1 /* $NetBSD: kern_exit.c,v 1.298 2023/10/08 12:38:58 ad Exp $ */
2
3 /*-
4 * Copyright (c) 1998, 1999, 2006, 2007, 2008, 2020, 2023
5 * The NetBSD Foundation, Inc.
6 * All rights reserved.
7 *
8 * This code is derived from software contributed to The NetBSD Foundation
9 * by Jason R. Thorpe of the Numerical Aerospace Simulation Facility,
10 * NASA Ames Research Center, and by Andrew Doran.
11 *
12 * Redistribution and use in source and binary forms, with or without
13 * modification, are permitted provided that the following conditions
14 * are met:
15 * 1. Redistributions of source code must retain the above copyright
16 * notice, this list of conditions and the following disclaimer.
17 * 2. Redistributions in binary form must reproduce the above copyright
18 * notice, this list of conditions and the following disclaimer in the
19 * documentation and/or other materials provided with the distribution.
20 *
21 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
22 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
23 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
24 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
25 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
26 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
27 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
28 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
29 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
30 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
31 * POSSIBILITY OF SUCH DAMAGE.
32 */
33
34 /*
35 * Copyright (c) 1982, 1986, 1989, 1991, 1993
36 * The Regents of the University of California. All rights reserved.
37 * (c) UNIX System Laboratories, Inc.
38 * All or some portions of this file are derived from material licensed
39 * to the University of California by American Telephone and Telegraph
40 * Co. or Unix System Laboratories, Inc. and are reproduced herein with
41 * the permission of UNIX System Laboratories, Inc.
42 *
43 * Redistribution and use in source and binary forms, with or without
44 * modification, are permitted provided that the following conditions
45 * are met:
46 * 1. Redistributions of source code must retain the above copyright
47 * notice, this list of conditions and the following disclaimer.
48 * 2. Redistributions in binary form must reproduce the above copyright
49 * notice, this list of conditions and the following disclaimer in the
50 * documentation and/or other materials provided with the distribution.
51 * 3. Neither the name of the University nor the names of its contributors
52 * may be used to endorse or promote products derived from this software
53 * without specific prior written permission.
54 *
55 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
56 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
57 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
58 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
59 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
60 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
61 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
62 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
63 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
64 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
65 * SUCH DAMAGE.
66 *
67 * @(#)kern_exit.c 8.10 (Berkeley) 2/23/95
68 */
69
70 #include <sys/cdefs.h>
71 __KERNEL_RCSID(0, "$NetBSD: kern_exit.c,v 1.298 2023/10/08 12:38:58 ad Exp $");
72
73 #include "opt_ktrace.h"
74 #include "opt_dtrace.h"
75 #include "opt_sysv.h"
76
77 #include <sys/param.h>
78 #include <sys/systm.h>
79 #include <sys/ioctl.h>
80 #include <sys/tty.h>
81 #include <sys/time.h>
82 #include <sys/resource.h>
83 #include <sys/kernel.h>
84 #include <sys/proc.h>
85 #include <sys/buf.h>
86 #include <sys/wait.h>
87 #include <sys/file.h>
88 #include <sys/fstrans.h>
89 #include <sys/vnode.h>
90 #include <sys/syslog.h>
91 #include <sys/pool.h>
92 #include <sys/uidinfo.h>
93 #include <sys/ptrace.h>
94 #include <sys/acct.h>
95 #include <sys/filedesc.h>
96 #include <sys/ras.h>
97 #include <sys/signalvar.h>
98 #include <sys/sched.h>
99 #include <sys/mount.h>
100 #include <sys/syscallargs.h>
101 #include <sys/kauth.h>
102 #include <sys/sleepq.h>
103 #include <sys/lock.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 #include <sys/psref.h>
111
112 #include <uvm/uvm_extern.h>
113
114 #ifdef DEBUG_EXIT
115 int debug_exit = 0;
116 #define DPRINTF(x) if (debug_exit) printf x
117 #else
118 #define DPRINTF(x)
119 #endif
120
121 static int find_stopped_child(struct proc *, idtype_t, id_t, int,
122 struct proc **, struct wrusage *, siginfo_t *);
123 static void proc_free(struct proc *, struct wrusage *);
124
125 /*
126 * DTrace SDT provider definitions
127 */
128 SDT_PROVIDER_DECLARE(proc);
129 SDT_PROBE_DEFINE1(proc, kernel, , exit, "int");
130
131 /*
132 * Fill in the appropriate signal information, and signal the parent.
133 */
134 /* XXX noclone works around a gcc 4.5 bug on arm */
135 static void __noclone
exit_psignal(struct proc * p,struct proc * pp,ksiginfo_t * ksi)136 exit_psignal(struct proc *p, struct proc *pp, ksiginfo_t *ksi)
137 {
138
139 KSI_INIT(ksi);
140 if ((ksi->ksi_signo = P_EXITSIG(p)) == SIGCHLD) {
141 if (p->p_xsig) {
142 if (p->p_sflag & PS_COREDUMP)
143 ksi->ksi_code = CLD_DUMPED;
144 else
145 ksi->ksi_code = CLD_KILLED;
146 ksi->ksi_status = p->p_xsig;
147 } else {
148 ksi->ksi_code = CLD_EXITED;
149 ksi->ksi_status = p->p_xexit;
150 }
151 } else {
152 ksi->ksi_code = SI_USER;
153 ksi->ksi_status = p->p_xsig;
154 }
155 /*
156 * We fill those in, even for non-SIGCHLD.
157 * It's safe to access p->p_cred unlocked here.
158 */
159 ksi->ksi_pid = p->p_pid;
160 ksi->ksi_uid = kauth_cred_geteuid(p->p_cred);
161 /* XXX: is this still valid? */
162 ksi->ksi_utime = p->p_stats->p_ru.ru_utime.tv_sec;
163 ksi->ksi_stime = p->p_stats->p_ru.ru_stime.tv_sec;
164 }
165
166 /*
167 * exit --
168 * Death of process.
169 */
170 int
sys_exit(struct lwp * l,const struct sys_exit_args * uap,register_t * retval)171 sys_exit(struct lwp *l, const struct sys_exit_args *uap, register_t *retval)
172 {
173 /* {
174 syscallarg(int) rval;
175 } */
176 struct proc *p = l->l_proc;
177
178 /* Don't call exit1() multiple times in the same process. */
179 mutex_enter(p->p_lock);
180 if (p->p_sflag & PS_WEXIT) {
181 mutex_exit(p->p_lock);
182 lwp_exit(l);
183 }
184
185 /* exit1() will release the mutex. */
186 exit1(l, SCARG(uap, rval), 0);
187 /* NOTREACHED */
188 return (0);
189 }
190
191 /*
192 * Exit: deallocate address space and other resources, change proc state
193 * to zombie, and unlink proc from allproc and parent's lists. Save exit
194 * status and rusage for wait(). Check for child processes and orphan them.
195 *
196 * Must be called with p->p_lock held. Does not return.
197 */
198 void
exit1(struct lwp * l,int exitcode,int signo)199 exit1(struct lwp *l, int exitcode, int signo)
200 {
201 struct proc *p, *child, *next_child, *old_parent, *new_parent;
202 struct pgrp *pgrp;
203 ksiginfo_t ksi;
204 ksiginfoq_t kq;
205 int wakeinit;
206
207 p = l->l_proc;
208
209 /* Verify that we hold no locks other than p->p_lock. */
210 LOCKDEBUG_BARRIER(p->p_lock, 0);
211
212 /* XXX Temporary: something is leaking kernel_lock. */
213 KERNEL_UNLOCK_ALL(l, NULL);
214
215 KASSERT(mutex_owned(p->p_lock));
216 KASSERT(p->p_vmspace != NULL);
217
218 if (__predict_false(p == initproc)) {
219 panic("init died (signal %d, exit %d)", signo, exitcode);
220 }
221
222 p->p_sflag |= PS_WEXIT;
223
224 /*
225 * Force all other LWPs to exit before we do. Only then can we
226 * begin to tear down the rest of the process state.
227 */
228 if (p->p_nlwps > 1) {
229 exit_lwps(l);
230 }
231
232 ksiginfo_queue_init(&kq);
233
234 /*
235 * If we have been asked to stop on exit, do so now.
236 */
237 if (__predict_false(p->p_sflag & PS_STOPEXIT)) {
238 KASSERT(l->l_blcnt == 0);
239 sigclearall(p, &contsigmask, &kq);
240
241 if (!mutex_tryenter(&proc_lock)) {
242 mutex_exit(p->p_lock);
243 mutex_enter(&proc_lock);
244 mutex_enter(p->p_lock);
245 }
246 p->p_waited = 0;
247 p->p_pptr->p_nstopchild++;
248 p->p_stat = SSTOP;
249 mutex_exit(&proc_lock);
250 lwp_lock(l);
251 p->p_nrlwps--;
252 l->l_stat = LSSTOP;
253 lwp_unlock(l);
254 mutex_exit(p->p_lock);
255 lwp_lock(l);
256 spc_lock(l->l_cpu);
257 mi_switch(l);
258 mutex_enter(p->p_lock);
259 }
260
261 /*
262 * Bin any remaining signals and mark the process as dying so it will
263 * not be found for, e.g. signals.
264 */
265 sigfillset(&p->p_sigctx.ps_sigignore);
266 sigclearall(p, NULL, &kq);
267 p->p_stat = SDYING;
268
269 /*
270 * Perform any required thread cleanup. Do this early so
271 * anyone wanting to look us up by our global thread ID
272 * will fail to find us.
273 *
274 * N.B. this will unlock p->p_lock on our behalf.
275 */
276 lwp_thread_cleanup(l);
277
278 ksiginfo_queue_drain(&kq);
279
280 /* Destroy any lwpctl info. */
281 if (p->p_lwpctl != NULL)
282 lwp_ctl_exit();
283
284 /*
285 * Drain all remaining references that procfs, ptrace and others may
286 * have on the process.
287 */
288 rw_enter(&p->p_reflock, RW_WRITER);
289
290 DPRINTF(("%s: %d.%d exiting.\n", __func__, p->p_pid, l->l_lid));
291
292 ptimers_free(p, TIMERS_ALL);
293 #if defined(__HAVE_RAS)
294 ras_purgeall();
295 #endif
296
297 /*
298 * Close open files, release open-file table and free signal
299 * actions. This may block!
300 */
301 fd_free();
302 cwdfree(p->p_cwdi);
303 p->p_cwdi = NULL;
304 doexithooks(p);
305 sigactsfree(p->p_sigacts);
306
307 /*
308 * Write out accounting data.
309 */
310 (void)acct_process(l);
311
312 #ifdef KTRACE
313 /*
314 * Release trace file.
315 */
316 if (p->p_tracep != NULL) {
317 mutex_enter(&ktrace_lock);
318 ktrderef(p);
319 mutex_exit(&ktrace_lock);
320 }
321 #endif
322
323 p->p_xexit = exitcode;
324 p->p_xsig = signo;
325
326 /*
327 * If emulation has process exit hook, call it now.
328 * Set the exit status now so that the exit hook has
329 * an opportunity to tweak it (COMPAT_LINUX requires
330 * this for thread group emulation)
331 */
332 if (p->p_emul->e_proc_exit)
333 (*p->p_emul->e_proc_exit)(p);
334
335 /*
336 * Free the VM resources we're still holding on to.
337 * We must do this from a valid thread because doing
338 * so may block. This frees vmspace, which we don't
339 * need anymore. The only remaining lwp is the one
340 * we run at this moment, nothing runs in userland
341 * anymore.
342 */
343 ruspace(p); /* Update our vm resource use */
344 uvm_proc_exit(p);
345
346 /*
347 * Stop profiling.
348 */
349 if (__predict_false((p->p_stflag & PST_PROFIL) != 0)) {
350 mutex_spin_enter(&p->p_stmutex);
351 stopprofclock(p);
352 mutex_spin_exit(&p->p_stmutex);
353 }
354
355 /*
356 * If parent is waiting for us to exit or exec, PL_PPWAIT is set; we
357 * wake up the parent early to avoid deadlock. We can do this once
358 * the VM resources are released.
359 */
360 mutex_enter(&proc_lock);
361 if (p->p_lflag & PL_PPWAIT) {
362 lwp_t *lp;
363
364 l->l_lwpctl = NULL; /* was on loan from blocked parent */
365 p->p_lflag &= ~PL_PPWAIT;
366
367 lp = p->p_vforklwp;
368 p->p_vforklwp = NULL;
369 lp->l_vforkwaiting = false;
370 cv_broadcast(&lp->l_waitcv);
371 }
372
373 if (SESS_LEADER(p)) {
374 struct vnode *vprele = NULL, *vprevoke = NULL;
375 struct session *sp = p->p_session;
376 struct tty *tp;
377
378 if (sp->s_ttyvp) {
379 /*
380 * Controlling process.
381 * Signal foreground pgrp,
382 * drain controlling terminal
383 * and revoke access to controlling terminal.
384 */
385 tp = sp->s_ttyp;
386 mutex_spin_enter(&tty_lock);
387 if (tp->t_session == sp) {
388 /* we can't guarantee the revoke will do this */
389 pgrp = tp->t_pgrp;
390 tp->t_pgrp = NULL;
391 tp->t_session = NULL;
392 mutex_spin_exit(&tty_lock);
393 if (pgrp != NULL) {
394 pgsignal(pgrp, SIGHUP, 1);
395 }
396 mutex_exit(&proc_lock);
397 (void) ttywait(tp);
398 mutex_enter(&proc_lock);
399
400 /* The tty could have been revoked. */
401 vprevoke = sp->s_ttyvp;
402 } else
403 mutex_spin_exit(&tty_lock);
404 vprele = sp->s_ttyvp;
405 sp->s_ttyvp = NULL;
406 /*
407 * s_ttyp is not zero'd; we use this to indicate
408 * that the session once had a controlling terminal.
409 * (for logging and informational purposes)
410 */
411 }
412 sp->s_leader = NULL;
413
414 if (vprevoke != NULL || vprele != NULL) {
415 if (vprevoke != NULL) {
416 /* Releases proc_lock. */
417 proc_sessrele(sp);
418 VOP_REVOKE(vprevoke, REVOKEALL);
419 } else
420 mutex_exit(&proc_lock);
421 if (vprele != NULL)
422 vrele(vprele);
423 mutex_enter(&proc_lock);
424 }
425 }
426 fixjobc(p, p->p_pgrp, 0);
427
428 /* Release fstrans private data. */
429 fstrans_lwp_dtor(l);
430
431 /*
432 * Finalize the last LWP's specificdata, as well as the
433 * specificdata for the proc itself.
434 */
435 lwp_finispecific(l);
436 proc_finispecific(p);
437
438 /*
439 * Reset p_opptr pointer of all former children which got
440 * traced by another process and were reparented. We reset
441 * it to NULL here; the trace detach code then reparents
442 * the child to initproc. We only check allproc list, since
443 * eventual former children on zombproc list won't reference
444 * p_opptr anymore.
445 */
446 if (__predict_false(p->p_slflag & PSL_CHTRACED)) {
447 struct proc *q;
448 PROCLIST_FOREACH(q, &allproc) {
449 if (q->p_opptr == p)
450 q->p_opptr = NULL;
451 }
452 PROCLIST_FOREACH(q, &zombproc) {
453 if (q->p_opptr == p)
454 q->p_opptr = NULL;
455 }
456 }
457
458 /*
459 * Give orphaned children to init(8).
460 */
461 child = LIST_FIRST(&p->p_children);
462 wakeinit = (child != NULL);
463 for (; child != NULL; child = next_child) {
464 next_child = LIST_NEXT(child, p_sibling);
465
466 /*
467 * Traced processes are killed since their existence
468 * means someone is screwing up. Since we reset the
469 * trace flags, the logic in sys_wait4() would not be
470 * triggered to reparent the process to its
471 * original parent, so we must do this here.
472 */
473 if (__predict_false(child->p_slflag & PSL_TRACED)) {
474 mutex_enter(p->p_lock);
475 child->p_slflag &=
476 ~(PSL_TRACED|PSL_SYSCALL);
477 mutex_exit(p->p_lock);
478 if (child->p_opptr != child->p_pptr) {
479 struct proc *t = child->p_opptr;
480 proc_reparent(child, t ? t : initproc);
481 child->p_opptr = NULL;
482 } else
483 proc_reparent(child, initproc);
484 killproc(child, "orphaned traced process");
485 } else
486 proc_reparent(child, initproc);
487 }
488
489 /*
490 * Move proc from allproc to zombproc, it's now nearly ready to be
491 * collected by parent.
492 */
493 LIST_REMOVE(l, l_list);
494 LIST_REMOVE(p, p_list);
495 LIST_INSERT_HEAD(&zombproc, p, p_list);
496
497 /*
498 * Mark the process as dead. We must do this before we signal
499 * the parent.
500 */
501 p->p_stat = SDEAD;
502
503 /*
504 * Let anyone watching this DTrace probe know what we're
505 * on our way out.
506 */
507 SDT_PROBE(proc, kernel, , exit,
508 ((p->p_sflag & PS_COREDUMP) ? CLD_DUMPED :
509 (p->p_xsig ? CLD_KILLED : CLD_EXITED)),
510 0,0,0,0);
511
512 /* Put in front of parent's sibling list for parent to collect it */
513 old_parent = p->p_pptr;
514 old_parent->p_nstopchild++;
515 if (LIST_FIRST(&old_parent->p_children) != p) {
516 /* Put child where it can be found quickly */
517 LIST_REMOVE(p, p_sibling);
518 LIST_INSERT_HEAD(&old_parent->p_children, p, p_sibling);
519 }
520
521 /*
522 * Notify parent that we're gone. If parent has the P_NOCLDWAIT
523 * flag set, notify init instead (and hope it will handle
524 * this situation).
525 */
526 if (old_parent->p_flag & (PK_NOCLDWAIT|PK_CLDSIGIGN)) {
527 proc_reparent(p, initproc);
528 wakeinit = 1;
529
530 /*
531 * If this was the last child of our parent, notify
532 * parent, so in case he was wait(2)ing, he will
533 * continue.
534 */
535 if (LIST_FIRST(&old_parent->p_children) == NULL)
536 cv_broadcast(&old_parent->p_waitcv);
537 }
538
539 /* Reload parent pointer, since p may have been reparented above */
540 new_parent = p->p_pptr;
541
542 if (__predict_false(p->p_exitsig != 0)) {
543 exit_psignal(p, new_parent, &ksi);
544 kpsignal(new_parent, &ksi, NULL);
545 }
546
547 /* Calculate the final rusage info. */
548 calcru(p, &p->p_stats->p_ru.ru_utime, &p->p_stats->p_ru.ru_stime,
549 NULL, NULL);
550
551 callout_destroy(&l->l_timeout_ch);
552
553 /*
554 * Release any PCU resources before becoming a zombie.
555 */
556 pcu_discard_all(l);
557
558 /*
559 * Notify other processes tracking us with a knote that
560 * we're exiting.
561 *
562 * N.B. we do this here because the process is now SDEAD,
563 * and thus cannot have any more knotes attached. Also,
564 * knote_proc_exit() expects that p->p_lock is already
565 * held (and will assert so).
566 */
567 mutex_enter(p->p_lock);
568 if (!SLIST_EMPTY(&p->p_klist)) {
569 knote_proc_exit(p);
570 }
571
572 /* Free the LWP ID */
573 proc_free_lwpid(p, l->l_lid);
574 lwp_drainrefs(l);
575 lwp_lock(l);
576 l->l_prflag &= ~LPR_DETACHED;
577 l->l_stat = LSZOMB;
578 lwp_unlock(l);
579 KASSERT(curlwp == l);
580 KASSERT(p->p_nrlwps == 1);
581 KASSERT(p->p_nlwps == 1);
582 p->p_stat = SZOMB;
583 p->p_nrlwps--;
584 p->p_nzlwps++;
585 p->p_ndlwps = 0;
586 mutex_exit(p->p_lock);
587
588 /*
589 * Signal the parent to collect us, and drop the proclist lock.
590 * Drop debugger/procfs lock; no new references can be gained.
591 */
592 rw_exit(&p->p_reflock);
593 cv_broadcast(&p->p_pptr->p_waitcv);
594 mutex_exit(&proc_lock);
595 if (wakeinit)
596 cv_broadcast(&initproc->p_waitcv);
597
598 /*
599 * NOTE: WE ARE NO LONGER ALLOWED TO SLEEP!
600 */
601
602 /*
603 * Give machine-dependent code a chance to free any MD LWP
604 * resources. This must be done before uvm_lwp_exit(), in
605 * case these resources are in the PCB.
606 */
607 cpu_lwp_free(l, 1);
608
609 /* Switch away into oblivion. */
610 lwp_lock(l);
611 spc_lock(l->l_cpu);
612 mi_switch(l);
613 panic("exit1");
614 }
615
616 void
exit_lwps(struct lwp * l)617 exit_lwps(struct lwp *l)
618 {
619 proc_t *p = l->l_proc;
620 lwp_t *l2;
621
622 retry:
623 KASSERT(mutex_owned(p->p_lock));
624
625 /*
626 * Interrupt LWPs in interruptable sleep, unsuspend suspended
627 * LWPs and then wait for everyone else to finish.
628 */
629 LIST_FOREACH(l2, &p->p_lwps, l_sibling) {
630 if (l2 == l)
631 continue;
632 lwp_lock(l2);
633 l2->l_flag |= LW_WEXIT;
634 lwp_need_userret(l2);
635 if ((l2->l_stat == LSSLEEP && (l2->l_flag & LW_SINTR)) ||
636 l2->l_stat == LSSUSPENDED || l2->l_stat == LSSTOP) {
637 l2->l_flag &= ~LW_DBGSUSPEND;
638 /* setrunnable() will release the lock. */
639 setrunnable(l2);
640 continue;
641 }
642 lwp_unlock(l2);
643 }
644
645 /*
646 * Wait for every LWP to exit. Note: LWPs can get suspended/slept
647 * behind us or there may even be new LWPs created. Therefore, a
648 * full retry is required on error.
649 */
650 while (p->p_nlwps > 1) {
651 if (lwp_wait(l, 0, NULL, true)) {
652 goto retry;
653 }
654 }
655
656 KASSERT(p->p_nlwps == 1);
657 }
658
659 int
do_sys_waitid(idtype_t idtype,id_t id,int * pid,int * status,int options,struct wrusage * wru,siginfo_t * si)660 do_sys_waitid(idtype_t idtype, id_t id, int *pid, int *status, int options,
661 struct wrusage *wru, siginfo_t *si)
662 {
663 proc_t *child;
664 int error;
665
666
667 if (wru != NULL)
668 memset(wru, 0, sizeof(*wru));
669 if (si != NULL)
670 memset(si, 0, sizeof(*si));
671
672 mutex_enter(&proc_lock);
673 error = find_stopped_child(curproc, idtype, id, options, &child,
674 wru, si);
675 if (child == NULL) {
676 mutex_exit(&proc_lock);
677 *pid = 0;
678 *status = 0;
679 return error;
680 }
681 *pid = child->p_pid;
682
683 if (child->p_stat == SZOMB) {
684 /* Child is exiting */
685 *status = P_WAITSTATUS(child);
686 /* proc_free() will release the proc_lock. */
687 if (options & WNOWAIT) {
688 mutex_exit(&proc_lock);
689 } else {
690 proc_free(child, wru);
691 }
692 } else {
693 /* Don't mark SIGCONT if we are being stopped */
694 *status = (child->p_xsig == SIGCONT && child->p_stat != SSTOP) ?
695 W_CONTCODE() : W_STOPCODE(child->p_xsig);
696 mutex_exit(&proc_lock);
697 }
698 return 0;
699 }
700
701 int
do_sys_wait(int * pid,int * status,int options,struct rusage * ru)702 do_sys_wait(int *pid, int *status, int options, struct rusage *ru)
703 {
704 idtype_t idtype;
705 id_t id;
706 int ret;
707 struct wrusage wru;
708
709 /*
710 * Translate the special pid values into the (idtype, pid)
711 * pair for wait6. The WAIT_MYPGRP case is handled by
712 * find_stopped_child() on its own.
713 */
714 if (*pid == WAIT_ANY) {
715 idtype = P_ALL;
716 id = 0;
717 } else if (*pid < 0) {
718 idtype = P_PGID;
719 id = (id_t)-*pid;
720 } else {
721 idtype = P_PID;
722 id = (id_t)*pid;
723 }
724 options |= WEXITED | WTRAPPED;
725 ret = do_sys_waitid(idtype, id, pid, status, options, ru ? &wru : NULL,
726 NULL);
727 if (ru)
728 *ru = wru.wru_self;
729 return ret;
730 }
731
732 int
sys___wait450(struct lwp * l,const struct sys___wait450_args * uap,register_t * retval)733 sys___wait450(struct lwp *l, const struct sys___wait450_args *uap,
734 register_t *retval)
735 {
736 /* {
737 syscallarg(int) pid;
738 syscallarg(int *) status;
739 syscallarg(int) options;
740 syscallarg(struct rusage *) rusage;
741 } */
742 int error, status, pid = SCARG(uap, pid);
743 struct rusage ru;
744
745 error = do_sys_wait(&pid, &status, SCARG(uap, options),
746 SCARG(uap, rusage) != NULL ? &ru : NULL);
747
748 retval[0] = pid;
749 if (pid == 0) {
750 return error;
751 }
752 if (SCARG(uap, status)) {
753 error = copyout(&status, SCARG(uap, status), sizeof(status));
754 }
755 if (SCARG(uap, rusage) && error == 0) {
756 error = copyout(&ru, SCARG(uap, rusage), sizeof(ru));
757 }
758 return error;
759 }
760
761 int
sys_wait6(struct lwp * l,const struct sys_wait6_args * uap,register_t * retval)762 sys_wait6(struct lwp *l, const struct sys_wait6_args *uap, register_t *retval)
763 {
764 /* {
765 syscallarg(idtype_t) idtype;
766 syscallarg(id_t) id;
767 syscallarg(int *) status;
768 syscallarg(int) options;
769 syscallarg(struct wrusage *) wru;
770 syscallarg(siginfo_t *) si;
771 } */
772 struct wrusage wru, *wrup;
773 siginfo_t si, *sip;
774 idtype_t idtype;
775 int pid;
776 id_t id;
777 int error, status;
778
779 idtype = SCARG(uap, idtype);
780 id = SCARG(uap, id);
781
782 if (SCARG(uap, wru) != NULL)
783 wrup = &wru;
784 else
785 wrup = NULL;
786
787 if (SCARG(uap, info) != NULL)
788 sip = &si;
789 else
790 sip = NULL;
791
792 /*
793 * We expect all callers of wait6() to know about WEXITED and
794 * WTRAPPED.
795 */
796 error = do_sys_waitid(idtype, id, &pid, &status, SCARG(uap, options),
797 wrup, sip);
798
799 retval[0] = pid; /* tell userland who it was */
800
801 #if 0
802 /*
803 * should we copyout if there was no process, hence no useful data?
804 * We don't for an old style wait4() (etc) but I believe
805 * FreeBSD does for wait6(), so a tossup... Go with FreeBSD for now.
806 */
807 if (pid == 0)
808 return error;
809 #endif
810
811 if (SCARG(uap, status) != NULL && error == 0)
812 error = copyout(&status, SCARG(uap, status), sizeof(status));
813 if (SCARG(uap, wru) != NULL && error == 0)
814 error = copyout(&wru, SCARG(uap, wru), sizeof(wru));
815 if (SCARG(uap, info) != NULL && error == 0)
816 error = copyout(&si, SCARG(uap, info), sizeof(si));
817 return error;
818 }
819
820
821 /*
822 * Find a process that matches the provided criteria, and fill siginfo
823 * and resources if found.
824 * Returns:
825 * -1: Not found, abort early
826 * 0: Not matched
827 * 1: Matched, there might be more matches
828 * 2: This is the only match
829 */
830 static int
match_process(const struct proc * pp,struct proc ** q,idtype_t idtype,id_t id,int options,struct wrusage * wrusage,siginfo_t * siginfo)831 match_process(const struct proc *pp, struct proc **q, idtype_t idtype, id_t id,
832 int options, struct wrusage *wrusage, siginfo_t *siginfo)
833 {
834 struct rusage *rup;
835 struct proc *p = *q;
836 int rv = 1;
837
838 switch (idtype) {
839 case P_ALL:
840 mutex_enter(p->p_lock);
841 break;
842 case P_PID:
843 if (p->p_pid != (pid_t)id) {
844 p = *q = proc_find_raw((pid_t)id);
845 if (p == NULL || p->p_stat == SIDL || p->p_pptr != pp) {
846 *q = NULL;
847 return -1;
848 }
849 }
850 mutex_enter(p->p_lock);
851 rv++;
852 break;
853 case P_PGID:
854 if (p->p_pgid != (pid_t)id)
855 return 0;
856 mutex_enter(p->p_lock);
857 break;
858 case P_SID:
859 if (p->p_session->s_sid != (pid_t)id)
860 return 0;
861 mutex_enter(p->p_lock);
862 break;
863 case P_UID:
864 mutex_enter(p->p_lock);
865 if (kauth_cred_geteuid(p->p_cred) != (uid_t)id) {
866 mutex_exit(p->p_lock);
867 return 0;
868 }
869 break;
870 case P_GID:
871 mutex_enter(p->p_lock);
872 if (kauth_cred_getegid(p->p_cred) != (gid_t)id) {
873 mutex_exit(p->p_lock);
874 return 0;
875 }
876 break;
877 case P_CID:
878 case P_PSETID:
879 case P_CPUID:
880 /* XXX: Implement me */
881 default:
882 return 0;
883 }
884
885 if ((options & WEXITED) == 0 && p->p_stat == SZOMB) {
886 mutex_exit(p->p_lock);
887 return 0;
888 }
889
890 if (siginfo != NULL) {
891 siginfo->si_errno = 0;
892
893 /*
894 * SUSv4 requires that the si_signo value is always
895 * SIGCHLD. Obey it despite the rfork(2) interface
896 * allows to request other signal for child exit
897 * notification.
898 */
899 siginfo->si_signo = SIGCHLD;
900
901 /*
902 * This is still a rough estimate. We will fix the
903 * cases TRAPPED, STOPPED, and CONTINUED later.
904 */
905 if (p->p_sflag & PS_COREDUMP) {
906 siginfo->si_code = CLD_DUMPED;
907 siginfo->si_status = p->p_xsig;
908 } else if (p->p_xsig) {
909 siginfo->si_code = CLD_KILLED;
910 siginfo->si_status = p->p_xsig;
911 } else {
912 siginfo->si_code = CLD_EXITED;
913 siginfo->si_status = p->p_xexit;
914 }
915
916 siginfo->si_pid = p->p_pid;
917 siginfo->si_uid = kauth_cred_geteuid(p->p_cred);
918 siginfo->si_utime = p->p_stats->p_ru.ru_utime.tv_sec;
919 siginfo->si_stime = p->p_stats->p_ru.ru_stime.tv_sec;
920 }
921
922 /*
923 * There should be no reason to limit resources usage info to
924 * exited processes only. A snapshot about any resources used
925 * by a stopped process may be exactly what is needed.
926 */
927 if (wrusage != NULL) {
928 rup = &wrusage->wru_self;
929 *rup = p->p_stats->p_ru;
930 calcru(p, &rup->ru_utime, &rup->ru_stime, NULL, NULL);
931
932 rup = &wrusage->wru_children;
933 *rup = p->p_stats->p_cru;
934 calcru(p, &rup->ru_utime, &rup->ru_stime, NULL, NULL);
935 }
936
937 mutex_exit(p->p_lock);
938 return rv;
939 }
940
941 /*
942 * Determine if there are existing processes being debugged
943 * that used to be (and sometime later will be again) children
944 * of a specific parent (while matching wait criteria)
945 */
946 static bool
debugged_child_exists(idtype_t idtype,id_t id,int options,siginfo_t * si,const struct proc * parent)947 debugged_child_exists(idtype_t idtype, id_t id, int options, siginfo_t *si,
948 const struct proc *parent)
949 {
950 struct proc *pp;
951
952 /*
953 * If we are searching for a specific pid, we can optimise a little
954 */
955 if (idtype == P_PID) {
956 /*
957 * Check the specific process to see if its real parent is us
958 */
959 pp = proc_find_raw((pid_t)id);
960 if (pp != NULL && pp->p_stat != SIDL && pp->p_opptr == parent) {
961 /*
962 * using P_ALL here avoids match_process() doing the
963 * same work that we just did, but incorrectly for
964 * this scenario.
965 */
966 if (match_process(parent, &pp, P_ALL, id, options,
967 NULL, si))
968 return true;
969 }
970 return false;
971 }
972
973 /*
974 * For the hard cases, just look everywhere to see if some
975 * stolen (reparented) process is really our lost child.
976 * Then check if that process could satisfy the wait conditions.
977 */
978
979 /*
980 * XXX inefficient, but hopefully fairly rare.
981 * XXX should really use a list of reparented processes.
982 */
983 PROCLIST_FOREACH(pp, &allproc) {
984 if (pp->p_stat == SIDL) /* XXX impossible ?? */
985 continue;
986 if (pp->p_opptr == parent &&
987 match_process(parent, &pp, idtype, id, options, NULL, si))
988 return true;
989 }
990 PROCLIST_FOREACH(pp, &zombproc) {
991 if (pp->p_stat == SIDL) /* XXX impossible ?? */
992 continue;
993 if (pp->p_opptr == parent &&
994 match_process(parent, &pp, idtype, id, options, NULL, si))
995 return true;
996 }
997
998 return false;
999 }
1000
1001 /*
1002 * Scan list of child processes for a child process that has stopped or
1003 * exited. Used by sys_wait4 and 'compat' equivalents.
1004 *
1005 * Must be called with the proc_lock held, and may release while waiting.
1006 */
1007 static int
find_stopped_child(struct proc * parent,idtype_t idtype,id_t id,int options,struct proc ** child_p,struct wrusage * wru,siginfo_t * si)1008 find_stopped_child(struct proc *parent, idtype_t idtype, id_t id, int options,
1009 struct proc **child_p, struct wrusage *wru, siginfo_t *si)
1010 {
1011 struct proc *child, *dead;
1012 int error;
1013
1014 KASSERT(mutex_owned(&proc_lock));
1015
1016 if (options & ~WALLOPTS) {
1017 *child_p = NULL;
1018 return EINVAL;
1019 }
1020
1021 if ((options & WSELECTOPTS) == 0) {
1022 /*
1023 * We will be unable to find any matching processes,
1024 * because there are no known events to look for.
1025 * Prefer to return error instead of blocking
1026 * indefinitely.
1027 */
1028 *child_p = NULL;
1029 return EINVAL;
1030 }
1031
1032 if ((pid_t)id == WAIT_MYPGRP && (idtype == P_PID || idtype == P_PGID)) {
1033 id = (id_t)parent->p_pgid;
1034 idtype = P_PGID;
1035 }
1036
1037 for (;;) {
1038 error = ECHILD;
1039 dead = NULL;
1040
1041 LIST_FOREACH(child, &parent->p_children, p_sibling) {
1042 int rv = match_process(parent, &child, idtype, id,
1043 options, wru, si);
1044 if (rv == -1)
1045 break;
1046 if (rv == 0)
1047 continue;
1048
1049 /*
1050 * Wait for processes with p_exitsig != SIGCHLD
1051 * processes only if WALTSIG is set; wait for
1052 * processes with p_exitsig == SIGCHLD only
1053 * if WALTSIG is clear.
1054 */
1055 if (((options & WALLSIG) == 0) &&
1056 (options & WALTSIG ? child->p_exitsig == SIGCHLD
1057 : P_EXITSIG(child) != SIGCHLD)){
1058 if (rv == 2) {
1059 child = NULL;
1060 break;
1061 }
1062 continue;
1063 }
1064
1065 error = 0;
1066 if ((options & WNOZOMBIE) == 0) {
1067 if (child->p_stat == SZOMB)
1068 break;
1069 if (child->p_stat == SDEAD) {
1070 /*
1071 * We may occasionally arrive here
1072 * after receiving a signal, but
1073 * immediately before the child
1074 * process is zombified. The wait
1075 * will be short, so avoid returning
1076 * to userspace.
1077 */
1078 dead = child;
1079 }
1080 }
1081
1082 if ((options & WCONTINUED) != 0 &&
1083 child->p_xsig == SIGCONT &&
1084 (child->p_sflag & PS_CONTINUED)) {
1085 if ((options & WNOWAIT) == 0) {
1086 child->p_sflag &= ~PS_CONTINUED;
1087 child->p_waited = 1;
1088 parent->p_nstopchild--;
1089 }
1090 if (si) {
1091 si->si_status = child->p_xsig;
1092 si->si_code = CLD_CONTINUED;
1093 }
1094 break;
1095 }
1096
1097 if ((options & (WTRAPPED|WSTOPPED)) != 0 &&
1098 child->p_stat == SSTOP &&
1099 child->p_waited == 0 &&
1100 ((child->p_slflag & PSL_TRACED) ||
1101 options & (WUNTRACED|WSTOPPED))) {
1102 if ((options & WNOWAIT) == 0) {
1103 child->p_waited = 1;
1104 parent->p_nstopchild--;
1105 }
1106 if (si) {
1107 si->si_status = child->p_xsig;
1108 si->si_code =
1109 (child->p_slflag & PSL_TRACED) ?
1110 CLD_TRAPPED : CLD_STOPPED;
1111 }
1112 break;
1113 }
1114 if (parent->p_nstopchild == 0 || rv == 2) {
1115 child = NULL;
1116 break;
1117 }
1118 }
1119
1120 /*
1121 * If we found nothing, but we are the bereaved parent
1122 * of a stolen child, look and see if that child (or
1123 * one of them) meets our search criteria. If so, then
1124 * we cannot succeed, but we can hang (wait...),
1125 * or if WNOHANG, return 0 instead of ECHILD
1126 */
1127 if (child == NULL && error == ECHILD &&
1128 (parent->p_slflag & PSL_CHTRACED) &&
1129 debugged_child_exists(idtype, id, options, si, parent))
1130 error = 0;
1131
1132 if (child != NULL || error != 0 ||
1133 ((options & WNOHANG) != 0 && dead == NULL)) {
1134 *child_p = child;
1135 return error;
1136 }
1137
1138 /*
1139 * Wait for another child process to stop.
1140 */
1141 error = cv_wait_sig(&parent->p_waitcv, &proc_lock);
1142
1143 if (error != 0) {
1144 *child_p = NULL;
1145 return error;
1146 }
1147 }
1148 }
1149
1150 /*
1151 * Free a process after parent has taken all the state info. Must be called
1152 * with the proclist lock held, and will release before returning.
1153 *
1154 * *ru is returned to the caller, and must be freed by the caller.
1155 */
1156 static void
proc_free(struct proc * p,struct wrusage * wru)1157 proc_free(struct proc *p, struct wrusage *wru)
1158 {
1159 struct proc *parent = p->p_pptr;
1160 struct lwp *l;
1161 ksiginfo_t ksi;
1162 kauth_cred_t cred1, cred2;
1163 uid_t uid;
1164
1165 KASSERT(mutex_owned(&proc_lock));
1166 KASSERT(p->p_nlwps == 1);
1167 KASSERT(p->p_nzlwps == 1);
1168 KASSERT(p->p_nrlwps == 0);
1169 KASSERT(p->p_stat == SZOMB);
1170
1171 /*
1172 * If we got the child via ptrace(2) or procfs, and
1173 * the parent is different (meaning the process was
1174 * attached, rather than run as a child), then we need
1175 * to give it back to the old parent, and send the
1176 * parent the exit signal. The rest of the cleanup
1177 * will be done when the old parent waits on the child.
1178 */
1179 if ((p->p_slflag & PSL_TRACED) != 0 && p->p_opptr != parent) {
1180 mutex_enter(p->p_lock);
1181 p->p_slflag &= ~(PSL_TRACED|PSL_SYSCALL);
1182 mutex_exit(p->p_lock);
1183 parent = (p->p_opptr == NULL) ? initproc : p->p_opptr;
1184 proc_reparent(p, parent);
1185 p->p_opptr = NULL;
1186 if (p->p_exitsig != 0) {
1187 exit_psignal(p, parent, &ksi);
1188 kpsignal(parent, &ksi, NULL);
1189 }
1190 cv_broadcast(&parent->p_waitcv);
1191 mutex_exit(&proc_lock);
1192 return;
1193 }
1194
1195 sched_proc_exit(parent, p);
1196
1197 /*
1198 * Add child times of exiting process onto its own times.
1199 * This cannot be done any earlier else it might get done twice.
1200 */
1201 l = LIST_FIRST(&p->p_lwps);
1202 ruadd(&p->p_stats->p_ru, &l->l_ru);
1203 ruadd(&p->p_stats->p_ru, &p->p_stats->p_cru);
1204 ruadd(&parent->p_stats->p_cru, &p->p_stats->p_ru);
1205 if (wru != NULL) {
1206 wru->wru_self = p->p_stats->p_ru;
1207 wru->wru_children = p->p_stats->p_cru;
1208 }
1209 p->p_xsig = 0;
1210 p->p_xexit = 0;
1211
1212 /*
1213 * At this point we are going to start freeing the final resources.
1214 * If anyone tries to access the proc structure after here they will
1215 * get a shock - bits are missing. Attempt to make it hard! We
1216 * don't bother with any further locking past this point.
1217 */
1218 p->p_stat = SIDL; /* not even a zombie any more */
1219 LIST_REMOVE(p, p_list); /* off zombproc */
1220 parent->p_nstopchild--;
1221 LIST_REMOVE(p, p_sibling);
1222
1223 /*
1224 * Let pid be reallocated.
1225 */
1226 proc_free_pid(p->p_pid);
1227 atomic_dec_uint(&nprocs);
1228
1229 /*
1230 * Unlink process from its process group.
1231 * Releases the proc_lock.
1232 */
1233 proc_leavepgrp(p);
1234
1235 /*
1236 * Delay release until after lwp_free.
1237 */
1238 cred2 = l->l_cred;
1239
1240 /*
1241 * Free the last LWP's resources.
1242 *
1243 * lwp_free ensures the LWP is no longer running on another CPU.
1244 */
1245 lwp_free(l, false, true);
1246
1247 /*
1248 * Now no one except us can reach the process p.
1249 */
1250
1251 /*
1252 * Decrement the count of procs running with this uid.
1253 */
1254 cred1 = p->p_cred;
1255 uid = kauth_cred_getuid(cred1);
1256 (void)chgproccnt(uid, -1);
1257
1258 /*
1259 * Release substructures.
1260 */
1261
1262 lim_free(p->p_limit);
1263 pstatsfree(p->p_stats);
1264 kauth_cred_free(cred1);
1265 kauth_cred_free(cred2);
1266
1267 /*
1268 * Release reference to text vnode
1269 */
1270 if (p->p_textvp)
1271 vrele(p->p_textvp);
1272 kmem_strfree(p->p_path);
1273
1274 mutex_destroy(&p->p_auxlock);
1275 mutex_obj_free(p->p_lock);
1276 mutex_destroy(&p->p_stmutex);
1277 cv_destroy(&p->p_waitcv);
1278 cv_destroy(&p->p_lwpcv);
1279 rw_destroy(&p->p_reflock);
1280
1281 proc_free_mem(p);
1282 }
1283
1284 /*
1285 * Change the parent of a process for tracing purposes.
1286 */
1287 void
proc_changeparent(struct proc * t,struct proc * p)1288 proc_changeparent(struct proc *t, struct proc *p)
1289 {
1290 SET(t->p_slflag, PSL_TRACED);
1291 t->p_opptr = t->p_pptr;
1292 if (t->p_pptr == p)
1293 return;
1294 struct proc *parent = t->p_pptr;
1295
1296 if (parent->p_lock < t->p_lock) {
1297 if (!mutex_tryenter(parent->p_lock)) {
1298 mutex_exit(t->p_lock);
1299 mutex_enter(parent->p_lock);
1300 mutex_enter(t->p_lock);
1301 }
1302 } else if (parent->p_lock > t->p_lock) {
1303 mutex_enter(parent->p_lock);
1304 }
1305 parent->p_slflag |= PSL_CHTRACED;
1306 proc_reparent(t, p);
1307 if (parent->p_lock != t->p_lock)
1308 mutex_exit(parent->p_lock);
1309 }
1310
1311 /*
1312 * make process 'parent' the new parent of process 'child'.
1313 *
1314 * Must be called with proc_lock held.
1315 */
1316 void
proc_reparent(struct proc * child,struct proc * parent)1317 proc_reparent(struct proc *child, struct proc *parent)
1318 {
1319
1320 KASSERT(mutex_owned(&proc_lock));
1321
1322 if (child->p_pptr == parent)
1323 return;
1324
1325 if (child->p_stat == SZOMB || child->p_stat == SDEAD ||
1326 (child->p_stat == SSTOP && !child->p_waited)) {
1327 child->p_pptr->p_nstopchild--;
1328 parent->p_nstopchild++;
1329 }
1330 if (parent == initproc) {
1331 child->p_exitsig = SIGCHLD;
1332 child->p_ppid = parent->p_pid;
1333 }
1334
1335 LIST_REMOVE(child, p_sibling);
1336 LIST_INSERT_HEAD(&parent->p_children, child, p_sibling);
1337 child->p_pptr = parent;
1338 }
1339