 * 621 * => Must be called with proc_lock held. 622 */ 623 proc_t * 624 proc_find_raw(pid_t pid) 625 { 626 struct pid_table *pt; 627 proc_t *p; 628 629 KASSERT(mutex_owned(proc_lock)); 630 pt = &pid_table[pid & pid_tbl_mask]; 631 p = pt->pt_proc; 632 if (__predict_false(!P_VALID(p) || pt->pt_pid != pid)) { 633 return NULL; 634 } 635 return p; 636 } 637 638 proc_t * 639 proc_find(pid_t pid) 640 { 641 proc_t *p; 642 643 p = proc_find_raw(pid); 644 if (__predict_false(p == NULL)) { 645 return NULL; 646 } 647 648 /* 649 * Only allow live processes to be found by PID. 650 * XXX: p_stat might change, since unlocked. 651 */ 652 if (__predict_true(p->p_stat == SACTIVE || p->p_stat == SSTOP)) { 653 return p; 654 } 655 return NULL; 656 } 657 658 /* 659 * pgrp_find: locate a process group by the ID. 660 * 661 * => Must be called with proc_lock held. 662 */ 663 struct pgrp * 664 pgrp_find(pid_t pgid) 665 { 666 struct pgrp *pg; 667 668 KASSERT(mutex_owned(proc_lock)); 669 670 pg = pid_table[pgid & pid_tbl_mask].pt_pgrp; 671 672 /* 673 * Cannot look up a process group that only exists because the 674 * session has not died yet (traditional). 675 */ 676 if (pg == NULL || pg->pg_id != pgid || LIST_EMPTY(&pg->pg_members)) { 677 return NULL; 678 } 679 return pg; 680 } 681 682 static void 683 expand_pid_table(void) 684 { 685 size_t pt_size, tsz; 686 struct pid_table *n_pt, *new_pt; 687 struct proc *proc; 688 struct pgrp *pgrp; 689 pid_t pid, rpid; 690 u_int i; 691 uint new_pt_mask; 692 693 pt_size = pid_tbl_mask + 1; 694 tsz = pt_size * 2 * sizeof(struct pid_table); 695 new_pt = kmem_alloc(tsz, KM_SLEEP); 696 new_pt_mask = pt_size * 2 - 1; 697 698 mutex_enter(proc_lock); 699 if (pt_size != pid_tbl_mask + 1) { 700 /* Another process beat us to it... */ 701 mutex_exit(proc_lock); 702 kmem_free(new_pt, tsz); 703 return; 704 } 705 706 /* 707 * Copy entries from old table into new one. 708 * If 'pid' is 'odd' we need to place in the upper half, 709 * even pid's to the lower half. 710 * Free items stay in the low half so we don't have to 711 * fixup the reference to them. 712 * We stuff free items on the front of the freelist 713 * because we can't write to unmodified entries. 714 * Processing the table backwards maintains a semblance 715 * of issuing pid numbers that increase with time. 716 */ 717 i = pt_size - 1; 718 n_pt = new_pt + i; 719 for (; ; i--, n_pt--) { 720 proc = pid_table[i].pt_proc; 721 pgrp = pid_table[i].pt_pgrp; 722 if (!P_VALID(proc)) { 723 /* Up 'use count' so that link is valid */ 724 pid = (P_NEXT(proc) + pt_size) & ~pt_size; 725 rpid = 0; 726 proc = P_FREE(pid); 727 if (pgrp) 728 pid = pgrp->pg_id; 729 } else { 730 pid = pid_table[i].pt_pid; 731 rpid = pid; 732 } 733 734 /* Save entry in appropriate half of table */ 735 n_pt[pid & pt_size].pt_proc = proc; 736 n_pt[pid & pt_size].pt_pgrp = pgrp; 737 n_pt[pid & pt_size].pt_pid = rpid; 738 739 /* Put other piece on start of free list */ 740 pid = (pid ^ pt_size) & ~pid_tbl_mask; 741 n_pt[pid & pt_size].pt_proc = 742 P_FREE((pid & ~pt_size) | next_free_pt); 743 n_pt[pid & pt_size].pt_pgrp = 0; 744 n_pt[pid & pt_size].pt_pid = 0; 745 746 next_free_pt = i | (pid & pt_size); 747 if (i == 0) 748 break; 749 } 750 751 /* Save old table size and switch tables */ 752 tsz = pt_size * sizeof(struct pid_table); 753 n_pt = pid_table; 754 pid_table = new_pt; 755 pid_tbl_mask = new_pt_mask; 756 757 /* 758 * pid_max starts as PID_MAX (= 30000), once we have 16384 759 * allocated pids we need it to be larger! 760 */ 761 if (pid_tbl_mask > PID_MAX) { 762 pid_max = pid_tbl_mask * 2 + 1; 763 pid_alloc_lim |= pid_alloc_lim << 1; 764 } else 765 pid_alloc_lim <<= 1; /* doubles number of free slots... */ 766 767 mutex_exit(proc_lock); 768 kmem_free(n_pt, tsz); 769 } 770 771 struct proc * 772 proc_alloc(void) 773 { 774 struct proc *p; 775 776 p = pool_cache_get(proc_cache, PR_WAITOK); 777 p->p_stat = SIDL; /* protect against others */ 778 proc_initspecific(p); 779 kdtrace_proc_ctor(NULL, p); 780 p->p_pid = -1; 781 proc_alloc_pid(p); 782 return p; 783 } 784 785 /* 786 * proc_alloc_pid: allocate PID and record the given proc 'p' so that 787 * proc_find_raw() can find it by the PID. 788 */ 789 790 pid_t 791 proc_alloc_pid(struct proc *p) 792 { 793 struct pid_table *pt; 794 pid_t pid; 795 int nxt; 796 797 for (;;expand_pid_table()) { 798 if (__predict_false(pid_alloc_cnt >= pid_alloc_lim)) 799 /* ensure pids cycle through 2000+ values */ 800 continue; 801 mutex_enter(proc_lock); 802 pt = &pid_table[next_free_pt]; 803 #ifdef DIAGNOSTIC 804 if (__predict_false(P_VALID(pt->pt_proc) || pt->pt_pgrp)) 805 panic("proc_alloc: slot busy"); 806 #endif 807 nxt = P_NEXT(pt->pt_proc); 808 if (nxt & pid_tbl_mask) 809 break; 810 /* Table full - expand (NB last entry not used....) */ 811 mutex_exit(proc_lock); 812 } 813 814 /* pid is 'saved use count' + 'size' + entry */ 815 pid = (nxt & ~pid_tbl_mask) + pid_tbl_mask + 1 + next_free_pt; 816 if ((uint)pid > (uint)pid_max) 817 pid &= pid_tbl_mask; 818 next_free_pt = nxt & pid_tbl_mask; 819 820 /* Grab table slot */ 821 pt->pt_proc = p; 822 823 KASSERT(pt->pt_pid == 0); 824 pt->pt_pid = pid; 825 if (p->p_pid == -1) { 826 p->p_pid = pid; 827 } 828 pid_alloc_cnt++; 829 mutex_exit(proc_lock); 830 831 return pid; 832 } 833 834 /* 835 * Free a process id - called from proc_free (in kern_exit.c) 836 * 837 * Called with the proc_lock held. 838 */ 839 void 840 proc_free_pid(pid_t pid) 841 { 842 struct pid_table *pt; 843 844 KASSERT(mutex_owned(proc_lock)); 845 846 pt = &pid_table[pid & pid_tbl_mask]; 847 848 /* save pid use count in slot */ 849 pt->pt_proc = P_FREE(pid & ~pid_tbl_mask); 850 KASSERT(pt->pt_pid == pid); 851 pt->pt_pid = 0; 852 853 if (pt->pt_pgrp == NULL) { 854 /* link last freed entry onto ours */ 855 pid &= pid_tbl_mask; 856 pt = &pid_table[last_free_pt]; 857 pt->pt_proc = P_FREE(P_NEXT(pt->pt_proc) | pid); 858 pt->pt_pid = 0; 859 last_free_pt = pid; 860 pid_alloc_cnt--; 861 } 862 863 atomic_dec_uint(&nprocs); 864 } 865 866 void 867 proc_free_mem(struct proc *p) 868 { 869 870 kdtrace_proc_dtor(NULL, p); 871 pool_cache_put(proc_cache, p); 872 } 873 874 /* 875 * proc_enterpgrp: move p to a new or existing process group (and session). 876 * 877 * If we are creating a new pgrp, the pgid should equal 878 * the calling process' pid. 879 * If is only valid to enter a process group that is in the session 880 * of the process. 881 * Also mksess should only be set if we are creating a process group 882 * 883 * Only called from sys_setsid, sys_setpgid and posix_spawn/spawn_return. 884 */ 885 int 886 proc_enterpgrp(struct proc *curp, pid_t pid, pid_t pgid, bool mksess) 887 { 888 struct pgrp *new_pgrp, *pgrp; 889 struct session *sess; 890 struct proc *p; 891 int rval; 892 pid_t pg_id = NO_PGID; 893 894 sess = mksess ? kmem_alloc(sizeof(*sess), KM_SLEEP) : NULL; 895 896 /* Allocate data areas we might need before doing any validity checks */ 897 mutex_enter(proc_lock); /* Because pid_table might change */ 898 if (pid_table[pgid & pid_tbl_mask].pt_pgrp == 0) { 899 mutex_exit(proc_lock); 900 new_pgrp = kmem_alloc(sizeof(*new_pgrp), KM_SLEEP); 901 mutex_enter(proc_lock); 902 } else 903 new_pgrp = NULL; 904 rval = EPERM; /* most common error (to save typing) */ 905 906 /* Check pgrp exists or can be created */ 907 pgrp = pid_table[pgid & pid_tbl_mask].pt_pgrp; 908 if (pgrp != NULL && pgrp->pg_id != pgid) 909 goto done; 910 911 /* Can only set another process under restricted circumstances. */ 912 if (pid != curp->p_pid) { 913 /* Must exist and be one of our children... */ 914 p = proc_find(pid); 915 if (p == NULL || !p_inferior(p, curp)) { 916 rval = ESRCH; 917 goto done; 918 } 919 /* ... in the same session... */ 920 if (sess != NULL || p->p_session != curp->p_session) 921 goto done; 922 /* ... existing pgid must be in same session ... */ 923 if (pgrp != NULL && pgrp->pg_session != p->p_session) 924 goto done; 925 /* ... and not done an exec. */ 926 if (p->p_flag & PK_EXEC) { 927 rval = EACCES; 928 goto done; 929 } 930 } else { 931 /* ... setsid() cannot re-enter a pgrp */ 932 if (mksess && (curp->p_pgid == curp->p_pid || 933 pgrp_find(curp->p_pid))) 934 goto done; 935 p = curp; 936 } 937 938 /* Changing the process group/session of a session 939 leader is definitely off limits. */ 940 if (SESS_LEADER(p)) { 941 if (sess == NULL && p->p_pgrp == pgrp) 942 /* unless it's a definite noop */ 943 rval = 0; 944 goto done; 945 } 946 947 /* Can only create a process group with id of process */ 948 if (pgrp == NULL && pgid != pid) 949 goto done; 950 951 /* Can only create a session if creating pgrp */ 952 if (sess != NULL && pgrp != NULL) 953 goto done; 954 955 /* Check we allocated memory for a pgrp... */ 956 if (pgrp == NULL && new_pgrp == NULL) 957 goto done; 958 959 /* Don't attach to 'zombie' pgrp */ 960 if (pgrp != NULL && LIST_EMPTY(&pgrp->pg_members)) 961 goto done; 962 963 /* Expect to succeed now */ 964 rval = 0; 965 966 if (pgrp == p->p_pgrp) 967 /* nothing to do */ 968 goto done; 969 970 /* Ok all setup, link up required structures */ 971 972 if (pgrp == NULL) { 973 pgrp = new_pgrp; 974 new_pgrp = NULL; 975 if (sess != NULL) { 976 sess->s_sid = p->p_pid; 977 sess->s_leader = p; 978 sess->s_count = 1; 979 sess->s_ttyvp = NULL; 980 sess->s_ttyp = NULL; 981 sess->s_flags = p->p_session->s_flags & ~S_LOGIN_SET; 982 memcpy(sess->s_login, p->p_session->s_login, 983 sizeof(sess->s_login)); 984 p->p_lflag &= ~PL_CONTROLT; 985 } else { 986 sess = p->p_pgrp->pg_session; 987 proc_sesshold(sess); 988 } 989 pgrp->pg_session = sess; 990 sess = NULL; 991 992 pgrp->pg_id = pgid; 993 LIST_INIT(&pgrp->pg_members); 994 #ifdef DIAGNOSTIC 995 if (__predict_false(pid_table[pgid & pid_tbl_mask].pt_pgrp)) 996 panic("enterpgrp: pgrp table slot in use"); 997 if (__predict_false(mksess && p != curp)) 998 panic("enterpgrp: mksession and p != curproc"); 999 #endif 1000 pid_table[pgid & pid_tbl_mask].pt_pgrp = pgrp; 1001 pgrp->pg_jobc = 0; 1002 } 1003 1004 /* 1005 * Adjust eligibility of affected pgrps to participate in job control. 1006 * Increment eligibility counts before decrementing, otherwise we 1007 * could reach 0 spuriously during the first call. 1008 */ 1009 fixjobc(p, pgrp, 1); 1010 fixjobc(p, p->p_pgrp, 0); 1011 1012 /* Interlock with ttread(). */ 1013 mutex_spin_enter(&tty_lock); 1014 1015 /* Move process to requested group. */ 1016 LIST_REMOVE(p, p_pglist); 1017 if (LIST_EMPTY(&p->p_pgrp->pg_members)) 1018 /* defer delete until we've dumped the lock */ 1019 pg_id = p->p_pgrp->pg_id; 1020 p->p_pgrp = pgrp; 1021 LIST_INSERT_HEAD(&pgrp->pg_members, p, p_pglist); 1022 1023 /* Done with the swap; we can release the tty mutex. */ 1024 mutex_spin_exit(&tty_lock); 1025 1026 done: 1027 if (pg_id != NO_PGID) { 1028 /* Releases proc_lock. */ 1029 pg_delete(pg_id); 1030 } else { 1031 mutex_exit(proc_lock); 1032 } 1033 if (sess != NULL) 1034 kmem_free(sess, sizeof(*sess)); 1035 if (new_pgrp != NULL) 1036 kmem_free(new_pgrp, sizeof(*new_pgrp)); 1037 #ifdef DEBUG_PGRP 1038 if (__predict_false(rval)) 1039 printf("enterpgrp(%d,%d,%d), curproc %d, rval %d\n", 1040 pid, pgid, mksess, curp->p_pid, rval); 1041 #endif 1042 return rval; 1043 } 1044 1045 /* 1046 * proc_leavepgrp: remove a process from its process group. 1047 * => must be called with the proc_lock held, which will be released; 1048 */ 1049 void 1050 proc_leavepgrp(struct proc *p) 1051 { 1052 struct pgrp *pgrp; 1053 1054 KASSERT(mutex_owned(proc_lock)); 1055 1056 /* Interlock with ttread() */ 1057 mutex_spin_enter(&tty_lock); 1058 pgrp = p->p_pgrp; 1059 LIST_REMOVE(p, p_pglist); 1060 p->p_pgrp = NULL; 1061 mutex_spin_exit(&tty_lock); 1062 1063 if (LIST_EMPTY(&pgrp->pg_members)) { 1064 /* Releases proc_lock. */ 1065 pg_delete(pgrp->pg_id); 1066 } else { 1067 mutex_exit(proc_lock); 1068 } 1069 } 1070 1071 /* 1072 * pg_remove: remove a process group from the table. 1073 * => must be called with the proc_lock held; 1074 * => returns process group to free; 1075 */ 1076 static struct pgrp * 1077 pg_remove(pid_t pg_id) 1078 { 1079 struct pgrp *pgrp; 1080 struct pid_table *pt; 1081 1082 KASSERT(mutex_owned(proc_lock)); 1083 1084 pt = &pid_table[pg_id & pid_tbl_mask]; 1085 pgrp = pt->pt_pgrp; 1086 1087 KASSERT(pgrp != NULL); 1088 KASSERT(pgrp->pg_id == pg_id); 1089 KASSERT(LIST_EMPTY(&pgrp->pg_members)); 1090 1091 pt->pt_pgrp = NULL; 1092 1093 if (!P_VALID(pt->pt_proc)) { 1094 /* Orphaned pgrp, put slot onto free list. */ 1095 KASSERT((P_NEXT(pt->pt_proc) & pid_tbl_mask) == 0); 1096 pg_id &= pid_tbl_mask; 1097 pt = &pid_table[last_free_pt]; 1098 pt->pt_proc = P_FREE(P_NEXT(pt->pt_proc) | pg_id); 1099 KASSERT(pt->pt_pid == 0); 1100 last_free_pt = pg_id; 1101 pid_alloc_cnt--; 1102 } 1103 return pgrp; 1104 } 1105 1106 /* 1107 * pg_delete: delete and free a process group. 1108 * => must be called with the proc_lock held, which will be released. 1109 */ 1110 static void 1111 pg_delete(pid_t pg_id) 1112 { 1113 struct pgrp *pg; 1114 struct tty *ttyp; 1115 struct session *ss; 1116 1117 KASSERT(mutex_owned(proc_lock)); 1118 1119 pg = pid_table[pg_id & pid_tbl_mask].pt_pgrp; 1120 if (pg == NULL || pg->pg_id != pg_id || !LIST_EMPTY(&pg->pg_members)) { 1121 mutex_exit(proc_lock); 1122 return; 1123 } 1124 1125 ss = pg->pg_session; 1126 1127 /* Remove reference (if any) from tty to this process group */ 1128 mutex_spin_enter(&tty_lock); 1129 ttyp = ss->s_ttyp; 1130 if (ttyp != NULL && ttyp->t_pgrp == pg) { 1131 ttyp->t_pgrp = NULL; 1132 KASSERT(ttyp->t_session == ss); 1133 } 1134 mutex_spin_exit(&tty_lock); 1135 1136 /* 1137 * The leading process group in a session is freed by proc_sessrele(), 1138 * if last reference. Note: proc_sessrele() releases proc_lock. 1139 */ 1140 pg = (ss->s_sid != pg->pg_id) ? pg_remove(pg_id) : NULL; 1141 proc_sessrele(ss); 1142 1143 if (pg != NULL) { 1144 /* Free it, if was not done by proc_sessrele(). */ 1145 kmem_free(pg, sizeof(struct pgrp)); 1146 } 1147 } 1148 1149 /* 1150 * Adjust pgrp jobc counters when specified process changes process group. 1151 * We count the number of processes in each process group that "qualify" 1152 * the group for terminal job control (those with a parent in a different 1153 * process group of the same session). If that count reaches zero, the 1154 * process group becomes orphaned. Check both the specified process' 1155 * process group and that of its children. 1156 * entering == 0 => p is leaving specified group. 1157 * entering == 1 => p is entering specified group. 1158 * 1159 * Call with proc_lock held. 1160 */ 1161 void 1162 fixjobc(struct proc *p, struct pgrp *pgrp, int entering) 1163 { 1164 struct pgrp *hispgrp; 1165 struct session *mysession = pgrp->pg_session; 1166 struct proc *child; 1167 1168 KASSERT(mutex_owned(proc_lock)); 1169 1170 /* 1171 * Check p's parent to see whether p qualifies its own process 1172 * group; if so, adjust count for p's process group. 1173 */ 1174 hispgrp = p->p_pptr->p_pgrp; 1175 if (hispgrp != pgrp && hispgrp->pg_session == mysession) { 1176 if (entering) { 1177 pgrp->pg_jobc++; 1178 p->p_lflag &= ~PL_ORPHANPG; 1179 } else if (--pgrp->pg_jobc == 0) 1180 orphanpg(pgrp); 1181 } 1182 1183 /* 1184 * Check this process' children to see whether they qualify 1185 * their process groups; if so, adjust counts for children's 1186 * process groups. 1187 */ 1188 LIST_FOREACH(child, &p->p_children, p_sibling) { 1189 hispgrp = child->p_pgrp; 1190 if (hispgrp != pgrp && hispgrp->pg_session == mysession && 1191 !P_ZOMBIE(child)) { 1192 if (entering) { 1193 child->p_lflag &= ~PL_ORPHANPG; 1194 hispgrp->pg_jobc++; 1195 } else if (--hispgrp->pg_jobc == 0) 1196 orphanpg(hispgrp); 1197 } 1198 } 1199 } 1200 1201 /* 1202 * A process group has become orphaned; 1203 * if there are any stopped processes in the group, 1204 * hang-up all process in that group. 1205 * 1206 * Call with proc_lock held. 1207 */ 1208 static void 1209 orphanpg(struct pgrp *pg) 1210 { 1211 struct proc *p; 1212 1213 KASSERT(mutex_owned(proc_lock)); 1214 1215 LIST_FOREACH(p, &pg->pg_members, p_pglist) { 1216 if (p->p_stat == SSTOP) { 1217 p->p_lflag |= PL_ORPHANPG; 1218 psignal(p, SIGHUP); 1219 psignal(p, SIGCONT); 1220 } 1221 } 1222 } 1223 1224 #ifdef DDB 1225 #include <ddb/db_output.h> 1226 void pidtbl_dump(void); 1227 void 1228 pidtbl_dump(void) 1229 { 1230 struct pid_table *pt; 1231 struct proc *p; 1232 struct pgrp *pgrp; 1233 int id; 1234 1235 db_printf("pid table %p size %x, next %x, last %x\n", 1236 pid_table, pid_tbl_mask+1, 1237 next_free_pt, last_free_pt); 1238 for (pt = pid_table, id = 0; id <= pid_tbl_mask; id++, pt++) { 1239 p = pt->pt_proc; 1240 if (!P_VALID(p) && !pt->pt_pgrp) 1241 continue; 1242 db_printf(" id %x: ", id); 1243 if (P_VALID(p)) 1244 db_printf("slotpid %d proc %p id %d (0x%x) %s\n", 1245 pt->pt_pid, p, p->p_pid, p->p_pid, p->p_comm); 1246 else 1247 db_printf("next %x use %x\n", 1248 P_NEXT(p) & pid_tbl_mask, 1249 P_NEXT(p) & ~pid_tbl_mask); 1250 if ((pgrp = pt->pt_pgrp)) { 1251 db_printf("\tsession %p, sid %d, count %d, login %s\n", 1252 pgrp->pg_session, pgrp->pg_session->s_sid, 1253 pgrp->pg_session->s_count, 1254 pgrp->pg_session->s_login); 1255 db_printf("\tpgrp %p, pg_id %d, pg_jobc %d, members %p\n", 1256 pgrp, pgrp->pg_id, pgrp->pg_jobc, 1257 LIST_FIRST(&pgrp->pg_members)); 1258 LIST_FOREACH(p, &pgrp->pg_members, p_pglist) { 1259 db_printf("\t\tpid %d addr %p pgrp %p %s\n", 1260 p->p_pid, p, p->p_pgrp, p->p_comm); 1261 } 1262 } 1263 } 1264 } 1265 #endif /* DDB */ 1266 1267 #ifdef KSTACK_CHECK_MAGIC 1268 1269 #define KSTACK_MAGIC 0xdeadbeaf 1270 1271 /* XXX should be per process basis? */ 1272 static int kstackleftmin = KSTACK_SIZE; 1273 static int kstackleftthres = KSTACK_SIZE / 8; 1274 1275 void 1276 kstack_setup_magic(const struct lwp *l) 1277 { 1278 uint32_t *ip; 1279 uint32_t const *end; 1280 1281 KASSERT(l != NULL); 1282 KASSERT(l != &lwp0); 1283 1284 /* 1285 * fill all the stack with magic number 1286 * so that later modification on it can be detected. 1287 */ 1288 ip = (uint32_t *)KSTACK_LOWEST_ADDR(l); 1289 end = (uint32_t *)((char *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE); 1290 for (; ip < end; ip++) { 1291 *ip = KSTACK_MAGIC; 1292 } 1293 } 1294 1295 void 1296 kstack_check_magic(const struct lwp *l) 1297 { 1298 uint32_t const *ip, *end; 1299 int stackleft; 1300 1301 KASSERT(l != NULL); 1302 1303 /* don't check proc0 */ /*XXX*/ 1304 if (l == &lwp0) 1305 return; 1306 1307 #ifdef __MACHINE_STACK_GROWS_UP 1308 /* stack grows upwards (eg. hppa) */ 1309 ip = (uint32_t *)((void *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE); 1310 end = (uint32_t *)KSTACK_LOWEST_ADDR(l); 1311 for (ip--; ip >= end; ip--) 1312 if (*ip != KSTACK_MAGIC) 1313 break; 1314 1315 stackleft = (void *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE - (void *)ip; 1316 #else /* __MACHINE_STACK_GROWS_UP */ 1317 /* stack grows downwards (eg. i386) */ 1318 ip = (uint32_t *)KSTACK_LOWEST_ADDR(l); 1319 end = (uint32_t *)((char *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE); 1320 for (; ip < end; ip++) 1321 if (*ip != KSTACK_MAGIC) 1322 break; 1323 1324 stackleft = ((const char *)ip) - (const char *)KSTACK_LOWEST_ADDR(l); 1325 #endif /* __MACHINE_STACK_GROWS_UP */ 1326 1327 if (kstackleftmin > stackleft) { 1328 kstackleftmin = stackleft; 1329 if (stackleft < kstackleftthres) 1330 printf("warning: kernel stack left %d bytes" 1331 "(pid %u:lid %u)\n", stackleft, 1332 (u_int)l->l_proc->p_pid, (u_int)l->l_lid); 1333 } 1334 1335 if (stackleft <= 0) { 1336 panic("magic on the top of kernel stack changed for " 1337 "pid %u, lid %u: maybe kernel stack overflow", 1338 (u_int)l->l_proc->p_pid, (u_int)l->l_lid); 1339 } 1340 } 1341 #endif /* KSTACK_CHECK_MAGIC */ 1342 1343 int 1344 proclist_foreach_call(struct proclist *list, 1345 int (*callback)(struct proc *, void *arg), void *arg) 1346 { 1347 struct proc marker; 1348 struct proc *p; 1349 int ret = 0; 1350 1351 marker.p_flag = PK_MARKER; 1352 mutex_enter(proc_lock); 1353 for (p = LIST_FIRST(list); ret == 0 && p != NULL;) { 1354 if (p->p_flag & PK_MARKER) { 1355 p = LIST_NEXT(p, p_list); 1356 continue; 1357 } 1358 LIST_INSERT_AFTER(p, &marker, p_list); 1359 ret = (*callback)(p, arg); 1360 KASSERT(mutex_owned(proc_lock)); 1361 p = LIST_NEXT(&marker, p_list); 1362 LIST_REMOVE(&marker, p_list); 1363 } 1364 mutex_exit(proc_lock); 1365 1366 return ret; 1367 } 1368 1369 int 1370 proc_vmspace_getref(struct proc *p, struct vmspace **vm) 1371 { 1372 1373 /* XXXCDC: how should locking work here? */ 1374 1375 /* curproc exception is for coredump. */ 1376 1377 if ((p != curproc && (p->p_sflag & PS_WEXIT) != 0) || 1378 (p->p_vmspace->vm_refcnt < 1)) { /* XXX */ 1379 return EFAULT; 1380 } 1381 1382 uvmspace_addref(p->p_vmspace); 1383 *vm = p->p_vmspace; 1384 1385 return 0; 1386 } 1387 1388 /* 1389 * Acquire a write lock on the process credential. 1390 */ 1391 void 1392 proc_crmod_enter(void) 1393 { 1394 struct lwp *l = curlwp; 1395 struct proc *p = l->l_proc; 1396 kauth_cred_t oc; 1397 1398 /* Reset what needs to be reset in plimit. */ 1399 if (p->p_limit->pl_corename != defcorename) { 1400 lim_setcorename(p, defcorename, 0); 1401 } 1402 1403 mutex_enter(p->p_lock); 1404 1405 /* Ensure the LWP cached credentials are up to date. */ 1406 if ((oc = l->l_cred) != p->p_cred) { 1407 kauth_cred_hold(p->p_cred); 1408 l->l_cred = p->p_cred; 1409 kauth_cred_free(oc); 1410 } 1411 } 1412 1413 /* 1414 * Set in a new process credential, and drop the write lock. The credential 1415 * must have a reference already. Optionally, free a no-longer required 1416 * credential. The scheduler also needs to inspect p_cred, so we also 1417 * briefly acquire the sched state mutex. 1418 */ 1419 void 1420 proc_crmod_leave(kauth_cred_t scred, kauth_cred_t fcred, bool sugid) 1421 { 1422 struct lwp *l = curlwp, *l2; 1423 struct proc *p = l->l_proc; 1424 kauth_cred_t oc; 1425 1426 KASSERT(mutex_owned(p->p_lock)); 1427 1428 /* Is there a new credential to set in? */ 1429 if (scred != NULL) { 1430 p->p_cred = scred; 1431 LIST_FOREACH(l2, &p->p_lwps, l_sibling) { 1432 if (l2 != l) 1433 l2->l_prflag |= LPR_CRMOD; 1434 } 1435 1436 /* Ensure the LWP cached credentials are up to date. */ 1437 if ((oc = l->l_cred) != scred) { 1438 kauth_cred_hold(scred); 1439 l->l_cred = scred; 1440 } 1441 } else 1442 oc = NULL; /* XXXgcc */ 1443 1444 if (sugid) { 1445 /* 1446 * Mark process as having changed credentials, stops 1447 * tracing etc. 1448 */ 1449 p->p_flag |= PK_SUGID; 1450 } 1451 1452 mutex_exit(p->p_lock); 1453 1454 /* If there is a credential to be released, free it now. */ 1455 if (fcred != NULL) { 1456 KASSERT(scred != NULL); 1457 kauth_cred_free(fcred); 1458 if (oc != scred) 1459 kauth_cred_free(oc); 1460 } 1461 } 1462 1463 /* 1464 * proc_specific_key_create -- 1465 * Create a key for subsystem proc-specific data. 1466 */ 1467 int 1468 proc_specific_key_create(specificdata_key_t *keyp, specificdata_dtor_t dtor) 1469 { 1470 1471 return (specificdata_key_create(proc_specificdata_domain, keyp, dtor)); 1472 } 1473 1474 /* 1475 * proc_specific_key_delete -- 1476 * Delete a key for subsystem proc-specific data. 1477 */ 1478 void 1479 proc_specific_key_delete(specificdata_key_t key) 1480 { 1481 1482 specificdata_key_delete(proc_specificdata_domain, key); 1483 } 1484 1485 /* 1486 * proc_initspecific -- 1487 * Initialize a proc's specificdata container. 1488 */ 1489 void 1490 proc_initspecific(struct proc *p) 1491 { 1492 int error __diagused; 1493 1494 error = specificdata_init(proc_specificdata_domain, &p->p_specdataref); 1495 KASSERT(error == 0); 1496 } 1497 1498 /* 1499 * proc_finispecific -- 1500 * Finalize a proc's specificdata container. 1501 */ 1502 void 1503 proc_finispecific(struct proc *p) 1504 { 1505 1506 specificdata_fini(proc_specificdata_domain, &p->p_specdataref); 1507 } 1508 1509 /* 1510 * proc_getspecific -- 1511 * Return proc-specific data corresponding to the specified key. 1512 */ 1513 void * 1514 proc_getspecific(struct proc *p, specificdata_key_t key) 1515 { 1516 1517 return (specificdata_getspecific(proc_specificdata_domain, 1518 &p->p_specdataref, key)); 1519 } 1520 1521 /* 1522 * proc_setspecific -- 1523 * Set proc-specific data corresponding to the specified key. 1524 */ 1525 void 1526 proc_setspecific(struct proc *p, specificdata_key_t key, void *data) 1527 { 1528 1529 specificdata_setspecific(proc_specificdata_domain, 1530 &p->p_specdataref, key, data); 1531 } 1532 1533 int 1534 proc_uidmatch(kauth_cred_t cred, kauth_cred_t target) 1535 { 1536 int r = 0; 1537 1538 if (kauth_cred_getuid(cred) != kauth_cred_getuid(target) || 1539 kauth_cred_getuid(cred) != kauth_cred_getsvuid(target)) { 1540 /* 1541 * suid proc of ours or proc not ours 1542 */ 1543 r = EPERM; 1544 } else if (kauth_cred_getgid(target) != kauth_cred_getsvgid(target)) { 1545 /* 1546 * sgid proc has sgid back to us temporarily 1547 */ 1548 r = EPERM; 1549 } else { 1550 /* 1551 * our rgid must be in target's group list (ie, 1552 * sub-processes started by a sgid process) 1553 */ 1554 int ismember = 0; 1555 1556 if (kauth_cred_ismember_gid(cred, 1557 kauth_cred_getgid(target), &ismember) != 0 || 1558 !ismember) 1559 r = EPERM; 1560 } 1561 1562 return (r); 1563 } 1564 1565 /* 1566 * sysctl stuff 1567 */ 1568 1569 #define KERN_PROCSLOP (5 * sizeof(struct kinfo_proc)) 1570 1571 static const u_int sysctl_flagmap[] = { 1572 PK_ADVLOCK, P_ADVLOCK, 1573 PK_EXEC, P_EXEC, 1574 PK_NOCLDWAIT, P_NOCLDWAIT, 1575 PK_32, P_32, 1576 PK_CLDSIGIGN, P_CLDSIGIGN, 1577 PK_SUGID, P_SUGID, 1578 0 1579 }; 1580 1581 static const u_int sysctl_sflagmap[] = { 1582 PS_NOCLDSTOP, P_NOCLDSTOP, 1583 PS_WEXIT, P_WEXIT, 1584 PS_STOPFORK, P_STOPFORK, 1585 PS_STOPEXEC, P_STOPEXEC, 1586 PS_STOPEXIT, P_STOPEXIT, 1587 0 1588 }; 1589 1590 static const u_int sysctl_slflagmap[] = { 1591 PSL_TRACED, P_TRACED, 1592 PSL_CHTRACED, P_CHTRACED, 1593 PSL_SYSCALL, P_SYSCALL, 1594 0 1595 }; 1596 1597 static const u_int sysctl_lflagmap[] = { 1598 PL_CONTROLT, P_CONTROLT, 1599 PL_PPWAIT, P_PPWAIT, 1600 0 1601 }; 1602 1603 static const u_int sysctl_stflagmap[] = { 1604 PST_PROFIL, P_PROFIL, 1605 0 1606 1607 }; 1608 1609 /* used by kern_lwp also */ 1610 const u_int sysctl_lwpflagmap[] = { 1611 LW_SINTR, L_SINTR, 1612 LW_SYSTEM, L_SYSTEM, 1613 0 1614 }; 1615 1616 /* 1617 * Find the most ``active'' lwp of a process and return it for ps display 1618 * purposes 1619 */ 1620 static struct lwp * 1621 proc_active_lwp(struct proc *p) 1622 { 1623 static const int ostat[] = { 1624 0, 1625 2, /* LSIDL */ 1626 6, /* LSRUN */ 1627 5, /* LSSLEEP */ 1628 4, /* LSSTOP */ 1629 0, /* LSZOMB */ 1630 1, /* LSDEAD */ 1631 7, /* LSONPROC */ 1632 3 /* LSSUSPENDED */ 1633 }; 1634 1635 struct lwp *l, *lp = NULL; 1636 LIST_FOREACH(l, &p->p_lwps, l_sibling) { 1637 KASSERT(l->l_stat >= 0 && l->l_stat < __arraycount(ostat)); 1638 if (lp == NULL || 1639 ostat[l->l_stat] > ostat[lp->l_stat] || 1640 (ostat[l->l_stat] == ostat[lp->l_stat] && 1641 l->l_cpticks > lp->l_cpticks)) { 1642 lp = l; 1643 continue; 1644 } 1645 } 1646 return lp; 1647 } 1648 1649 static int 1650 sysctl_doeproc(SYSCTLFN_ARGS) 1651 { 1652 union { 1653 struct kinfo_proc kproc; 1654 struct kinfo_proc2 kproc2; 1655 } *kbuf; 1656 struct proc *p, *next, *marker; 1657 char *where, *dp; 1658 int type, op, arg, error; 1659 u_int elem_size, kelem_size, elem_count; 1660 size_t buflen, needed; 1661 bool match, zombie, mmmbrains; 1662 const bool allowaddr = get_expose_address(curproc); 1663 1664 if (namelen == 1 && name[0] == CTL_QUERY) 1665 return (sysctl_query(SYSCTLFN_CALL(rnode))); 1666 1667 dp = where = oldp; 1668 buflen = where != NULL ? *oldlenp : 0; 1669 error = 0; 1670 needed = 0; 1671 type = rnode->sysctl_num; 1672 1673 if (type == KERN_PROC) { 1674 if (namelen == 0) 1675 return EINVAL; 1676 switch (op = name[0]) { 1677 case KERN_PROC_ALL: 1678 if (namelen != 1) 1679 return EINVAL; 1680 arg = 0; 1681 break; 1682 default: 1683 if (namelen != 2) 1684 return EINVAL; 1685 arg = name[1]; 1686 break; 1687 } 1688 elem_count = 0; /* Hush little compiler, don't you cry */ 1689 kelem_size = elem_size = sizeof(kbuf->kproc); 1690 } else { 1691 if (namelen != 4) 1692 return EINVAL; 1693 op = name[0]; 1694 arg = name[1]; 1695 elem_size = name[2]; 1696 elem_count = name[3]; 1697 kelem_size = sizeof(kbuf->kproc2); 1698 } 1699 1700 sysctl_unlock(); 1701 1702 kbuf = kmem_zalloc(sizeof(*kbuf), KM_SLEEP); 1703 marker = kmem_alloc(sizeof(*marker), KM_SLEEP); 1704 marker->p_flag = PK_MARKER; 1705 1706 mutex_enter(proc_lock); 1707 /* 1708 * Start with zombies to prevent reporting processes twice, in case they 1709 * are dying and being moved from the list of alive processes to zombies. 1710 */ 1711 mmmbrains = true; 1712 for (p = LIST_FIRST(&zombproc);; p = next) { 1713 if (p == NULL) { 1714 if (mmmbrains) { 1715 p = LIST_FIRST(&allproc); 1716 mmmbrains = false; 1717 } 1718 if (p == NULL) 1719 break; 1720 } 1721 next = LIST_NEXT(p, p_list); 1722 if ((p->p_flag & PK_MARKER) != 0) 1723 continue; 1724 1725 /* 1726 * Skip embryonic processes. 1727 */ 1728 if (p->p_stat == SIDL) 1729 continue; 1730 1731 mutex_enter(p->p_lock); 1732 error = kauth_authorize_process(l->l_cred, 1733 KAUTH_PROCESS_CANSEE, p, 1734 KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_EPROC), NULL, NULL); 1735 if (error != 0) { 1736 mutex_exit(p->p_lock); 1737 continue; 1738 } 1739 1740 /* 1741 * Hande all the operations in one switch on the cost of 1742 * algorithm complexity is on purpose. The win splitting this 1743 * function into several similar copies makes maintenance burden 1744 * burden, code grow and boost is neglible in practical systems. 1745 */ 1746 switch (op) { 1747 case KERN_PROC_PID: 1748 match = (p->p_pid == (pid_t)arg); 1749 break; 1750 1751 case KERN_PROC_PGRP: 1752 match = (p->p_pgrp->pg_id == (pid_t)arg); 1753 break; 1754 1755 case KERN_PROC_SESSION: 1756 match = (p->p_session->s_sid == (pid_t)arg); 1757 break; 1758 1759 case KERN_PROC_TTY: 1760 match = true; 1761 if (arg == (int) KERN_PROC_TTY_REVOKE) { 1762 if ((p->p_lflag & PL_CONTROLT) == 0 || 1763 p->p_session->s_ttyp == NULL || 1764 p->p_session->s_ttyvp != NULL) { 1765 match = false; 1766 } 1767 } else if ((p->p_lflag & PL_CONTROLT) == 0 || 1768 p->p_session->s_ttyp == NULL) { 1769 if ((dev_t)arg != KERN_PROC_TTY_NODEV) { 1770 match = false; 1771 } 1772 } else if (p->p_session->s_ttyp->t_dev != (dev_t)arg) { 1773 match = false; 1774 } 1775 break; 1776 1777 case KERN_PROC_UID: 1778 match = (kauth_cred_geteuid(p->p_cred) == (uid_t)arg); 1779 break; 1780 1781 case KERN_PROC_RUID: 1782 match = (kauth_cred_getuid(p->p_cred) == (uid_t)arg); 1783 break; 1784 1785 case KERN_PROC_GID: 1786 match = (kauth_cred_getegid(p->p_cred) == (uid_t)arg); 1787 break; 1788 1789 case KERN_PROC_RGID: 1790 match = (kauth_cred_getgid(p->p_cred) == (uid_t)arg); 1791 break; 1792 1793 case KERN_PROC_ALL: 1794 match = true; 1795 /* allow everything */ 1796 break; 1797 1798 default: 1799 error = EINVAL; 1800 mutex_exit(p->p_lock); 1801 goto cleanup; 1802 } 1803 if (!match) { 1804 mutex_exit(p->p_lock); 1805 continue; 1806 } 1807 1808 /* 1809 * Grab a hold on the process. 1810 */ 1811 if (mmmbrains) { 1812 zombie = true; 1813 } else { 1814 zombie = !rw_tryenter(&p->p_reflock, RW_READER); 1815 } 1816 if (zombie) { 1817 LIST_INSERT_AFTER(p, marker, p_list); 1818 } 1819 1820 if (buflen >= elem_size && 1821 (type == KERN_PROC || elem_count > 0)) { 1822 ruspace(p); /* Update process vm resource use */ 1823 1824 if (type == KERN_PROC) { 1825 fill_proc(p, &kbuf->kproc.kp_proc, allowaddr); 1826 fill_eproc(p, &kbuf->kproc.kp_eproc, zombie, 1827 allowaddr); 1828 } else { 1829 fill_kproc2(p, &kbuf->kproc2, zombie, 1830 allowaddr); 1831 elem_count--; 1832 } 1833 mutex_exit(p->p_lock); 1834 mutex_exit(proc_lock); 1835 /* 1836 * Copy out elem_size, but not larger than kelem_size 1837 */ 1838 error = sysctl_copyout(l, kbuf, dp, 1839 uimin(kelem_size, elem_size)); 1840 mutex_enter(proc_lock); 1841 if (error) { 1842 goto bah; 1843 } 1844 dp += elem_size; 1845 buflen -= elem_size; 1846 } else { 1847 mutex_exit(p->p_lock); 1848 } 1849 needed += elem_size; 1850 1851 /* 1852 * Release reference to process. 1853 */ 1854 if (zombie) { 1855 next = LIST_NEXT(marker, p_list); 1856 LIST_REMOVE(marker, p_list); 1857 } else { 1858 rw_exit(&p->p_reflock); 1859 next = LIST_NEXT(p, p_list); 1860 } 1861 1862 /* 1863 * Short-circuit break quickly! 1864 */ 1865 if (op == KERN_PROC_PID) 1866 break; 1867 } 1868 mutex_exit(proc_lock); 1869 1870 if (where != NULL) { 1871 *oldlenp = dp - where; 1872 if (needed > *oldlenp) { 1873 error = ENOMEM; 1874 goto out; 1875 } 1876 } else { 1877 needed += KERN_PROCSLOP; 1878 *oldlenp = needed; 1879 } 1880 kmem_free(kbuf, sizeof(*kbuf)); 1881 kmem_free(marker, sizeof(*marker)); 1882 sysctl_relock(); 1883 return 0; 1884 bah: 1885 if (zombie) 1886 LIST_REMOVE(marker, p_list); 1887 else 1888 rw_exit(&p->p_reflock); 1889 cleanup: 1890 mutex_exit(proc_lock); 1891 out: 1892 kmem_free(kbuf, sizeof(*kbuf)); 1893 kmem_free(marker, sizeof(*marker)); 1894 sysctl_relock(); 1895 return error; 1896 } 1897 1898 int 1899 copyin_psstrings(struct proc *p, struct ps_strings *arginfo) 1900 { 1901 #if !defined(_RUMPKERNEL) 1902 int retval; 1903 1904 if (p->p_flag & PK_32) { 1905 MODULE_HOOK_CALL(kern_proc32_copyin_hook, (p, arginfo), 1906 enosys(), retval); 1907 return retval; 1908 } 1909 #endif /* !defined(_RUMPKERNEL) */ 1910 1911 return copyin_proc(p, (void *)p->p_psstrp, arginfo, sizeof(*arginfo)); 1912 } 1913 1914 static int 1915 copy_procargs_sysctl_cb(void *cookie_, const void *src, size_t off, size_t len) 1916 { 1917 void **cookie = cookie_; 1918 struct lwp *l = cookie[0]; 1919 char *dst = cookie[1]; 1920 1921 return sysctl_copyout(l, src, dst + off, len); 1922 } 1923 1924 /* 1925 * sysctl helper routine for kern.proc_args pseudo-subtree. 1926 */ 1927 static int 1928 sysctl_kern_proc_args(SYSCTLFN_ARGS) 1929 { 1930 struct ps_strings pss; 1931 struct proc *p; 1932 pid_t pid; 1933 int type, error; 1934 void *cookie[2]; 1935 1936 if (namelen == 1 && name[0] == CTL_QUERY) 1937 return (sysctl_query(SYSCTLFN_CALL(rnode))); 1938 1939 if (newp != NULL || namelen != 2) 1940 return (EINVAL); 1941 pid = name[0]; 1942 type = name[1]; 1943 1944 switch (type) { 1945 case KERN_PROC_PATHNAME: 1946 sysctl_unlock(); 1947 error = fill_pathname(l, pid, oldp, oldlenp); 1948 sysctl_relock(); 1949 return error; 1950 1951 case KERN_PROC_CWD: 1952 sysctl_unlock(); 1953 error = fill_cwd(l, pid, oldp, oldlenp); 1954 sysctl_relock(); 1955 return error; 1956 1957 case KERN_PROC_ARGV: 1958 case KERN_PROC_NARGV: 1959 case KERN_PROC_ENV: 1960 case KERN_PROC_NENV: 1961 /* ok */ 1962 break; 1963 default: 1964 return (EINVAL); 1965 } 1966 1967 sysctl_unlock(); 1968 1969 /* check pid */ 1970 mutex_enter(proc_lock); 1971 if ((p = proc_find(pid)) == NULL) { 1972 error = EINVAL; 1973 goto out_locked; 1974 } 1975 mutex_enter(p->p_lock); 1976 1977 /* Check permission. */ 1978 if (type == KERN_PROC_ARGV || type == KERN_PROC_NARGV) 1979 error = kauth_authorize_process(l->l_cred, KAUTH_PROCESS_CANSEE, 1980 p, KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_ARGS), NULL, NULL); 1981 else if (type == KERN_PROC_ENV || type == KERN_PROC_NENV) 1982 error = kauth_authorize_process(l->l_cred, KAUTH_PROCESS_CANSEE, 1983 p, KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_ENV), NULL, NULL); 1984 else 1985 error = EINVAL; /* XXXGCC */ 1986 if (error) { 1987 mutex_exit(p->p_lock); 1988 goto out_locked; 1989 } 1990 1991 if (oldp == NULL) { 1992 if (type == KERN_PROC_NARGV || type == KERN_PROC_NENV) 1993 *oldlenp = sizeof (int); 1994 else 1995 *oldlenp = ARG_MAX; /* XXX XXX XXX */ 1996 error = 0; 1997 mutex_exit(p->p_lock); 1998 goto out_locked; 1999 } 2000 2001 /* 2002 * Zombies don't have a stack, so we can't read their psstrings. 2003 * System processes also don't have a user stack. 2004 */ 2005 if (P_ZOMBIE(p) || (p->p_flag & PK_SYSTEM) != 0) { 2006 error = EINVAL; 2007 mutex_exit(p->p_lock); 2008 goto out_locked; 2009 } 2010 2011 error = rw_tryenter(&p->p_reflock, RW_READER) ? 0 : EBUSY; 2012 mutex_exit(p->p_lock); 2013 if (error) { 2014 goto out_locked; 2015 } 2016 mutex_exit(proc_lock); 2017 2018 if (type == KERN_PROC_NARGV || type == KERN_PROC_NENV) { 2019 int value; 2020 if ((error = copyin_psstrings(p, &pss)) == 0) { 2021 if (type == KERN_PROC_NARGV) 2022 value = pss.ps_nargvstr; 2023 else 2024 value = pss.ps_nenvstr; 2025 error = sysctl_copyout(l, &value, oldp, sizeof(value)); 2026 *oldlenp = sizeof(value); 2027 } 2028 } else { 2029 cookie[0] = l; 2030 cookie[1] = oldp; 2031 error = copy_procargs(p, type, oldlenp, 2032 copy_procargs_sysctl_cb, cookie); 2033 } 2034 rw_exit(&p->p_reflock); 2035 sysctl_relock(); 2036 return error; 2037 2038 out_locked: 2039 mutex_exit(proc_lock); 2040 sysctl_relock(); 2041 return error; 2042 } 2043 2044 int 2045 copy_procargs(struct proc *p, int oid, size_t *limit, 2046 int (*cb)(void *, const void *, size_t, size_t), void *cookie) 2047 { 2048 struct ps_strings pss; 2049 size_t len, i, loaded, entry_len; 2050 struct uio auio; 2051 struct iovec aiov; 2052 int error, argvlen; 2053 char *arg; 2054 char **argv; 2055 vaddr_t user_argv; 2056 struct vmspace *vmspace; 2057 2058 /* 2059 * Allocate a temporary buffer to hold the argument vector and 2060 * the arguments themselve. 2061 */ 2062 arg = kmem_alloc(PAGE_SIZE, KM_SLEEP); 2063 argv = kmem_alloc(PAGE_SIZE, KM_SLEEP); 2064 2065 /* 2066 * Lock the process down in memory. 2067 */ 2068 vmspace = p->p_vmspace; 2069 uvmspace_addref(vmspace); 2070 2071 /* 2072 * Read in the ps_strings structure. 2073 */ 2074 if ((error = copyin_psstrings(p, &pss)) != 0) 2075 goto done; 2076 2077 /* 2078 * Now read the address of the argument vector. 2079 */ 2080 switch (oid) { 2081 case KERN_PROC_ARGV: 2082 user_argv = (uintptr_t)pss.ps_argvstr; 2083 argvlen = pss.ps_nargvstr; 2084 break; 2085 case KERN_PROC_ENV: 2086 user_argv = (uintptr_t)pss.ps_envstr; 2087 argvlen = pss.ps_nenvstr; 2088 break; 2089 default: 2090 error = EINVAL; 2091 goto done; 2092 } 2093 2094 if (argvlen < 0) { 2095 error = EIO; 2096 goto done; 2097 } 2098 2099 2100 /* 2101 * Now copy each string. 2102 */ 2103 len = 0; /* bytes written to user buffer */ 2104 loaded = 0; /* bytes from argv already processed */ 2105 i = 0; /* To make compiler happy */ 2106 entry_len = PROC_PTRSZ(p); 2107 2108 for (; argvlen; --argvlen) { 2109 int finished = 0; 2110 vaddr_t base; 2111 size_t xlen; 2112 int j; 2113 2114 if (loaded == 0) { 2115 size_t rem = entry_len * argvlen; 2116 loaded = MIN(rem, PAGE_SIZE); 2117 error = copyin_vmspace(vmspace, 2118 (const void *)user_argv, argv, loaded); 2119 if (error) 2120 break; 2121 user_argv += loaded; 2122 i = 0; 2123 } 2124 2125 #if !defined(_RUMPKERNEL) 2126 if (p->p_flag & PK_32) 2127 MODULE_HOOK_CALL(kern_proc32_base_hook, 2128 (argv, i++), 0, base); 2129 else 2130 #endif /* !defined(_RUMPKERNEL) */ 2131 base = (vaddr_t)argv[i++]; 2132 loaded -= entry_len; 2133 2134 /* 2135 * The program has messed around with its arguments, 2136 * possibly deleting some, and replacing them with 2137 * NULL's. Treat this as the last argument and not 2138 * a failure. 2139 */ 2140 if (base == 0) 2141 break; 2142 2143 while (!finished) { 2144 xlen = PAGE_SIZE - (base & PAGE_MASK); 2145 2146 aiov.iov_base = arg; 2147 aiov.iov_len = PAGE_SIZE; 2148 auio.uio_iov = &aiov; 2149 auio.uio_iovcnt = 1; 2150 auio.uio_offset = base; 2151 auio.uio_resid = xlen; 2152 auio.uio_rw = UIO_READ; 2153 UIO_SETUP_SYSSPACE(&auio); 2154 error = uvm_io(&vmspace->vm_map, &auio, 0); 2155 if (error) 2156 goto done; 2157 2158 /* Look for the end of the string */ 2159 for (j = 0; j < xlen; j++) { 2160 if (arg[j] == '\0') { 2161 xlen = j + 1; 2162 finished = 1; 2163 break; 2164 } 2165 } 2166 2167 /* Check for user buffer overflow */ 2168 if (len + xlen > *limit) { 2169 finished = 1; 2170 if (len > *limit) 2171 xlen = 0; 2172 else 2173 xlen = *limit - len; 2174 } 2175 2176 /* Copyout the page */ 2177 error = (*cb)(cookie, arg, len, xlen); 2178 if (error) 2179 goto done; 2180 2181 len += xlen; 2182 base += xlen; 2183 } 2184 } 2185 *limit = len; 2186 2187 done: 2188 kmem_free(argv, PAGE_SIZE); 2189 kmem_free(arg, PAGE_SIZE); 2190 uvmspace_free(vmspace); 2191 return error; 2192 } 2193 2194 /* 2195 * Fill in a proc structure for the specified process. 2196 */ 2197 static void 2198 fill_proc(const struct proc *psrc, struct proc *p, bool allowaddr) 2199 { 2200 COND_SET_VALUE(p->p_list, psrc->p_list, allowaddr); 2201 COND_SET_VALUE(p->p_auxlock, psrc->p_auxlock, allowaddr); 2202 COND_SET_VALUE(p->p_lock, psrc->p_lock, allowaddr); 2203 COND_SET_VALUE(p->p_stmutex, psrc->p_stmutex, allowaddr); 2204 COND_SET_VALUE(p->p_reflock, psrc->p_reflock, allowaddr); 2205 COND_SET_VALUE(p->p_waitcv, psrc->p_waitcv, allowaddr); 2206 COND_SET_VALUE(p->p_lwpcv, psrc->p_lwpcv, allowaddr); 2207 COND_SET_VALUE(p->p_cred, psrc->p_cred, allowaddr); 2208 COND_SET_VALUE(p->p_fd, psrc->p_fd, allowaddr); 2209 COND_SET_VALUE(p->p_cwdi, psrc->p_cwdi, allowaddr); 2210 COND_SET_VALUE(p->p_stats, psrc->p_stats, allowaddr); 2211 COND_SET_VALUE(p->p_limit, psrc->p_limit, allowaddr); 2212 COND_SET_VALUE(p->p_vmspace, psrc->p_vmspace, allowaddr); 2213 COND_SET_VALUE(p->p_sigacts, psrc->p_sigacts, allowaddr); 2214 COND_SET_VALUE(p->p_aio, psrc->p_aio, allowaddr); 2215 p->p_mqueue_cnt = psrc->p_mqueue_cnt; 2216 COND_SET_VALUE(p->p_specdataref, psrc->p_specdataref, allowaddr); 2217 p->p_exitsig = psrc->p_exitsig; 2218 p->p_flag = psrc->p_flag; 2219 p->p_sflag = psrc->p_sflag; 2220 p->p_slflag = psrc->p_slflag; 2221 p->p_lflag = psrc->p_lflag; 2222 p->p_stflag = psrc->p_stflag; 2223 p->p_stat = psrc->p_stat; 2224 p->p_trace_enabled = psrc->p_trace_enabled; 2225 p->p_pid = psrc->p_pid; 2226 COND_SET_VALUE(p->p_pglist, psrc->p_pglist, allowaddr); 2227 COND_SET_VALUE(p->p_pptr, psrc->p_pptr, allowaddr); 2228 COND_SET_VALUE(p->p_sibling, psrc->p_sibling, allowaddr); 2229 COND_SET_VALUE(p->p_children, psrc->p_children, allowaddr); 2230 COND_SET_VALUE(p->p_lwps, psrc->p_lwps, allowaddr); 2231 COND_SET_VALUE(p->p_raslist, psrc->p_raslist, allowaddr); 2232 p->p_nlwps = psrc->p_nlwps; 2233 p->p_nzlwps = psrc->p_nzlwps; 2234 p->p_nrlwps = psrc->p_nrlwps; 2235 p->p_nlwpwait = psrc->p_nlwpwait; 2236 p->p_ndlwps = psrc->p_ndlwps; 2237 p->p_nlwpid = psrc->p_nlwpid; 2238 p->p_nstopchild = psrc->p_nstopchild; 2239 p->p_waited = psrc->p_waited; 2240 COND_SET_VALUE(p->p_zomblwp, psrc->p_zomblwp, allowaddr); 2241 COND_SET_VALUE(p->p_vforklwp, psrc->p_vforklwp, allowaddr); 2242 COND_SET_VALUE(p->p_sched_info, psrc->p_sched_info, allowaddr); 2243 p->p_estcpu = psrc->p_estcpu; 2244 p->p_estcpu_inherited = psrc->p_estcpu_inherited; 2245 p->p_forktime = psrc->p_forktime; 2246 p->p_pctcpu = psrc->p_pctcpu; 2247 COND_SET_VALUE(p->p_opptr, psrc->p_opptr, allowaddr); 2248 COND_SET_VALUE(p->p_timers, psrc->p_timers, allowaddr); 2249 p->p_rtime = psrc->p_rtime; 2250 p->p_uticks = psrc->p_uticks; 2251 p->p_sticks = psrc->p_sticks; 2252 p->p_iticks = psrc->p_iticks; 2253 p->p_xutime = psrc->p_xutime; 2254 p->p_xstime = psrc->p_xstime; 2255 p->p_traceflag = psrc->p_traceflag; 2256 COND_SET_VALUE(p->p_tracep, psrc->p_tracep, allowaddr); 2257 COND_SET_VALUE(p->p_textvp, psrc->p_textvp, allowaddr); 2258 COND_SET_VALUE(p->p_emul, psrc->p_emul, allowaddr); 2259 COND_SET_VALUE(p->p_emuldata, psrc->p_emuldata, allowaddr); 2260 COND_SET_VALUE(p->p_execsw, psrc->p_execsw, allowaddr); 2261 COND_SET_VALUE(p->p_klist, psrc->p_klist, allowaddr); 2262 COND_SET_VALUE(p->p_sigwaiters, psrc->p_sigwaiters, allowaddr); 2263 COND_SET_VALUE(p->p_sigpend, psrc->p_sigpend, allowaddr); 2264 COND_SET_VALUE(p->p_lwpctl, psrc->p_lwpctl, allowaddr); 2265 p->p_ppid = psrc->p_ppid; 2266 p->p_oppid = psrc->p_oppid; 2267 COND_SET_VALUE(p->p_path, psrc->p_path, allowaddr); 2268 COND_SET_VALUE(p->p_sigctx, psrc->p_sigctx, allowaddr); 2269 p->p_nice = psrc->p_nice; 2270 memcpy(p->p_comm, psrc->p_comm, sizeof(p->p_comm)); 2271 COND_SET_VALUE(p->p_pgrp, psrc->p_pgrp, allowaddr); 2272 COND_SET_VALUE(p->p_psstrp, psrc->p_psstrp, allowaddr); 2273 p->p_pax = psrc->p_pax; 2274 p->p_xexit = psrc->p_xexit; 2275 p->p_xsig = psrc->p_xsig; 2276 p->p_acflag = psrc->p_acflag; 2277 COND_SET_VALUE(p->p_md, psrc->p_md, allowaddr); 2278 p->p_stackbase = psrc->p_stackbase; 2279 COND_SET_VALUE(p->p_dtrace, psrc->p_dtrace, allowaddr); 2280 } 2281 2282 /* 2283 * Fill in an eproc structure for the specified process. 2284 */ 2285 void 2286 fill_eproc(struct proc *p, struct eproc *ep, bool zombie, bool allowaddr) 2287 { 2288 struct tty *tp; 2289 struct lwp *l; 2290 2291 KASSERT(mutex_owned(proc_lock)); 2292 KASSERT(mutex_owned(p->p_lock)); 2293 2294 COND_SET_VALUE(ep->e_paddr, p, allowaddr); 2295 COND_SET_VALUE(ep->e_sess, p->p_session, allowaddr); 2296 if (p->p_cred) { 2297 kauth_cred_topcred(p->p_cred, &ep->e_pcred); 2298 kauth_cred_toucred(p->p_cred, &ep->e_ucred); 2299 } 2300 if (p->p_stat != SIDL && !P_ZOMBIE(p) && !zombie) { 2301 struct vmspace *vm = p->p_vmspace; 2302 2303 ep->e_vm.vm_rssize = vm_resident_count(vm); 2304 ep->e_vm.vm_tsize = vm->vm_tsize; 2305 ep->e_vm.vm_dsize = vm->vm_dsize; 2306 ep->e_vm.vm_ssize = vm->vm_ssize; 2307 ep->e_vm.vm_map.size = vm->vm_map.size; 2308 2309 /* Pick the primary (first) LWP */ 2310 l = proc_active_lwp(p); 2311 KASSERT(l != NULL); 2312 lwp_lock(l); 2313 if (l->l_wchan) 2314 strncpy(ep->e_wmesg, l->l_wmesg, WMESGLEN); 2315 lwp_unlock(l); 2316 } 2317 ep->e_ppid = p->p_ppid; 2318 if (p->p_pgrp && p->p_session) { 2319 ep->e_pgid = p->p_pgrp->pg_id; 2320 ep->e_jobc = p->p_pgrp->pg_jobc; 2321 ep->e_sid = p->p_session->s_sid; 2322 if ((p->p_lflag & PL_CONTROLT) && 2323 (tp = p->p_session->s_ttyp)) { 2324 ep->e_tdev = tp->t_dev; 2325 ep->e_tpgid = tp->t_pgrp ? tp->t_pgrp->pg_id : NO_PGID; 2326 COND_SET_VALUE(ep->e_tsess, tp->t_session, allowaddr); 2327 } else 2328 ep->e_tdev = (uint32_t)NODEV; 2329 ep->e_flag = p->p_session->s_ttyvp ? EPROC_CTTY : 0; 2330 if (SESS_LEADER(p)) 2331 ep->e_flag |= EPROC_SLEADER; 2332 strncpy(ep->e_login, p->p_session->s_login, MAXLOGNAME); 2333 } 2334 ep->e_xsize = ep->e_xrssize = 0; 2335 ep->e_xccount = ep->e_xswrss = 0; 2336 } 2337 2338 /* 2339 * Fill in a kinfo_proc2 structure for the specified process. 2340 */ 2341 void 2342 fill_kproc2(struct proc *p, struct kinfo_proc2 *ki, bool zombie, bool allowaddr) 2343 { 2344 struct tty *tp; 2345 struct lwp *l, *l2; 2346 struct timeval ut, st, rt; 2347 sigset_t ss1, ss2; 2348 struct rusage ru; 2349 struct vmspace *vm; 2350 2351 KASSERT(mutex_owned(proc_lock)); 2352 KASSERT(mutex_owned(p->p_lock)); 2353 2354 sigemptyset(&ss1); 2355 sigemptyset(&ss2); 2356 2357 COND_SET_VALUE(ki->p_paddr, PTRTOUINT64(p), allowaddr); 2358 COND_SET_VALUE(ki->p_fd, PTRTOUINT64(p->p_fd), allowaddr); 2359 COND_SET_VALUE(ki->p_cwdi, PTRTOUINT64(p->p_cwdi), allowaddr); 2360 COND_SET_VALUE(ki->p_stats, PTRTOUINT64(p->p_stats), allowaddr); 2361 COND_SET_VALUE(ki->p_limit, PTRTOUINT64(p->p_limit), allowaddr); 2362 COND_SET_VALUE(ki->p_vmspace, PTRTOUINT64(p->p_vmspace), allowaddr); 2363 COND_SET_VALUE(ki->p_sigacts, PTRTOUINT64(p->p_sigacts), allowaddr); 2364 COND_SET_VALUE(ki->p_sess, PTRTOUINT64(p->p_session), allowaddr); 2365 ki->p_tsess = 0; /* may be changed if controlling tty below */ 2366 COND_SET_VALUE(ki->p_ru, PTRTOUINT64(&p->p_stats->p_ru), allowaddr); 2367 ki->p_eflag = 0; 2368 ki->p_exitsig = p->p_exitsig; 2369 ki->p_flag = L_INMEM; /* Process never swapped out */ 2370 ki->p_flag |= sysctl_map_flags(sysctl_flagmap, p->p_flag); 2371 ki->p_flag |= sysctl_map_flags(sysctl_sflagmap, p->p_sflag); 2372 ki->p_flag |= sysctl_map_flags(sysctl_slflagmap, p->p_slflag); 2373 ki->p_flag |= sysctl_map_flags(sysctl_lflagmap, p->p_lflag); 2374 ki->p_flag |= sysctl_map_flags(sysctl_stflagmap, p->p_stflag); 2375 ki->p_pid = p->p_pid; 2376 ki->p_ppid = p->p_ppid; 2377 ki->p_uid = kauth_cred_geteuid(p->p_cred); 2378 ki->p_ruid = kauth_cred_getuid(p->p_cred); 2379 ki->p_gid = kauth_cred_getegid(p->p_cred); 2380 ki->p_rgid = kauth_cred_getgid(p->p_cred); 2381 ki->p_svuid = kauth_cred_getsvuid(p->p_cred); 2382 ki->p_svgid = kauth_cred_getsvgid(p->p_cred); 2383 ki->p_ngroups = kauth_cred_ngroups(p->p_cred); 2384 kauth_cred_getgroups(p->p_cred, ki->p_groups, 2385 uimin(ki->p_ngroups, sizeof(ki->p_groups) / sizeof(ki->p_groups[0])), 2386 UIO_SYSSPACE); 2387 2388 ki->p_uticks = p->p_uticks; 2389 ki->p_sticks = p->p_sticks; 2390 ki->p_iticks = p->p_iticks; 2391 ki->p_tpgid = NO_PGID; /* may be changed if controlling tty below */ 2392 COND_SET_VALUE(ki->p_tracep, PTRTOUINT64(p->p_tracep), allowaddr); 2393 ki->p_traceflag = p->p_traceflag; 2394 2395 memcpy(&ki->p_sigignore, &p->p_sigctx.ps_sigignore,sizeof(ki_sigset_t)); 2396 memcpy(&ki->p_sigcatch, &p->p_sigctx.ps_sigcatch, sizeof(ki_sigset_t)); 2397 2398 ki->p_cpticks = 0; 2399 ki->p_pctcpu = p->p_pctcpu; 2400 ki->p_estcpu = 0; 2401 ki->p_stat = p->p_stat; /* Will likely be overridden by LWP status */ 2402 ki->p_realstat = p->p_stat; 2403 ki->p_nice = p->p_nice; 2404 ki->p_xstat = P_WAITSTATUS(p); 2405 ki->p_acflag = p->p_acflag; 2406 2407 strncpy(ki->p_comm, p->p_comm, 2408 uimin(sizeof(ki->p_comm), sizeof(p->p_comm))); 2409 strncpy(ki->p_ename, p->p_emul->e_name, sizeof(ki->p_ename)); 2410 2411 ki->p_nlwps = p->p_nlwps; 2412 ki->p_realflag = ki->p_flag; 2413 2414 if (p->p_stat != SIDL && !P_ZOMBIE(p) && !zombie) { 2415 vm = p->p_vmspace; 2416 ki->p_vm_rssize = vm_resident_count(vm); 2417 ki->p_vm_tsize = vm->vm_tsize; 2418 ki->p_vm_dsize = vm->vm_dsize; 2419 ki->p_vm_ssize = vm->vm_ssize; 2420 ki->p_vm_vsize = atop(vm->vm_map.size); 2421 /* 2422 * Since the stack is initially mapped mostly with 2423 * PROT_NONE and grown as needed, adjust the "mapped size" 2424 * to skip the unused stack portion. 2425 */ 2426 ki->p_vm_msize = 2427 atop(vm->vm_map.size) - vm->vm_issize + vm->vm_ssize; 2428 2429 /* Pick the primary (first) LWP */ 2430 l = proc_active_lwp(p); 2431 KASSERT(l != NULL); 2432 lwp_lock(l); 2433 ki->p_nrlwps = p->p_nrlwps; 2434 ki->p_forw = 0; 2435 ki->p_back = 0; 2436 COND_SET_VALUE(ki->p_addr, PTRTOUINT64(l->l_addr), allowaddr); 2437 ki->p_stat = l->l_stat; 2438 ki->p_flag |= sysctl_map_flags(sysctl_lwpflagmap, l->l_flag); 2439 ki->p_swtime = l->l_swtime; 2440 ki->p_slptime = l->l_slptime; 2441 if (l->l_stat == LSONPROC) 2442 ki->p_schedflags = l->l_cpu->ci_schedstate.spc_flags; 2443 else 2444 ki->p_schedflags = 0; 2445 ki->p_priority = lwp_eprio(l); 2446 ki->p_usrpri = l->l_priority; 2447 if (l->l_wchan) 2448 strncpy(ki->p_wmesg, l->l_wmesg, sizeof(ki->p_wmesg)); 2449 COND_SET_VALUE(ki->p_wchan, PTRTOUINT64(l->l_wchan), allowaddr); 2450 ki->p_cpuid = cpu_index(l->l_cpu); 2451 lwp_unlock(l); 2452 LIST_FOREACH(l, &p->p_lwps, l_sibling) { 2453 /* This is hardly correct, but... */ 2454 sigplusset(&l->l_sigpend.sp_set, &ss1); 2455 sigplusset(&l->l_sigmask, &ss2); 2456 ki->p_cpticks += l->l_cpticks; 2457 ki->p_pctcpu += l->l_pctcpu; 2458 ki->p_estcpu += l->l_estcpu; 2459 } 2460 } 2461 sigplusset(&p->p_sigpend.sp_set, &ss2); 2462 memcpy(&ki->p_siglist, &ss1, sizeof(ki_sigset_t)); 2463 memcpy(&ki->p_sigmask, &ss2, sizeof(ki_sigset_t)); 2464 2465 if (p->p_session != NULL) { 2466 ki->p_sid = p->p_session->s_sid; 2467 ki->p__pgid = p->p_pgrp->pg_id; 2468 if (p->p_session->s_ttyvp) 2469 ki->p_eflag |= EPROC_CTTY; 2470 if (SESS_LEADER(p)) 2471 ki->p_eflag |= EPROC_SLEADER; 2472 strncpy(ki->p_login, p->p_session->s_login, 2473 uimin(sizeof ki->p_login - 1, sizeof p->p_session->s_login)); 2474 ki->p_jobc = p->p_pgrp->pg_jobc; 2475 if ((p->p_lflag & PL_CONTROLT) && (tp = p->p_session->s_ttyp)) { 2476 ki->p_tdev = tp->t_dev; 2477 ki->p_tpgid = tp->t_pgrp ? tp->t_pgrp->pg_id : NO_PGID; 2478 COND_SET_VALUE(ki->p_tsess, PTRTOUINT64(tp->t_session), 2479 allowaddr); 2480 } else { 2481 ki->p_tdev = (int32_t)NODEV; 2482 } 2483 } 2484 2485 if (!P_ZOMBIE(p) && !zombie) { 2486 ki->p_uvalid = 1; 2487 ki->p_ustart_sec = p->p_stats->p_start.tv_sec; 2488 ki->p_ustart_usec = p->p_stats->p_start.tv_usec; 2489 2490 calcru(p, &ut, &st, NULL, &rt); 2491 ki->p_rtime_sec = rt.tv_sec; 2492 ki->p_rtime_usec = rt.tv_usec; 2493 ki->p_uutime_sec = ut.tv_sec; 2494 ki->p_uutime_usec = ut.tv_usec; 2495 ki->p_ustime_sec = st.tv_sec; 2496 ki->p_ustime_usec = st.tv_usec; 2497 2498 memcpy(&ru, &p->p_stats->p_ru, sizeof(ru)); 2499 ki->p_uru_nvcsw = 0; 2500 ki->p_uru_nivcsw = 0; 2501 LIST_FOREACH(l2, &p->p_lwps, l_sibling) { 2502 ki->p_uru_nvcsw += (l2->l_ncsw - l2->l_nivcsw); 2503 ki->p_uru_nivcsw += l2->l_nivcsw; 2504 ruadd(&ru, &l2->l_ru); 2505 } 2506 ki->p_uru_maxrss = ru.ru_maxrss; 2507 ki->p_uru_ixrss = ru.ru_ixrss; 2508 ki->p_uru_idrss = ru.ru_idrss; 2509 ki->p_uru_isrss = ru.ru_isrss; 2510 ki->p_uru_minflt = ru.ru_minflt; 2511 ki->p_uru_majflt = ru.ru_majflt; 2512 ki->p_uru_nswap = ru.ru_nswap; 2513 ki->p_uru_inblock = ru.ru_inblock; 2514 ki->p_uru_oublock = ru.ru_oublock; 2515 ki->p_uru_msgsnd = ru.ru_msgsnd; 2516 ki->p_uru_msgrcv = ru.ru_msgrcv; 2517 ki->p_uru_nsignals = ru.ru_nsignals; 2518 2519 timeradd(&p->p_stats->p_cru.ru_utime, 2520 &p->p_stats->p_cru.ru_stime, &ut); 2521 ki->p_uctime_sec = ut.tv_sec; 2522 ki->p_uctime_usec = ut.tv_usec; 2523 } 2524 } 2525 2526 2527 int 2528 proc_find_locked(struct lwp *l, struct proc **p, pid_t pid) 2529 { 2530 int error; 2531 2532 mutex_enter(proc_lock); 2533 if (pid == -1) 2534 *p = l->l_proc; 2535 else 2536 *p = proc_find(pid); 2537 2538 if (*p == NULL) { 2539 if (pid != -1) 2540 mutex_exit(proc_lock); 2541 return ESRCH; 2542 } 2543 if (pid != -1) 2544 mutex_enter((*p)->p_lock); 2545 mutex_exit(proc_lock); 2546 2547 error = kauth_authorize_process(l->l_cred, 2548 KAUTH_PROCESS_CANSEE, *p, 2549 KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_ENTRY), NULL, NULL); 2550 if (error) { 2551 if (pid != -1) 2552 mutex_exit((*p)->p_lock); 2553 } 2554 return error; 2555 } 2556 2557 static int 2558 fill_pathname(struct lwp *l, pid_t pid, void *oldp, size_t *oldlenp) 2559 { 2560 int error; 2561 struct proc *p; 2562 2563 if ((error = proc_find_locked(l, &p, pid)) != 0) 2564 return error; 2565 2566 if (p->p_path == NULL) { 2567 if (pid != -1) 2568 mutex_exit(p->p_lock); 2569 return ENOENT; 2570 } 2571 2572 size_t len = strlen(p->p_path) + 1; 2573 if (oldp != NULL) { 2574 size_t copylen = uimin(len, *oldlenp); 2575 error = sysctl_copyout(l, p->p_path, oldp, copylen); 2576 if (error == 0 && *oldlenp < len) 2577 error = ENOSPC; 2578 } 2579 *oldlenp = len; 2580 if (pid != -1) 2581 mutex_exit(p->p_lock); 2582 return error; 2583 } 2584 2585 static int 2586 fill_cwd(struct lwp *l, pid_t pid, void *oldp, size_t *oldlenp) 2587 { 2588 int error; 2589 struct proc *p; 2590 char *path; 2591 char *bp, *bend; 2592 struct cwdinfo *cwdi; 2593 struct vnode *vp; 2594 size_t len, lenused; 2595 2596 if ((error = proc_find_locked(l, &p, pid)) != 0) 2597 return error; 2598 2599 len = MAXPATHLEN * 4; 2600 2601 path = kmem_alloc(len, KM_SLEEP); 2602 2603 bp = &path[len]; 2604 bend = bp; 2605 *(--bp) = '\0'; 2606 2607 cwdi = p->p_cwdi; 2608 rw_enter(&cwdi->cwdi_lock, RW_READER); 2609 vp = cwdi->cwdi_cdir; 2610 error = getcwd_common(vp, NULL, &bp, path, len/2, 0, l); 2611 rw_exit(&cwdi->cwdi_lock); 2612 2613 if (error) 2614 goto out; 2615 2616 lenused = bend - bp; 2617 2618 if (oldp != NULL) { 2619 size_t copylen = uimin(lenused, *oldlenp); 2620 error = sysctl_copyout(l, bp, oldp, copylen); 2621 if (error == 0 && *oldlenp < lenused) 2622 error = ENOSPC; 2623 } 2624 *oldlenp = lenused; 2625 out: 2626 if (pid != -1) 2627 mutex_exit(p->p_lock); 2628 kmem_free(path, len); 2629 return error; 2630 } 2631 2632 int 2633 proc_getauxv(struct proc *p, void **buf, size_t *len) 2634 { 2635 struct ps_strings pss; 2636 int error; 2637 void *uauxv, *kauxv; 2638 size_t size; 2639 2640 if ((error = copyin_psstrings(p, &pss)) != 0) 2641 return error; 2642 if (pss.ps_envstr == NULL) 2643 return EIO; 2644 2645 size = p->p_execsw->es_arglen; 2646 if (size == 0) 2647 return EIO; 2648 2649 size_t ptrsz = PROC_PTRSZ(p); 2650 uauxv = (void *)((char *)pss.ps_envstr + (pss.ps_nenvstr + 1) * ptrsz); 2651 2652 kauxv = kmem_alloc(size, KM_SLEEP); 2653 2654 error = copyin_proc(p, uauxv, kauxv, size); 2655 if (error) { 2656 kmem_free(kauxv, size); 2657 return error; 2658 } 2659 2660 *buf = kauxv; 2661 *len = size; 2662 2663 return 0; 2664 } 2665 2666 2667 static int 2668 sysctl_security_expose_address(SYSCTLFN_ARGS) 2669 { 2670 int expose_address, error; 2671 struct sysctlnode node; 2672 2673 node = *rnode; 2674 node.sysctl_data = &expose_address; 2675 expose_address = *(int *)rnode->sysctl_data; 2676 error = sysctl_lookup(SYSCTLFN_CALL(&node)); 2677 if (error || newp == NULL) 2678 return error; 2679 2680 if (kauth_authorize_system(l->l_cred, KAUTH_SYSTEM_KERNADDR, 2681 0, NULL, NULL, NULL)) 2682 return EPERM; 2683 2684 switch (expose_address) { 2685 case 0: 2686 case 1: 2687 case 2: 2688 break; 2689 default: 2690 return EINVAL; 2691 } 2692 2693 *(int *)rnode->sysctl_data = expose_address; 2694 2695 return 0; 2696 } 2697 2698 bool 2699 get_expose_address(struct proc *p) 2700 { 2701 /* allow only if sysctl variable is set or privileged */ 2702 return kauth_authorize_process(kauth_cred_get(), KAUTH_PROCESS_CANSEE, 2703 p, KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_KPTR), NULL, NULL) == 0; 2704 } 2705
Indexes created Sun Sep 21 16:09:51 GMT 2025