1 /* $NetBSD: kern_exit.c,v 1.244 2014/05/05 15:45:32 christos Exp $ */ 2 3 /*- 4 * Copyright (c) 1998, 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 * (c) UNIX System Laboratories, Inc. 37 * All or some portions of this file are derived from material licensed 38 * to the University of California by American Telephone and Telegraph 39 * Co. or Unix System Laboratories, Inc. and are reproduced herein with 40 * the permission of UNIX System Laboratories, Inc. 41 * 42 * Redistribution and use in source and binary forms, with or without 43 * modification, are permitted provided that the following conditions 44 * are met: 45 * 1. Redistributions of source code must retain the above copyright 46 * notice, this list of conditions and the following disclaimer. 47 * 2. Redistributions in binary form must reproduce the above copyright 48 * notice, this list of conditions and the following disclaimer in the 49 * documentation and/or other materials provided with the distribution. 50 * 3. Neither the name of the University nor the names of its contributors 51 * may be used to endorse or promote products derived from this software 52 * without specific prior written permission. 53 * 54 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 55 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 56 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 57 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 58 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 59 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 60 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 61 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 62 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 63 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 64 * SUCH DAMAGE. 65 * 66 * @(#)kern_exit.c 8.10 (Berkeley) 2/23/95 67 */ 68 69 #include <sys/cdefs.h> 70 __KERNEL_RCSID(0, "$NetBSD: kern_exit.c,v 1.244 2014/05/05 15:45:32 christos Exp $"); 71 72 #include "opt_ktrace.h" 73 #include "opt_perfctrs.h" 74 #include "opt_sysv.h" 75 76 #include <sys/param.h> 77 #include <sys/systm.h> 78 #include <sys/ioctl.h> 79 #include <sys/tty.h> 80 #include <sys/time.h> 81 #include <sys/resource.h> 82 #include <sys/kernel.h> 83 #include <sys/proc.h> 84 #include <sys/buf.h> 85 #include <sys/wait.h> 86 #include <sys/file.h> 87 #include <sys/vnode.h> 88 #include <sys/syslog.h> 89 #include <sys/pool.h> 90 #include <sys/uidinfo.h> 91 #if defined(PERFCTRS) 92 #include <sys/pmc.h> 93 #endif 94 #include <sys/ptrace.h> 95 #include <sys/acct.h> 96 #include <sys/filedesc.h> 97 #include <sys/ras.h> 98 #include <sys/signalvar.h> 99 #include <sys/sched.h> 100 #include <sys/mount.h> 101 #include <sys/syscallargs.h> 102 #include <sys/kauth.h> 103 #include <sys/sleepq.h> 104 #include <sys/lockdebug.h> 105 #include <sys/ktrace.h> 106 #include <sys/cpu.h> 107 #include <sys/lwpctl.h> 108 #include <sys/atomic.h> 109 #include <sys/sdt.h> 110 111 #include <uvm/uvm_extern.h> 112 113 #ifdef DEBUG_EXIT 114 int debug_exit = 0; 115 #define DPRINTF(x) if (debug_exit) printf x 116 #else 117 #define DPRINTF(x) 118 #endif 119 120 static int find_stopped_child(struct proc *, pid_t, int, struct proc **, int *); 121 static void proc_free(struct proc *, struct rusage *); 122 123 /* 124 * DTrace SDT provider definitions 125 */ 126 SDT_PROBE_DEFINE(proc,,,exit,exit, 127 "int", NULL, /* reason */ 128 NULL, NULL, NULL, NULL, 129 NULL, NULL, NULL, NULL); 130 /* 131 * Fill in the appropriate signal information, and signal the parent. 132 */ 133 /* XXX noclone works around a gcc 4.5 bug on arm */ 134 static void __noclone 135 exit_psignal(struct proc *p, struct proc *pp, ksiginfo_t *ksi) 136 { 137 138 KSI_INIT(ksi); 139 if ((ksi->ksi_signo = P_EXITSIG(p)) == SIGCHLD) { 140 if (WIFSIGNALED(p->p_xstat)) { 141 if (WCOREDUMP(p->p_xstat)) 142 ksi->ksi_code = CLD_DUMPED; 143 else 144 ksi->ksi_code = CLD_KILLED; 145 } else { 146 ksi->ksi_code = CLD_EXITED; 147 } 148 } 149 /* 150 * We fill those in, even for non-SIGCHLD. 151 * It's safe to access p->p_cred unlocked here. 152 */ 153 ksi->ksi_pid = p->p_pid; 154 ksi->ksi_uid = kauth_cred_geteuid(p->p_cred); 155 ksi->ksi_status = p->p_xstat; 156 /* XXX: is this still valid? */ 157 ksi->ksi_utime = p->p_stats->p_ru.ru_utime.tv_sec; 158 ksi->ksi_stime = p->p_stats->p_ru.ru_stime.tv_sec; 159 } 160 161 /* 162 * exit -- 163 * Death of process. 164 */ 165 int 166 sys_exit(struct lwp *l, const struct sys_exit_args *uap, register_t *retval) 167 { 168 /* { 169 syscallarg(int) rval; 170 } */ 171 struct proc *p = l->l_proc; 172 173 /* Don't call exit1() multiple times in the same process. */ 174 mutex_enter(p->p_lock); 175 if (p->p_sflag & PS_WEXIT) { 176 mutex_exit(p->p_lock); 177 lwp_exit(l); 178 } 179 180 /* exit1() will release the mutex. */ 181 exit1(l, W_EXITCODE(SCARG(uap, rval), 0)); 182 /* NOTREACHED */ 183 return (0); 184 } 185 186 /* 187 * Exit: deallocate address space and other resources, change proc state 188 * to zombie, and unlink proc from allproc and parent's lists. Save exit 189 * status and rusage for wait(). Check for child processes and orphan them. 190 * 191 * Must be called with p->p_lock held. Does not return. 192 */ 193 void 194 exit1(struct lwp *l, int rv) 195 { 196 struct proc *p, *child, *next_child, *old_parent, *new_parent; 197 struct pgrp *pgrp; 198 ksiginfo_t ksi; 199 ksiginfoq_t kq; 200 int wakeinit; 201 202 p = l->l_proc; 203 204 KASSERT(mutex_owned(p->p_lock)); 205 KASSERT(p->p_vmspace != NULL); 206 207 if (__predict_false(p == initproc)) { 208 panic("init died (signal %d, exit %d)", 209 WTERMSIG(rv), WEXITSTATUS(rv)); 210 } 211 212 p->p_sflag |= PS_WEXIT; 213 214 /* 215 * Force all other LWPs to exit before we do. Only then can we 216 * begin to tear down the rest of the process state. 217 */ 218 if (p->p_nlwps > 1) { 219 exit_lwps(l); 220 } 221 222 ksiginfo_queue_init(&kq); 223 224 /* 225 * If we have been asked to stop on exit, do so now. 226 */ 227 if (__predict_false(p->p_sflag & PS_STOPEXIT)) { 228 KERNEL_UNLOCK_ALL(l, &l->l_biglocks); 229 sigclearall(p, &contsigmask, &kq); 230 p->p_waited = 0; 231 membar_producer(); 232 p->p_stat = SSTOP; 233 lwp_lock(l); 234 p->p_nrlwps--; 235 l->l_stat = LSSTOP; 236 lwp_unlock(l); 237 mutex_exit(p->p_lock); 238 lwp_lock(l); 239 mi_switch(l); 240 KERNEL_LOCK(l->l_biglocks, l); 241 mutex_enter(p->p_lock); 242 } 243 244 /* 245 * Bin any remaining signals and mark the process as dying so it will 246 * not be found for, e.g. signals. 247 */ 248 sigfillset(&p->p_sigctx.ps_sigignore); 249 sigclearall(p, NULL, &kq); 250 p->p_stat = SDYING; 251 mutex_exit(p->p_lock); 252 ksiginfo_queue_drain(&kq); 253 254 /* Destroy any lwpctl info. */ 255 if (p->p_lwpctl != NULL) 256 lwp_ctl_exit(); 257 258 /* 259 * Drain all remaining references that procfs, ptrace and others may 260 * have on the process. 261 */ 262 rw_enter(&p->p_reflock, RW_WRITER); 263 264 DPRINTF(("exit1: %d.%d exiting.\n", p->p_pid, l->l_lid)); 265 266 timers_free(p, TIMERS_ALL); 267 #if defined(__HAVE_RAS) 268 ras_purgeall(); 269 #endif 270 271 /* 272 * Close open files, release open-file table and free signal 273 * actions. This may block! 274 */ 275 fd_free(); 276 cwdfree(p->p_cwdi); 277 p->p_cwdi = NULL; 278 doexithooks(p); 279 sigactsfree(p->p_sigacts); 280 281 /* 282 * Write out accounting data. 283 */ 284 (void)acct_process(l); 285 286 #ifdef KTRACE 287 /* 288 * Release trace file. 289 */ 290 if (p->p_tracep != NULL) { 291 mutex_enter(&ktrace_lock); 292 ktrderef(p); 293 mutex_exit(&ktrace_lock); 294 } 295 #endif 296 297 /* 298 * If emulation has process exit hook, call it now. 299 * Set the exit status now so that the exit hook has 300 * an opportunity to tweak it (COMPAT_LINUX requires 301 * this for thread group emulation) 302 */ 303 p->p_xstat = rv; 304 if (p->p_emul->e_proc_exit) 305 (*p->p_emul->e_proc_exit)(p); 306 307 /* 308 * Free the VM resources we're still holding on to. 309 * We must do this from a valid thread because doing 310 * so may block. This frees vmspace, which we don't 311 * need anymore. The only remaining lwp is the one 312 * we run at this moment, nothing runs in userland 313 * anymore. 314 */ 315 uvm_proc_exit(p); 316 317 /* 318 * Stop profiling. 319 */ 320 if (__predict_false((p->p_stflag & PST_PROFIL) != 0)) { 321 mutex_spin_enter(&p->p_stmutex); 322 stopprofclock(p); 323 mutex_spin_exit(&p->p_stmutex); 324 } 325 326 /* 327 * If parent is waiting for us to exit or exec, PL_PPWAIT is set; we 328 * wake up the parent early to avoid deadlock. We can do this once 329 * the VM resources are released. 330 */ 331 mutex_enter(proc_lock); 332 if (p->p_lflag & PL_PPWAIT) { 333 #if 0 334 lwp_t *lp; 335 336 l->l_lwpctl = NULL; /* was on loan from blocked parent */ 337 p->p_lflag &= ~PL_PPWAIT; 338 339 lp = p->p_vforklwp; 340 p->p_vforklwp = NULL; 341 lp->l_pflag &= ~LP_VFORKWAIT; /* XXX */ 342 cv_broadcast(&lp->l_waitcv); 343 #else 344 l->l_lwpctl = NULL; /* was on loan from blocked parent */ 345 p->p_lflag &= ~PL_PPWAIT; 346 cv_broadcast(&p->p_pptr->p_waitcv); 347 #endif 348 } 349 350 if (SESS_LEADER(p)) { 351 struct vnode *vprele = NULL, *vprevoke = NULL; 352 struct session *sp = p->p_session; 353 struct tty *tp; 354 355 if (sp->s_ttyvp) { 356 /* 357 * Controlling process. 358 * Signal foreground pgrp, 359 * drain controlling terminal 360 * and revoke access to controlling terminal. 361 */ 362 tp = sp->s_ttyp; 363 mutex_spin_enter(&tty_lock); 364 if (tp->t_session == sp) { 365 /* we can't guarantee the revoke will do this */ 366 pgrp = tp->t_pgrp; 367 tp->t_pgrp = NULL; 368 tp->t_session = NULL; 369 mutex_spin_exit(&tty_lock); 370 if (pgrp != NULL) { 371 pgsignal(pgrp, SIGHUP, 1); 372 } 373 mutex_exit(proc_lock); 374 (void) ttywait(tp); 375 mutex_enter(proc_lock); 376 377 /* The tty could have been revoked. */ 378 vprevoke = sp->s_ttyvp; 379 } else 380 mutex_spin_exit(&tty_lock); 381 vprele = sp->s_ttyvp; 382 sp->s_ttyvp = NULL; 383 /* 384 * s_ttyp is not zero'd; we use this to indicate 385 * that the session once had a controlling terminal. 386 * (for logging and informational purposes) 387 */ 388 } 389 sp->s_leader = NULL; 390 391 if (vprevoke != NULL || vprele != NULL) { 392 if (vprevoke != NULL) { 393 /* Releases proc_lock. */ 394 proc_sessrele(sp); 395 VOP_REVOKE(vprevoke, REVOKEALL); 396 } else 397 mutex_exit(proc_lock); 398 if (vprele != NULL) 399 vrele(vprele); 400 mutex_enter(proc_lock); 401 } 402 } 403 fixjobc(p, p->p_pgrp, 0); 404 405 /* 406 * Finalize the last LWP's specificdata, as well as the 407 * specificdata for the proc itself. 408 */ 409 lwp_finispecific(l); 410 proc_finispecific(p); 411 412 /* 413 * Notify interested parties of our demise. 414 */ 415 KNOTE(&p->p_klist, NOTE_EXIT); 416 417 SDT_PROBE(proc,,,exit, 418 (WCOREDUMP(rv) ? CLD_DUMPED : 419 (WIFSIGNALED(rv) ? CLD_KILLED : CLD_EXITED)), 420 0,0,0,0); 421 422 #if PERFCTRS 423 /* 424 * Save final PMC information in parent process & clean up. 425 */ 426 if (PMC_ENABLED(p)) { 427 pmc_save_context(p); 428 pmc_accumulate(p->p_pptr, p); 429 pmc_process_exit(p); 430 } 431 #endif 432 433 /* 434 * Reset p_opptr pointer of all former children which got 435 * traced by another process and were reparented. We reset 436 * it to NULL here; the trace detach code then reparents 437 * the child to initproc. We only check allproc list, since 438 * eventual former children on zombproc list won't reference 439 * p_opptr anymore. 440 */ 441 if (__predict_false(p->p_slflag & PSL_CHTRACED)) { 442 struct proc *q; 443 PROCLIST_FOREACH(q, &allproc) { 444 if (q->p_opptr == p) 445 q->p_opptr = NULL; 446 } 447 } 448 449 /* 450 * Give orphaned children to init(8). 451 */ 452 child = LIST_FIRST(&p->p_children); 453 wakeinit = (child != NULL); 454 for (; child != NULL; child = next_child) { 455 next_child = LIST_NEXT(child, p_sibling); 456 457 /* 458 * Traced processes are killed since their existence 459 * means someone is screwing up. Since we reset the 460 * trace flags, the logic in sys_wait4() would not be 461 * triggered to reparent the process to its 462 * original parent, so we must do this here. 463 */ 464 if (__predict_false(child->p_slflag & PSL_TRACED)) { 465 mutex_enter(p->p_lock); 466 child->p_slflag &= 467 ~(PSL_TRACED|PSL_FSTRACE|PSL_SYSCALL); 468 mutex_exit(p->p_lock); 469 if (child->p_opptr != child->p_pptr) { 470 struct proc *t = child->p_opptr; 471 proc_reparent(child, t ? t : initproc); 472 child->p_opptr = NULL; 473 } else 474 proc_reparent(child, initproc); 475 killproc(child, "orphaned traced process"); 476 } else 477 proc_reparent(child, initproc); 478 } 479 480 /* 481 * Move proc from allproc to zombproc, it's now nearly ready to be 482 * collected by parent. 483 */ 484 LIST_REMOVE(l, l_list); 485 LIST_REMOVE(p, p_list); 486 LIST_INSERT_HEAD(&zombproc, p, p_list); 487 488 /* 489 * Mark the process as dead. We must do this before we signal 490 * the parent. 491 */ 492 p->p_stat = SDEAD; 493 494 /* Put in front of parent's sibling list for parent to collect it */ 495 old_parent = p->p_pptr; 496 old_parent->p_nstopchild++; 497 if (LIST_FIRST(&old_parent->p_children) != p) { 498 /* Put child where it can be found quickly */ 499 LIST_REMOVE(p, p_sibling); 500 LIST_INSERT_HEAD(&old_parent->p_children, p, p_sibling); 501 } 502 503 /* 504 * Notify parent that we're gone. If parent has the P_NOCLDWAIT 505 * flag set, notify init instead (and hope it will handle 506 * this situation). 507 */ 508 if (old_parent->p_flag & (PK_NOCLDWAIT|PK_CLDSIGIGN)) { 509 proc_reparent(p, initproc); 510 wakeinit = 1; 511 512 /* 513 * If this was the last child of our parent, notify 514 * parent, so in case he was wait(2)ing, he will 515 * continue. 516 */ 517 if (LIST_FIRST(&old_parent->p_children) == NULL) 518 cv_broadcast(&old_parent->p_waitcv); 519 } 520 521 /* Reload parent pointer, since p may have been reparented above */ 522 new_parent = p->p_pptr; 523 524 if (__predict_false((p->p_slflag & PSL_FSTRACE) == 0 && 525 p->p_exitsig != 0)) { 526 exit_psignal(p, new_parent, &ksi); 527 kpsignal(new_parent, &ksi, NULL); 528 } 529 530 /* Calculate the final rusage info. */ 531 calcru(p, &p->p_stats->p_ru.ru_utime, &p->p_stats->p_ru.ru_stime, 532 NULL, NULL); 533 534 if (wakeinit) 535 cv_broadcast(&initproc->p_waitcv); 536 537 callout_destroy(&l->l_timeout_ch); 538 539 /* 540 * Release any PCU resources before becoming a zombie. 541 */ 542 pcu_discard_all(l); 543 544 mutex_enter(p->p_lock); 545 /* Free the linux lwp id */ 546 if ((l->l_pflag & LP_PIDLID) != 0 && l->l_lid != p->p_pid) 547 proc_free_pid(l->l_lid); 548 lwp_drainrefs(l); 549 lwp_lock(l); 550 l->l_prflag &= ~LPR_DETACHED; 551 l->l_stat = LSZOMB; 552 lwp_unlock(l); 553 KASSERT(curlwp == l); 554 KASSERT(p->p_nrlwps == 1); 555 KASSERT(p->p_nlwps == 1); 556 p->p_stat = SZOMB; 557 p->p_nrlwps--; 558 p->p_nzlwps++; 559 p->p_ndlwps = 0; 560 mutex_exit(p->p_lock); 561 562 /* 563 * Signal the parent to collect us, and drop the proclist lock. 564 * Drop debugger/procfs lock; no new references can be gained. 565 */ 566 cv_broadcast(&p->p_pptr->p_waitcv); 567 rw_exit(&p->p_reflock); 568 mutex_exit(proc_lock); 569 570 /* Verify that we hold no locks other than the kernel lock. */ 571 LOCKDEBUG_BARRIER(&kernel_lock, 0); 572 573 /* 574 * NOTE: WE ARE NO LONGER ALLOWED TO SLEEP! 575 */ 576 577 /* 578 * Give machine-dependent code a chance to free any MD LWP 579 * resources. This must be done before uvm_lwp_exit(), in 580 * case these resources are in the PCB. 581 */ 582 cpu_lwp_free(l, 1); 583 584 pmap_deactivate(l); 585 586 /* This process no longer needs to hold the kernel lock. */ 587 #ifdef notyet 588 /* XXXSMP hold in lwp_userret() */ 589 KERNEL_UNLOCK_LAST(l); 590 #else 591 KERNEL_UNLOCK_ALL(l, NULL); 592 #endif 593 594 lwp_exit_switchaway(l); 595 } 596 597 void 598 exit_lwps(struct lwp *l) 599 { 600 proc_t *p = l->l_proc; 601 lwp_t *l2; 602 int nlocks; 603 604 KERNEL_UNLOCK_ALL(l, &nlocks); 605 retry: 606 KASSERT(mutex_owned(p->p_lock)); 607 608 /* 609 * Interrupt LWPs in interruptable sleep, unsuspend suspended 610 * LWPs and then wait for everyone else to finish. 611 */ 612 LIST_FOREACH(l2, &p->p_lwps, l_sibling) { 613 if (l2 == l) 614 continue; 615 lwp_lock(l2); 616 l2->l_flag |= LW_WEXIT; 617 if ((l2->l_stat == LSSLEEP && (l2->l_flag & LW_SINTR)) || 618 l2->l_stat == LSSUSPENDED || l2->l_stat == LSSTOP) { 619 /* setrunnable() will release the lock. */ 620 setrunnable(l2); 621 continue; 622 } 623 lwp_unlock(l2); 624 } 625 626 /* 627 * Wait for every LWP to exit. Note: LWPs can get suspended/slept 628 * behind us or there may even be new LWPs created. Therefore, a 629 * full retry is required on error. 630 */ 631 while (p->p_nlwps > 1) { 632 if (lwp_wait(l, 0, NULL, true)) { 633 goto retry; 634 } 635 } 636 637 KERNEL_LOCK(nlocks, l); 638 KASSERT(p->p_nlwps == 1); 639 } 640 641 int 642 do_sys_wait(int *pid, int *status, int options, struct rusage *ru) 643 { 644 proc_t *child; 645 int error; 646 647 if (ru != NULL) { 648 memset(ru, 0, sizeof(*ru)); 649 } 650 mutex_enter(proc_lock); 651 error = find_stopped_child(curproc, *pid, options, &child, status); 652 if (child == NULL) { 653 mutex_exit(proc_lock); 654 *pid = 0; 655 return error; 656 } 657 *pid = child->p_pid; 658 659 if (child->p_stat == SZOMB) { 660 /* proc_free() will release the proc_lock. */ 661 if (options & WNOWAIT) { 662 mutex_exit(proc_lock); 663 } else { 664 proc_free(child, ru); 665 } 666 } else { 667 /* Child state must have been SSTOP. */ 668 mutex_exit(proc_lock); 669 *status = W_STOPCODE(*status); 670 } 671 return 0; 672 } 673 674 int 675 sys___wait450(struct lwp *l, const struct sys___wait450_args *uap, 676 register_t *retval) 677 { 678 /* { 679 syscallarg(int) pid; 680 syscallarg(int *) status; 681 syscallarg(int) options; 682 syscallarg(struct rusage *) rusage; 683 } */ 684 int error, status, pid = SCARG(uap, pid); 685 struct rusage ru; 686 687 error = do_sys_wait(&pid, &status, SCARG(uap, options), 688 SCARG(uap, rusage) != NULL ? &ru : NULL); 689 690 retval[0] = pid; 691 if (pid == 0) { 692 return error; 693 } 694 if (SCARG(uap, status)) { 695 error = copyout(&status, SCARG(uap, status), sizeof(status)); 696 } 697 if (SCARG(uap, rusage) && error == 0) { 698 error = copyout(&ru, SCARG(uap, rusage), sizeof(ru)); 699 } 700 return error; 701 } 702 703 /* 704 * Scan list of child processes for a child process that has stopped or 705 * exited. Used by sys_wait4 and 'compat' equivalents. 706 * 707 * Must be called with the proc_lock held, and may release while waiting. 708 */ 709 static int 710 find_stopped_child(struct proc *parent, pid_t pid, int options, 711 struct proc **child_p, int *status_p) 712 { 713 struct proc *child, *dead; 714 int error; 715 716 KASSERT(mutex_owned(proc_lock)); 717 718 if (options & ~(WUNTRACED|WNOHANG|WALTSIG|WALLSIG) 719 && !(options & WOPTSCHECKED)) { 720 *child_p = NULL; 721 return EINVAL; 722 } 723 724 if (pid == 0 && !(options & WOPTSCHECKED)) 725 pid = -parent->p_pgid; 726 727 for (;;) { 728 error = ECHILD; 729 dead = NULL; 730 731 LIST_FOREACH(child, &parent->p_children, p_sibling) { 732 if (pid >= 0) { 733 if (child->p_pid != pid) { 734 child = proc_find_raw(pid); 735 if (child == NULL || 736 child->p_stat == SIDL || 737 child->p_pptr != parent) { 738 child = NULL; 739 break; 740 } 741 } 742 } else if (pid != WAIT_ANY && child->p_pgid != -pid) { 743 /* Child not in correct pgrp */ 744 continue; 745 } 746 747 /* 748 * Wait for processes with p_exitsig != SIGCHLD 749 * processes only if WALTSIG is set; wait for 750 * processes with p_exitsig == SIGCHLD only 751 * if WALTSIG is clear. 752 */ 753 if (((options & WALLSIG) == 0) && 754 (options & WALTSIG ? child->p_exitsig == SIGCHLD 755 : P_EXITSIG(child) != SIGCHLD)){ 756 if (child->p_pid == pid) { 757 child = NULL; 758 break; 759 } 760 continue; 761 } 762 763 error = 0; 764 if ((options & WNOZOMBIE) == 0) { 765 if (child->p_stat == SZOMB) 766 break; 767 if (child->p_stat == SDEAD) { 768 /* 769 * We may occasionally arrive here 770 * after receiving a signal, but 771 * immediately before the child 772 * process is zombified. The wait 773 * will be short, so avoid returning 774 * to userspace. 775 */ 776 dead = child; 777 } 778 } 779 780 if (child->p_stat == SSTOP && 781 child->p_waited == 0 && 782 (child->p_slflag & PSL_TRACED || 783 options & WUNTRACED)) { 784 if ((options & WNOWAIT) == 0) { 785 child->p_waited = 1; 786 parent->p_nstopchild--; 787 } 788 break; 789 } 790 if (parent->p_nstopchild == 0 || child->p_pid == pid) { 791 child = NULL; 792 break; 793 } 794 } 795 796 if (child != NULL || error != 0 || 797 ((options & WNOHANG) != 0 && dead == NULL)) { 798 if (child != NULL) { 799 *status_p = child->p_xstat; 800 } 801 *child_p = child; 802 return error; 803 } 804 805 /* 806 * Wait for another child process to stop. 807 */ 808 error = cv_wait_sig(&parent->p_waitcv, proc_lock); 809 810 if (error != 0) { 811 *child_p = NULL; 812 return error; 813 } 814 } 815 } 816 817 /* 818 * Free a process after parent has taken all the state info. Must be called 819 * with the proclist lock held, and will release before returning. 820 * 821 * *ru is returned to the caller, and must be freed by the caller. 822 */ 823 static void 824 proc_free(struct proc *p, struct rusage *ru) 825 { 826 struct proc *parent = p->p_pptr; 827 struct lwp *l; 828 ksiginfo_t ksi; 829 kauth_cred_t cred1, cred2; 830 uid_t uid; 831 832 KASSERT(mutex_owned(proc_lock)); 833 KASSERT(p->p_nlwps == 1); 834 KASSERT(p->p_nzlwps == 1); 835 KASSERT(p->p_nrlwps == 0); 836 KASSERT(p->p_stat == SZOMB); 837 838 /* 839 * If we got the child via ptrace(2) or procfs, and 840 * the parent is different (meaning the process was 841 * attached, rather than run as a child), then we need 842 * to give it back to the old parent, and send the 843 * parent the exit signal. The rest of the cleanup 844 * will be done when the old parent waits on the child. 845 */ 846 if ((p->p_slflag & PSL_TRACED) != 0 && p->p_opptr != parent) { 847 mutex_enter(p->p_lock); 848 p->p_slflag &= ~(PSL_TRACED|PSL_FSTRACE|PSL_SYSCALL); 849 mutex_exit(p->p_lock); 850 parent = (p->p_opptr == NULL) ? initproc : p->p_opptr; 851 proc_reparent(p, parent); 852 p->p_opptr = NULL; 853 if (p->p_exitsig != 0) { 854 exit_psignal(p, parent, &ksi); 855 kpsignal(parent, &ksi, NULL); 856 } 857 cv_broadcast(&parent->p_waitcv); 858 mutex_exit(proc_lock); 859 return; 860 } 861 862 sched_proc_exit(parent, p); 863 864 /* 865 * Add child times of exiting process onto its own times. 866 * This cannot be done any earlier else it might get done twice. 867 */ 868 l = LIST_FIRST(&p->p_lwps); 869 p->p_stats->p_ru.ru_nvcsw += (l->l_ncsw - l->l_nivcsw); 870 p->p_stats->p_ru.ru_nivcsw += l->l_nivcsw; 871 ruadd(&p->p_stats->p_ru, &l->l_ru); 872 ruadd(&p->p_stats->p_ru, &p->p_stats->p_cru); 873 ruadd(&parent->p_stats->p_cru, &p->p_stats->p_ru); 874 if (ru != NULL) 875 *ru = p->p_stats->p_ru; 876 p->p_xstat = 0; 877 878 /* 879 * At this point we are going to start freeing the final resources. 880 * If anyone tries to access the proc structure after here they will 881 * get a shock - bits are missing. Attempt to make it hard! We 882 * don't bother with any further locking past this point. 883 */ 884 p->p_stat = SIDL; /* not even a zombie any more */ 885 LIST_REMOVE(p, p_list); /* off zombproc */ 886 parent->p_nstopchild--; 887 LIST_REMOVE(p, p_sibling); 888 889 /* 890 * Let pid be reallocated. 891 */ 892 proc_free_pid(p->p_pid); 893 894 /* 895 * Unlink process from its process group. 896 * Releases the proc_lock. 897 */ 898 proc_leavepgrp(p); 899 900 /* 901 * Delay release until after lwp_free. 902 */ 903 cred2 = l->l_cred; 904 905 /* 906 * Free the last LWP's resources. 907 * 908 * lwp_free ensures the LWP is no longer running on another CPU. 909 */ 910 lwp_free(l, false, true); 911 912 /* 913 * Now no one except us can reach the process p. 914 */ 915 916 /* 917 * Decrement the count of procs running with this uid. 918 */ 919 cred1 = p->p_cred; 920 uid = kauth_cred_getuid(cred1); 921 (void)chgproccnt(uid, -1); 922 923 /* 924 * Release substructures. 925 */ 926 927 lim_free(p->p_limit); 928 pstatsfree(p->p_stats); 929 kauth_cred_free(cred1); 930 kauth_cred_free(cred2); 931 932 /* 933 * Release reference to text vnode 934 */ 935 if (p->p_textvp) 936 vrele(p->p_textvp); 937 938 mutex_destroy(&p->p_auxlock); 939 mutex_obj_free(p->p_lock); 940 mutex_destroy(&p->p_stmutex); 941 cv_destroy(&p->p_waitcv); 942 cv_destroy(&p->p_lwpcv); 943 rw_destroy(&p->p_reflock); 944 945 proc_free_mem(p); 946 } 947 948 /* 949 * make process 'parent' the new parent of process 'child'. 950 * 951 * Must be called with proc_lock held. 952 */ 953 void 954 proc_reparent(struct proc *child, struct proc *parent) 955 { 956 957 KASSERT(mutex_owned(proc_lock)); 958 959 if (child->p_pptr == parent) 960 return; 961 962 if (child->p_stat == SZOMB || 963 (child->p_stat == SSTOP && !child->p_waited)) { 964 child->p_pptr->p_nstopchild--; 965 parent->p_nstopchild++; 966 } 967 if (parent == initproc) 968 child->p_exitsig = SIGCHLD; 969 970 LIST_REMOVE(child, p_sibling); 971 LIST_INSERT_HEAD(&parent->p_children, child, p_sibling); 972 child->p_pptr = parent; 973 child->p_ppid = parent->p_pid; 974 } 975