/*      $NetBSD: kern_exit.c,v 1.300 2025/03/16 15:52:18 riastradh Exp $        */

/*-
* Copyright (c) 1998, 1999, 2006, 2007, 2008, 2020, 2023
*     The NetBSD Foundation, Inc.
* All rights reserved.
*
* This code is derived from software contributed to The NetBSD Foundation
* by Jason R. Thorpe of the Numerical Aerospace Simulation Facility,
* NASA Ames Research Center, and by Andrew Doran.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
*    notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
*    notice, this list of conditions and the following disclaimer in the
*    documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
* ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
* TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
* BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*/

/*
* Copyright (c) 1982, 1986, 1989, 1991, 1993
*      The Regents of the University of California.  All rights reserved.
* (c) UNIX System Laboratories, Inc.
* All or some portions of this file are derived from material licensed
* to the University of California by American Telephone and Telegraph
* Co. or Unix System Laboratories, Inc. and are reproduced herein with
* the permission of UNIX System Laboratories, Inc.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
*    notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
*    notice, this list of conditions and the following disclaimer in the
*    documentation and/or other materials provided with the distribution.
* 3. Neither the name of the University nor the names of its contributors
*    may be used to endorse or promote products derived from this software
*    without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
*      @(#)kern_exit.c 8.10 (Berkeley) 2/23/95
*/

#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: kern_exit.c,v 1.300 2025/03/16 15:52:18 riastradh Exp $");

#include "opt_ktrace.h"
#include "opt_dtrace.h"
#include "opt_sysv.h"

#include <sys/param.h>
#include <sys/types.h>

#include <sys/acct.h>
#include <sys/atomic.h>
#include <sys/buf.h>
#include <sys/cpu.h>
#include <sys/file.h>
#include <sys/filedesc.h>
#include <sys/fstrans.h>
#include <sys/ioctl.h>
#include <sys/kauth.h>
#include <sys/kernel.h>
#include <sys/ktrace.h>
#include <sys/lock.h>
#include <sys/lockdebug.h>
#include <sys/lwpctl.h>
#include <sys/mount.h>
#include <sys/pool.h>
#include <sys/proc.h>
#include <sys/psref.h>
#include <sys/ptrace.h>
#include <sys/ras.h>
#include <sys/resource.h>
#include <sys/sched.h>
#include <sys/sdt.h>
#include <sys/signalvar.h>
#include <sys/sleepq.h>
#include <sys/syscallargs.h>
#include <sys/syslog.h>
#include <sys/systm.h>
#include <sys/time.h>
#include <sys/tty.h>
#include <sys/uidinfo.h>
#include <sys/vnode.h>
#include <sys/wait.h>

#include <uvm/uvm_extern.h>

#ifdef DEBUG_EXIT
int debug_exit = 0;
#define DPRINTF(x) if (debug_exit) printf x
#else
#define DPRINTF(x)
#endif

static int find_stopped_child(struct proc *, idtype_t, id_t, int,
   struct proc **, struct wrusage *, siginfo_t *);
static void proc_free(struct proc *, struct wrusage *);

/*
* DTrace SDT provider definitions
*/
SDT_PROVIDER_DECLARE(proc);
SDT_PROBE_DEFINE1(proc, kernel, , exit, "int");

/*
* Fill in the appropriate signal information, and signal the parent.
*/
/* XXX noclone works around a gcc 4.5 bug on arm */
static void __noclone
exit_psignal(struct proc *p, struct proc *pp, ksiginfo_t *ksi)
{

       KSI_INIT(ksi);
       if ((ksi->ksi_signo = P_EXITSIG(p)) == SIGCHLD) {
               if (p->p_xsig) {
                       if (p->p_sflag & PS_COREDUMP)
                               ksi->ksi_code = CLD_DUMPED;
                       else
                               ksi->ksi_code = CLD_KILLED;
                       ksi->ksi_status = p->p_xsig;
               } else {
                       ksi->ksi_code = CLD_EXITED;
                       ksi->ksi_status = p->p_xexit;
               }
       } else {
               ksi->ksi_code = SI_USER;
               ksi->ksi_status = p->p_xsig;
       }
       /*
        * We fill those in, even for non-SIGCHLD.
        * It's safe to access p->p_cred unlocked here.
        */
       ksi->ksi_pid = p->p_pid;
       ksi->ksi_uid = kauth_cred_geteuid(p->p_cred);
       /* XXX: is this still valid? */
       ksi->ksi_utime = p->p_stats->p_ru.ru_utime.tv_sec;
       ksi->ksi_stime = p->p_stats->p_ru.ru_stime.tv_sec;
}

/*
* exit --
*      Death of process.
*/
int
sys_exit(struct lwp *l, const struct sys_exit_args *uap, register_t *retval)
{
       /* {
               syscallarg(int) rval;
       } */
       struct proc *p = l->l_proc;

       /* Don't call exit1() multiple times in the same process. */
       mutex_enter(p->p_lock);
       if (p->p_sflag & PS_WEXIT) {
               mutex_exit(p->p_lock);
               lwp_exit(l);
       }

       /* exit1() will release the mutex. */
       exit1(l, SCARG(uap, rval), 0);
       /* NOTREACHED */
       return (0);
}

/*
* Exit: deallocate address space and other resources, change proc state
* to zombie, and unlink proc from allproc and parent's lists.  Save exit
* status and rusage for wait().  Check for child processes and orphan them.
*
* Must be called with p->p_lock held.  Does not return.
*/
void
exit1(struct lwp *l, int exitcode, int signo)
{
       struct proc     *p, *child, *next_child, *old_parent, *new_parent;
       struct pgrp     *pgrp;
       ksiginfo_t      ksi;
       ksiginfoq_t     kq;
       int             wakeinit;

       p = l->l_proc;

       /* Verify that we hold no locks other than p->p_lock. */
       LOCKDEBUG_BARRIER(p->p_lock, 0);

       /* XXX Temporary: something is leaking kernel_lock. */
       KERNEL_UNLOCK_ALL(l, NULL);

       KASSERT(mutex_owned(p->p_lock));
       KASSERT(p->p_vmspace != NULL);

       if (__predict_false(p == initproc)) {
               panic("init died (signal %d, exit %d)", signo, exitcode);
       }

       p->p_sflag |= PS_WEXIT;

       /*
        * Force all other LWPs to exit before we do.  Only then can we
        * begin to tear down the rest of the process state.
        */
       if (p->p_nlwps > 1) {
               exit_lwps(l);
       }

       ksiginfo_queue_init(&kq);

       /*
        * If we have been asked to stop on exit, do so now.
        */
       if (__predict_false(p->p_sflag & PS_STOPEXIT)) {
               KASSERT(l->l_blcnt == 0);
               sigclearall(p, &contsigmask, &kq);

               if (!mutex_tryenter(&proc_lock)) {
                       mutex_exit(p->p_lock);
                       mutex_enter(&proc_lock);
                       mutex_enter(p->p_lock);
               }
               p->p_waited = 0;
               p->p_pptr->p_nstopchild++;
               p->p_stat = SSTOP;
               mutex_exit(&proc_lock);
               lwp_lock(l);
               p->p_nrlwps--;
               l->l_stat = LSSTOP;
               lwp_unlock(l);
               mutex_exit(p->p_lock);
               lwp_lock(l);
               spc_lock(l->l_cpu);
               mi_switch(l);
               mutex_enter(p->p_lock);
       }

       /*
        * Bin any remaining signals and mark the process as dying so it will
        * not be found for, e.g. signals.
        */
       sigfillset(&p->p_sigctx.ps_sigignore);
       sigclearall(p, NULL, &kq);
       p->p_stat = SDYING;

       /*
        * Perform any required thread cleanup.  Do this early so
        * anyone wanting to look us up by our global thread ID
        * will fail to find us.
        *
        * N.B. this will unlock p->p_lock on our behalf.
        */
       lwp_thread_cleanup(l);

       ksiginfo_queue_drain(&kq);

       /* Destroy any lwpctl info. */
       if (p->p_lwpctl != NULL)
               lwp_ctl_exit();

       /*
        * Drain all remaining references that procfs, ptrace and others may
        * have on the process.
        */
       rw_enter(&p->p_reflock, RW_WRITER);

       DPRINTF(("%s: %d.%d exiting.\n", __func__, p->p_pid, l->l_lid));

       ptimers_free(p, TIMERS_ALL);
#if defined(__HAVE_RAS)
       ras_purgeall();
#endif

       /*
        * Close open files, release open-file table and free signal
        * actions.  This may block!
        */
       fd_free();
       cwdfree(p->p_cwdi);
       p->p_cwdi = NULL;
       doexithooks(p);
       sigactsfree(p->p_sigacts);

       /*
        * Write out accounting data.
        */
       (void)acct_process(l);

#ifdef KTRACE
       /*
        * Release trace file.
        */
       if (p->p_tracep != NULL) {
               mutex_enter(&ktrace_lock);
               ktrderef(p);
               mutex_exit(&ktrace_lock);
       }
#endif

       p->p_xexit = exitcode;
       p->p_xsig = signo;

       /*
        * If emulation has process exit hook, call it now.
        * Set the exit status now so that the exit hook has
        * an opportunity to tweak it (COMPAT_LINUX requires
        * this for thread group emulation)
        */
       if (p->p_emul->e_proc_exit)
               (*p->p_emul->e_proc_exit)(p);

       /*
        * Free the VM resources we're still holding on to.
        * We must do this from a valid thread because doing
        * so may block. This frees vmspace, which we don't
        * need anymore. The only remaining lwp is the one
        * we run at this moment, nothing runs in userland
        * anymore.
        */
       ruspace(p);     /* Update our vm resource use */
       uvm_proc_exit(p);

       /*
        * Stop profiling.
        */
       if (__predict_false((p->p_stflag & PST_PROFIL) != 0)) {
               mutex_spin_enter(&p->p_stmutex);
               stopprofclock(p);
               mutex_spin_exit(&p->p_stmutex);
       }

       /*
        * If parent is waiting for us to exit or exec, PL_PPWAIT is set; we
        * wake up the parent early to avoid deadlock.  We can do this once
        * the VM resources are released.
        */
       mutex_enter(&proc_lock);
       if (p->p_lflag & PL_PPWAIT) {
               lwp_t *lp;

               l->l_lwpctl = NULL; /* was on loan from blocked parent */
               p->p_lflag &= ~PL_PPWAIT;

               lp = p->p_vforklwp;
               p->p_vforklwp = NULL;
               lp->l_vforkwaiting = false;
               cv_broadcast(&lp->l_waitcv);
       }

       if (SESS_LEADER(p)) {
               struct vnode *vprele = NULL, *vprevoke = NULL;
               struct session *sp = p->p_session;
               struct tty *tp;

               if (sp->s_ttyvp) {
                       /*
                        * Controlling process.
                        * Signal foreground pgrp,
                        * drain controlling terminal
                        * and revoke access to controlling terminal.
                        */
                       tp = sp->s_ttyp;
                       mutex_spin_enter(&tty_lock);
                       if (tp->t_session == sp) {
                               /* we can't guarantee the revoke will do this */
                               pgrp = tp->t_pgrp;
                               tp->t_pgrp = NULL;
                               tp->t_session = NULL;
                               mutex_spin_exit(&tty_lock);
                               if (pgrp != NULL) {
                                       pgsignal(pgrp, SIGHUP, 1);
                               }
                               mutex_exit(&proc_lock);
                               (void) ttywait(tp);
                               mutex_enter(&proc_lock);

                               /* The tty could have been revoked. */
                               vprevoke = sp->s_ttyvp;
                       } else
                               mutex_spin_exit(&tty_lock);
                       vprele = sp->s_ttyvp;
                       sp->s_ttyvp = NULL;
                       /*
                        * s_ttyp is not zero'd; we use this to indicate
                        * that the session once had a controlling terminal.
                        * (for logging and informational purposes)
                        */
               }
               sp->s_leader = NULL;

               if (vprevoke != NULL || vprele != NULL) {
                       if (vprevoke != NULL) {
                               /* Releases proc_lock. */
                               proc_sessrele(sp);
                               VOP_REVOKE(vprevoke, REVOKEALL);
                       } else
                               mutex_exit(&proc_lock);
                       if (vprele != NULL)
                               vrele(vprele);
                       mutex_enter(&proc_lock);
               }
       }
       fixjobc(p, p->p_pgrp, 0);

       /* Release fstrans private data. */
       fstrans_lwp_dtor(l);

       /*
        * Finalize the last LWP's specificdata, as well as the
        * specificdata for the proc itself.
        */
       lwp_finispecific(l);
       proc_finispecific(p);

       /*
        * Reset p_opptr pointer of all former children which got
        * traced by another process and were reparented. We reset
        * it to NULL here; the trace detach code then reparents
        * the child to initproc. We only check allproc list, since
        * eventual former children on zombproc list won't reference
        * p_opptr anymore.
        */
       if (__predict_false(p->p_slflag & PSL_CHTRACED)) {
               struct proc *q;
               PROCLIST_FOREACH(q, &allproc) {
                       if (q->p_opptr == p)
                               q->p_opptr = NULL;
               }
               PROCLIST_FOREACH(q, &zombproc) {
                       if (q->p_opptr == p)
                               q->p_opptr = NULL;
               }
       }

       /*
        * Give orphaned children to init(8).
        */
       child = LIST_FIRST(&p->p_children);
       wakeinit = (child != NULL);
       for (; child != NULL; child = next_child) {
               next_child = LIST_NEXT(child, p_sibling);

               /*
                * Traced processes are killed since their existence
                * means someone is screwing up. Since we reset the
                * trace flags, the logic in sys_wait4() would not be
                * triggered to reparent the process to its
                * original parent, so we must do this here.
                */
               if (__predict_false(child->p_slflag & PSL_TRACED)) {
                       mutex_enter(p->p_lock);
                       child->p_slflag &=
                           ~(PSL_TRACED|PSL_SYSCALL);
                       mutex_exit(p->p_lock);
                       if (child->p_opptr != child->p_pptr) {
                               struct proc *t = child->p_opptr;
                               proc_reparent(child, t ? t : initproc);
                               child->p_opptr = NULL;
                       } else
                               proc_reparent(child, initproc);
                       killproc(child, "orphaned traced process");
               } else
                       proc_reparent(child, initproc);
       }

       /*
        * Move proc from allproc to zombproc, it's now nearly ready to be
        * collected by parent.
        */
       LIST_REMOVE(l, l_list);
       LIST_REMOVE(p, p_list);
       LIST_INSERT_HEAD(&zombproc, p, p_list);

       /*
        * Mark the process as dead.  We must do this before we signal
        * the parent.
        */
       p->p_stat = SDEAD;

       /*
        * Let anyone watching this DTrace probe know what we're
        * on our way out.
        */
       SDT_PROBE(proc, kernel, , exit,
               ((p->p_sflag & PS_COREDUMP) ? CLD_DUMPED :
                (p->p_xsig ? CLD_KILLED : CLD_EXITED)),
               0,0,0,0);

       /* Put in front of parent's sibling list for parent to collect it */
       old_parent = p->p_pptr;
       old_parent->p_nstopchild++;
       if (LIST_FIRST(&old_parent->p_children) != p) {
               /* Put child where it can be found quickly */
               LIST_REMOVE(p, p_sibling);
               LIST_INSERT_HEAD(&old_parent->p_children, p, p_sibling);
       }

       /*
        * Notify parent that we're gone.  If parent has the P_NOCLDWAIT
        * flag set, notify init instead (and hope it will handle
        * this situation).
        */
       if (old_parent->p_flag & (PK_NOCLDWAIT|PK_CLDSIGIGN)) {
               proc_reparent(p, initproc);
               wakeinit = 1;

               /*
                * If this was the last child of our parent, notify
                * parent, so in case he was wait(2)ing, he will
                * continue.
                */
               if (LIST_FIRST(&old_parent->p_children) == NULL)
                       cv_broadcast(&old_parent->p_waitcv);
       }

       /* Reload parent pointer, since p may have been reparented above */
       new_parent = p->p_pptr;

       if (__predict_false(p->p_exitsig != 0)) {
               exit_psignal(p, new_parent, &ksi);
               kpsignal(new_parent, &ksi, NULL);
       }

       /* Calculate the final rusage info.  */
       calcru(p, &p->p_stats->p_ru.ru_utime, &p->p_stats->p_ru.ru_stime,
           NULL, NULL);

       callout_destroy(&l->l_timeout_ch);

       /*
        * Release any PCU resources before becoming a zombie.
        */
       pcu_discard_all(l);

       /*
        * Notify other processes tracking us with a knote that
        * we're exiting.
        *
        * N.B. we do this here because the process is now SDEAD,
        * and thus cannot have any more knotes attached.  Also,
        * knote_proc_exit() expects that p->p_lock is already
        * held (and will assert so).
        */
       mutex_enter(p->p_lock);
       if (!SLIST_EMPTY(&p->p_klist)) {
               knote_proc_exit(p);
       }

       /* Free the LWP ID */
       proc_free_lwpid(p, l->l_lid);
       lwp_drainrefs(l);
       lwp_lock(l);
       l->l_prflag &= ~LPR_DETACHED;
       l->l_stat = LSZOMB;
       lwp_unlock(l);
       KASSERT(curlwp == l);
       KASSERT(p->p_nrlwps == 1);
       KASSERT(p->p_nlwps == 1);
       p->p_stat = SZOMB;
       p->p_nrlwps--;
       p->p_nzlwps++;
       p->p_ndlwps = 0;
       mutex_exit(p->p_lock);

       /*
        * Signal the parent to collect us, and drop the proclist lock.
        * Drop debugger/procfs lock; no new references can be gained.
        */
       rw_exit(&p->p_reflock);
       cv_broadcast(&p->p_pptr->p_waitcv);
       mutex_exit(&proc_lock);
       if (wakeinit)
               cv_broadcast(&initproc->p_waitcv);

       /*
        * NOTE: WE ARE NO LONGER ALLOWED TO SLEEP!
        */

       /*
        * Give machine-dependent code a chance to free any MD LWP
        * resources.  This must be done before uvm_lwp_exit(), in
        * case these resources are in the PCB.
        */
       cpu_lwp_free(l, 1);

       /* Switch away into oblivion. */
       lwp_lock(l);
       spc_lock(l->l_cpu);
       mi_switch(l);
       panic("exit1");
}

void
exit_lwps(struct lwp *l)
{
       proc_t *p = l->l_proc;
       lwp_t *l2;

retry:
       KASSERT(mutex_owned(p->p_lock));

       /*
        * Interrupt LWPs in interruptable sleep, unsuspend suspended
        * LWPs and then wait for everyone else to finish.
        */
       LIST_FOREACH(l2, &p->p_lwps, l_sibling) {
               if (l2 == l)
                       continue;
               lwp_lock(l2);
               l2->l_flag |= LW_WEXIT;
               lwp_need_userret(l2);
               if ((l2->l_stat == LSSLEEP && (l2->l_flag & LW_SINTR)) ||
                   l2->l_stat == LSSUSPENDED || l2->l_stat == LSSTOP) {
                       l2->l_flag &= ~LW_DBGSUSPEND;
                       /* setrunnable() will release the lock. */
                       setrunnable(l2);
                       continue;
               }
               lwp_unlock(l2);
       }

       /*
        * Wait for every LWP to exit.  Note: LWPs can get suspended/slept
        * behind us or there may even be new LWPs created.  Therefore, a
        * full retry is required on error.
        */
       while (p->p_nlwps > 1) {
               if (lwp_wait(l, 0, NULL, true)) {
                       goto retry;
               }
       }

       KASSERT(p->p_nlwps == 1);
}

int
do_sys_waitid(idtype_t idtype, id_t id, int *pid, int *status, int options,
   struct wrusage *wru, siginfo_t *si)
{
       proc_t *child;
       int error;


       if (wru != NULL)
               memset(wru, 0, sizeof(*wru));
       if (si != NULL)
               memset(si, 0, sizeof(*si));

       mutex_enter(&proc_lock);
       error = find_stopped_child(curproc, idtype, id, options, &child,
           wru, si);
       if (child == NULL) {
               mutex_exit(&proc_lock);
               *pid = 0;
               *status = 0;
               return error;
       }
       *pid = child->p_pid;

       if (child->p_stat == SZOMB) {
               /* Child is exiting */
               *status = P_WAITSTATUS(child);
               /* proc_free() will release the proc_lock. */
               if (options & WNOWAIT) {
                       mutex_exit(&proc_lock);
               } else {
                       proc_free(child, wru);
               }
       } else {
               /* Don't mark SIGCONT if we are being stopped */
               *status = (child->p_xsig == SIGCONT && child->p_stat != SSTOP) ?
                   W_CONTCODE() : W_STOPCODE(child->p_xsig);
               mutex_exit(&proc_lock);
       }
       return 0;
}

int
do_sys_wait(int *pid, int *status, int options, struct rusage *ru)
{
       idtype_t idtype;
       id_t id;
       int ret;
       struct wrusage wru;

       /*
        * Translate the special pid values into the (idtype, pid)
        * pair for wait6. The WAIT_MYPGRP case is handled by
        * find_stopped_child() on its own.
        */
       if (*pid == WAIT_ANY) {
               idtype = P_ALL;
               id = 0;
       } else if (*pid < 0) {
               idtype = P_PGID;
               id = (id_t)-*pid;
       } else {
               idtype = P_PID;
               id = (id_t)*pid;
       }
       options |= WEXITED | WTRAPPED;
       ret = do_sys_waitid(idtype, id, pid, status, options, ru ? &wru : NULL,
           NULL);
       if (ru)
               *ru = wru.wru_self;
       return ret;
}

int
sys___wait450(struct lwp *l, const struct sys___wait450_args *uap,
   register_t *retval)
{
       /* {
               syscallarg(int)                 pid;
               syscallarg(int *)               status;
               syscallarg(int)                 options;
               syscallarg(struct rusage *)     rusage;
       } */
       int error, status, pid = SCARG(uap, pid);
       struct rusage ru;

       error = do_sys_wait(&pid, &status, SCARG(uap, options),
           SCARG(uap, rusage) != NULL ? &ru : NULL);

       retval[0] = pid;
       if (pid == 0) {
               return error;
       }
       if (SCARG(uap, status)) {
               error = copyout(&status, SCARG(uap, status), sizeof(status));
       }
       if (SCARG(uap, rusage) && error == 0) {
               error = copyout(&ru, SCARG(uap, rusage), sizeof(ru));
       }
       return error;
}

int
sys_wait6(struct lwp *l, const struct sys_wait6_args *uap, register_t *retval)
{
       /* {
               syscallarg(idtype_t)            idtype;
               syscallarg(id_t)                id;
               syscallarg(int *)               status;
               syscallarg(int)                 options;
               syscallarg(struct wrusage *)    wru;
               syscallarg(siginfo_t *)         si;
       } */
       struct wrusage wru, *wrup;
       siginfo_t si, *sip;
       idtype_t idtype;
       int pid;
       id_t id;
       int error, status;

       idtype = SCARG(uap, idtype);
       id = SCARG(uap, id);

       if (SCARG(uap, wru) != NULL)
               wrup = &wru;
       else
               wrup = NULL;

       if (SCARG(uap, info) != NULL)
               sip = &si;
       else
               sip = NULL;

       /*
        *  We expect all callers of wait6() to know about WEXITED and
        *  WTRAPPED.
        */
       error = do_sys_waitid(idtype, id, &pid, &status, SCARG(uap, options),
           wrup, sip);

       retval[0] = pid;        /* tell userland who it was */

#if 0
       /*
        * should we copyout if there was no process, hence no useful data?
        * We don't for an old style wait4() (etc) but I believe
        * FreeBSD does for wait6(), so a tossup...  Go with FreeBSD for now.
        */
       if (pid == 0)
               return error;
#endif

       if (SCARG(uap, status) != NULL && error == 0)
               error = copyout(&status, SCARG(uap, status), sizeof(status));
       if (SCARG(uap, wru) != NULL && error == 0)
               error = copyout(&wru, SCARG(uap, wru), sizeof(wru));
       if (SCARG(uap, info) != NULL && error == 0)
               error = copyout(&si, SCARG(uap, info), sizeof(si));
       return error;
}


/*
* Find a process that matches the provided criteria, and fill siginfo
* and resources if found.
* Returns:
*      -1:     Not found, abort early
*       0:     Not matched
*       1:     Matched, there might be more matches
*       2:     This is the only match
*/
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)
{
       struct rusage *rup;
       struct proc *p = *q;
       int rv = 1;

       switch (idtype) {
       case P_ALL:
               mutex_enter(p->p_lock);
               break;
       case P_PID:
               if (p->p_pid != (pid_t)id) {
                       p = *q = proc_find_raw((pid_t)id);
                       if (p == NULL || p->p_stat == SIDL || p->p_pptr != pp) {
                               *q = NULL;
                               return -1;
                       }
               }
               mutex_enter(p->p_lock);
               rv++;
               break;
       case P_PGID:
               if (p->p_pgid != (pid_t)id)
                       return 0;
               mutex_enter(p->p_lock);
               break;
       case P_SID:
               if (p->p_session->s_sid != (pid_t)id)
                       return 0;
               mutex_enter(p->p_lock);
               break;
       case P_UID:
               mutex_enter(p->p_lock);
               if (kauth_cred_geteuid(p->p_cred) != (uid_t)id) {
                       mutex_exit(p->p_lock);
                       return 0;
               }
               break;
       case P_GID:
               mutex_enter(p->p_lock);
               if (kauth_cred_getegid(p->p_cred) != (gid_t)id) {
                       mutex_exit(p->p_lock);
                       return 0;
               }
               break;
       case P_CID:
       case P_PSETID:
       case P_CPUID:
               /* XXX: Implement me */
       default:
               return 0;
       }

       if ((options & WEXITED) == 0 && p->p_stat == SZOMB) {
               mutex_exit(p->p_lock);
               return 0;
       }

       if (siginfo != NULL) {
               siginfo->si_errno = 0;

               /*
                * SUSv4 requires that the si_signo value is always
                * SIGCHLD. Obey it despite the rfork(2) interface
                * allows to request other signal for child exit
                * notification.
                */
               siginfo->si_signo = SIGCHLD;

               /*
                *  This is still a rough estimate.  We will fix the
                *  cases TRAPPED, STOPPED, and CONTINUED later.
                */
               if (p->p_sflag & PS_COREDUMP) {
                       siginfo->si_code = CLD_DUMPED;
                       siginfo->si_status = p->p_xsig;
               } else if (p->p_xsig) {
                       siginfo->si_code = CLD_KILLED;
                       siginfo->si_status = p->p_xsig;
               } else {
                       siginfo->si_code = CLD_EXITED;
                       siginfo->si_status = p->p_xexit;
               }

               siginfo->si_pid = p->p_pid;
               siginfo->si_uid = kauth_cred_geteuid(p->p_cred);
               siginfo->si_utime = p->p_stats->p_ru.ru_utime.tv_sec;
               siginfo->si_stime = p->p_stats->p_ru.ru_stime.tv_sec;
       }

       /*
        * There should be no reason to limit resources usage info to
        * exited processes only.  A snapshot about any resources used
        * by a stopped process may be exactly what is needed.
        */
       if (wrusage != NULL) {
               rup = &wrusage->wru_self;
               *rup = p->p_stats->p_ru;
               calcru(p, &rup->ru_utime, &rup->ru_stime, NULL, NULL);

               rup = &wrusage->wru_children;
               *rup = p->p_stats->p_cru;
               calcru(p, &rup->ru_utime, &rup->ru_stime, NULL, NULL);
       }

       mutex_exit(p->p_lock);
       return rv;
}

/*
* Determine if there are existing processes being debugged
* that used to be (and sometime later will be again) children
* of a specific parent (while matching wait criteria)
*/
static bool
debugged_child_exists(idtype_t idtype, id_t id, int options, siginfo_t *si,
   const struct proc *parent)
{
       struct proc *pp;

       /*
        * If we are searching for a specific pid, we can optimise a little
        */
       if (idtype == P_PID) {
               /*
                * Check the specific process to see if its real parent is us
                */
               pp = proc_find_raw((pid_t)id);
               if (pp != NULL && pp->p_stat != SIDL && pp->p_opptr == parent) {
                       /*
                        * using P_ALL here avoids match_process() doing the
                        * same work that we just did, but incorrectly for
                        * this scenario.
                        */
                       if (match_process(parent, &pp, P_ALL, id, options,
                           NULL, si))
                               return true;
               }
               return false;
       }

       /*
        * For the hard cases, just look everywhere to see if some
        * stolen (reparented) process is really our lost child.
        * Then check if that process could satisfy the wait conditions.
        */

       /*
        * XXX inefficient, but hopefully fairly rare.
        * XXX should really use a list of reparented processes.
        */
       PROCLIST_FOREACH(pp, &allproc) {
               if (pp->p_stat == SIDL)         /* XXX impossible ?? */
                       continue;
               if (pp->p_opptr == parent &&
                   match_process(parent, &pp, idtype, id, options, NULL, si))
                       return true;
       }
       PROCLIST_FOREACH(pp, &zombproc) {
               if (pp->p_stat == SIDL)         /* XXX impossible ?? */
                       continue;
               if (pp->p_opptr == parent &&
                   match_process(parent, &pp, idtype, id, options, NULL, si))
                       return true;
       }

       return false;
}

/*
* Scan list of child processes for a child process that has stopped or
* exited.  Used by sys_wait4 and 'compat' equivalents.
*
* Must be called with the proc_lock held, and may release while waiting.
*/
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)
{
       struct proc *child, *dead;
       int error;

       KASSERT(mutex_owned(&proc_lock));

       if (options & ~WALLOPTS) {
               *child_p = NULL;
               return SET_ERROR(EINVAL);
       }

       if ((options & WSELECTOPTS) == 0) {
               /*
                * We will be unable to find any matching processes,
                * because there are no known events to look for.
                * Prefer to return error instead of blocking
                * indefinitely.
                */
               *child_p = NULL;
               return SET_ERROR(EINVAL);
       }

       if ((pid_t)id == WAIT_MYPGRP && (idtype == P_PID || idtype == P_PGID)) {
               id = (id_t)parent->p_pgid;
               idtype = P_PGID;
       }

       for (;;) {
               error = ECHILD;
               dead = NULL;

               LIST_FOREACH(child, &parent->p_children, p_sibling) {
                       int rv = match_process(parent, &child, idtype, id,
                           options, wru, si);
                       if (rv == -1)
                               break;
                       if (rv == 0)
                               continue;

                       /*
                        * Wait for processes with p_exitsig != SIGCHLD
                        * processes only if WALTSIG is set; wait for
                        * processes with p_exitsig == SIGCHLD only
                        * if WALTSIG is clear.
                        */
                       if (((options & WALLSIG) == 0) &&
                           (options & WALTSIG ? child->p_exitsig == SIGCHLD
                                               : P_EXITSIG(child) != SIGCHLD)){
                               if (rv == 2) {
                                       child = NULL;
                                       break;
                               }
                               continue;
                       }

                       error = 0;
                       if ((options & WNOZOMBIE) == 0) {
                               if (child->p_stat == SZOMB)
                                       break;
                               if (child->p_stat == SDEAD) {
                                       /*
                                        * We may occasionally arrive here
                                        * after receiving a signal, but
                                        * immediately before the child
                                        * process is zombified.  The wait
                                        * will be short, so avoid returning
                                        * to userspace.
                                        */
                                       dead = child;
                               }
                       }

                       if ((options & WCONTINUED) != 0 &&
                           child->p_xsig == SIGCONT &&
                           (child->p_sflag & PS_CONTINUED)) {
                               if ((options & WNOWAIT) == 0) {
                                       child->p_sflag &= ~PS_CONTINUED;
                                       child->p_waited = 1;
                                       parent->p_nstopchild--;
                               }
                               if (si) {
                                       si->si_status = child->p_xsig;
                                       si->si_code = CLD_CONTINUED;
                               }
                               break;
                       }

                       if ((options & (WTRAPPED|WSTOPPED)) != 0 &&
                           child->p_stat == SSTOP &&
                           child->p_waited == 0 &&
                           ((child->p_slflag & PSL_TRACED) ||
                           options & (WUNTRACED|WSTOPPED))) {
                               if ((options & WNOWAIT) == 0) {
                                       child->p_waited = 1;
                                       parent->p_nstopchild--;
                               }
                               if (si) {
                                       si->si_status = child->p_xsig;
                                       si->si_code =
                                           (child->p_slflag & PSL_TRACED) ?
                                           CLD_TRAPPED : CLD_STOPPED;
                               }
                               break;
                       }
                       if (parent->p_nstopchild == 0 || rv == 2) {
                               child = NULL;
                               break;
                       }
               }

               /*
                * If we found nothing, but we are the bereaved parent
                * of a stolen child, look and see if that child (or
                * one of them) meets our search criteria.   If so, then
                * we cannot succeed, but we can hang (wait...),
                * or if WNOHANG, return 0 instead of ECHILD
                */
               if (child == NULL && error == ECHILD &&
                   (parent->p_slflag & PSL_CHTRACED) &&
                   debugged_child_exists(idtype, id, options, si, parent))
                       error = 0;

               if (child != NULL || error != 0 ||
                   ((options & WNOHANG) != 0 && dead == NULL)) {
                       *child_p = child;
                       return SET_ERROR(error);
               }

               /*
                * Wait for another child process to stop.
                */
               error = cv_wait_sig(&parent->p_waitcv, &proc_lock);

               if (error != 0) {
                       *child_p = NULL;
                       return error;
               }
       }
}

/*
* Free a process after parent has taken all the state info.  Must be called
* with the proclist lock held, and will release before returning.
*
* *ru is returned to the caller, and must be freed by the caller.
*/
static void
proc_free(struct proc *p, struct wrusage *wru)
{
       struct proc *parent = p->p_pptr;
       struct lwp *l;
       ksiginfo_t ksi;
       kauth_cred_t cred1, cred2;
       uid_t uid;

       KASSERT(mutex_owned(&proc_lock));
       KASSERT(p->p_nlwps == 1);
       KASSERT(p->p_nzlwps == 1);
       KASSERT(p->p_nrlwps == 0);
       KASSERT(p->p_stat == SZOMB);

       /*
        * If we got the child via ptrace(2) or procfs, and
        * the parent is different (meaning the process was
        * attached, rather than run as a child), then we need
        * to give it back to the old parent, and send the
        * parent the exit signal.  The rest of the cleanup
        * will be done when the old parent waits on the child.
        */
       if ((p->p_slflag & PSL_TRACED) != 0 && p->p_opptr != parent) {
               mutex_enter(p->p_lock);
               p->p_slflag &= ~(PSL_TRACED|PSL_SYSCALL);
               mutex_exit(p->p_lock);
               parent = (p->p_opptr == NULL) ? initproc : p->p_opptr;
               proc_reparent(p, parent);
               p->p_opptr = NULL;
               if (p->p_exitsig != 0) {
                       exit_psignal(p, parent, &ksi);
                       kpsignal(parent, &ksi, NULL);
               }
               cv_broadcast(&parent->p_waitcv);
               mutex_exit(&proc_lock);
               return;
       }

       sched_proc_exit(parent, p);

       /*
        * Add child times of exiting process onto its own times.
        * This cannot be done any earlier else it might get done twice.
        */
       l = LIST_FIRST(&p->p_lwps);
       ruadd(&p->p_stats->p_ru, &l->l_ru);
       ruadd(&p->p_stats->p_ru, &p->p_stats->p_cru);
       ruadd(&parent->p_stats->p_cru, &p->p_stats->p_ru);
       if (wru != NULL) {
               wru->wru_self = p->p_stats->p_ru;
               wru->wru_children = p->p_stats->p_cru;
       }
       p->p_xsig = 0;
       p->p_xexit = 0;

       /*
        * At this point we are going to start freeing the final resources.
        * If anyone tries to access the proc structure after here they will
        * get a shock - bits are missing.  Attempt to make it hard!  We
        * don't bother with any further locking past this point.
        */
       p->p_stat = SIDL;               /* not even a zombie any more */
       LIST_REMOVE(p, p_list); /* off zombproc */
       parent->p_nstopchild--;
       LIST_REMOVE(p, p_sibling);

       /*
        * Let pid be reallocated.
        */
       proc_free_pid(p->p_pid);
       atomic_dec_uint(&nprocs);

       /*
        * Unlink process from its process group.
        * Releases the proc_lock.
        */
       proc_leavepgrp(p);

       /*
        * Delay release until after lwp_free.
        */
       cred2 = l->l_cred;

       /*
        * Free the last LWP's resources.
        *
        * lwp_free ensures the LWP is no longer running on another CPU.
        */
       lwp_free(l, false, true);

       /*
        * Now no one except us can reach the process p.
        */

       /*
        * Decrement the count of procs running with this uid.
        */
       cred1 = p->p_cred;
       uid = kauth_cred_getuid(cred1);
       (void)chgproccnt(uid, -1);

       /*
        * Release substructures.
        */

       lim_free(p->p_limit);
       pstatsfree(p->p_stats);
       kauth_cred_free(cred1);
       kauth_cred_free(cred2);

       /*
        * Release reference to text vnode
        */
       if (p->p_textvp)
               vrele(p->p_textvp);
       kmem_strfree(p->p_path);

       mutex_destroy(&p->p_auxlock);
       mutex_obj_free(p->p_lock);
       mutex_destroy(&p->p_stmutex);
       cv_destroy(&p->p_waitcv);
       cv_destroy(&p->p_lwpcv);
       rw_destroy(&p->p_reflock);

       proc_free_mem(p);
}

/*
* Change the parent of a process for tracing purposes.
*/
void
proc_changeparent(struct proc *t, struct proc *p)
{
       SET(t->p_slflag, PSL_TRACED);
       t->p_opptr = t->p_pptr;
       if (t->p_pptr == p)
               return;
       struct proc *parent = t->p_pptr;

       if (parent->p_lock < t->p_lock) {
               if (!mutex_tryenter(parent->p_lock)) {
                       mutex_exit(t->p_lock);
                       mutex_enter(parent->p_lock);
                       mutex_enter(t->p_lock);
               }
       } else if (parent->p_lock > t->p_lock) {
               mutex_enter(parent->p_lock);
       }
       parent->p_slflag |= PSL_CHTRACED;
       proc_reparent(t, p);
       if (parent->p_lock != t->p_lock)
               mutex_exit(parent->p_lock);
}

/*
* make process 'parent' the new parent of process 'child'.
*
* Must be called with proc_lock held.
*/
void
proc_reparent(struct proc *child, struct proc *parent)
{

       KASSERT(mutex_owned(&proc_lock));

       if (child->p_pptr == parent)
               return;

       if (child->p_stat == SZOMB || child->p_stat == SDEAD ||
           (child->p_stat == SSTOP && !child->p_waited)) {
               child->p_pptr->p_nstopchild--;
               parent->p_nstopchild++;
       }
       if (parent == initproc) {
               child->p_exitsig = SIGCHLD;
               child->p_ppid = parent->p_pid;
       }

       LIST_REMOVE(child, p_sibling);
       LIST_INSERT_HEAD(&parent->p_children, child, p_sibling);
       child->p_pptr = parent;
}