Home | History | Annotate | Line # | Download | only in kern
kern_proc.c revision 1.221
      1 /*	$NetBSD: kern_proc.c,v 1.221 2018/11/24 19:22:17 christos Exp $	*/
      2 
      3 /*-
      4  * Copyright (c) 1999, 2006, 2007, 2008 The NetBSD Foundation, Inc.
      5  * All rights reserved.
      6  *
      7  * This code is derived from software contributed to The NetBSD Foundation
      8  * by Jason R. Thorpe of the Numerical Aerospace Simulation Facility,
      9  * NASA Ames Research Center, and by Andrew Doran.
     10  *
     11  * Redistribution and use in source and binary forms, with or without
     12  * modification, are permitted provided that the following conditions
     13  * are met:
     14  * 1. Redistributions of source code must retain the above copyright
     15  *    notice, this list of conditions and the following disclaimer.
     16  * 2. Redistributions in binary form must reproduce the above copyright
     17  *    notice, this list of conditions and the following disclaimer in the
     18  *    documentation and/or other materials provided with the distribution.
     19  *
     20  * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
     21  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
     22  * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
     23  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
     24  * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
     25  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
     26  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
     27  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
     28  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
     29  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
     30  * POSSIBILITY OF SUCH DAMAGE.
     31  */
     32 
     33 /*
     34  * Copyright (c) 1982, 1986, 1989, 1991, 1993
     35  *	The Regents of the University of California.  All rights reserved.
     36  *
     37  * Redistribution and use in source and binary forms, with or without
     38  * modification, are permitted provided that the following conditions
     39  * are met:
     40  * 1. Redistributions of source code must retain the above copyright
     41  *    notice, this list of conditions and the following disclaimer.
     42  * 2. Redistributions in binary form must reproduce the above copyright
     43  *    notice, this list of conditions and the following disclaimer in the
     44  *    documentation and/or other materials provided with the distribution.
     45  * 3. Neither the name of the University nor the names of its contributors
     46  *    may be used to endorse or promote products derived from this software
     47  *    without specific prior written permission.
     48  *
     49  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
     50  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     51  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     52  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
     53  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     54  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     55  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     56  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     57  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     58  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     59  * SUCH DAMAGE.
     60  *
     61  *	@(#)kern_proc.c	8.7 (Berkeley) 2/14/95
     62  */
     63 
     64 #include <sys/cdefs.h>
     65 __KERNEL_RCSID(0, "$NetBSD: kern_proc.c,v 1.221 2018/11/24 19:22:17 christos Exp $");
     66 
     67 #ifdef _KERNEL_OPT
     68 #include "opt_kstack.h"
     69 #include "opt_maxuprc.h"
     70 #include "opt_dtrace.h"
     71 #include "opt_compat_netbsd32.h"
     72 #endif
     73 
     74 #if defined(__HAVE_COMPAT_NETBSD32) && !defined(COMPAT_NETBSD32) \
     75     && !defined(_RUMPKERNEL)
     76 #define COMPAT_NETBSD32
     77 #endif
     78 
     79 #include <sys/param.h>
     80 #include <sys/systm.h>
     81 #include <sys/kernel.h>
     82 #include <sys/proc.h>
     83 #include <sys/resourcevar.h>
     84 #include <sys/buf.h>
     85 #include <sys/acct.h>
     86 #include <sys/wait.h>
     87 #include <sys/file.h>
     88 #include <ufs/ufs/quota.h>
     89 #include <sys/uio.h>
     90 #include <sys/pool.h>
     91 #include <sys/pset.h>
     92 #include <sys/mbuf.h>
     93 #include <sys/ioctl.h>
     94 #include <sys/tty.h>
     95 #include <sys/signalvar.h>
     96 #include <sys/ras.h>
     97 #include <sys/filedesc.h>
     98 #include <sys/syscall_stats.h>
     99 #include <sys/kauth.h>
    100 #include <sys/sleepq.h>
    101 #include <sys/atomic.h>
    102 #include <sys/kmem.h>
    103 #include <sys/namei.h>
    104 #include <sys/dtrace_bsd.h>
    105 #include <sys/sysctl.h>
    106 #include <sys/exec.h>
    107 #include <sys/cpu.h>
    108 
    109 #include <uvm/uvm_extern.h>
    110 #include <uvm/uvm.h>
    111 
    112 #ifdef COMPAT_NETBSD32
    113 #include <compat/netbsd32/netbsd32.h>
    114 #endif
    115 
    116 /*
    117  * Process lists.
    118  */
    119 
    120 struct proclist		allproc		__cacheline_aligned;
    121 struct proclist		zombproc	__cacheline_aligned;
    122 
    123 kmutex_t *		proc_lock	__cacheline_aligned;
    124 
    125 /*
    126  * pid to proc lookup is done by indexing the pid_table array.
    127  * Since pid numbers are only allocated when an empty slot
    128  * has been found, there is no need to search any lists ever.
    129  * (an orphaned pgrp will lock the slot, a session will lock
    130  * the pgrp with the same number.)
    131  * If the table is too small it is reallocated with twice the
    132  * previous size and the entries 'unzipped' into the two halves.
    133  * A linked list of free entries is passed through the pt_proc
    134  * field of 'free' items - set odd to be an invalid ptr.
    135  */
    136 
    137 struct pid_table {
    138 	struct proc	*pt_proc;
    139 	struct pgrp	*pt_pgrp;
    140 	pid_t		pt_pid;
    141 };
    142 #if 1	/* strongly typed cast - should be a noop */
    143 static inline uint p2u(struct proc *p) { return (uint)(uintptr_t)p; }
    144 #else
    145 #define p2u(p) ((uint)p)
    146 #endif
    147 #define P_VALID(p) (!(p2u(p) & 1))
    148 #define P_NEXT(p) (p2u(p) >> 1)
    149 #define P_FREE(pid) ((struct proc *)(uintptr_t)((pid) << 1 | 1))
    150 
    151 /*
    152  * Table of process IDs (PIDs).
    153  */
    154 static struct pid_table *pid_table	__read_mostly;
    155 
    156 #define	INITIAL_PID_TABLE_SIZE		(1 << 5)
    157 
    158 /* Table mask, threshold for growing and number of allocated PIDs. */
    159 static u_int		pid_tbl_mask	__read_mostly;
    160 static u_int		pid_alloc_lim	__read_mostly;
    161 static u_int		pid_alloc_cnt	__cacheline_aligned;
    162 
    163 /* Next free, last free and maximum PIDs. */
    164 static u_int		next_free_pt	__cacheline_aligned;
    165 static u_int		last_free_pt	__cacheline_aligned;
    166 static pid_t		pid_max		__read_mostly;
    167 
    168 /* Components of the first process -- never freed. */
    169 
    170 extern struct emul emul_netbsd;	/* defined in kern_exec.c */
    171 
    172 struct session session0 = {
    173 	.s_count = 1,
    174 	.s_sid = 0,
    175 };
    176 struct pgrp pgrp0 = {
    177 	.pg_members = LIST_HEAD_INITIALIZER(&pgrp0.pg_members),
    178 	.pg_session = &session0,
    179 };
    180 filedesc_t filedesc0;
    181 struct cwdinfo cwdi0 = {
    182 	.cwdi_cmask = CMASK,
    183 	.cwdi_refcnt = 1,
    184 };
    185 struct plimit limit0;
    186 struct pstats pstat0;
    187 struct vmspace vmspace0;
    188 struct sigacts sigacts0;
    189 struct proc proc0 = {
    190 	.p_lwps = LIST_HEAD_INITIALIZER(&proc0.p_lwps),
    191 	.p_sigwaiters = LIST_HEAD_INITIALIZER(&proc0.p_sigwaiters),
    192 	.p_nlwps = 1,
    193 	.p_nrlwps = 1,
    194 	.p_nlwpid = 1,		/* must match lwp0.l_lid */
    195 	.p_pgrp = &pgrp0,
    196 	.p_comm = "system",
    197 	/*
    198 	 * Set P_NOCLDWAIT so that kernel threads are reparented to init(8)
    199 	 * when they exit.  init(8) can easily wait them out for us.
    200 	 */
    201 	.p_flag = PK_SYSTEM | PK_NOCLDWAIT,
    202 	.p_stat = SACTIVE,
    203 	.p_nice = NZERO,
    204 	.p_emul = &emul_netbsd,
    205 	.p_cwdi = &cwdi0,
    206 	.p_limit = &limit0,
    207 	.p_fd = &filedesc0,
    208 	.p_vmspace = &vmspace0,
    209 	.p_stats = &pstat0,
    210 	.p_sigacts = &sigacts0,
    211 #ifdef PROC0_MD_INITIALIZERS
    212 	PROC0_MD_INITIALIZERS
    213 #endif
    214 };
    215 kauth_cred_t cred0;
    216 
    217 static const int	nofile	= NOFILE;
    218 static const int	maxuprc	= MAXUPRC;
    219 
    220 static int sysctl_doeproc(SYSCTLFN_PROTO);
    221 static int sysctl_kern_proc_args(SYSCTLFN_PROTO);
    222 
    223 /*
    224  * The process list descriptors, used during pid allocation and
    225  * by sysctl.  No locking on this data structure is needed since
    226  * it is completely static.
    227  */
    228 const struct proclist_desc proclists[] = {
    229 	{ &allproc	},
    230 	{ &zombproc	},
    231 	{ NULL		},
    232 };
    233 
    234 static struct pgrp *	pg_remove(pid_t);
    235 static void		pg_delete(pid_t);
    236 static void		orphanpg(struct pgrp *);
    237 
    238 static specificdata_domain_t proc_specificdata_domain;
    239 
    240 static pool_cache_t proc_cache;
    241 
    242 static kauth_listener_t proc_listener;
    243 
    244 static void fill_proc(const struct proc *, struct proc *);
    245 static int fill_pathname(struct lwp *, pid_t, void *, size_t *);
    246 
    247 static int
    248 proc_listener_cb(kauth_cred_t cred, kauth_action_t action, void *cookie,
    249     void *arg0, void *arg1, void *arg2, void *arg3)
    250 {
    251 	struct proc *p;
    252 	int result;
    253 
    254 	result = KAUTH_RESULT_DEFER;
    255 	p = arg0;
    256 
    257 	switch (action) {
    258 	case KAUTH_PROCESS_CANSEE: {
    259 		enum kauth_process_req req;
    260 
    261 		req = (enum kauth_process_req)arg1;
    262 
    263 		switch (req) {
    264 		case KAUTH_REQ_PROCESS_CANSEE_ARGS:
    265 		case KAUTH_REQ_PROCESS_CANSEE_ENTRY:
    266 		case KAUTH_REQ_PROCESS_CANSEE_OPENFILES:
    267 		case KAUTH_REQ_PROCESS_CANSEE_EPROC:
    268 			result = KAUTH_RESULT_ALLOW;
    269 			break;
    270 
    271 		case KAUTH_REQ_PROCESS_CANSEE_ENV:
    272 			if (kauth_cred_getuid(cred) !=
    273 			    kauth_cred_getuid(p->p_cred) ||
    274 			    kauth_cred_getuid(cred) !=
    275 			    kauth_cred_getsvuid(p->p_cred))
    276 				break;
    277 
    278 			result = KAUTH_RESULT_ALLOW;
    279 
    280 			break;
    281 
    282 		case KAUTH_REQ_PROCESS_CANSEE_KPTR:
    283 		default:
    284 			break;
    285 		}
    286 
    287 		break;
    288 		}
    289 
    290 	case KAUTH_PROCESS_FORK: {
    291 		int lnprocs = (int)(unsigned long)arg2;
    292 
    293 		/*
    294 		 * Don't allow a nonprivileged user to use the last few
    295 		 * processes. The variable lnprocs is the current number of
    296 		 * processes, maxproc is the limit.
    297 		 */
    298 		if (__predict_false((lnprocs >= maxproc - 5)))
    299 			break;
    300 
    301 		result = KAUTH_RESULT_ALLOW;
    302 
    303 		break;
    304 		}
    305 
    306 	case KAUTH_PROCESS_CORENAME:
    307 	case KAUTH_PROCESS_STOPFLAG:
    308 		if (proc_uidmatch(cred, p->p_cred) == 0)
    309 			result = KAUTH_RESULT_ALLOW;
    310 
    311 		break;
    312 
    313 	default:
    314 		break;
    315 	}
    316 
    317 	return result;
    318 }
    319 
    320 static int
    321 proc_ctor(void *arg __unused, void *obj, int flags __unused)
    322 {
    323 	memset(obj, 0, sizeof(struct proc));
    324 	return 0;
    325 }
    326 
    327 /*
    328  * Initialize global process hashing structures.
    329  */
    330 void
    331 procinit(void)
    332 {
    333 	const struct proclist_desc *pd;
    334 	u_int i;
    335 #define	LINK_EMPTY ((PID_MAX + INITIAL_PID_TABLE_SIZE) & ~(INITIAL_PID_TABLE_SIZE - 1))
    336 
    337 	for (pd = proclists; pd->pd_list != NULL; pd++)
    338 		LIST_INIT(pd->pd_list);
    339 
    340 	proc_lock = mutex_obj_alloc(MUTEX_DEFAULT, IPL_NONE);
    341 	pid_table = kmem_alloc(INITIAL_PID_TABLE_SIZE
    342 	    * sizeof(struct pid_table), KM_SLEEP);
    343 	pid_tbl_mask = INITIAL_PID_TABLE_SIZE - 1;
    344 	pid_max = PID_MAX;
    345 
    346 	/* Set free list running through table...
    347 	   Preset 'use count' above PID_MAX so we allocate pid 1 next. */
    348 	for (i = 0; i <= pid_tbl_mask; i++) {
    349 		pid_table[i].pt_proc = P_FREE(LINK_EMPTY + i + 1);
    350 		pid_table[i].pt_pgrp = 0;
    351 		pid_table[i].pt_pid = 0;
    352 	}
    353 	/* slot 0 is just grabbed */
    354 	next_free_pt = 1;
    355 	/* Need to fix last entry. */
    356 	last_free_pt = pid_tbl_mask;
    357 	pid_table[last_free_pt].pt_proc = P_FREE(LINK_EMPTY);
    358 	/* point at which we grow table - to avoid reusing pids too often */
    359 	pid_alloc_lim = pid_tbl_mask - 1;
    360 #undef LINK_EMPTY
    361 
    362 	proc_specificdata_domain = specificdata_domain_create();
    363 	KASSERT(proc_specificdata_domain != NULL);
    364 
    365 	proc_cache = pool_cache_init(sizeof(struct proc), 0, 0, 0,
    366 	    "procpl", NULL, IPL_NONE, proc_ctor, NULL, NULL);
    367 
    368 	proc_listener = kauth_listen_scope(KAUTH_SCOPE_PROCESS,
    369 	    proc_listener_cb, NULL);
    370 }
    371 
    372 void
    373 procinit_sysctl(void)
    374 {
    375 	static struct sysctllog *clog;
    376 
    377 	sysctl_createv(&clog, 0, NULL, NULL,
    378 		       CTLFLAG_PERMANENT,
    379 		       CTLTYPE_NODE, "proc",
    380 		       SYSCTL_DESCR("System-wide process information"),
    381 		       sysctl_doeproc, 0, NULL, 0,
    382 		       CTL_KERN, KERN_PROC, CTL_EOL);
    383 	sysctl_createv(&clog, 0, NULL, NULL,
    384 		       CTLFLAG_PERMANENT,
    385 		       CTLTYPE_NODE, "proc2",
    386 		       SYSCTL_DESCR("Machine-independent process information"),
    387 		       sysctl_doeproc, 0, NULL, 0,
    388 		       CTL_KERN, KERN_PROC2, CTL_EOL);
    389 	sysctl_createv(&clog, 0, NULL, NULL,
    390 		       CTLFLAG_PERMANENT,
    391 		       CTLTYPE_NODE, "proc_args",
    392 		       SYSCTL_DESCR("Process argument information"),
    393 		       sysctl_kern_proc_args, 0, NULL, 0,
    394 		       CTL_KERN, KERN_PROC_ARGS, CTL_EOL);
    395 
    396 	/*
    397 	  "nodes" under these:
    398 
    399 	  KERN_PROC_ALL
    400 	  KERN_PROC_PID pid
    401 	  KERN_PROC_PGRP pgrp
    402 	  KERN_PROC_SESSION sess
    403 	  KERN_PROC_TTY tty
    404 	  KERN_PROC_UID uid
    405 	  KERN_PROC_RUID uid
    406 	  KERN_PROC_GID gid
    407 	  KERN_PROC_RGID gid
    408 
    409 	  all in all, probably not worth the effort...
    410 	*/
    411 }
    412 
    413 /*
    414  * Initialize process 0.
    415  */
    416 void
    417 proc0_init(void)
    418 {
    419 	struct proc *p;
    420 	struct pgrp *pg;
    421 	struct rlimit *rlim;
    422 	rlim_t lim;
    423 	int i;
    424 
    425 	p = &proc0;
    426 	pg = &pgrp0;
    427 
    428 	mutex_init(&p->p_stmutex, MUTEX_DEFAULT, IPL_HIGH);
    429 	mutex_init(&p->p_auxlock, MUTEX_DEFAULT, IPL_NONE);
    430 	p->p_lock = mutex_obj_alloc(MUTEX_DEFAULT, IPL_NONE);
    431 
    432 	rw_init(&p->p_reflock);
    433 	cv_init(&p->p_waitcv, "wait");
    434 	cv_init(&p->p_lwpcv, "lwpwait");
    435 
    436 	LIST_INSERT_HEAD(&p->p_lwps, &lwp0, l_sibling);
    437 
    438 	pid_table[0].pt_proc = p;
    439 	LIST_INSERT_HEAD(&allproc, p, p_list);
    440 
    441 	pid_table[0].pt_pgrp = pg;
    442 	LIST_INSERT_HEAD(&pg->pg_members, p, p_pglist);
    443 
    444 #ifdef __HAVE_SYSCALL_INTERN
    445 	(*p->p_emul->e_syscall_intern)(p);
    446 #endif
    447 
    448 	/* Create credentials. */
    449 	cred0 = kauth_cred_alloc();
    450 	p->p_cred = cred0;
    451 
    452 	/* Create the CWD info. */
    453 	rw_init(&cwdi0.cwdi_lock);
    454 
    455 	/* Create the limits structures. */
    456 	mutex_init(&limit0.pl_lock, MUTEX_DEFAULT, IPL_NONE);
    457 
    458 	rlim = limit0.pl_rlimit;
    459 	for (i = 0; i < __arraycount(limit0.pl_rlimit); i++) {
    460 		rlim[i].rlim_cur = RLIM_INFINITY;
    461 		rlim[i].rlim_max = RLIM_INFINITY;
    462 	}
    463 
    464 	rlim[RLIMIT_NOFILE].rlim_max = maxfiles;
    465 	rlim[RLIMIT_NOFILE].rlim_cur = maxfiles < nofile ? maxfiles : nofile;
    466 
    467 	rlim[RLIMIT_NPROC].rlim_max = maxproc;
    468 	rlim[RLIMIT_NPROC].rlim_cur = maxproc < maxuprc ? maxproc : maxuprc;
    469 
    470 	lim = MIN(VM_MAXUSER_ADDRESS, ctob((rlim_t)uvmexp.free));
    471 	rlim[RLIMIT_RSS].rlim_max = lim;
    472 	rlim[RLIMIT_MEMLOCK].rlim_max = lim;
    473 	rlim[RLIMIT_MEMLOCK].rlim_cur = lim / 3;
    474 
    475 	rlim[RLIMIT_NTHR].rlim_max = maxlwp;
    476 	rlim[RLIMIT_NTHR].rlim_cur = maxlwp < maxuprc ? maxlwp : maxuprc;
    477 
    478 	/* Note that default core name has zero length. */
    479 	limit0.pl_corename = defcorename;
    480 	limit0.pl_cnlen = 0;
    481 	limit0.pl_refcnt = 1;
    482 	limit0.pl_writeable = false;
    483 	limit0.pl_sv_limit = NULL;
    484 
    485 	/* Configure virtual memory system, set vm rlimits. */
    486 	uvm_init_limits(p);
    487 
    488 	/* Initialize file descriptor table for proc0. */
    489 	fd_init(&filedesc0);
    490 
    491 	/*
    492 	 * Initialize proc0's vmspace, which uses the kernel pmap.
    493 	 * All kernel processes (which never have user space mappings)
    494 	 * share proc0's vmspace, and thus, the kernel pmap.
    495 	 */
    496 	uvmspace_init(&vmspace0, pmap_kernel(), round_page(VM_MIN_ADDRESS),
    497 	    trunc_page(VM_MAXUSER_ADDRESS),
    498 #ifdef __USE_TOPDOWN_VM
    499 	    true
    500 #else
    501 	    false
    502 #endif
    503 	    );
    504 
    505 	/* Initialize signal state for proc0. XXX IPL_SCHED */
    506 	mutex_init(&p->p_sigacts->sa_mutex, MUTEX_DEFAULT, IPL_SCHED);
    507 	siginit(p);
    508 
    509 	proc_initspecific(p);
    510 	kdtrace_proc_ctor(NULL, p);
    511 }
    512 
    513 /*
    514  * Session reference counting.
    515  */
    516 
    517 void
    518 proc_sesshold(struct session *ss)
    519 {
    520 
    521 	KASSERT(mutex_owned(proc_lock));
    522 	ss->s_count++;
    523 }
    524 
    525 void
    526 proc_sessrele(struct session *ss)
    527 {
    528 
    529 	KASSERT(mutex_owned(proc_lock));
    530 	/*
    531 	 * We keep the pgrp with the same id as the session in order to
    532 	 * stop a process being given the same pid.  Since the pgrp holds
    533 	 * a reference to the session, it must be a 'zombie' pgrp by now.
    534 	 */
    535 	if (--ss->s_count == 0) {
    536 		struct pgrp *pg;
    537 
    538 		pg = pg_remove(ss->s_sid);
    539 		mutex_exit(proc_lock);
    540 
    541 		kmem_free(pg, sizeof(struct pgrp));
    542 		kmem_free(ss, sizeof(struct session));
    543 	} else {
    544 		mutex_exit(proc_lock);
    545 	}
    546 }
    547 
    548 /*
    549  * Check that the specified process group is in the session of the
    550  * specified process.
    551  * Treats -ve ids as process ids.
    552  * Used to validate TIOCSPGRP requests.
    553  */
    554 int
    555 pgid_in_session(struct proc *p, pid_t pg_id)
    556 {
    557 	struct pgrp *pgrp;
    558 	struct session *session;
    559 	int error;
    560 
    561 	mutex_enter(proc_lock);
    562 	if (pg_id < 0) {
    563 		struct proc *p1 = proc_find(-pg_id);
    564 		if (p1 == NULL) {
    565 			error = EINVAL;
    566 			goto fail;
    567 		}
    568 		pgrp = p1->p_pgrp;
    569 	} else {
    570 		pgrp = pgrp_find(pg_id);
    571 		if (pgrp == NULL) {
    572 			error = EINVAL;
    573 			goto fail;
    574 		}
    575 	}
    576 	session = pgrp->pg_session;
    577 	error = (session != p->p_pgrp->pg_session) ? EPERM : 0;
    578 fail:
    579 	mutex_exit(proc_lock);
    580 	return error;
    581 }
    582 
    583 /*
    584  * p_inferior: is p an inferior of q?
    585  */
    586 static inline bool
    587 p_inferior(struct proc *p, struct proc *q)
    588 {
    589 
    590 	KASSERT(mutex_owned(proc_lock));
    591 
    592 	for (; p != q; p = p->p_pptr)
    593 		if (p->p_pid == 0)
    594 			return false;
    595 	return true;
    596 }
    597 
    598 /*
    599  * proc_find: locate a process by the ID.
    600  *
    601  * => Must be called with proc_lock held.
    602  */
    603 proc_t *
    604 proc_find_raw(pid_t pid)
    605 {
    606 	struct pid_table *pt;
    607 	proc_t *p;
    608 
    609 	KASSERT(mutex_owned(proc_lock));
    610 	pt = &pid_table[pid & pid_tbl_mask];
    611 	p = pt->pt_proc;
    612 	if (__predict_false(!P_VALID(p) || pt->pt_pid != pid)) {
    613 		return NULL;
    614 	}
    615 	return p;
    616 }
    617 
    618 proc_t *
    619 proc_find(pid_t pid)
    620 {
    621 	proc_t *p;
    622 
    623 	p = proc_find_raw(pid);
    624 	if (__predict_false(p == NULL)) {
    625 		return NULL;
    626 	}
    627 
    628 	/*
    629 	 * Only allow live processes to be found by PID.
    630 	 * XXX: p_stat might change, since unlocked.
    631 	 */
    632 	if (__predict_true(p->p_stat == SACTIVE || p->p_stat == SSTOP)) {
    633 		return p;
    634 	}
    635 	return NULL;
    636 }
    637 
    638 /*
    639  * pgrp_find: locate a process group by the ID.
    640  *
    641  * => Must be called with proc_lock held.
    642  */
    643 struct pgrp *
    644 pgrp_find(pid_t pgid)
    645 {
    646 	struct pgrp *pg;
    647 
    648 	KASSERT(mutex_owned(proc_lock));
    649 
    650 	pg = pid_table[pgid & pid_tbl_mask].pt_pgrp;
    651 
    652 	/*
    653 	 * Cannot look up a process group that only exists because the
    654 	 * session has not died yet (traditional).
    655 	 */
    656 	if (pg == NULL || pg->pg_id != pgid || LIST_EMPTY(&pg->pg_members)) {
    657 		return NULL;
    658 	}
    659 	return pg;
    660 }
    661 
    662 static void
    663 expand_pid_table(void)
    664 {
    665 	size_t pt_size, tsz;
    666 	struct pid_table *n_pt, *new_pt;
    667 	struct proc *proc;
    668 	struct pgrp *pgrp;
    669 	pid_t pid, rpid;
    670 	u_int i;
    671 	uint new_pt_mask;
    672 
    673 	pt_size = pid_tbl_mask + 1;
    674 	tsz = pt_size * 2 * sizeof(struct pid_table);
    675 	new_pt = kmem_alloc(tsz, KM_SLEEP);
    676 	new_pt_mask = pt_size * 2 - 1;
    677 
    678 	mutex_enter(proc_lock);
    679 	if (pt_size != pid_tbl_mask + 1) {
    680 		/* Another process beat us to it... */
    681 		mutex_exit(proc_lock);
    682 		kmem_free(new_pt, tsz);
    683 		return;
    684 	}
    685 
    686 	/*
    687 	 * Copy entries from old table into new one.
    688 	 * If 'pid' is 'odd' we need to place in the upper half,
    689 	 * even pid's to the lower half.
    690 	 * Free items stay in the low half so we don't have to
    691 	 * fixup the reference to them.
    692 	 * We stuff free items on the front of the freelist
    693 	 * because we can't write to unmodified entries.
    694 	 * Processing the table backwards maintains a semblance
    695 	 * of issuing pid numbers that increase with time.
    696 	 */
    697 	i = pt_size - 1;
    698 	n_pt = new_pt + i;
    699 	for (; ; i--, n_pt--) {
    700 		proc = pid_table[i].pt_proc;
    701 		pgrp = pid_table[i].pt_pgrp;
    702 		if (!P_VALID(proc)) {
    703 			/* Up 'use count' so that link is valid */
    704 			pid = (P_NEXT(proc) + pt_size) & ~pt_size;
    705 			rpid = 0;
    706 			proc = P_FREE(pid);
    707 			if (pgrp)
    708 				pid = pgrp->pg_id;
    709 		} else {
    710 			pid = pid_table[i].pt_pid;
    711 			rpid = pid;
    712 		}
    713 
    714 		/* Save entry in appropriate half of table */
    715 		n_pt[pid & pt_size].pt_proc = proc;
    716 		n_pt[pid & pt_size].pt_pgrp = pgrp;
    717 		n_pt[pid & pt_size].pt_pid = rpid;
    718 
    719 		/* Put other piece on start of free list */
    720 		pid = (pid ^ pt_size) & ~pid_tbl_mask;
    721 		n_pt[pid & pt_size].pt_proc =
    722 			P_FREE((pid & ~pt_size) | next_free_pt);
    723 		n_pt[pid & pt_size].pt_pgrp = 0;
    724 		n_pt[pid & pt_size].pt_pid = 0;
    725 
    726 		next_free_pt = i | (pid & pt_size);
    727 		if (i == 0)
    728 			break;
    729 	}
    730 
    731 	/* Save old table size and switch tables */
    732 	tsz = pt_size * sizeof(struct pid_table);
    733 	n_pt = pid_table;
    734 	pid_table = new_pt;
    735 	pid_tbl_mask = new_pt_mask;
    736 
    737 	/*
    738 	 * pid_max starts as PID_MAX (= 30000), once we have 16384
    739 	 * allocated pids we need it to be larger!
    740 	 */
    741 	if (pid_tbl_mask > PID_MAX) {
    742 		pid_max = pid_tbl_mask * 2 + 1;
    743 		pid_alloc_lim |= pid_alloc_lim << 1;
    744 	} else
    745 		pid_alloc_lim <<= 1;	/* doubles number of free slots... */
    746 
    747 	mutex_exit(proc_lock);
    748 	kmem_free(n_pt, tsz);
    749 }
    750 
    751 struct proc *
    752 proc_alloc(void)
    753 {
    754 	struct proc *p;
    755 
    756 	p = pool_cache_get(proc_cache, PR_WAITOK);
    757 	p->p_stat = SIDL;			/* protect against others */
    758 	proc_initspecific(p);
    759 	kdtrace_proc_ctor(NULL, p);
    760 	p->p_pid = -1;
    761 	proc_alloc_pid(p);
    762 	return p;
    763 }
    764 
    765 /*
    766  * proc_alloc_pid: allocate PID and record the given proc 'p' so that
    767  * proc_find_raw() can find it by the PID.
    768  */
    769 
    770 pid_t
    771 proc_alloc_pid(struct proc *p)
    772 {
    773 	struct pid_table *pt;
    774 	pid_t pid;
    775 	int nxt;
    776 
    777 	for (;;expand_pid_table()) {
    778 		if (__predict_false(pid_alloc_cnt >= pid_alloc_lim))
    779 			/* ensure pids cycle through 2000+ values */
    780 			continue;
    781 		mutex_enter(proc_lock);
    782 		pt = &pid_table[next_free_pt];
    783 #ifdef DIAGNOSTIC
    784 		if (__predict_false(P_VALID(pt->pt_proc) || pt->pt_pgrp))
    785 			panic("proc_alloc: slot busy");
    786 #endif
    787 		nxt = P_NEXT(pt->pt_proc);
    788 		if (nxt & pid_tbl_mask)
    789 			break;
    790 		/* Table full - expand (NB last entry not used....) */
    791 		mutex_exit(proc_lock);
    792 	}
    793 
    794 	/* pid is 'saved use count' + 'size' + entry */
    795 	pid = (nxt & ~pid_tbl_mask) + pid_tbl_mask + 1 + next_free_pt;
    796 	if ((uint)pid > (uint)pid_max)
    797 		pid &= pid_tbl_mask;
    798 	next_free_pt = nxt & pid_tbl_mask;
    799 
    800 	/* Grab table slot */
    801 	pt->pt_proc = p;
    802 
    803 	KASSERT(pt->pt_pid == 0);
    804 	pt->pt_pid = pid;
    805 	if (p->p_pid == -1) {
    806 		p->p_pid = pid;
    807 	}
    808 	pid_alloc_cnt++;
    809 	mutex_exit(proc_lock);
    810 
    811 	return pid;
    812 }
    813 
    814 /*
    815  * Free a process id - called from proc_free (in kern_exit.c)
    816  *
    817  * Called with the proc_lock held.
    818  */
    819 void
    820 proc_free_pid(pid_t pid)
    821 {
    822 	struct pid_table *pt;
    823 
    824 	KASSERT(mutex_owned(proc_lock));
    825 
    826 	pt = &pid_table[pid & pid_tbl_mask];
    827 
    828 	/* save pid use count in slot */
    829 	pt->pt_proc = P_FREE(pid & ~pid_tbl_mask);
    830 	KASSERT(pt->pt_pid == pid);
    831 	pt->pt_pid = 0;
    832 
    833 	if (pt->pt_pgrp == NULL) {
    834 		/* link last freed entry onto ours */
    835 		pid &= pid_tbl_mask;
    836 		pt = &pid_table[last_free_pt];
    837 		pt->pt_proc = P_FREE(P_NEXT(pt->pt_proc) | pid);
    838 		pt->pt_pid = 0;
    839 		last_free_pt = pid;
    840 		pid_alloc_cnt--;
    841 	}
    842 
    843 	atomic_dec_uint(&nprocs);
    844 }
    845 
    846 void
    847 proc_free_mem(struct proc *p)
    848 {
    849 
    850 	kdtrace_proc_dtor(NULL, p);
    851 	pool_cache_put(proc_cache, p);
    852 }
    853 
    854 /*
    855  * proc_enterpgrp: move p to a new or existing process group (and session).
    856  *
    857  * If we are creating a new pgrp, the pgid should equal
    858  * the calling process' pid.
    859  * If is only valid to enter a process group that is in the session
    860  * of the process.
    861  * Also mksess should only be set if we are creating a process group
    862  *
    863  * Only called from sys_setsid, sys_setpgid and posix_spawn/spawn_return.
    864  */
    865 int
    866 proc_enterpgrp(struct proc *curp, pid_t pid, pid_t pgid, bool mksess)
    867 {
    868 	struct pgrp *new_pgrp, *pgrp;
    869 	struct session *sess;
    870 	struct proc *p;
    871 	int rval;
    872 	pid_t pg_id = NO_PGID;
    873 
    874 	sess = mksess ? kmem_alloc(sizeof(*sess), KM_SLEEP) : NULL;
    875 
    876 	/* Allocate data areas we might need before doing any validity checks */
    877 	mutex_enter(proc_lock);		/* Because pid_table might change */
    878 	if (pid_table[pgid & pid_tbl_mask].pt_pgrp == 0) {
    879 		mutex_exit(proc_lock);
    880 		new_pgrp = kmem_alloc(sizeof(*new_pgrp), KM_SLEEP);
    881 		mutex_enter(proc_lock);
    882 	} else
    883 		new_pgrp = NULL;
    884 	rval = EPERM;	/* most common error (to save typing) */
    885 
    886 	/* Check pgrp exists or can be created */
    887 	pgrp = pid_table[pgid & pid_tbl_mask].pt_pgrp;
    888 	if (pgrp != NULL && pgrp->pg_id != pgid)
    889 		goto done;
    890 
    891 	/* Can only set another process under restricted circumstances. */
    892 	if (pid != curp->p_pid) {
    893 		/* Must exist and be one of our children... */
    894 		p = proc_find(pid);
    895 		if (p == NULL || !p_inferior(p, curp)) {
    896 			rval = ESRCH;
    897 			goto done;
    898 		}
    899 		/* ... in the same session... */
    900 		if (sess != NULL || p->p_session != curp->p_session)
    901 			goto done;
    902 		/* ... existing pgid must be in same session ... */
    903 		if (pgrp != NULL && pgrp->pg_session != p->p_session)
    904 			goto done;
    905 		/* ... and not done an exec. */
    906 		if (p->p_flag & PK_EXEC) {
    907 			rval = EACCES;
    908 			goto done;
    909 		}
    910 	} else {
    911 		/* ... setsid() cannot re-enter a pgrp */
    912 		if (mksess && (curp->p_pgid == curp->p_pid ||
    913 		    pgrp_find(curp->p_pid)))
    914 			goto done;
    915 		p = curp;
    916 	}
    917 
    918 	/* Changing the process group/session of a session
    919 	   leader is definitely off limits. */
    920 	if (SESS_LEADER(p)) {
    921 		if (sess == NULL && p->p_pgrp == pgrp)
    922 			/* unless it's a definite noop */
    923 			rval = 0;
    924 		goto done;
    925 	}
    926 
    927 	/* Can only create a process group with id of process */
    928 	if (pgrp == NULL && pgid != pid)
    929 		goto done;
    930 
    931 	/* Can only create a session if creating pgrp */
    932 	if (sess != NULL && pgrp != NULL)
    933 		goto done;
    934 
    935 	/* Check we allocated memory for a pgrp... */
    936 	if (pgrp == NULL && new_pgrp == NULL)
    937 		goto done;
    938 
    939 	/* Don't attach to 'zombie' pgrp */
    940 	if (pgrp != NULL && LIST_EMPTY(&pgrp->pg_members))
    941 		goto done;
    942 
    943 	/* Expect to succeed now */
    944 	rval = 0;
    945 
    946 	if (pgrp == p->p_pgrp)
    947 		/* nothing to do */
    948 		goto done;
    949 
    950 	/* Ok all setup, link up required structures */
    951 
    952 	if (pgrp == NULL) {
    953 		pgrp = new_pgrp;
    954 		new_pgrp = NULL;
    955 		if (sess != NULL) {
    956 			sess->s_sid = p->p_pid;
    957 			sess->s_leader = p;
    958 			sess->s_count = 1;
    959 			sess->s_ttyvp = NULL;
    960 			sess->s_ttyp = NULL;
    961 			sess->s_flags = p->p_session->s_flags & ~S_LOGIN_SET;
    962 			memcpy(sess->s_login, p->p_session->s_login,
    963 			    sizeof(sess->s_login));
    964 			p->p_lflag &= ~PL_CONTROLT;
    965 		} else {
    966 			sess = p->p_pgrp->pg_session;
    967 			proc_sesshold(sess);
    968 		}
    969 		pgrp->pg_session = sess;
    970 		sess = NULL;
    971 
    972 		pgrp->pg_id = pgid;
    973 		LIST_INIT(&pgrp->pg_members);
    974 #ifdef DIAGNOSTIC
    975 		if (__predict_false(pid_table[pgid & pid_tbl_mask].pt_pgrp))
    976 			panic("enterpgrp: pgrp table slot in use");
    977 		if (__predict_false(mksess && p != curp))
    978 			panic("enterpgrp: mksession and p != curproc");
    979 #endif
    980 		pid_table[pgid & pid_tbl_mask].pt_pgrp = pgrp;
    981 		pgrp->pg_jobc = 0;
    982 	}
    983 
    984 	/*
    985 	 * Adjust eligibility of affected pgrps to participate in job control.
    986 	 * Increment eligibility counts before decrementing, otherwise we
    987 	 * could reach 0 spuriously during the first call.
    988 	 */
    989 	fixjobc(p, pgrp, 1);
    990 	fixjobc(p, p->p_pgrp, 0);
    991 
    992 	/* Interlock with ttread(). */
    993 	mutex_spin_enter(&tty_lock);
    994 
    995 	/* Move process to requested group. */
    996 	LIST_REMOVE(p, p_pglist);
    997 	if (LIST_EMPTY(&p->p_pgrp->pg_members))
    998 		/* defer delete until we've dumped the lock */
    999 		pg_id = p->p_pgrp->pg_id;
   1000 	p->p_pgrp = pgrp;
   1001 	LIST_INSERT_HEAD(&pgrp->pg_members, p, p_pglist);
   1002 
   1003 	/* Done with the swap; we can release the tty mutex. */
   1004 	mutex_spin_exit(&tty_lock);
   1005 
   1006     done:
   1007 	if (pg_id != NO_PGID) {
   1008 		/* Releases proc_lock. */
   1009 		pg_delete(pg_id);
   1010 	} else {
   1011 		mutex_exit(proc_lock);
   1012 	}
   1013 	if (sess != NULL)
   1014 		kmem_free(sess, sizeof(*sess));
   1015 	if (new_pgrp != NULL)
   1016 		kmem_free(new_pgrp, sizeof(*new_pgrp));
   1017 #ifdef DEBUG_PGRP
   1018 	if (__predict_false(rval))
   1019 		printf("enterpgrp(%d,%d,%d), curproc %d, rval %d\n",
   1020 			pid, pgid, mksess, curp->p_pid, rval);
   1021 #endif
   1022 	return rval;
   1023 }
   1024 
   1025 /*
   1026  * proc_leavepgrp: remove a process from its process group.
   1027  *  => must be called with the proc_lock held, which will be released;
   1028  */
   1029 void
   1030 proc_leavepgrp(struct proc *p)
   1031 {
   1032 	struct pgrp *pgrp;
   1033 
   1034 	KASSERT(mutex_owned(proc_lock));
   1035 
   1036 	/* Interlock with ttread() */
   1037 	mutex_spin_enter(&tty_lock);
   1038 	pgrp = p->p_pgrp;
   1039 	LIST_REMOVE(p, p_pglist);
   1040 	p->p_pgrp = NULL;
   1041 	mutex_spin_exit(&tty_lock);
   1042 
   1043 	if (LIST_EMPTY(&pgrp->pg_members)) {
   1044 		/* Releases proc_lock. */
   1045 		pg_delete(pgrp->pg_id);
   1046 	} else {
   1047 		mutex_exit(proc_lock);
   1048 	}
   1049 }
   1050 
   1051 /*
   1052  * pg_remove: remove a process group from the table.
   1053  *  => must be called with the proc_lock held;
   1054  *  => returns process group to free;
   1055  */
   1056 static struct pgrp *
   1057 pg_remove(pid_t pg_id)
   1058 {
   1059 	struct pgrp *pgrp;
   1060 	struct pid_table *pt;
   1061 
   1062 	KASSERT(mutex_owned(proc_lock));
   1063 
   1064 	pt = &pid_table[pg_id & pid_tbl_mask];
   1065 	pgrp = pt->pt_pgrp;
   1066 
   1067 	KASSERT(pgrp != NULL);
   1068 	KASSERT(pgrp->pg_id == pg_id);
   1069 	KASSERT(LIST_EMPTY(&pgrp->pg_members));
   1070 
   1071 	pt->pt_pgrp = NULL;
   1072 
   1073 	if (!P_VALID(pt->pt_proc)) {
   1074 		/* Orphaned pgrp, put slot onto free list. */
   1075 		KASSERT((P_NEXT(pt->pt_proc) & pid_tbl_mask) == 0);
   1076 		pg_id &= pid_tbl_mask;
   1077 		pt = &pid_table[last_free_pt];
   1078 		pt->pt_proc = P_FREE(P_NEXT(pt->pt_proc) | pg_id);
   1079 		KASSERT(pt->pt_pid == 0);
   1080 		last_free_pt = pg_id;
   1081 		pid_alloc_cnt--;
   1082 	}
   1083 	return pgrp;
   1084 }
   1085 
   1086 /*
   1087  * pg_delete: delete and free a process group.
   1088  *  => must be called with the proc_lock held, which will be released.
   1089  */
   1090 static void
   1091 pg_delete(pid_t pg_id)
   1092 {
   1093 	struct pgrp *pg;
   1094 	struct tty *ttyp;
   1095 	struct session *ss;
   1096 
   1097 	KASSERT(mutex_owned(proc_lock));
   1098 
   1099 	pg = pid_table[pg_id & pid_tbl_mask].pt_pgrp;
   1100 	if (pg == NULL || pg->pg_id != pg_id || !LIST_EMPTY(&pg->pg_members)) {
   1101 		mutex_exit(proc_lock);
   1102 		return;
   1103 	}
   1104 
   1105 	ss = pg->pg_session;
   1106 
   1107 	/* Remove reference (if any) from tty to this process group */
   1108 	mutex_spin_enter(&tty_lock);
   1109 	ttyp = ss->s_ttyp;
   1110 	if (ttyp != NULL && ttyp->t_pgrp == pg) {
   1111 		ttyp->t_pgrp = NULL;
   1112 		KASSERT(ttyp->t_session == ss);
   1113 	}
   1114 	mutex_spin_exit(&tty_lock);
   1115 
   1116 	/*
   1117 	 * The leading process group in a session is freed by proc_sessrele(),
   1118 	 * if last reference.  Note: proc_sessrele() releases proc_lock.
   1119 	 */
   1120 	pg = (ss->s_sid != pg->pg_id) ? pg_remove(pg_id) : NULL;
   1121 	proc_sessrele(ss);
   1122 
   1123 	if (pg != NULL) {
   1124 		/* Free it, if was not done by proc_sessrele(). */
   1125 		kmem_free(pg, sizeof(struct pgrp));
   1126 	}
   1127 }
   1128 
   1129 /*
   1130  * Adjust pgrp jobc counters when specified process changes process group.
   1131  * We count the number of processes in each process group that "qualify"
   1132  * the group for terminal job control (those with a parent in a different
   1133  * process group of the same session).  If that count reaches zero, the
   1134  * process group becomes orphaned.  Check both the specified process'
   1135  * process group and that of its children.
   1136  * entering == 0 => p is leaving specified group.
   1137  * entering == 1 => p is entering specified group.
   1138  *
   1139  * Call with proc_lock held.
   1140  */
   1141 void
   1142 fixjobc(struct proc *p, struct pgrp *pgrp, int entering)
   1143 {
   1144 	struct pgrp *hispgrp;
   1145 	struct session *mysession = pgrp->pg_session;
   1146 	struct proc *child;
   1147 
   1148 	KASSERT(mutex_owned(proc_lock));
   1149 
   1150 	/*
   1151 	 * Check p's parent to see whether p qualifies its own process
   1152 	 * group; if so, adjust count for p's process group.
   1153 	 */
   1154 	hispgrp = p->p_pptr->p_pgrp;
   1155 	if (hispgrp != pgrp && hispgrp->pg_session == mysession) {
   1156 		if (entering) {
   1157 			pgrp->pg_jobc++;
   1158 			p->p_lflag &= ~PL_ORPHANPG;
   1159 		} else if (--pgrp->pg_jobc == 0)
   1160 			orphanpg(pgrp);
   1161 	}
   1162 
   1163 	/*
   1164 	 * Check this process' children to see whether they qualify
   1165 	 * their process groups; if so, adjust counts for children's
   1166 	 * process groups.
   1167 	 */
   1168 	LIST_FOREACH(child, &p->p_children, p_sibling) {
   1169 		hispgrp = child->p_pgrp;
   1170 		if (hispgrp != pgrp && hispgrp->pg_session == mysession &&
   1171 		    !P_ZOMBIE(child)) {
   1172 			if (entering) {
   1173 				child->p_lflag &= ~PL_ORPHANPG;
   1174 				hispgrp->pg_jobc++;
   1175 			} else if (--hispgrp->pg_jobc == 0)
   1176 				orphanpg(hispgrp);
   1177 		}
   1178 	}
   1179 }
   1180 
   1181 /*
   1182  * A process group has become orphaned;
   1183  * if there are any stopped processes in the group,
   1184  * hang-up all process in that group.
   1185  *
   1186  * Call with proc_lock held.
   1187  */
   1188 static void
   1189 orphanpg(struct pgrp *pg)
   1190 {
   1191 	struct proc *p;
   1192 
   1193 	KASSERT(mutex_owned(proc_lock));
   1194 
   1195 	LIST_FOREACH(p, &pg->pg_members, p_pglist) {
   1196 		if (p->p_stat == SSTOP) {
   1197 			p->p_lflag |= PL_ORPHANPG;
   1198 			psignal(p, SIGHUP);
   1199 			psignal(p, SIGCONT);
   1200 		}
   1201 	}
   1202 }
   1203 
   1204 #ifdef DDB
   1205 #include <ddb/db_output.h>
   1206 void pidtbl_dump(void);
   1207 void
   1208 pidtbl_dump(void)
   1209 {
   1210 	struct pid_table *pt;
   1211 	struct proc *p;
   1212 	struct pgrp *pgrp;
   1213 	int id;
   1214 
   1215 	db_printf("pid table %p size %x, next %x, last %x\n",
   1216 		pid_table, pid_tbl_mask+1,
   1217 		next_free_pt, last_free_pt);
   1218 	for (pt = pid_table, id = 0; id <= pid_tbl_mask; id++, pt++) {
   1219 		p = pt->pt_proc;
   1220 		if (!P_VALID(p) && !pt->pt_pgrp)
   1221 			continue;
   1222 		db_printf("  id %x: ", id);
   1223 		if (P_VALID(p))
   1224 			db_printf("slotpid %d proc %p id %d (0x%x) %s\n",
   1225 				pt->pt_pid, p, p->p_pid, p->p_pid, p->p_comm);
   1226 		else
   1227 			db_printf("next %x use %x\n",
   1228 				P_NEXT(p) & pid_tbl_mask,
   1229 				P_NEXT(p) & ~pid_tbl_mask);
   1230 		if ((pgrp = pt->pt_pgrp)) {
   1231 			db_printf("\tsession %p, sid %d, count %d, login %s\n",
   1232 			    pgrp->pg_session, pgrp->pg_session->s_sid,
   1233 			    pgrp->pg_session->s_count,
   1234 			    pgrp->pg_session->s_login);
   1235 			db_printf("\tpgrp %p, pg_id %d, pg_jobc %d, members %p\n",
   1236 			    pgrp, pgrp->pg_id, pgrp->pg_jobc,
   1237 			    LIST_FIRST(&pgrp->pg_members));
   1238 			LIST_FOREACH(p, &pgrp->pg_members, p_pglist) {
   1239 				db_printf("\t\tpid %d addr %p pgrp %p %s\n",
   1240 				    p->p_pid, p, p->p_pgrp, p->p_comm);
   1241 			}
   1242 		}
   1243 	}
   1244 }
   1245 #endif /* DDB */
   1246 
   1247 #ifdef KSTACK_CHECK_MAGIC
   1248 
   1249 #define	KSTACK_MAGIC	0xdeadbeaf
   1250 
   1251 /* XXX should be per process basis? */
   1252 static int	kstackleftmin = KSTACK_SIZE;
   1253 static int	kstackleftthres = KSTACK_SIZE / 8;
   1254 
   1255 void
   1256 kstack_setup_magic(const struct lwp *l)
   1257 {
   1258 	uint32_t *ip;
   1259 	uint32_t const *end;
   1260 
   1261 	KASSERT(l != NULL);
   1262 	KASSERT(l != &lwp0);
   1263 
   1264 	/*
   1265 	 * fill all the stack with magic number
   1266 	 * so that later modification on it can be detected.
   1267 	 */
   1268 	ip = (uint32_t *)KSTACK_LOWEST_ADDR(l);
   1269 	end = (uint32_t *)((char *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE);
   1270 	for (; ip < end; ip++) {
   1271 		*ip = KSTACK_MAGIC;
   1272 	}
   1273 }
   1274 
   1275 void
   1276 kstack_check_magic(const struct lwp *l)
   1277 {
   1278 	uint32_t const *ip, *end;
   1279 	int stackleft;
   1280 
   1281 	KASSERT(l != NULL);
   1282 
   1283 	/* don't check proc0 */ /*XXX*/
   1284 	if (l == &lwp0)
   1285 		return;
   1286 
   1287 #ifdef __MACHINE_STACK_GROWS_UP
   1288 	/* stack grows upwards (eg. hppa) */
   1289 	ip = (uint32_t *)((void *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE);
   1290 	end = (uint32_t *)KSTACK_LOWEST_ADDR(l);
   1291 	for (ip--; ip >= end; ip--)
   1292 		if (*ip != KSTACK_MAGIC)
   1293 			break;
   1294 
   1295 	stackleft = (void *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE - (void *)ip;
   1296 #else /* __MACHINE_STACK_GROWS_UP */
   1297 	/* stack grows downwards (eg. i386) */
   1298 	ip = (uint32_t *)KSTACK_LOWEST_ADDR(l);
   1299 	end = (uint32_t *)((char *)KSTACK_LOWEST_ADDR(l) + KSTACK_SIZE);
   1300 	for (; ip < end; ip++)
   1301 		if (*ip != KSTACK_MAGIC)
   1302 			break;
   1303 
   1304 	stackleft = ((const char *)ip) - (const char *)KSTACK_LOWEST_ADDR(l);
   1305 #endif /* __MACHINE_STACK_GROWS_UP */
   1306 
   1307 	if (kstackleftmin > stackleft) {
   1308 		kstackleftmin = stackleft;
   1309 		if (stackleft < kstackleftthres)
   1310 			printf("warning: kernel stack left %d bytes"
   1311 			    "(pid %u:lid %u)\n", stackleft,
   1312 			    (u_int)l->l_proc->p_pid, (u_int)l->l_lid);
   1313 	}
   1314 
   1315 	if (stackleft <= 0) {
   1316 		panic("magic on the top of kernel stack changed for "
   1317 		    "pid %u, lid %u: maybe kernel stack overflow",
   1318 		    (u_int)l->l_proc->p_pid, (u_int)l->l_lid);
   1319 	}
   1320 }
   1321 #endif /* KSTACK_CHECK_MAGIC */
   1322 
   1323 int
   1324 proclist_foreach_call(struct proclist *list,
   1325     int (*callback)(struct proc *, void *arg), void *arg)
   1326 {
   1327 	struct proc marker;
   1328 	struct proc *p;
   1329 	int ret = 0;
   1330 
   1331 	marker.p_flag = PK_MARKER;
   1332 	mutex_enter(proc_lock);
   1333 	for (p = LIST_FIRST(list); ret == 0 && p != NULL;) {
   1334 		if (p->p_flag & PK_MARKER) {
   1335 			p = LIST_NEXT(p, p_list);
   1336 			continue;
   1337 		}
   1338 		LIST_INSERT_AFTER(p, &marker, p_list);
   1339 		ret = (*callback)(p, arg);
   1340 		KASSERT(mutex_owned(proc_lock));
   1341 		p = LIST_NEXT(&marker, p_list);
   1342 		LIST_REMOVE(&marker, p_list);
   1343 	}
   1344 	mutex_exit(proc_lock);
   1345 
   1346 	return ret;
   1347 }
   1348 
   1349 int
   1350 proc_vmspace_getref(struct proc *p, struct vmspace **vm)
   1351 {
   1352 
   1353 	/* XXXCDC: how should locking work here? */
   1354 
   1355 	/* curproc exception is for coredump. */
   1356 
   1357 	if ((p != curproc && (p->p_sflag & PS_WEXIT) != 0) ||
   1358 	    (p->p_vmspace->vm_refcnt < 1)) { /* XXX */
   1359 		return EFAULT;
   1360 	}
   1361 
   1362 	uvmspace_addref(p->p_vmspace);
   1363 	*vm = p->p_vmspace;
   1364 
   1365 	return 0;
   1366 }
   1367 
   1368 /*
   1369  * Acquire a write lock on the process credential.
   1370  */
   1371 void
   1372 proc_crmod_enter(void)
   1373 {
   1374 	struct lwp *l = curlwp;
   1375 	struct proc *p = l->l_proc;
   1376 	kauth_cred_t oc;
   1377 
   1378 	/* Reset what needs to be reset in plimit. */
   1379 	if (p->p_limit->pl_corename != defcorename) {
   1380 		lim_setcorename(p, defcorename, 0);
   1381 	}
   1382 
   1383 	mutex_enter(p->p_lock);
   1384 
   1385 	/* Ensure the LWP cached credentials are up to date. */
   1386 	if ((oc = l->l_cred) != p->p_cred) {
   1387 		kauth_cred_hold(p->p_cred);
   1388 		l->l_cred = p->p_cred;
   1389 		kauth_cred_free(oc);
   1390 	}
   1391 }
   1392 
   1393 /*
   1394  * Set in a new process credential, and drop the write lock.  The credential
   1395  * must have a reference already.  Optionally, free a no-longer required
   1396  * credential.  The scheduler also needs to inspect p_cred, so we also
   1397  * briefly acquire the sched state mutex.
   1398  */
   1399 void
   1400 proc_crmod_leave(kauth_cred_t scred, kauth_cred_t fcred, bool sugid)
   1401 {
   1402 	struct lwp *l = curlwp, *l2;
   1403 	struct proc *p = l->l_proc;
   1404 	kauth_cred_t oc;
   1405 
   1406 	KASSERT(mutex_owned(p->p_lock));
   1407 
   1408 	/* Is there a new credential to set in? */
   1409 	if (scred != NULL) {
   1410 		p->p_cred = scred;
   1411 		LIST_FOREACH(l2, &p->p_lwps, l_sibling) {
   1412 			if (l2 != l)
   1413 				l2->l_prflag |= LPR_CRMOD;
   1414 		}
   1415 
   1416 		/* Ensure the LWP cached credentials are up to date. */
   1417 		if ((oc = l->l_cred) != scred) {
   1418 			kauth_cred_hold(scred);
   1419 			l->l_cred = scred;
   1420 		}
   1421 	} else
   1422 		oc = NULL;	/* XXXgcc */
   1423 
   1424 	if (sugid) {
   1425 		/*
   1426 		 * Mark process as having changed credentials, stops
   1427 		 * tracing etc.
   1428 		 */
   1429 		p->p_flag |= PK_SUGID;
   1430 	}
   1431 
   1432 	mutex_exit(p->p_lock);
   1433 
   1434 	/* If there is a credential to be released, free it now. */
   1435 	if (fcred != NULL) {
   1436 		KASSERT(scred != NULL);
   1437 		kauth_cred_free(fcred);
   1438 		if (oc != scred)
   1439 			kauth_cred_free(oc);
   1440 	}
   1441 }
   1442 
   1443 /*
   1444  * proc_specific_key_create --
   1445  *	Create a key for subsystem proc-specific data.
   1446  */
   1447 int
   1448 proc_specific_key_create(specificdata_key_t *keyp, specificdata_dtor_t dtor)
   1449 {
   1450 
   1451 	return (specificdata_key_create(proc_specificdata_domain, keyp, dtor));
   1452 }
   1453 
   1454 /*
   1455  * proc_specific_key_delete --
   1456  *	Delete a key for subsystem proc-specific data.
   1457  */
   1458 void
   1459 proc_specific_key_delete(specificdata_key_t key)
   1460 {
   1461 
   1462 	specificdata_key_delete(proc_specificdata_domain, key);
   1463 }
   1464 
   1465 /*
   1466  * proc_initspecific --
   1467  *	Initialize a proc's specificdata container.
   1468  */
   1469 void
   1470 proc_initspecific(struct proc *p)
   1471 {
   1472 	int error __diagused;
   1473 
   1474 	error = specificdata_init(proc_specificdata_domain, &p->p_specdataref);
   1475 	KASSERT(error == 0);
   1476 }
   1477 
   1478 /*
   1479  * proc_finispecific --
   1480  *	Finalize a proc's specificdata container.
   1481  */
   1482 void
   1483 proc_finispecific(struct proc *p)
   1484 {
   1485 
   1486 	specificdata_fini(proc_specificdata_domain, &p->p_specdataref);
   1487 }
   1488 
   1489 /*
   1490  * proc_getspecific --
   1491  *	Return proc-specific data corresponding to the specified key.
   1492  */
   1493 void *
   1494 proc_getspecific(struct proc *p, specificdata_key_t key)
   1495 {
   1496 
   1497 	return (specificdata_getspecific(proc_specificdata_domain,
   1498 					 &p->p_specdataref, key));
   1499 }
   1500 
   1501 /*
   1502  * proc_setspecific --
   1503  *	Set proc-specific data corresponding to the specified key.
   1504  */
   1505 void
   1506 proc_setspecific(struct proc *p, specificdata_key_t key, void *data)
   1507 {
   1508 
   1509 	specificdata_setspecific(proc_specificdata_domain,
   1510 				 &p->p_specdataref, key, data);
   1511 }
   1512 
   1513 int
   1514 proc_uidmatch(kauth_cred_t cred, kauth_cred_t target)
   1515 {
   1516 	int r = 0;
   1517 
   1518 	if (kauth_cred_getuid(cred) != kauth_cred_getuid(target) ||
   1519 	    kauth_cred_getuid(cred) != kauth_cred_getsvuid(target)) {
   1520 		/*
   1521 		 * suid proc of ours or proc not ours
   1522 		 */
   1523 		r = EPERM;
   1524 	} else if (kauth_cred_getgid(target) != kauth_cred_getsvgid(target)) {
   1525 		/*
   1526 		 * sgid proc has sgid back to us temporarily
   1527 		 */
   1528 		r = EPERM;
   1529 	} else {
   1530 		/*
   1531 		 * our rgid must be in target's group list (ie,
   1532 		 * sub-processes started by a sgid process)
   1533 		 */
   1534 		int ismember = 0;
   1535 
   1536 		if (kauth_cred_ismember_gid(cred,
   1537 		    kauth_cred_getgid(target), &ismember) != 0 ||
   1538 		    !ismember)
   1539 			r = EPERM;
   1540 	}
   1541 
   1542 	return (r);
   1543 }
   1544 
   1545 /*
   1546  * sysctl stuff
   1547  */
   1548 
   1549 #define KERN_PROCSLOP	(5 * sizeof(struct kinfo_proc))
   1550 
   1551 static const u_int sysctl_flagmap[] = {
   1552 	PK_ADVLOCK, P_ADVLOCK,
   1553 	PK_EXEC, P_EXEC,
   1554 	PK_NOCLDWAIT, P_NOCLDWAIT,
   1555 	PK_32, P_32,
   1556 	PK_CLDSIGIGN, P_CLDSIGIGN,
   1557 	PK_SUGID, P_SUGID,
   1558 	0
   1559 };
   1560 
   1561 static const u_int sysctl_sflagmap[] = {
   1562 	PS_NOCLDSTOP, P_NOCLDSTOP,
   1563 	PS_WEXIT, P_WEXIT,
   1564 	PS_STOPFORK, P_STOPFORK,
   1565 	PS_STOPEXEC, P_STOPEXEC,
   1566 	PS_STOPEXIT, P_STOPEXIT,
   1567 	0
   1568 };
   1569 
   1570 static const u_int sysctl_slflagmap[] = {
   1571 	PSL_TRACED, P_TRACED,
   1572 	PSL_CHTRACED, P_CHTRACED,
   1573 	PSL_SYSCALL, P_SYSCALL,
   1574 	0
   1575 };
   1576 
   1577 static const u_int sysctl_lflagmap[] = {
   1578 	PL_CONTROLT, P_CONTROLT,
   1579 	PL_PPWAIT, P_PPWAIT,
   1580 	0
   1581 };
   1582 
   1583 static const u_int sysctl_stflagmap[] = {
   1584 	PST_PROFIL, P_PROFIL,
   1585 	0
   1586 
   1587 };
   1588 
   1589 /* used by kern_lwp also */
   1590 const u_int sysctl_lwpflagmap[] = {
   1591 	LW_SINTR, L_SINTR,
   1592 	LW_SYSTEM, L_SYSTEM,
   1593 	0
   1594 };
   1595 
   1596 /*
   1597  * Find the most ``active'' lwp of a process and return it for ps display
   1598  * purposes
   1599  */
   1600 static struct lwp *
   1601 proc_active_lwp(struct proc *p)
   1602 {
   1603 	static const int ostat[] = {
   1604 		0,
   1605 		2,	/* LSIDL */
   1606 		6,	/* LSRUN */
   1607 		5,	/* LSSLEEP */
   1608 		4,	/* LSSTOP */
   1609 		0,	/* LSZOMB */
   1610 		1,	/* LSDEAD */
   1611 		7,	/* LSONPROC */
   1612 		3	/* LSSUSPENDED */
   1613 	};
   1614 
   1615 	struct lwp *l, *lp = NULL;
   1616 	LIST_FOREACH(l, &p->p_lwps, l_sibling) {
   1617 		KASSERT(l->l_stat >= 0 && l->l_stat < __arraycount(ostat));
   1618 		if (lp == NULL ||
   1619 		    ostat[l->l_stat] > ostat[lp->l_stat] ||
   1620 		    (ostat[l->l_stat] == ostat[lp->l_stat] &&
   1621 		    l->l_cpticks > lp->l_cpticks)) {
   1622 			lp = l;
   1623 			continue;
   1624 		}
   1625 	}
   1626 	return lp;
   1627 }
   1628 
   1629 static int
   1630 sysctl_doeproc(SYSCTLFN_ARGS)
   1631 {
   1632 	union {
   1633 		struct kinfo_proc kproc;
   1634 		struct kinfo_proc2 kproc2;
   1635 	} *kbuf;
   1636 	struct proc *p, *next, *marker;
   1637 	char *where, *dp;
   1638 	int type, op, arg, error;
   1639 	u_int elem_size, kelem_size, elem_count;
   1640 	size_t buflen, needed;
   1641 	bool match, zombie, mmmbrains;
   1642 
   1643 	if (namelen == 1 && name[0] == CTL_QUERY)
   1644 		return (sysctl_query(SYSCTLFN_CALL(rnode)));
   1645 
   1646 	dp = where = oldp;
   1647 	buflen = where != NULL ? *oldlenp : 0;
   1648 	error = 0;
   1649 	needed = 0;
   1650 	type = rnode->sysctl_num;
   1651 
   1652 	if (type == KERN_PROC) {
   1653 		if (namelen == 0)
   1654 			return EINVAL;
   1655 		switch (op = name[0]) {
   1656 		case KERN_PROC_ALL:
   1657 			if (namelen != 1)
   1658 				return EINVAL;
   1659 			arg = 0;
   1660 			break;
   1661 		default:
   1662 			if (namelen != 2)
   1663 				return EINVAL;
   1664 			arg = name[1];
   1665 			break;
   1666 		}
   1667 		elem_count = 0;	/* Hush little compiler, don't you cry */
   1668 		kelem_size = elem_size = sizeof(kbuf->kproc);
   1669 	} else {
   1670 		if (namelen != 4)
   1671 			return EINVAL;
   1672 		op = name[0];
   1673 		arg = name[1];
   1674 		elem_size = name[2];
   1675 		elem_count = name[3];
   1676 		kelem_size = sizeof(kbuf->kproc2);
   1677 	}
   1678 
   1679 	sysctl_unlock();
   1680 
   1681 	kbuf = kmem_zalloc(sizeof(*kbuf), KM_SLEEP);
   1682 	marker = kmem_alloc(sizeof(*marker), KM_SLEEP);
   1683 	marker->p_flag = PK_MARKER;
   1684 
   1685 	mutex_enter(proc_lock);
   1686 	/*
   1687 	 * Start with zombies to prevent reporting processes twice, in case they
   1688 	 * are dying and being moved from the list of alive processes to zombies.
   1689 	 */
   1690 	mmmbrains = true;
   1691 	for (p = LIST_FIRST(&zombproc);; p = next) {
   1692 		if (p == NULL) {
   1693 			if (mmmbrains) {
   1694 				p = LIST_FIRST(&allproc);
   1695 				mmmbrains = false;
   1696 			}
   1697 			if (p == NULL)
   1698 				break;
   1699 		}
   1700 		next = LIST_NEXT(p, p_list);
   1701 		if ((p->p_flag & PK_MARKER) != 0)
   1702 			continue;
   1703 
   1704 		/*
   1705 		 * Skip embryonic processes.
   1706 		 */
   1707 		if (p->p_stat == SIDL)
   1708 			continue;
   1709 
   1710 		mutex_enter(p->p_lock);
   1711 		error = kauth_authorize_process(l->l_cred,
   1712 		    KAUTH_PROCESS_CANSEE, p,
   1713 		    KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_EPROC), NULL, NULL);
   1714 		if (error != 0) {
   1715 			mutex_exit(p->p_lock);
   1716 			continue;
   1717 		}
   1718 
   1719 		/*
   1720 		 * Hande all the operations in one switch on the cost of
   1721 		 * algorithm complexity is on purpose. The win splitting this
   1722 		 * function into several similar copies makes maintenance burden
   1723 		 * burden, code grow and boost is neglible in practical systems.
   1724 		 */
   1725 		switch (op) {
   1726 		case KERN_PROC_PID:
   1727 			match = (p->p_pid == (pid_t)arg);
   1728 			break;
   1729 
   1730 		case KERN_PROC_PGRP:
   1731 			match = (p->p_pgrp->pg_id == (pid_t)arg);
   1732 			break;
   1733 
   1734 		case KERN_PROC_SESSION:
   1735 			match = (p->p_session->s_sid == (pid_t)arg);
   1736 			break;
   1737 
   1738 		case KERN_PROC_TTY:
   1739 			match = true;
   1740 			if (arg == (int) KERN_PROC_TTY_REVOKE) {
   1741 				if ((p->p_lflag & PL_CONTROLT) == 0 ||
   1742 				    p->p_session->s_ttyp == NULL ||
   1743 				    p->p_session->s_ttyvp != NULL) {
   1744 				    	match = false;
   1745 				}
   1746 			} else if ((p->p_lflag & PL_CONTROLT) == 0 ||
   1747 			    p->p_session->s_ttyp == NULL) {
   1748 				if ((dev_t)arg != KERN_PROC_TTY_NODEV) {
   1749 					match = false;
   1750 				}
   1751 			} else if (p->p_session->s_ttyp->t_dev != (dev_t)arg) {
   1752 				match = false;
   1753 			}
   1754 			break;
   1755 
   1756 		case KERN_PROC_UID:
   1757 			match = (kauth_cred_geteuid(p->p_cred) == (uid_t)arg);
   1758 			break;
   1759 
   1760 		case KERN_PROC_RUID:
   1761 			match = (kauth_cred_getuid(p->p_cred) == (uid_t)arg);
   1762 			break;
   1763 
   1764 		case KERN_PROC_GID:
   1765 			match = (kauth_cred_getegid(p->p_cred) == (uid_t)arg);
   1766 			break;
   1767 
   1768 		case KERN_PROC_RGID:
   1769 			match = (kauth_cred_getgid(p->p_cred) == (uid_t)arg);
   1770 			break;
   1771 
   1772 		case KERN_PROC_ALL:
   1773 			match = true;
   1774 			/* allow everything */
   1775 			break;
   1776 
   1777 		default:
   1778 			error = EINVAL;
   1779 			mutex_exit(p->p_lock);
   1780 			goto cleanup;
   1781 		}
   1782 		if (!match) {
   1783 			mutex_exit(p->p_lock);
   1784 			continue;
   1785 		}
   1786 
   1787 		/*
   1788 		 * Grab a hold on the process.
   1789 		 */
   1790 		if (mmmbrains) {
   1791 			zombie = true;
   1792 		} else {
   1793 			zombie = !rw_tryenter(&p->p_reflock, RW_READER);
   1794 		}
   1795 		if (zombie) {
   1796 			LIST_INSERT_AFTER(p, marker, p_list);
   1797 		}
   1798 
   1799 		if (buflen >= elem_size &&
   1800 		    (type == KERN_PROC || elem_count > 0)) {
   1801 			if (type == KERN_PROC) {
   1802 				fill_proc(p, &kbuf->kproc.kp_proc);
   1803 				fill_eproc(p, &kbuf->kproc.kp_eproc, zombie);
   1804 			} else {
   1805 				fill_kproc2(p, &kbuf->kproc2, zombie);
   1806 				elem_count--;
   1807 			}
   1808 			mutex_exit(p->p_lock);
   1809 			mutex_exit(proc_lock);
   1810 			/*
   1811 			 * Copy out elem_size, but not larger than kelem_size
   1812 			 */
   1813 			error = sysctl_copyout(l, kbuf, dp,
   1814 			    uimin(kelem_size, elem_size));
   1815 			mutex_enter(proc_lock);
   1816 			if (error) {
   1817 				goto bah;
   1818 			}
   1819 			dp += elem_size;
   1820 			buflen -= elem_size;
   1821 		} else {
   1822 			mutex_exit(p->p_lock);
   1823 		}
   1824 		needed += elem_size;
   1825 
   1826 		/*
   1827 		 * Release reference to process.
   1828 		 */
   1829 	 	if (zombie) {
   1830 			next = LIST_NEXT(marker, p_list);
   1831  			LIST_REMOVE(marker, p_list);
   1832 		} else {
   1833 			rw_exit(&p->p_reflock);
   1834 			next = LIST_NEXT(p, p_list);
   1835 		}
   1836 
   1837 		/*
   1838 		 * Short-circuit break quickly!
   1839 		 */
   1840 		if (op == KERN_PROC_PID)
   1841                 	break;
   1842 	}
   1843 	mutex_exit(proc_lock);
   1844 
   1845 	if (where != NULL) {
   1846 		*oldlenp = dp - where;
   1847 		if (needed > *oldlenp) {
   1848 			error = ENOMEM;
   1849 			goto out;
   1850 		}
   1851 	} else {
   1852 		needed += KERN_PROCSLOP;
   1853 		*oldlenp = needed;
   1854 	}
   1855 	kmem_free(kbuf, sizeof(*kbuf));
   1856 	kmem_free(marker, sizeof(*marker));
   1857 	sysctl_relock();
   1858 	return 0;
   1859  bah:
   1860  	if (zombie)
   1861  		LIST_REMOVE(marker, p_list);
   1862 	else
   1863 		rw_exit(&p->p_reflock);
   1864  cleanup:
   1865 	mutex_exit(proc_lock);
   1866  out:
   1867 	kmem_free(kbuf, sizeof(*kbuf));
   1868 	kmem_free(marker, sizeof(*marker));
   1869 	sysctl_relock();
   1870 	return error;
   1871 }
   1872 
   1873 int
   1874 copyin_psstrings(struct proc *p, struct ps_strings *arginfo)
   1875 {
   1876 
   1877 #ifdef COMPAT_NETBSD32
   1878 	if (p->p_flag & PK_32) {
   1879 		struct ps_strings32 arginfo32;
   1880 
   1881 		int error = copyin_proc(p, (void *)p->p_psstrp, &arginfo32,
   1882 		    sizeof(arginfo32));
   1883 		if (error)
   1884 			return error;
   1885 		arginfo->ps_argvstr = (void *)(uintptr_t)arginfo32.ps_argvstr;
   1886 		arginfo->ps_nargvstr = arginfo32.ps_nargvstr;
   1887 		arginfo->ps_envstr = (void *)(uintptr_t)arginfo32.ps_envstr;
   1888 		arginfo->ps_nenvstr = arginfo32.ps_nenvstr;
   1889 		return 0;
   1890 	}
   1891 #endif
   1892 	return copyin_proc(p, (void *)p->p_psstrp, arginfo, sizeof(*arginfo));
   1893 }
   1894 
   1895 static int
   1896 copy_procargs_sysctl_cb(void *cookie_, const void *src, size_t off, size_t len)
   1897 {
   1898 	void **cookie = cookie_;
   1899 	struct lwp *l = cookie[0];
   1900 	char *dst = cookie[1];
   1901 
   1902 	return sysctl_copyout(l, src, dst + off, len);
   1903 }
   1904 
   1905 /*
   1906  * sysctl helper routine for kern.proc_args pseudo-subtree.
   1907  */
   1908 static int
   1909 sysctl_kern_proc_args(SYSCTLFN_ARGS)
   1910 {
   1911 	struct ps_strings pss;
   1912 	struct proc *p;
   1913 	pid_t pid;
   1914 	int type, error;
   1915 	void *cookie[2];
   1916 
   1917 	if (namelen == 1 && name[0] == CTL_QUERY)
   1918 		return (sysctl_query(SYSCTLFN_CALL(rnode)));
   1919 
   1920 	if (newp != NULL || namelen != 2)
   1921 		return (EINVAL);
   1922 	pid = name[0];
   1923 	type = name[1];
   1924 
   1925 	switch (type) {
   1926 	case KERN_PROC_PATHNAME:
   1927 		sysctl_unlock();
   1928 		error = fill_pathname(l, pid, oldp, oldlenp);
   1929 		sysctl_relock();
   1930 		return error;
   1931 
   1932 	case KERN_PROC_ARGV:
   1933 	case KERN_PROC_NARGV:
   1934 	case KERN_PROC_ENV:
   1935 	case KERN_PROC_NENV:
   1936 		/* ok */
   1937 		break;
   1938 	default:
   1939 		return (EINVAL);
   1940 	}
   1941 
   1942 	sysctl_unlock();
   1943 
   1944 	/* check pid */
   1945 	mutex_enter(proc_lock);
   1946 	if ((p = proc_find(pid)) == NULL) {
   1947 		error = EINVAL;
   1948 		goto out_locked;
   1949 	}
   1950 	mutex_enter(p->p_lock);
   1951 
   1952 	/* Check permission. */
   1953 	if (type == KERN_PROC_ARGV || type == KERN_PROC_NARGV)
   1954 		error = kauth_authorize_process(l->l_cred, KAUTH_PROCESS_CANSEE,
   1955 		    p, KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_ARGS), NULL, NULL);
   1956 	else if (type == KERN_PROC_ENV || type == KERN_PROC_NENV)
   1957 		error = kauth_authorize_process(l->l_cred, KAUTH_PROCESS_CANSEE,
   1958 		    p, KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_ENV), NULL, NULL);
   1959 	else
   1960 		error = EINVAL; /* XXXGCC */
   1961 	if (error) {
   1962 		mutex_exit(p->p_lock);
   1963 		goto out_locked;
   1964 	}
   1965 
   1966 	if (oldp == NULL) {
   1967 		if (type == KERN_PROC_NARGV || type == KERN_PROC_NENV)
   1968 			*oldlenp = sizeof (int);
   1969 		else
   1970 			*oldlenp = ARG_MAX;	/* XXX XXX XXX */
   1971 		error = 0;
   1972 		mutex_exit(p->p_lock);
   1973 		goto out_locked;
   1974 	}
   1975 
   1976 	/*
   1977 	 * Zombies don't have a stack, so we can't read their psstrings.
   1978 	 * System processes also don't have a user stack.
   1979 	 */
   1980 	if (P_ZOMBIE(p) || (p->p_flag & PK_SYSTEM) != 0) {
   1981 		error = EINVAL;
   1982 		mutex_exit(p->p_lock);
   1983 		goto out_locked;
   1984 	}
   1985 
   1986 	error = rw_tryenter(&p->p_reflock, RW_READER) ? 0 : EBUSY;
   1987 	mutex_exit(p->p_lock);
   1988 	if (error) {
   1989 		goto out_locked;
   1990 	}
   1991 	mutex_exit(proc_lock);
   1992 
   1993 	if (type == KERN_PROC_NARGV || type == KERN_PROC_NENV) {
   1994 		int value;
   1995 		if ((error = copyin_psstrings(p, &pss)) == 0) {
   1996 			if (type == KERN_PROC_NARGV)
   1997 				value = pss.ps_nargvstr;
   1998 			else
   1999 				value = pss.ps_nenvstr;
   2000 			error = sysctl_copyout(l, &value, oldp, sizeof(value));
   2001 			*oldlenp = sizeof(value);
   2002 		}
   2003 	} else {
   2004 		cookie[0] = l;
   2005 		cookie[1] = oldp;
   2006 		error = copy_procargs(p, type, oldlenp,
   2007 		    copy_procargs_sysctl_cb, cookie);
   2008 	}
   2009 	rw_exit(&p->p_reflock);
   2010 	sysctl_relock();
   2011 	return error;
   2012 
   2013 out_locked:
   2014 	mutex_exit(proc_lock);
   2015 	sysctl_relock();
   2016 	return error;
   2017 }
   2018 
   2019 int
   2020 copy_procargs(struct proc *p, int oid, size_t *limit,
   2021     int (*cb)(void *, const void *, size_t, size_t), void *cookie)
   2022 {
   2023 	struct ps_strings pss;
   2024 	size_t len, i, loaded, entry_len;
   2025 	struct uio auio;
   2026 	struct iovec aiov;
   2027 	int error, argvlen;
   2028 	char *arg;
   2029 	char **argv;
   2030 	vaddr_t user_argv;
   2031 	struct vmspace *vmspace;
   2032 
   2033 	/*
   2034 	 * Allocate a temporary buffer to hold the argument vector and
   2035 	 * the arguments themselve.
   2036 	 */
   2037 	arg = kmem_alloc(PAGE_SIZE, KM_SLEEP);
   2038 	argv = kmem_alloc(PAGE_SIZE, KM_SLEEP);
   2039 
   2040 	/*
   2041 	 * Lock the process down in memory.
   2042 	 */
   2043 	vmspace = p->p_vmspace;
   2044 	uvmspace_addref(vmspace);
   2045 
   2046 	/*
   2047 	 * Read in the ps_strings structure.
   2048 	 */
   2049 	if ((error = copyin_psstrings(p, &pss)) != 0)
   2050 		goto done;
   2051 
   2052 	/*
   2053 	 * Now read the address of the argument vector.
   2054 	 */
   2055 	switch (oid) {
   2056 	case KERN_PROC_ARGV:
   2057 		user_argv = (uintptr_t)pss.ps_argvstr;
   2058 		argvlen = pss.ps_nargvstr;
   2059 		break;
   2060 	case KERN_PROC_ENV:
   2061 		user_argv = (uintptr_t)pss.ps_envstr;
   2062 		argvlen = pss.ps_nenvstr;
   2063 		break;
   2064 	default:
   2065 		error = EINVAL;
   2066 		goto done;
   2067 	}
   2068 
   2069 	if (argvlen < 0) {
   2070 		error = EIO;
   2071 		goto done;
   2072 	}
   2073 
   2074 
   2075 	/*
   2076 	 * Now copy each string.
   2077 	 */
   2078 	len = 0; /* bytes written to user buffer */
   2079 	loaded = 0; /* bytes from argv already processed */
   2080 	i = 0; /* To make compiler happy */
   2081 	entry_len = PROC_PTRSZ(p);
   2082 
   2083 	for (; argvlen; --argvlen) {
   2084 		int finished = 0;
   2085 		vaddr_t base;
   2086 		size_t xlen;
   2087 		int j;
   2088 
   2089 		if (loaded == 0) {
   2090 			size_t rem = entry_len * argvlen;
   2091 			loaded = MIN(rem, PAGE_SIZE);
   2092 			error = copyin_vmspace(vmspace,
   2093 			    (const void *)user_argv, argv, loaded);
   2094 			if (error)
   2095 				break;
   2096 			user_argv += loaded;
   2097 			i = 0;
   2098 		}
   2099 
   2100 #ifdef COMPAT_NETBSD32
   2101 		if (p->p_flag & PK_32) {
   2102 			netbsd32_charp *argv32;
   2103 
   2104 			argv32 = (netbsd32_charp *)argv;
   2105 			base = (vaddr_t)NETBSD32PTR64(argv32[i++]);
   2106 		} else
   2107 #endif
   2108 			base = (vaddr_t)argv[i++];
   2109 		loaded -= entry_len;
   2110 
   2111 		/*
   2112 		 * The program has messed around with its arguments,
   2113 		 * possibly deleting some, and replacing them with
   2114 		 * NULL's. Treat this as the last argument and not
   2115 		 * a failure.
   2116 		 */
   2117 		if (base == 0)
   2118 			break;
   2119 
   2120 		while (!finished) {
   2121 			xlen = PAGE_SIZE - (base & PAGE_MASK);
   2122 
   2123 			aiov.iov_base = arg;
   2124 			aiov.iov_len = PAGE_SIZE;
   2125 			auio.uio_iov = &aiov;
   2126 			auio.uio_iovcnt = 1;
   2127 			auio.uio_offset = base;
   2128 			auio.uio_resid = xlen;
   2129 			auio.uio_rw = UIO_READ;
   2130 			UIO_SETUP_SYSSPACE(&auio);
   2131 			error = uvm_io(&vmspace->vm_map, &auio, 0);
   2132 			if (error)
   2133 				goto done;
   2134 
   2135 			/* Look for the end of the string */
   2136 			for (j = 0; j < xlen; j++) {
   2137 				if (arg[j] == '\0') {
   2138 					xlen = j + 1;
   2139 					finished = 1;
   2140 					break;
   2141 				}
   2142 			}
   2143 
   2144 			/* Check for user buffer overflow */
   2145 			if (len + xlen > *limit) {
   2146 				finished = 1;
   2147 				if (len > *limit)
   2148 					xlen = 0;
   2149 				else
   2150 					xlen = *limit - len;
   2151 			}
   2152 
   2153 			/* Copyout the page */
   2154 			error = (*cb)(cookie, arg, len, xlen);
   2155 			if (error)
   2156 				goto done;
   2157 
   2158 			len += xlen;
   2159 			base += xlen;
   2160 		}
   2161 	}
   2162 	*limit = len;
   2163 
   2164 done:
   2165 	kmem_free(argv, PAGE_SIZE);
   2166 	kmem_free(arg, PAGE_SIZE);
   2167 	uvmspace_free(vmspace);
   2168 	return error;
   2169 }
   2170 
   2171 /*
   2172  * Fill in a proc structure for the specified process.
   2173  */
   2174 static void
   2175 fill_proc(const struct proc *psrc, struct proc *p)
   2176 {
   2177 	const bool allowaddr = get_expose_address(curproc);
   2178 
   2179 	COND_SET_VALUE(p->p_list, psrc->p_list, allowaddr);
   2180 	COND_SET_VALUE(p->p_auxlock, psrc->p_auxlock, allowaddr);
   2181 	COND_SET_VALUE(p->p_lock, psrc->p_lock, allowaddr);
   2182 	COND_SET_VALUE(p->p_stmutex, psrc->p_stmutex, allowaddr);
   2183 	COND_SET_VALUE(p->p_reflock, psrc->p_reflock, allowaddr);
   2184 	COND_SET_VALUE(p->p_waitcv, psrc->p_waitcv, allowaddr);
   2185 	COND_SET_VALUE(p->p_lwpcv, psrc->p_lwpcv, allowaddr);
   2186 	COND_SET_VALUE(p->p_cred, psrc->p_cred, allowaddr);
   2187 	COND_SET_VALUE(p->p_fd, psrc->p_fd, allowaddr);
   2188 	COND_SET_VALUE(p->p_cwdi, psrc->p_cwdi, allowaddr);
   2189 	COND_SET_VALUE(p->p_stats, psrc->p_stats, allowaddr);
   2190 	COND_SET_VALUE(p->p_limit, psrc->p_limit, allowaddr);
   2191 	COND_SET_VALUE(p->p_vmspace, psrc->p_vmspace, allowaddr);
   2192 	COND_SET_VALUE(p->p_sigacts, psrc->p_sigacts, allowaddr);
   2193 	COND_SET_VALUE(p->p_aio, psrc->p_aio, allowaddr);
   2194 	p->p_mqueue_cnt = psrc->p_mqueue_cnt;
   2195 	COND_SET_VALUE(p->p_specdataref, psrc->p_specdataref, allowaddr);
   2196 	p->p_exitsig = psrc->p_exitsig;
   2197 	p->p_flag = psrc->p_flag;
   2198 	p->p_sflag = psrc->p_sflag;
   2199 	p->p_slflag = psrc->p_slflag;
   2200 	p->p_lflag = psrc->p_lflag;
   2201 	p->p_stflag = psrc->p_stflag;
   2202 	p->p_stat = psrc->p_stat;
   2203 	p->p_trace_enabled = psrc->p_trace_enabled;
   2204 	p->p_pid = psrc->p_pid;
   2205 	COND_SET_VALUE(p->p_pglist, psrc->p_pglist, allowaddr);
   2206 	COND_SET_VALUE(p->p_pptr, psrc->p_pptr, allowaddr);
   2207 	COND_SET_VALUE(p->p_sibling, psrc->p_sibling, allowaddr);
   2208 	COND_SET_VALUE(p->p_children, psrc->p_children, allowaddr);
   2209 	COND_SET_VALUE(p->p_lwps, psrc->p_lwps, allowaddr);
   2210 	COND_SET_VALUE(p->p_raslist, psrc->p_raslist, allowaddr);
   2211 	p->p_nlwps = psrc->p_nlwps;
   2212 	p->p_nzlwps = psrc->p_nzlwps;
   2213 	p->p_nrlwps = psrc->p_nrlwps;
   2214 	p->p_nlwpwait = psrc->p_nlwpwait;
   2215 	p->p_ndlwps = psrc->p_ndlwps;
   2216 	p->p_nlwpid = psrc->p_nlwpid;
   2217 	p->p_nstopchild = psrc->p_nstopchild;
   2218 	p->p_waited = psrc->p_waited;
   2219 	COND_SET_VALUE(p->p_zomblwp, psrc->p_zomblwp, allowaddr);
   2220 	COND_SET_VALUE(p->p_vforklwp, psrc->p_vforklwp, allowaddr);
   2221 	COND_SET_VALUE(p->p_sched_info, psrc->p_sched_info, allowaddr);
   2222 	p->p_estcpu = psrc->p_estcpu;
   2223 	p->p_estcpu_inherited = psrc->p_estcpu_inherited;
   2224 	p->p_forktime = psrc->p_forktime;
   2225 	p->p_pctcpu = psrc->p_pctcpu;
   2226 	COND_SET_VALUE(p->p_opptr, psrc->p_opptr, allowaddr);
   2227 	COND_SET_VALUE(p->p_timers, psrc->p_timers, allowaddr);
   2228 	p->p_rtime = psrc->p_rtime;
   2229 	p->p_uticks = psrc->p_uticks;
   2230 	p->p_sticks = psrc->p_sticks;
   2231 	p->p_iticks = psrc->p_iticks;
   2232 	p->p_xutime = psrc->p_xutime;
   2233 	p->p_xstime = psrc->p_xstime;
   2234 	p->p_traceflag = psrc->p_traceflag;
   2235 	COND_SET_VALUE(p->p_tracep, psrc->p_tracep, allowaddr);
   2236 	COND_SET_VALUE(p->p_textvp, psrc->p_textvp, allowaddr);
   2237 	COND_SET_VALUE(p->p_emul, psrc->p_emul, allowaddr);
   2238 	COND_SET_VALUE(p->p_emuldata, psrc->p_emuldata, allowaddr);
   2239 	COND_SET_VALUE(p->p_execsw, psrc->p_execsw, allowaddr);
   2240 	COND_SET_VALUE(p->p_klist, psrc->p_klist, allowaddr);
   2241 	COND_SET_VALUE(p->p_sigwaiters, psrc->p_sigwaiters, allowaddr);
   2242 	COND_SET_VALUE(p->p_sigpend, psrc->p_sigpend, allowaddr);
   2243 	COND_SET_VALUE(p->p_lwpctl, psrc->p_lwpctl, allowaddr);
   2244 	p->p_ppid = psrc->p_ppid;
   2245 	p->p_fpid = psrc->p_fpid;
   2246 	p->p_vfpid = psrc->p_vfpid;
   2247 	p->p_vfpid_done = psrc->p_vfpid_done;
   2248 	p->p_lwp_created = psrc->p_lwp_created;
   2249 	p->p_lwp_exited = psrc->p_lwp_exited;
   2250 	p->p_nsems = psrc->p_nsems;
   2251 	COND_SET_VALUE(p->p_path, psrc->p_path, allowaddr);
   2252 	COND_SET_VALUE(p->p_sigctx, psrc->p_sigctx, allowaddr);
   2253 	p->p_nice = psrc->p_nice;
   2254 	memcpy(p->p_comm, psrc->p_comm, sizeof(p->p_comm));
   2255 	COND_SET_VALUE(p->p_pgrp, psrc->p_pgrp, allowaddr);
   2256 	COND_SET_VALUE(p->p_psstrp, psrc->p_psstrp, allowaddr);
   2257 	p->p_pax = psrc->p_pax;
   2258 	p->p_xexit = psrc->p_xexit;
   2259 	p->p_xsig = psrc->p_xsig;
   2260 	p->p_acflag = psrc->p_acflag;
   2261 	COND_SET_VALUE(p->p_md, psrc->p_md, allowaddr);
   2262 	p->p_stackbase = psrc->p_stackbase;
   2263 	COND_SET_VALUE(p->p_dtrace, psrc->p_dtrace, allowaddr);
   2264 }
   2265 
   2266 /*
   2267  * Fill in an eproc structure for the specified process.
   2268  */
   2269 void
   2270 fill_eproc(struct proc *p, struct eproc *ep, bool zombie)
   2271 {
   2272 	struct tty *tp;
   2273 	struct lwp *l;
   2274 
   2275 	KASSERT(mutex_owned(proc_lock));
   2276 	KASSERT(mutex_owned(p->p_lock));
   2277 
   2278 	const bool allowaddr = get_expose_address(curproc);
   2279 
   2280 	COND_SET_VALUE(ep->e_paddr, p, allowaddr);
   2281 	COND_SET_VALUE(ep->e_sess, p->p_session, allowaddr);
   2282 	if (p->p_cred) {
   2283 		kauth_cred_topcred(p->p_cred, &ep->e_pcred);
   2284 		kauth_cred_toucred(p->p_cred, &ep->e_ucred);
   2285 	}
   2286 	if (p->p_stat != SIDL && !P_ZOMBIE(p) && !zombie) {
   2287 		struct vmspace *vm = p->p_vmspace;
   2288 
   2289 		ep->e_vm.vm_rssize = vm_resident_count(vm);
   2290 		ep->e_vm.vm_tsize = vm->vm_tsize;
   2291 		ep->e_vm.vm_dsize = vm->vm_dsize;
   2292 		ep->e_vm.vm_ssize = vm->vm_ssize;
   2293 		ep->e_vm.vm_map.size = vm->vm_map.size;
   2294 
   2295 		/* Pick the primary (first) LWP */
   2296 		l = proc_active_lwp(p);
   2297 		KASSERT(l != NULL);
   2298 		lwp_lock(l);
   2299 		if (l->l_wchan)
   2300 			strncpy(ep->e_wmesg, l->l_wmesg, WMESGLEN);
   2301 		lwp_unlock(l);
   2302 	}
   2303 	ep->e_ppid = p->p_ppid;
   2304 	if (p->p_pgrp && p->p_session) {
   2305 		ep->e_pgid = p->p_pgrp->pg_id;
   2306 		ep->e_jobc = p->p_pgrp->pg_jobc;
   2307 		ep->e_sid = p->p_session->s_sid;
   2308 		if ((p->p_lflag & PL_CONTROLT) &&
   2309 		    (tp = p->p_session->s_ttyp)) {
   2310 			ep->e_tdev = tp->t_dev;
   2311 			ep->e_tpgid = tp->t_pgrp ? tp->t_pgrp->pg_id : NO_PGID;
   2312 			COND_SET_VALUE(ep->e_tsess, tp->t_session, allowaddr);
   2313 		} else
   2314 			ep->e_tdev = (uint32_t)NODEV;
   2315 		ep->e_flag = p->p_session->s_ttyvp ? EPROC_CTTY : 0;
   2316 		if (SESS_LEADER(p))
   2317 			ep->e_flag |= EPROC_SLEADER;
   2318 		strncpy(ep->e_login, p->p_session->s_login, MAXLOGNAME);
   2319 	}
   2320 	ep->e_xsize = ep->e_xrssize = 0;
   2321 	ep->e_xccount = ep->e_xswrss = 0;
   2322 }
   2323 
   2324 /*
   2325  * Fill in a kinfo_proc2 structure for the specified process.
   2326  */
   2327 void
   2328 fill_kproc2(struct proc *p, struct kinfo_proc2 *ki, bool zombie)
   2329 {
   2330 	struct tty *tp;
   2331 	struct lwp *l, *l2;
   2332 	struct timeval ut, st, rt;
   2333 	sigset_t ss1, ss2;
   2334 	struct rusage ru;
   2335 	struct vmspace *vm;
   2336 
   2337 	KASSERT(mutex_owned(proc_lock));
   2338 	KASSERT(mutex_owned(p->p_lock));
   2339 
   2340 	const bool allowaddr = get_expose_address(curproc);
   2341 
   2342 	sigemptyset(&ss1);
   2343 	sigemptyset(&ss2);
   2344 
   2345 	COND_SET_VALUE(ki->p_paddr, PTRTOUINT64(p), allowaddr);
   2346 	COND_SET_VALUE(ki->p_fd, PTRTOUINT64(p->p_fd), allowaddr);
   2347 	COND_SET_VALUE(ki->p_cwdi, PTRTOUINT64(p->p_cwdi), allowaddr);
   2348 	COND_SET_VALUE(ki->p_stats, PTRTOUINT64(p->p_stats), allowaddr);
   2349 	COND_SET_VALUE(ki->p_limit, PTRTOUINT64(p->p_limit), allowaddr);
   2350 	COND_SET_VALUE(ki->p_vmspace, PTRTOUINT64(p->p_vmspace), allowaddr);
   2351 	COND_SET_VALUE(ki->p_sigacts, PTRTOUINT64(p->p_sigacts), allowaddr);
   2352 	COND_SET_VALUE(ki->p_sess, PTRTOUINT64(p->p_session), allowaddr);
   2353 	ki->p_tsess = 0;	/* may be changed if controlling tty below */
   2354 	COND_SET_VALUE(ki->p_ru, PTRTOUINT64(&p->p_stats->p_ru), allowaddr);
   2355 	ki->p_eflag = 0;
   2356 	ki->p_exitsig = p->p_exitsig;
   2357 	ki->p_flag = L_INMEM;   /* Process never swapped out */
   2358 	ki->p_flag |= sysctl_map_flags(sysctl_flagmap, p->p_flag);
   2359 	ki->p_flag |= sysctl_map_flags(sysctl_sflagmap, p->p_sflag);
   2360 	ki->p_flag |= sysctl_map_flags(sysctl_slflagmap, p->p_slflag);
   2361 	ki->p_flag |= sysctl_map_flags(sysctl_lflagmap, p->p_lflag);
   2362 	ki->p_flag |= sysctl_map_flags(sysctl_stflagmap, p->p_stflag);
   2363 	ki->p_pid = p->p_pid;
   2364 	ki->p_ppid = p->p_ppid;
   2365 	ki->p_uid = kauth_cred_geteuid(p->p_cred);
   2366 	ki->p_ruid = kauth_cred_getuid(p->p_cred);
   2367 	ki->p_gid = kauth_cred_getegid(p->p_cred);
   2368 	ki->p_rgid = kauth_cred_getgid(p->p_cred);
   2369 	ki->p_svuid = kauth_cred_getsvuid(p->p_cred);
   2370 	ki->p_svgid = kauth_cred_getsvgid(p->p_cred);
   2371 	ki->p_ngroups = kauth_cred_ngroups(p->p_cred);
   2372 	kauth_cred_getgroups(p->p_cred, ki->p_groups,
   2373 	    uimin(ki->p_ngroups, sizeof(ki->p_groups) / sizeof(ki->p_groups[0])),
   2374 	    UIO_SYSSPACE);
   2375 
   2376 	ki->p_uticks = p->p_uticks;
   2377 	ki->p_sticks = p->p_sticks;
   2378 	ki->p_iticks = p->p_iticks;
   2379 	ki->p_tpgid = NO_PGID;	/* may be changed if controlling tty below */
   2380 	COND_SET_VALUE(ki->p_tracep, PTRTOUINT64(p->p_tracep), allowaddr);
   2381 	ki->p_traceflag = p->p_traceflag;
   2382 
   2383 	memcpy(&ki->p_sigignore, &p->p_sigctx.ps_sigignore,sizeof(ki_sigset_t));
   2384 	memcpy(&ki->p_sigcatch, &p->p_sigctx.ps_sigcatch, sizeof(ki_sigset_t));
   2385 
   2386 	ki->p_cpticks = 0;
   2387 	ki->p_pctcpu = p->p_pctcpu;
   2388 	ki->p_estcpu = 0;
   2389 	ki->p_stat = p->p_stat; /* Will likely be overridden by LWP status */
   2390 	ki->p_realstat = p->p_stat;
   2391 	ki->p_nice = p->p_nice;
   2392 	ki->p_xstat = P_WAITSTATUS(p);
   2393 	ki->p_acflag = p->p_acflag;
   2394 
   2395 	strncpy(ki->p_comm, p->p_comm,
   2396 	    uimin(sizeof(ki->p_comm), sizeof(p->p_comm)));
   2397 	strncpy(ki->p_ename, p->p_emul->e_name, sizeof(ki->p_ename));
   2398 
   2399 	ki->p_nlwps = p->p_nlwps;
   2400 	ki->p_realflag = ki->p_flag;
   2401 
   2402 	if (p->p_stat != SIDL && !P_ZOMBIE(p) && !zombie) {
   2403 		vm = p->p_vmspace;
   2404 		ki->p_vm_rssize = vm_resident_count(vm);
   2405 		ki->p_vm_tsize = vm->vm_tsize;
   2406 		ki->p_vm_dsize = vm->vm_dsize;
   2407 		ki->p_vm_ssize = vm->vm_ssize;
   2408 		ki->p_vm_vsize = atop(vm->vm_map.size);
   2409 		/*
   2410 		 * Since the stack is initially mapped mostly with
   2411 		 * PROT_NONE and grown as needed, adjust the "mapped size"
   2412 		 * to skip the unused stack portion.
   2413 		 */
   2414 		ki->p_vm_msize =
   2415 		    atop(vm->vm_map.size) - vm->vm_issize + vm->vm_ssize;
   2416 
   2417 		/* Pick the primary (first) LWP */
   2418 		l = proc_active_lwp(p);
   2419 		KASSERT(l != NULL);
   2420 		lwp_lock(l);
   2421 		ki->p_nrlwps = p->p_nrlwps;
   2422 		ki->p_forw = 0;
   2423 		ki->p_back = 0;
   2424 		COND_SET_VALUE(ki->p_addr, PTRTOUINT64(l->l_addr), allowaddr);
   2425 		ki->p_stat = l->l_stat;
   2426 		ki->p_flag |= sysctl_map_flags(sysctl_lwpflagmap, l->l_flag);
   2427 		ki->p_swtime = l->l_swtime;
   2428 		ki->p_slptime = l->l_slptime;
   2429 		if (l->l_stat == LSONPROC)
   2430 			ki->p_schedflags = l->l_cpu->ci_schedstate.spc_flags;
   2431 		else
   2432 			ki->p_schedflags = 0;
   2433 		ki->p_priority = lwp_eprio(l);
   2434 		ki->p_usrpri = l->l_priority;
   2435 		if (l->l_wchan)
   2436 			strncpy(ki->p_wmesg, l->l_wmesg, sizeof(ki->p_wmesg));
   2437 		COND_SET_VALUE(ki->p_wchan, PTRTOUINT64(l->l_wchan), allowaddr);
   2438 		ki->p_cpuid = cpu_index(l->l_cpu);
   2439 		lwp_unlock(l);
   2440 		LIST_FOREACH(l, &p->p_lwps, l_sibling) {
   2441 			/* This is hardly correct, but... */
   2442 			sigplusset(&l->l_sigpend.sp_set, &ss1);
   2443 			sigplusset(&l->l_sigmask, &ss2);
   2444 			ki->p_cpticks += l->l_cpticks;
   2445 			ki->p_pctcpu += l->l_pctcpu;
   2446 			ki->p_estcpu += l->l_estcpu;
   2447 		}
   2448 	}
   2449 	sigplusset(&p->p_sigpend.sp_set, &ss2);
   2450 	memcpy(&ki->p_siglist, &ss1, sizeof(ki_sigset_t));
   2451 	memcpy(&ki->p_sigmask, &ss2, sizeof(ki_sigset_t));
   2452 
   2453 	if (p->p_session != NULL) {
   2454 		ki->p_sid = p->p_session->s_sid;
   2455 		ki->p__pgid = p->p_pgrp->pg_id;
   2456 		if (p->p_session->s_ttyvp)
   2457 			ki->p_eflag |= EPROC_CTTY;
   2458 		if (SESS_LEADER(p))
   2459 			ki->p_eflag |= EPROC_SLEADER;
   2460 		strncpy(ki->p_login, p->p_session->s_login,
   2461 		    uimin(sizeof ki->p_login - 1, sizeof p->p_session->s_login));
   2462 		ki->p_jobc = p->p_pgrp->pg_jobc;
   2463 		if ((p->p_lflag & PL_CONTROLT) && (tp = p->p_session->s_ttyp)) {
   2464 			ki->p_tdev = tp->t_dev;
   2465 			ki->p_tpgid = tp->t_pgrp ? tp->t_pgrp->pg_id : NO_PGID;
   2466 			COND_SET_VALUE(ki->p_tsess, PTRTOUINT64(tp->t_session),
   2467 			    allowaddr);
   2468 		} else {
   2469 			ki->p_tdev = (int32_t)NODEV;
   2470 		}
   2471 	}
   2472 
   2473 	if (!P_ZOMBIE(p) && !zombie) {
   2474 		ki->p_uvalid = 1;
   2475 		ki->p_ustart_sec = p->p_stats->p_start.tv_sec;
   2476 		ki->p_ustart_usec = p->p_stats->p_start.tv_usec;
   2477 
   2478 		calcru(p, &ut, &st, NULL, &rt);
   2479 		ki->p_rtime_sec = rt.tv_sec;
   2480 		ki->p_rtime_usec = rt.tv_usec;
   2481 		ki->p_uutime_sec = ut.tv_sec;
   2482 		ki->p_uutime_usec = ut.tv_usec;
   2483 		ki->p_ustime_sec = st.tv_sec;
   2484 		ki->p_ustime_usec = st.tv_usec;
   2485 
   2486 		memcpy(&ru, &p->p_stats->p_ru, sizeof(ru));
   2487 		ki->p_uru_nvcsw = 0;
   2488 		ki->p_uru_nivcsw = 0;
   2489 		LIST_FOREACH(l2, &p->p_lwps, l_sibling) {
   2490 			ki->p_uru_nvcsw += (l2->l_ncsw - l2->l_nivcsw);
   2491 			ki->p_uru_nivcsw += l2->l_nivcsw;
   2492 			ruadd(&ru, &l2->l_ru);
   2493 		}
   2494 		ki->p_uru_maxrss = ru.ru_maxrss;
   2495 		ki->p_uru_ixrss = ru.ru_ixrss;
   2496 		ki->p_uru_idrss = ru.ru_idrss;
   2497 		ki->p_uru_isrss = ru.ru_isrss;
   2498 		ki->p_uru_minflt = ru.ru_minflt;
   2499 		ki->p_uru_majflt = ru.ru_majflt;
   2500 		ki->p_uru_nswap = ru.ru_nswap;
   2501 		ki->p_uru_inblock = ru.ru_inblock;
   2502 		ki->p_uru_oublock = ru.ru_oublock;
   2503 		ki->p_uru_msgsnd = ru.ru_msgsnd;
   2504 		ki->p_uru_msgrcv = ru.ru_msgrcv;
   2505 		ki->p_uru_nsignals = ru.ru_nsignals;
   2506 
   2507 		timeradd(&p->p_stats->p_cru.ru_utime,
   2508 			 &p->p_stats->p_cru.ru_stime, &ut);
   2509 		ki->p_uctime_sec = ut.tv_sec;
   2510 		ki->p_uctime_usec = ut.tv_usec;
   2511 	}
   2512 }
   2513 
   2514 
   2515 int
   2516 proc_find_locked(struct lwp *l, struct proc **p, pid_t pid)
   2517 {
   2518 	int error;
   2519 
   2520 	mutex_enter(proc_lock);
   2521 	if (pid == -1)
   2522 		*p = l->l_proc;
   2523 	else
   2524 		*p = proc_find(pid);
   2525 
   2526 	if (*p == NULL) {
   2527 		if (pid != -1)
   2528 			mutex_exit(proc_lock);
   2529 		return ESRCH;
   2530 	}
   2531 	if (pid != -1)
   2532 		mutex_enter((*p)->p_lock);
   2533 	mutex_exit(proc_lock);
   2534 
   2535 	error = kauth_authorize_process(l->l_cred,
   2536 	    KAUTH_PROCESS_CANSEE, *p,
   2537 	    KAUTH_ARG(KAUTH_REQ_PROCESS_CANSEE_ENTRY), NULL, NULL);
   2538 	if (error) {
   2539 		if (pid != -1)
   2540 			mutex_exit((*p)->p_lock);
   2541 	}
   2542 	return error;
   2543 }
   2544 
   2545 static int
   2546 fill_pathname(struct lwp *l, pid_t pid, void *oldp, size_t *oldlenp)
   2547 {
   2548 	int error;
   2549 	struct proc *p;
   2550 
   2551 	if ((error = proc_find_locked(l, &p, pid)) != 0)
   2552 		return error;
   2553 
   2554 	if (p->p_path == NULL) {
   2555 		if (pid != -1)
   2556 			mutex_exit(p->p_lock);
   2557 		return ENOENT;
   2558 	}
   2559 
   2560 	size_t len = strlen(p->p_path) + 1;
   2561 	if (oldp != NULL) {
   2562 		size_t copylen = uimin(len, *oldlenp);
   2563 		error = sysctl_copyout(l, p->p_path, oldp, copylen);
   2564 		if (error == 0 && *oldlenp < len)
   2565 			error = ENOSPC;
   2566 	}
   2567 	*oldlenp = len;
   2568 	if (pid != -1)
   2569 		mutex_exit(p->p_lock);
   2570 	return error;
   2571 }
   2572 
   2573 int
   2574 proc_getauxv(struct proc *p, void **buf, size_t *len)
   2575 {
   2576 	struct ps_strings pss;
   2577 	int error;
   2578 	void *uauxv, *kauxv;
   2579 	size_t size;
   2580 
   2581 	if ((error = copyin_psstrings(p, &pss)) != 0)
   2582 		return error;
   2583 	if (pss.ps_envstr == NULL)
   2584 		return EIO;
   2585 
   2586 	size = p->p_execsw->es_arglen;
   2587 	if (size == 0)
   2588 		return EIO;
   2589 
   2590 	size_t ptrsz = PROC_PTRSZ(p);
   2591 	uauxv = (void *)((char *)pss.ps_envstr + (pss.ps_nenvstr + 1) * ptrsz);
   2592 
   2593 	kauxv = kmem_alloc(size, KM_SLEEP);
   2594 
   2595 	error = copyin_proc(p, uauxv, kauxv, size);
   2596 	if (error) {
   2597 		kmem_free(kauxv, size);
   2598 		return error;
   2599 	}
   2600 
   2601 	*buf = kauxv;
   2602 	*len = size;
   2603 
   2604 	return 0;
   2605 }
   2606