Home | History | Annotate | Line # | Download | only in kern
kern_sysctl.c revision 1.161
      1 /*	$NetBSD: kern_sysctl.c,v 1.161 2004/03/24 15:34:53 atatat Exp $	*/
      2 
      3 /*-
      4  * Copyright (c) 2003 The NetBSD Foundation, Inc.
      5  * All rights reserved.
      6  *
      7  * This code is derived from software contributed to The NetBSD Foundation
      8  * by Andrew Brown.
      9  *
     10  * Redistribution and use in source and binary forms, with or without
     11  * modification, are permitted provided that the following conditions
     12  * are met:
     13  * 1. Redistributions of source code must retain the above copyright
     14  *    notice, this list of conditions and the following disclaimer.
     15  * 2. Redistributions in binary form must reproduce the above copyright
     16  *    notice, this list of conditions and the following disclaimer in the
     17  *    documentation and/or other materials provided with the distribution.
     18  * 3. All advertising materials mentioning features or use of this software
     19  *    must display the following acknowledgement:
     20  *      This product includes software developed by the NetBSD
     21  *      Foundation, Inc. and its contributors.
     22  * 4. Neither the name of The NetBSD Foundation nor the names of its
     23  *    contributors may be used to endorse or promote products derived
     24  *    from this software without specific prior written permission.
     25  *
     26  * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
     27  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
     28  * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
     29  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
     30  * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
     31  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
     32  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
     33  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
     34  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
     35  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
     36  * POSSIBILITY OF SUCH DAMAGE.
     37  */
     38 
     39 /*-
     40  * Copyright (c) 1982, 1986, 1989, 1993
     41  *	The Regents of the University of California.  All rights reserved.
     42  *
     43  * This code is derived from software contributed to Berkeley by
     44  * Mike Karels at Berkeley Software Design, Inc.
     45  *
     46  * Redistribution and use in source and binary forms, with or without
     47  * modification, are permitted provided that the following conditions
     48  * are met:
     49  * 1. Redistributions of source code must retain the above copyright
     50  *    notice, this list of conditions and the following disclaimer.
     51  * 2. Redistributions in binary form must reproduce the above copyright
     52  *    notice, this list of conditions and the following disclaimer in the
     53  *    documentation and/or other materials provided with the distribution.
     54  * 3. Neither the name of the University nor the names of its contributors
     55  *    may be used to endorse or promote products derived from this software
     56  *    without specific prior written permission.
     57  *
     58  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
     59  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     60  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     61  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
     62  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     63  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     64  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     65  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     66  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     67  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     68  * SUCH DAMAGE.
     69  *
     70  *	@(#)kern_sysctl.c	8.9 (Berkeley) 5/20/95
     71  */
     72 
     73 /*
     74  * sysctl system call.
     75  */
     76 
     77 #include <sys/cdefs.h>
     78 __KERNEL_RCSID(0, "$NetBSD: kern_sysctl.c,v 1.161 2004/03/24 15:34:53 atatat Exp $");
     79 
     80 #include "opt_defcorename.h"
     81 #include "opt_insecure.h"
     82 #include "ksyms.h"
     83 
     84 #include <sys/param.h>
     85 #include <sys/sysctl.h>
     86 #include <sys/systm.h>
     87 #include <sys/buf.h>
     88 #include <sys/ksyms.h>
     89 #include <sys/malloc.h>
     90 #include <sys/mount.h>
     91 #include <sys/sa.h>
     92 #include <sys/syscallargs.h>
     93 #include <machine/stdarg.h>
     94 
     95 MALLOC_DEFINE(M_SYSCTLNODE, "sysctlnode", "sysctl node structures");
     96 MALLOC_DEFINE(M_SYSCTLDATA, "sysctldata", "misc sysctl data");
     97 
     98 static int sysctl_mmap(SYSCTLFN_RWPROTO);
     99 static int sysctl_alloc(struct sysctlnode *, int);
    100 static int sysctl_realloc(struct sysctlnode *);
    101 
    102 /*
    103  * the "root" of the new sysctl tree
    104  */
    105 static struct sysctlnode sysctl_root = {
    106 	.sysctl_flags = SYSCTL_VERSION|
    107 	    CTLFLAG_ROOT|CTLFLAG_READWRITE|
    108 	    CTLTYPE_NODE,
    109 	.sysctl_num = 0,
    110 	.sysctl_size = sizeof(struct sysctlnode),
    111 	.sysctl_name = "(root)",
    112 };
    113 
    114 /*
    115  * link set of functions that add nodes at boot time (see also
    116  * sysctl_buildtree())
    117  */
    118 __link_set_decl(sysctl_funcs, sysctl_setup_func);
    119 
    120 /*
    121  * The `sysctl_lock' is intended to serialize access to the sysctl
    122  * tree.  Given that it is now (a) dynamic, and (b) most consumers of
    123  * sysctl are going to be copying data out, the old `sysctl_memlock'
    124  * has been `upgraded' to simply guard the whole tree.
    125  *
    126  * The two new data here are to keep track of the locked chunk of
    127  * memory, if there is one, so that it can be released more easily
    128  * from anywhere.
    129  */
    130 struct lock sysctl_treelock;
    131 caddr_t sysctl_memaddr;
    132 size_t sysctl_memsize;
    133 
    134 /*
    135  * Attributes stored in the kernel.
    136  */
    137 char hostname[MAXHOSTNAMELEN];
    138 int hostnamelen;
    139 
    140 char domainname[MAXHOSTNAMELEN];
    141 int domainnamelen;
    142 
    143 long hostid;
    144 
    145 #ifdef INSECURE
    146 int securelevel = -1;
    147 #else
    148 int securelevel = 0;
    149 #endif
    150 
    151 #ifndef DEFCORENAME
    152 #define	DEFCORENAME	"%n.core"
    153 #endif
    154 char defcorename[MAXPATHLEN] = DEFCORENAME;
    155 
    156 /*
    157  * ********************************************************************
    158  * Section 0: Some simple glue
    159  * ********************************************************************
    160  * By wrapping copyin(), copyout(), and copyinstr() like this, we can
    161  * stop caring about who's calling us and simplify some code a bunch.
    162  * ********************************************************************
    163  */
    164 static inline int
    165 sysctl_copyin(const struct lwp *l, const void *uaddr, void *kaddr, size_t len)
    166 {
    167 
    168 	if (l != NULL)
    169 		return (copyin(uaddr, kaddr, len));
    170 	else
    171 		return (kcopy(uaddr, kaddr, len));
    172 }
    173 
    174 static inline int
    175 sysctl_copyout(const struct lwp *l, const void *kaddr, void *uaddr, size_t len)
    176 {
    177 
    178 	if (l != NULL)
    179 		return (copyout(kaddr, uaddr, len));
    180 	else
    181 		return (kcopy(kaddr, uaddr, len));
    182 }
    183 
    184 static inline int
    185 sysctl_copyinstr(const struct lwp *l, const void *uaddr, void *kaddr,
    186 		 size_t len, size_t *done)
    187 {
    188 
    189 	if (l != NULL)
    190 		return (copyinstr(uaddr, kaddr, len, done));
    191 	else
    192 		return (copystr(uaddr, kaddr, len, done));
    193 }
    194 
    195 /*
    196  * ********************************************************************
    197  * Initialize sysctl subsystem.
    198  * ********************************************************************
    199  */
    200 void
    201 sysctl_init(void)
    202 {
    203 	sysctl_setup_func **sysctl_setup, f;
    204 
    205 	lockinit(&sysctl_treelock, PRIBIO|PCATCH, "sysctl", 0, 0);
    206 
    207 	/*
    208 	 * dynamic mib numbers start here
    209 	 */
    210 	sysctl_root.sysctl_num = CREATE_BASE;
    211 
    212         __link_set_foreach(sysctl_setup, sysctl_funcs) {
    213 		/*
    214 		 * XXX - why do i have to coerce the pointers like this?
    215 		 */
    216 		f = (void*)*sysctl_setup;
    217 		(*f)(NULL);
    218 	}
    219 
    220 	/*
    221 	 * setting this means no more permanent nodes can be added,
    222 	 * trees that claim to be readonly at the root now are, and if
    223 	 * the main tree is readonly, *everything* is.
    224 	 */
    225 	sysctl_root.sysctl_flags |= CTLFLAG_PERMANENT;
    226 
    227 }
    228 
    229 /*
    230  * ********************************************************************
    231  * The main native sysctl system call itself.
    232  * ********************************************************************
    233  */
    234 int
    235 sys___sysctl(struct lwp *l, void *v, register_t *retval)
    236 {
    237 	struct sys___sysctl_args /* {
    238 		syscallarg(int *) name;
    239 		syscallarg(u_int) namelen;
    240 		syscallarg(void *) old;
    241 		syscallarg(size_t *) oldlenp;
    242 		syscallarg(void *) new;
    243 		syscallarg(size_t) newlen;
    244 	} */ *uap = v;
    245 	int error, nerror, name[CTL_MAXNAME];
    246 	size_t oldlen, savelen, *oldlenp;
    247 
    248 	/*
    249 	 * get oldlen
    250 	 */
    251 	oldlen = 0;
    252 	oldlenp = SCARG(uap, oldlenp);
    253 	if (oldlenp != NULL) {
    254 		error = copyin(oldlenp, &oldlen, sizeof(oldlen));
    255 		if (error)
    256 			return (error);
    257 	}
    258 	savelen = oldlen;
    259 
    260 	/*
    261 	 * top-level sysctl names may or may not be non-terminal, but
    262 	 * we don't care
    263 	 */
    264 	if (SCARG(uap, namelen) > CTL_MAXNAME || SCARG(uap, namelen) < 1)
    265 		return (EINVAL);
    266 	error = copyin(SCARG(uap, name), &name,
    267 		       SCARG(uap, namelen) * sizeof(int));
    268 	if (error)
    269 		return (error);
    270 
    271 	/*
    272 	 * wire old so that copyout() is less likely to fail?
    273 	 */
    274 	error = sysctl_lock(l, SCARG(uap, old), savelen);
    275 	if (error)
    276 		return (error);
    277 
    278 	/*
    279 	 * do sysctl work (NULL means main built-in default tree)
    280 	 */
    281 	error = sysctl_dispatch(&name[0], SCARG(uap, namelen),
    282 				SCARG(uap, old), &oldlen,
    283 				SCARG(uap, new), SCARG(uap, newlen),
    284 				&name[0], l, NULL);
    285 
    286 	/*
    287 	 * release the sysctl lock
    288 	 */
    289 	sysctl_unlock(l);
    290 
    291 	/*
    292 	 * set caller's oldlen to new value even in the face of an
    293 	 * error (if this gets an error and they didn't have one, they
    294 	 * get this one)
    295 	 */
    296 	if (oldlenp) {
    297 		nerror = copyout(&oldlen, oldlenp, sizeof(oldlen));
    298 		if (error == 0)
    299 			error = nerror;
    300 	}
    301 
    302 	/*
    303 	 * if the only problem is that we weren't given enough space,
    304 	 * that's an ENOMEM error
    305 	 */
    306 	if (error == 0 && SCARG(uap, old) != NULL && savelen < oldlen)
    307 		error = ENOMEM;
    308 
    309 	return (error);
    310 }
    311 
    312 /*
    313  * ********************************************************************
    314  * Section 1: How the tree is used
    315  * ********************************************************************
    316  * Implementations of sysctl for emulations should typically need only
    317  * these three functions in this order: lock the tree, dispatch
    318  * request into it, unlock the tree.
    319  * ********************************************************************
    320  */
    321 int
    322 sysctl_lock(struct lwp *l, void *oldp, size_t savelen)
    323 {
    324 	int error = 0;
    325 
    326 	error = lockmgr(&sysctl_treelock, LK_EXCLUSIVE, NULL);
    327 	if (error)
    328 		return (error);
    329 
    330 	if (l != NULL && oldp != NULL && savelen) {
    331 		error = uvm_vslock(l->l_proc, oldp, savelen, VM_PROT_WRITE);
    332 		if (error) {
    333 			(void) lockmgr(&sysctl_treelock, LK_RELEASE, NULL);
    334 			return (error);
    335 		}
    336 		sysctl_memaddr = oldp;
    337 		sysctl_memsize = savelen;
    338 	}
    339 
    340 	return (0);
    341 }
    342 
    343 /*
    344  * ********************************************************************
    345  * the main sysctl dispatch routine.  scans the given tree and picks a
    346  * function to call based on what it finds.
    347  * ********************************************************************
    348  */
    349 int
    350 sysctl_dispatch(SYSCTLFN_RWARGS)
    351 {
    352 	int error;
    353 	sysctlfn fn;
    354 	int ni;
    355 
    356 	if (rnode && SYSCTL_VERS(rnode->sysctl_flags) != SYSCTL_VERSION) {
    357 		printf("sysctl_dispatch: rnode %p wrong version\n", rnode);
    358 		return (EINVAL);
    359 	}
    360 
    361 	fn = NULL;
    362 	error = sysctl_locate(l, name, namelen, &rnode, &ni);
    363 
    364 	/*
    365 	 * the node we ended up at has a function, so call it.  it can
    366 	 * hand off to query or create if it wants to.
    367 	 */
    368 	if (rnode->sysctl_func != NULL)
    369 		fn = rnode->sysctl_func;
    370 
    371 	/*
    372 	 * we found the node they were looking for, so do a lookup.
    373 	 */
    374 	else if (error == 0)
    375 		fn = (sysctlfn)sysctl_lookup; /* XXX may write to rnode */
    376 
    377 	/*
    378 	 * prospective parent node found, but the terminal node was
    379 	 * not.  generic operations associate with the parent.
    380 	 */
    381 	else if (error == ENOENT && (ni + 1) == namelen && name[ni] < 0) {
    382 		switch (name[ni]) {
    383 		case CTL_QUERY:
    384 			fn = sysctl_query;
    385 			break;
    386 		case CTL_CREATE:
    387 #if NKSYMS > 0
    388 		case CTL_CREATESYM:
    389 #endif /* NKSYMS > 0 */
    390 			fn = (sysctlfn)sysctl_create; /* we own the rnode */
    391 			break;
    392 		case CTL_DESTROY:
    393 			fn = (sysctlfn)sysctl_destroy; /* we own the rnode */
    394 			break;
    395 		case CTL_MMAP:
    396 			fn = (sysctlfn)sysctl_mmap; /* we own the rnode */
    397 			break;
    398 		default:
    399 			error = EOPNOTSUPP;
    400 			break;
    401 		}
    402 	}
    403 
    404 	/*
    405 	 * after all of that, maybe we found someone who knows how to
    406 	 * get us what we want?
    407 	 */
    408 	if (fn != NULL)
    409 		error = (*fn)(name + ni, namelen - ni, oldp, oldlenp,
    410 			      newp, newlen, name, l, rnode);
    411 
    412 	else if (error == 0)
    413 		error = EOPNOTSUPP;
    414 
    415 	return (error);
    416 }
    417 
    418 /*
    419  * ********************************************************************
    420  * Releases the tree lock.  Note that if uvm_vslock() was called when
    421  * the lock was taken, we release that memory now.  By keeping track
    422  * of where and how much by ourselves, the lock can be released much
    423  * more easily from anywhere.
    424  * ********************************************************************
    425  */
    426 void
    427 sysctl_unlock(struct lwp *l)
    428 {
    429 
    430 	if (l != NULL && sysctl_memsize != 0) {
    431 		uvm_vsunlock(l->l_proc, sysctl_memaddr, sysctl_memsize);
    432 		sysctl_memsize = 0;
    433 	}
    434 
    435 	(void) lockmgr(&sysctl_treelock, LK_RELEASE, NULL);
    436 }
    437 
    438 /*
    439  * ********************************************************************
    440  * Section 2: The main tree interfaces
    441  * ********************************************************************
    442  * This is how sysctl_dispatch() does its work, and you can too, by
    443  * calling these routines from helpers (though typically only
    444  * sysctl_lookup() will be used).  The tree MUST BE LOCKED when these
    445  * are called.
    446  * ********************************************************************
    447  */
    448 
    449 /*
    450  * sysctl_locate -- Finds the node matching the given mib under the
    451  * given tree (via rv).  If no tree is given, we fall back to the
    452  * native tree.  The current process (via l) is used for access
    453  * control on the tree (some nodes may be traversable only by root) and
    454  * on return, nip will show how many numbers in the mib were consumed.
    455  */
    456 int
    457 sysctl_locate(struct lwp *l, const int *name, u_int namelen,
    458 	      struct sysctlnode **rnode, int *nip)
    459 {
    460 	struct sysctlnode *node, *pnode;
    461 	int tn, si, ni, error, alias;
    462 
    463 	/*
    464 	 * basic checks and setup
    465 	 */
    466 	if (*rnode == NULL)
    467 		*rnode = &sysctl_root;
    468 	if (nip)
    469 		*nip = 0;
    470 	if (namelen < 0)
    471 		return (EINVAL);
    472 	if (namelen == 0)
    473 		return (0);
    474 
    475 	/*
    476 	 * search starts from "root"
    477 	 */
    478 	pnode = *rnode;
    479 	if (SYSCTL_VERS(pnode->sysctl_flags) != SYSCTL_VERSION) {
    480 		printf("sysctl_locate: pnode %p wrong version\n", pnode);
    481 		return (EINVAL);
    482 	}
    483 	node = pnode->sysctl_child;
    484 	error = 0;
    485 
    486 	/*
    487 	 * scan for node to which new node should be attached
    488 	 */
    489 	for (ni = 0; ni < namelen; ni++) {
    490 		/*
    491 		 * walked off bottom of tree
    492 		 */
    493 		if (node == NULL) {
    494 			if (SYSCTL_TYPE(pnode->sysctl_flags) == CTLTYPE_NODE)
    495 				error = ENOENT;
    496 			else
    497 				error = ENOTDIR;
    498 			break;
    499 		}
    500 		/*
    501 		 * can anyone traverse this node or only root?
    502 		 */
    503 		if (l != NULL && (pnode->sysctl_flags & CTLFLAG_PRIVATE) &&
    504 		    (error = suser(l->l_proc->p_ucred, &l->l_proc->p_acflag))
    505 		    != 0)
    506 			return (error);
    507 		/*
    508 		 * find a child node with the right number
    509 		 */
    510 		tn = name[ni];
    511 		alias = 0;
    512 
    513 		si = 0;
    514 		/*
    515 		 * Note: ANYNUMBER only matches positive integers.
    516 		 * Since ANYNUMBER is only permitted on single-node
    517 		 * sub-trees (eg proc), check before the loop and skip
    518 		 * it if we can.
    519 		 */
    520 		if ((node[si].sysctl_flags & CTLFLAG_ANYNUMBER) && (tn >= 0))
    521 			goto foundit;
    522 		for (; si < pnode->sysctl_clen; si++) {
    523 			if (node[si].sysctl_num == tn) {
    524 				if (node[si].sysctl_flags & CTLFLAG_ALIAS) {
    525 					if (alias++ == 4)
    526 						break;
    527 					else {
    528 						tn = node[si].sysctl_alias;
    529 						si = -1;
    530 					}
    531 				}
    532 				else
    533 					goto foundit;
    534 			}
    535 		}
    536 		/*
    537 		 * if we ran off the end, it obviously doesn't exist
    538 		 */
    539 		error = ENOENT;
    540 		break;
    541 
    542 		/*
    543 		 * so far so good, move on down the line
    544 		 */
    545 	  foundit:
    546 		pnode = &node[si];
    547 		if (SYSCTL_TYPE(pnode->sysctl_flags) == CTLTYPE_NODE)
    548 			node = node[si].sysctl_child;
    549 		else
    550 			node = NULL;
    551 	}
    552 
    553 	*rnode = pnode;
    554 	if (nip)
    555 		*nip = ni;
    556 
    557 	return (error);
    558 }
    559 
    560 /*
    561  * sysctl_query -- The auto-discovery engine.  Copies out the
    562  * descriptions on nodes under the given node and handles overlay
    563  * trees.
    564  */
    565 int
    566 sysctl_query(SYSCTLFN_ARGS)
    567 {
    568 	int error, ni, elim;
    569 	size_t out, left, t;
    570 	struct sysctlnode *enode, *onode;
    571 
    572 	if (SYSCTL_VERS(rnode->sysctl_flags) != SYSCTL_VERSION) {
    573 		printf("sysctl_query: rnode %p wrong version\n", rnode);
    574 		return (EINVAL);
    575 	}
    576 
    577 	if (newp != NULL)
    578 		return (EPERM);
    579 	if (SYSCTL_TYPE(rnode->sysctl_flags) != CTLTYPE_NODE)
    580 		return (ENOTDIR);
    581 	if (namelen != 1 || name[0] != CTL_QUERY)
    582 		return (EINVAL);
    583 
    584 	error = 0;
    585 	out = 0;
    586 	left = *oldlenp;
    587 	elim = 0;
    588 	enode = NULL;
    589 
    590 	/*
    591 	 * process has overlay tree
    592 	 */
    593 	if (l && l->l_proc->p_emul->e_sysctlovly) {
    594 		enode = l->l_proc->p_emul->e_sysctlovly;
    595 		elim = (name - oname);
    596 		error = sysctl_locate(l, oname, elim, &enode, NULL);
    597 		if (error == 0) {
    598 			/* ah, found parent in overlay */
    599 			elim = enode->sysctl_clen;
    600 			enode = enode->sysctl_child;
    601 		}
    602 		else {
    603 			error = 0;
    604 			elim = 0;
    605 			enode = NULL;
    606 		}
    607 	}
    608 
    609 	for (ni = 0; ni < rnode->sysctl_clen; ni++) {
    610 		t = MIN(left, sizeof(struct sysctlnode));
    611 		onode = &rnode->sysctl_child[ni];
    612 		if (enode && enode->sysctl_num == onode->sysctl_num) {
    613 			if (SYSCTL_TYPE(enode->sysctl_flags) != CTLTYPE_NODE)
    614 				onode = enode;
    615 			if (--elim > 0)
    616 				enode++;
    617 			else
    618 				enode = NULL;
    619 		}
    620 		if (oldp != NULL && t > 0)
    621 			error = sysctl_copyout(l, onode, (char*)oldp + out, t);
    622 		if (error)
    623 			return (error);
    624 		out += sizeof(struct sysctlnode);
    625 		left -= t;
    626 	}
    627 
    628 	/*
    629 	 * overlay trees *MUST* be entirely consumed
    630 	 */
    631 	KASSERT(enode == NULL);
    632 
    633 	*oldlenp = out;
    634 
    635 	return (error);
    636 }
    637 
    638 #ifdef SYSCTL_DEBUG_CREATE
    639 #undef sysctl_create
    640 #endif /* SYSCTL_DEBUG_CREATE */
    641 
    642 /*
    643  * sysctl_create -- Adds a node (the description of which is taken
    644  * from newp) to the tree, returning a copy of it in the space pointed
    645  * to by oldp.  In the event that the requested slot is already taken
    646  * (either by name or by number), the offending node is returned
    647  * instead.  Yes, this is complex, but we want to make sure everything
    648  * is proper.
    649  */
    650 int
    651 sysctl_create(SYSCTLFN_RWARGS)
    652 {
    653 	struct sysctlnode nnode, *node, *pnode;
    654 	int error, ni, at, nm, type, sz, flags, rw, anum;
    655 	void *own;
    656 
    657 	error = 0;
    658 	own = NULL;
    659 	anum = -1;
    660 
    661 	if (SYSCTL_VERS(rnode->sysctl_flags) != SYSCTL_VERSION) {
    662 		printf("sysctl_create: rnode %p wrong version\n", rnode);
    663 		return (EINVAL);
    664 	}
    665 
    666 	if (namelen != 1 || (name[namelen - 1] != CTL_CREATE
    667 #if NKSYMS > 0
    668 			     && name[namelen - 1] != CTL_CREATESYM
    669 #endif /* NKSYMS > 0 */
    670 			     ))
    671 		return (EINVAL);
    672 
    673 	/*
    674 	 * processes can only add nodes at securelevel 0, must be
    675 	 * root, and can't add nodes to a parent that's not writeable
    676 	 */
    677 	if (l != NULL) {
    678 #ifndef SYSCTL_DISALLOW_CREATE
    679 		if (securelevel > 0)
    680 			return (EPERM);
    681 		error = suser(l->l_proc->p_ucred, &l->l_proc->p_acflag);
    682 		if (error)
    683 			return (error);
    684 		if (!(rnode->sysctl_flags & CTLFLAG_READWRITE))
    685 #endif /* SYSCTL_DISALLOW_CREATE */
    686 			return (EPERM);
    687 	}
    688 
    689 	/*
    690 	 * nothing can add a node if:
    691 	 * we've finished initial set up and
    692 	 * the tree itself is not writeable or
    693 	 * the entire sysctl system is not writeable
    694 	 */
    695 	if ((sysctl_root.sysctl_flags & CTLFLAG_PERMANENT) &&
    696 	    (!(sysctl_rootof(rnode)->sysctl_flags & CTLFLAG_READWRITE) ||
    697 	     !(sysctl_root.sysctl_flags & CTLFLAG_READWRITE)))
    698 		return (EPERM);
    699 
    700 	/*
    701 	 * it must be a "node", not a "int" or something
    702 	 */
    703 	if (SYSCTL_TYPE(rnode->sysctl_flags) != CTLTYPE_NODE)
    704 		return (ENOTDIR);
    705 	pnode = rnode;
    706 
    707 	if (newp == NULL || newlen != sizeof(struct sysctlnode))
    708 		return (EINVAL);
    709 	error = sysctl_copyin(l, newp, &nnode, MIN(sizeof(nnode), newlen));
    710 	if (error)
    711 		return (error);
    712 
    713 	/*
    714 	 * nodes passed in don't *have* parents
    715 	 */
    716 	if (nnode.sysctl_parent != NULL)
    717 		return (EINVAL);
    718 
    719 	/*
    720 	 * if we are indeed adding it, it should be a "good" name and
    721 	 * number
    722 	 */
    723 	nm = nnode.sysctl_num;
    724 #if NKSYMS > 0
    725 	if (nm == CTL_CREATESYM)
    726 		nm = CTL_CREATE;
    727 #endif /* NKSYMS > 0 */
    728 	if (nm < 0 && nm != CTL_CREATE)
    729 		return (EINVAL);
    730 	sz = 0;
    731 
    732 	/*
    733 	 * the name can't start with a digit
    734 	 */
    735 	if (nnode.sysctl_name[sz] >= '0' &&
    736 	    nnode.sysctl_name[sz] <= '9')
    737 		return (EINVAL);
    738 
    739 	/*
    740 	 * the name must be only alphanumerics or - or _, longer than
    741 	 * 0 bytes and less that SYSCTL_NAMELEN
    742 	 */
    743 	while (sz < SYSCTL_NAMELEN && nnode.sysctl_name[sz] != '\0') {
    744 		if ((nnode.sysctl_name[sz] >= '0' &&
    745 		     nnode.sysctl_name[sz] <= '9') ||
    746 		    (nnode.sysctl_name[sz] >= 'A' &&
    747 		     nnode.sysctl_name[sz] <= 'Z') ||
    748 		    (nnode.sysctl_name[sz] >= 'a' &&
    749 		     nnode.sysctl_name[sz] <= 'z') ||
    750 		    nnode.sysctl_name[sz] == '-' ||
    751 		    nnode.sysctl_name[sz] == '_')
    752 			sz++;
    753 		else
    754 			return (EINVAL);
    755 	}
    756 	if (sz == 0 || sz == SYSCTL_NAMELEN)
    757 		return (EINVAL);
    758 
    759 	/*
    760 	 * various checks revolve around size vs type, etc
    761 	 */
    762 	type = SYSCTL_TYPE(nnode.sysctl_flags);
    763 	flags = SYSCTL_FLAGS(nnode.sysctl_flags);
    764 	rw = (flags & CTLFLAG_READWRITE) ? B_WRITE : B_READ;
    765 	sz = nnode.sysctl_size;
    766 
    767 	/*
    768 	 * find out if there's a collision, and if so, let the caller
    769 	 * know what they collided with
    770 	 */
    771 	node = pnode->sysctl_child;
    772 	if (((flags & CTLFLAG_ANYNUMBER) && node) ||
    773 	    (node && node->sysctl_flags & CTLFLAG_ANYNUMBER))
    774 		return (EINVAL);
    775 	for (ni = at = 0; ni < pnode->sysctl_clen; ni++) {
    776 		if (nm == node[ni].sysctl_num ||
    777 		    strcmp(nnode.sysctl_name, node[ni].sysctl_name) == 0) {
    778 			if (oldp != NULL) {
    779 				/*
    780 				 * ignore error here, since we
    781 				 * are already fixed on EEXIST
    782 				 */
    783 				(void)sysctl_copyout(l, &node[ni], oldp,
    784 				     MIN(*oldlenp, sizeof(struct sysctlnode)));
    785 			}
    786 			*oldlenp = sizeof(struct sysctlnode);
    787 			return (EEXIST);
    788 		}
    789 		if (nm > node[ni].sysctl_num)
    790 			at++;
    791 	}
    792 
    793 	/*
    794 	 * use sysctl_ver to add to the tree iff it hasn't changed
    795 	 */
    796 	if (nnode.sysctl_ver != 0) {
    797 		/*
    798 		 * a specified value must match either the parent
    799 		 * node's version or the root node's version
    800 		 */
    801 		if (nnode.sysctl_ver != sysctl_rootof(rnode)->sysctl_ver &&
    802 		    nnode.sysctl_ver != rnode->sysctl_ver) {
    803 			return (EINVAL);
    804 		}
    805 	}
    806 
    807 	/*
    808 	 * only the kernel can assign functions to entries
    809 	 */
    810 	if (l != NULL && nnode.sysctl_func != NULL)
    811 		return (EPERM);
    812 
    813 	/*
    814 	 * only the kernel can create permanent entries, and only then
    815 	 * before the kernel is finished setting itself up
    816 	 */
    817 	if (l != NULL && (flags & ~SYSCTL_USERFLAGS))
    818 		return (EPERM);
    819 	if ((flags & CTLFLAG_PERMANENT) &
    820 	    (sysctl_root.sysctl_flags & CTLFLAG_PERMANENT))
    821 		return (EPERM);
    822 	if ((flags & (CTLFLAG_OWNDATA | CTLFLAG_IMMEDIATE)) ==
    823 	    (CTLFLAG_OWNDATA | CTLFLAG_IMMEDIATE))
    824 		return (EINVAL);
    825 	if ((flags & CTLFLAG_IMMEDIATE) &&
    826 	    type != CTLTYPE_INT && type != CTLTYPE_QUAD)
    827 		return (EINVAL);
    828 
    829 	/*
    830 	 * check size, or set it if unset and we can figure it out.
    831 	 * kernel created nodes are allowed to have a function instead
    832 	 * of a size (or a data pointer).
    833 	 */
    834 	switch (type) {
    835 	case CTLTYPE_NODE:
    836 		/*
    837 		 * only *i* can assert the size of a node
    838 		 */
    839 		if (flags & CTLFLAG_ALIAS) {
    840 			anum = nnode.sysctl_alias;
    841 			if (anum < 0)
    842 				return (EINVAL);
    843 			nnode.sysctl_alias = 0;
    844 		}
    845 		if (sz != 0 || nnode.sysctl_data != NULL)
    846 			return (EINVAL);
    847 		if (nnode.sysctl_csize != 0 ||
    848 		    nnode.sysctl_clen != 0 ||
    849 		    nnode.sysctl_child != 0)
    850 			return (EINVAL);
    851 		if (flags & CTLFLAG_OWNDATA)
    852 			return (EINVAL);
    853 		sz = sizeof(struct sysctlnode);
    854 		break;
    855 	case CTLTYPE_INT:
    856 		/*
    857 		 * since an int is an int, if the size is not given or
    858 		 * is wrong, we can "int-uit" it.
    859 		 */
    860 		if (sz != 0 && sz != sizeof(int))
    861 			return (EINVAL);
    862 		sz = sizeof(int);
    863 		break;
    864 	case CTLTYPE_STRING:
    865 		/*
    866 		 * strings are a little more tricky
    867 		 */
    868 		if (sz == 0) {
    869 			if (l == NULL) {
    870 				if (nnode.sysctl_func == NULL) {
    871 					if (nnode.sysctl_data == NULL)
    872 						return (EINVAL);
    873 					else
    874 						sz = strlen(nnode.sysctl_data) +
    875 						    1;
    876 				}
    877 			}
    878 			else if (nnode.sysctl_data == NULL &&
    879 				 flags & CTLFLAG_OWNDATA) {
    880 				return (EINVAL);
    881 			}
    882 			else {
    883 				char v[PAGE_SIZE], *e;
    884 				size_t s;
    885 
    886 				/*
    887 				 * we want a rough idea of what the
    888 				 * size is now
    889 				 */
    890 				e = nnode.sysctl_data;
    891 				do {
    892 					error = copyinstr(e, &v[0], sizeof(v),
    893 							  &s);
    894 					if (error) {
    895 						if (error != ENAMETOOLONG)
    896 							return (error);
    897 						e += PAGE_SIZE;
    898 						if ((e - 32 * PAGE_SIZE) >
    899 						    (char*)nnode.sysctl_data)
    900 							return (ERANGE);
    901 					}
    902 				} while (error != 0);
    903 				sz = s + (e - (char*)nnode.sysctl_data);
    904 			}
    905 		}
    906 		break;
    907 	case CTLTYPE_QUAD:
    908 		if (sz != 0 && sz != sizeof(u_quad_t))
    909 			return (EINVAL);
    910 		sz = sizeof(u_quad_t);
    911 		break;
    912 	case CTLTYPE_STRUCT:
    913 		if (sz == 0) {
    914 			if (l != NULL || nnode.sysctl_func == NULL)
    915 				return (EINVAL);
    916 			if (flags & CTLFLAG_OWNDATA)
    917 				return (EINVAL);
    918 		}
    919 		break;
    920 	default:
    921 		return (EINVAL);
    922 	}
    923 
    924 	/*
    925 	 * at this point, if sz is zero, we *must* have a
    926 	 * function to go with it and we can't own it.
    927 	 */
    928 
    929 	/*
    930 	 *  l  ptr own
    931 	 *  0   0   0  -> EINVAL (if no func)
    932 	 *  0   0   1  -> own
    933 	 *  0   1   0  -> kptr
    934 	 *  0   1   1  -> kptr
    935 	 *  1   0   0  -> EINVAL
    936 	 *  1   0   1  -> own
    937 	 *  1   1   0  -> kptr, no own (fault on lookup)
    938 	 *  1   1   1  -> uptr, own
    939 	 */
    940 	if (type != CTLTYPE_NODE) {
    941 		if (sz != 0) {
    942 			if (flags & CTLFLAG_OWNDATA) {
    943 				own = malloc(sz, M_SYSCTLDATA,
    944 					     M_WAITOK|M_CANFAIL);
    945 				if (nnode.sysctl_data == NULL)
    946 					memset(own, 0, sz);
    947 				else {
    948 					error = sysctl_copyin(l,
    949 					    nnode.sysctl_data, own, sz);
    950 					if (error != 0) {
    951 						FREE(own, M_SYSCTLDATA);
    952 						return (error);
    953 					}
    954 				}
    955 			}
    956 			else if ((nnode.sysctl_data != NULL) &&
    957 				 !(flags & CTLFLAG_IMMEDIATE)) {
    958 #if NKSYMS > 0
    959 				if (name[namelen - 1] == CTL_CREATESYM) {
    960 					char symname[128]; /* XXX enough? */
    961 					u_long symaddr;
    962 					size_t symlen;
    963 
    964 					error = sysctl_copyinstr(l,
    965 					    nnode.sysctl_data, symname,
    966 					    sizeof(symname), &symlen);
    967 					if (error)
    968 						return (error);
    969 					error = ksyms_getval_from_kernel(NULL,
    970 					    symname, &symaddr, KSYMS_EXTERN);
    971 					if (error)
    972 						return (error); /* EINVAL? */
    973 					nnode.sysctl_data = (void*)symaddr;
    974 				}
    975 #endif /* NKSYMS > 0 */
    976 				/*
    977 				 * Ideally, we'd like to verify here
    978 				 * that this address is acceptable,
    979 				 * but...
    980 				 *
    981 				 * - it might be valid now, only to
    982 				 *   become invalid later
    983 				 *
    984 				 * - it might be invalid only for the
    985 				 *   moment and valid later
    986 				 *
    987 				 * - or something else.
    988 				 *
    989 				 * Since we can't get a good answer,
    990 				 * we'll just accept the address as
    991 				 * given, and fault on individual
    992 				 * lookups.
    993 				 */
    994 			}
    995 		}
    996 		else if (nnode.sysctl_func == NULL)
    997 			return (EINVAL);
    998 	}
    999 
   1000 	/*
   1001 	 * a process can't assign a function to a node, and the kernel
   1002 	 * can't create a node that has no function or data.
   1003 	 * (XXX somewhat redundant check)
   1004 	 */
   1005 	if (l != NULL || nnode.sysctl_func == NULL) {
   1006 		if (type != CTLTYPE_NODE &&
   1007 		    nnode.sysctl_data == NULL &&
   1008 		    !(flags & CTLFLAG_IMMEDIATE) &&
   1009 		    own == NULL)
   1010 			return (EINVAL);
   1011 	}
   1012 
   1013 #ifdef SYSCTL_DISALLOW_KWRITE
   1014 	/*
   1015 	 * a process can't create a writable node unless it refers to
   1016 	 * new data.
   1017 	 */
   1018 	if (l != NULL && own == NULL && type != CTLTYPE_NODE &&
   1019 	    (flags & CTLFLAG_READWRITE) != CTLFLAG_READONLY &&
   1020 	    !(flags & CTLFLAG_IMMEDIATE))
   1021 		return (EPERM);
   1022 #endif /* SYSCTL_DISALLOW_KWRITE */
   1023 
   1024 	/*
   1025 	 * make sure there's somewhere to put the new stuff.
   1026 	 */
   1027 	if (pnode->sysctl_child == NULL) {
   1028 		if (flags & CTLFLAG_ANYNUMBER)
   1029 			error = sysctl_alloc(pnode, 1);
   1030 		else
   1031 			error = sysctl_alloc(pnode, 0);
   1032 		if (error)
   1033 			return (error);
   1034 	}
   1035 	node = pnode->sysctl_child;
   1036 
   1037 	/*
   1038 	 * no collisions, so pick a good dynamic number if we need to.
   1039 	 */
   1040 	if (nm == CTL_CREATE) {
   1041 		nm = ++sysctl_root.sysctl_num;
   1042 		for (ni = 0; ni < pnode->sysctl_clen; ni++) {
   1043 			if (nm == node[ni].sysctl_num) {
   1044 				nm++;
   1045 				ni = -1;
   1046 			}
   1047 			else if (nm > node[ni].sysctl_num)
   1048 				at = ni + 1;
   1049 		}
   1050 	}
   1051 
   1052 	/*
   1053 	 * oops...ran out of space
   1054 	 */
   1055 	if (pnode->sysctl_clen == pnode->sysctl_csize) {
   1056 		error = sysctl_realloc(pnode);
   1057 		if (error)
   1058 			return (error);
   1059 		node = pnode->sysctl_child;
   1060 	}
   1061 
   1062 	/*
   1063 	 * insert new node data
   1064 	 */
   1065 	if (at < pnode->sysctl_clen) {
   1066 		int t;
   1067 
   1068 		/*
   1069 		 * move the nodes that should come after the new one
   1070 		 */
   1071 		memmove(&node[at + 1], &node[at],
   1072 			(pnode->sysctl_clen - at) * sizeof(struct sysctlnode));
   1073 		memset(&node[at], 0, sizeof(struct sysctlnode));
   1074 		node[at].sysctl_parent = pnode;
   1075 		/*
   1076 		 * and...reparent any children of any moved nodes
   1077 		 */
   1078 		for (ni = at; ni <= pnode->sysctl_clen; ni++)
   1079 			if (SYSCTL_TYPE(node[ni].sysctl_flags) == CTLTYPE_NODE)
   1080 				for (t = 0; t < node[ni].sysctl_clen; t++)
   1081 					node[ni].sysctl_child[t].sysctl_parent =
   1082 						&node[ni];
   1083 	}
   1084 	node = &node[at];
   1085 	pnode->sysctl_clen++;
   1086 
   1087 	strlcpy(node->sysctl_name, nnode.sysctl_name,
   1088 		sizeof(node->sysctl_name));
   1089 	node->sysctl_num = nm;
   1090 	node->sysctl_size = sz;
   1091 	node->sysctl_flags = SYSCTL_VERSION|type|flags; /* XXX other trees */
   1092 	node->sysctl_csize = 0;
   1093 	node->sysctl_clen = 0;
   1094 	if (own) {
   1095 		node->sysctl_data = own;
   1096 		node->sysctl_flags |= CTLFLAG_OWNDATA;
   1097 	}
   1098 	else if (flags & CTLFLAG_ALIAS) {
   1099 		node->sysctl_alias = anum;
   1100 	}
   1101 	else if (flags & CTLFLAG_IMMEDIATE) {
   1102 		switch (type) {
   1103 		case CTLTYPE_INT:
   1104 			node->sysctl_idata = nnode.sysctl_idata;
   1105 			break;
   1106 		case CTLTYPE_QUAD:
   1107 			node->sysctl_qdata = nnode.sysctl_qdata;
   1108 			break;
   1109 		}
   1110 	}
   1111 	else {
   1112 		node->sysctl_data = nnode.sysctl_data;
   1113 		node->sysctl_flags &= ~CTLFLAG_OWNDATA;
   1114 	}
   1115         node->sysctl_func = nnode.sysctl_func;
   1116         node->sysctl_child = NULL;
   1117 	/* node->sysctl_parent should already be done */
   1118 
   1119 	/*
   1120 	 * update "version" on path to "root"
   1121 	 */
   1122 	for (; rnode->sysctl_parent != NULL; rnode = rnode->sysctl_parent)
   1123 		;
   1124 	pnode = node;
   1125 	for (nm = rnode->sysctl_ver + 1; pnode != NULL;
   1126 	     pnode = pnode->sysctl_parent)
   1127 		pnode->sysctl_ver = nm;
   1128 
   1129 	if (oldp != NULL)
   1130 		error = sysctl_copyout(l, node, oldp,
   1131 		    MIN(*oldlenp, sizeof(struct sysctlnode)));
   1132 	*oldlenp = sizeof(struct sysctlnode);
   1133 
   1134 	return (error);
   1135 }
   1136 
   1137 /*
   1138  * ********************************************************************
   1139  * A wrapper around sysctl_create() that prints the thing we're trying
   1140  * to add.
   1141  * ********************************************************************
   1142  */
   1143 #ifdef SYSCTL_DEBUG_CREATE
   1144 int _sysctl_create(SYSCTLFN_RWPROTO);
   1145 int
   1146 _sysctl_create(SYSCTLFN_RWARGS)
   1147 {
   1148 	const struct sysctlnode *node;
   1149 	int k, rc, ni, nl = namelen + (name - oname);
   1150 
   1151 	node = newp;
   1152 
   1153 	printf("namelen %d (", nl);
   1154 	for (ni = 0; ni < nl - 1; ni++)
   1155 		printf(" %d", oname[ni]);
   1156 	printf(" %d )\t[%s]\tflags %08x (%08x %d %zu)\n",
   1157 	       k = node->sysctl_num,
   1158 	       node->sysctl_name,
   1159 	       node->sysctl_flags,
   1160 	       SYSCTL_FLAGS(node->sysctl_flags),
   1161 	       SYSCTL_TYPE(node->sysctl_flags),
   1162 	       node->sysctl_size);
   1163 
   1164 	node = rnode;
   1165 	rc = sysctl_create(SYSCTLFN_CALL(rnode));
   1166 
   1167 	printf("sysctl_create(");
   1168 	for (ni = 0; ni < nl - 1; ni++)
   1169 		printf(" %d", oname[ni]);
   1170 	printf(" %d ) returned %d\n", k, rc);
   1171 
   1172 	return (rc);
   1173 }
   1174 #define sysctl_create _sysctl_create
   1175 #endif /* SYSCTL_DEBUG_CREATE */
   1176 
   1177 /*
   1178  * sysctl_destroy -- Removes a node (as described by newp) from the
   1179  * given tree, returning (if successful) a copy of the dead node in
   1180  * oldp.  Since we're removing stuff, there's not much to check.
   1181  */
   1182 int
   1183 sysctl_destroy(SYSCTLFN_RWARGS)
   1184 {
   1185 	struct sysctlnode *node, *pnode, onode, nnode;
   1186 	int ni, error;
   1187 
   1188 	if (SYSCTL_VERS(rnode->sysctl_flags) != SYSCTL_VERSION) {
   1189 		printf("sysctl_destroy: rnode %p wrong version\n", rnode);
   1190 		return (EINVAL);
   1191 	}
   1192 
   1193 	error = 0;
   1194 
   1195 	if (namelen != 1 || name[namelen - 1] != CTL_DESTROY)
   1196 		return (EINVAL);
   1197 
   1198 	/*
   1199 	 * processes can only destroy nodes at securelevel 0, must be
   1200 	 * root, and can't remove nodes from a parent that's not
   1201 	 * writeable
   1202 	 */
   1203 	if (l != NULL) {
   1204 #ifndef SYSCTL_DISALLOW_CREATE
   1205 		if (securelevel > 0)
   1206 			return (EPERM);
   1207 		error = suser(l->l_proc->p_ucred, &l->l_proc->p_acflag);
   1208 		if (error)
   1209 			return (error);
   1210 		if (!(rnode->sysctl_flags & CTLFLAG_READWRITE))
   1211 #endif /* SYSCTL_DISALLOW_CREATE */
   1212 			return (EPERM);
   1213 	}
   1214 
   1215 	/*
   1216 	 * nothing can remove a node if:
   1217 	 * the node is permanent (checked later) or
   1218 	 * the tree itself is not writeable or
   1219 	 * the entire sysctl system is not writeable
   1220 	 */
   1221 	if (!(sysctl_rootof(rnode)->sysctl_flags & CTLFLAG_READWRITE) ||
   1222 	    !(sysctl_root.sysctl_flags & CTLFLAG_READWRITE))
   1223 		return (EPERM);
   1224 
   1225 	if (newp == NULL)
   1226 		return (EINVAL);
   1227 	error = sysctl_copyin(l, newp, &nnode, MIN(sizeof(nnode), newlen));
   1228 	if (error)
   1229 		return (error);
   1230 	memset(&onode, 0, sizeof(struct sysctlnode));
   1231 
   1232 	node = rnode->sysctl_child;
   1233 	for (ni = 0; ni < rnode->sysctl_clen; ni++) {
   1234 		if (nnode.sysctl_num == node[ni].sysctl_num) {
   1235 			/*
   1236 			 * if name specified, must match
   1237 			 */
   1238 			if (nnode.sysctl_name[0] != '\0' &&
   1239 			    strcmp(nnode.sysctl_name, node[ni].sysctl_name))
   1240 				continue;
   1241 			/*
   1242 			 * if version specified, must match
   1243 			 */
   1244 			if (nnode.sysctl_ver != 0 &&
   1245 			    nnode.sysctl_ver != node[ni].sysctl_ver)
   1246 				continue;
   1247 			/*
   1248 			 * this must be the one
   1249 			 */
   1250 			break;
   1251 		}
   1252 	}
   1253 	if (ni == rnode->sysctl_clen)
   1254 		return (ENOENT);
   1255 	node = &node[ni];
   1256 	pnode = node->sysctl_parent;
   1257 
   1258 	/*
   1259 	 * if the kernel says permanent, it is, so there.  nyah.
   1260 	 */
   1261 	if (SYSCTL_FLAGS(node->sysctl_flags) & CTLFLAG_PERMANENT)
   1262 		return (EPERM);
   1263 
   1264 	/*
   1265 	 * can't delete non-empty nodes
   1266 	 */
   1267 	if (SYSCTL_TYPE(node->sysctl_flags) == CTLTYPE_NODE &&
   1268 	    node->sysctl_clen != 0)
   1269 		return (ENOTEMPTY);
   1270 
   1271 	/*
   1272 	 * if the node "owns" data, release it now
   1273 	 */
   1274 	if (node->sysctl_flags & CTLFLAG_OWNDATA) {
   1275 		if (node->sysctl_data != NULL)
   1276 			FREE(node->sysctl_data, M_SYSCTLDATA);
   1277 		node->sysctl_data = NULL;
   1278 	}
   1279 
   1280 	/*
   1281 	 * if the node to be removed is not the last one on the list,
   1282 	 * move the remaining nodes up, and reparent any grandchildren
   1283 	 */
   1284 	onode = *node;
   1285 	if (ni < pnode->sysctl_clen - 1) {
   1286 		int t;
   1287 
   1288 		memmove(&pnode->sysctl_child[ni], &pnode->sysctl_child[ni + 1],
   1289 			(pnode->sysctl_clen - ni - 1) *
   1290 			sizeof(struct sysctlnode));
   1291 		for (; ni < pnode->sysctl_clen - 1; ni++)
   1292 			if (SYSCTL_TYPE(pnode->sysctl_child[ni].sysctl_flags) ==
   1293 			    CTLTYPE_NODE)
   1294 				for (t = 0;
   1295 				     t < pnode->sysctl_child[ni].sysctl_clen;
   1296 				     t++)
   1297 					pnode->sysctl_child[ni].sysctl_child[t].
   1298 						sysctl_parent =
   1299 						&pnode->sysctl_child[ni];
   1300 		ni = pnode->sysctl_clen - 1;
   1301 		node = &pnode->sysctl_child[ni];
   1302 	}
   1303 
   1304 	/*
   1305 	 * reset the space we just vacated
   1306 	 */
   1307 	memset(node, 0, sizeof(struct sysctlnode));
   1308 	node->sysctl_parent = pnode;
   1309 	pnode->sysctl_clen--;
   1310 
   1311 	/*
   1312 	 * if this parent just lost its last child, nuke the creche
   1313 	 */
   1314 	if (pnode->sysctl_clen == 0) {
   1315 		FREE(pnode->sysctl_child, M_SYSCTLNODE);
   1316 		pnode->sysctl_csize = 0;
   1317 		pnode->sysctl_child = NULL;
   1318 	}
   1319 
   1320 	/*
   1321 	 * update "version" on path to "root"
   1322 	 */
   1323         for (; rnode->sysctl_parent != NULL; rnode = rnode->sysctl_parent)
   1324                 ;
   1325 	for (ni = rnode->sysctl_ver + 1; pnode != NULL;
   1326 	     pnode = pnode->sysctl_parent)
   1327 		pnode->sysctl_ver = ni;
   1328 
   1329 	if (oldp != NULL)
   1330 		error = sysctl_copyout(l, &onode, oldp,
   1331 		    MIN(*oldlenp, sizeof(struct sysctlnode)));
   1332 	*oldlenp = sizeof(struct sysctlnode);
   1333 
   1334 	return (error);
   1335 }
   1336 
   1337 /*
   1338  * sysctl_lookup -- Handles copyin/copyout of new and old values.
   1339  * Partial reads are globally allowed.  Only root can write to things
   1340  * unless the node says otherwise.
   1341  */
   1342 int
   1343 sysctl_lookup(SYSCTLFN_RWARGS)
   1344 {
   1345 	int error, rw;
   1346 	size_t sz, len;
   1347 	void *d;
   1348 
   1349 	if (SYSCTL_VERS(rnode->sysctl_flags) != SYSCTL_VERSION) {
   1350 		printf("sysctl_lookup: rnode %p wrong version\n", rnode);
   1351 		return (EINVAL);
   1352 	}
   1353 
   1354 	error = 0;
   1355 
   1356 	/*
   1357 	 * you can't "look up" a node.  you can "query" it, but you
   1358 	 * can't "look it up".
   1359 	 */
   1360 	if (SYSCTL_TYPE(rnode->sysctl_flags) == CTLTYPE_NODE || namelen != 0)
   1361 		return (EINVAL);
   1362 
   1363 	/*
   1364 	 * some nodes are private, so only root can look into them.
   1365 	 */
   1366 	if (l != NULL && (rnode->sysctl_flags & CTLFLAG_PRIVATE) &&
   1367 	    (error = suser(l->l_proc->p_ucred, &l->l_proc->p_acflag)) != 0)
   1368 		return (error);
   1369 
   1370 	/*
   1371 	 * if a node wants to be writable according to different rules
   1372 	 * other than "only root can write to stuff unless a flag is
   1373 	 * set", then it needs its own function which should have been
   1374 	 * called and not us.
   1375 	 */
   1376 	if (l != NULL && newp != NULL &&
   1377 	    !(rnode->sysctl_flags & CTLFLAG_ANYWRITE) &&
   1378 	    (error = suser(l->l_proc->p_ucred, &l->l_proc->p_acflag)) != 0)
   1379 		return (error);
   1380 
   1381 	/*
   1382 	 * is this node supposedly writable?
   1383 	 */
   1384 	rw = 0;
   1385 	switch (rnode->sysctl_flags & CTLFLAG_READWRITE) {
   1386 	    case CTLFLAG_READONLY1:
   1387 		rw = (securelevel < 1) ? 1 : 0;
   1388 		break;
   1389 	    case CTLFLAG_READONLY2:
   1390 		rw = (securelevel < 2) ? 1 : 0;
   1391 		break;
   1392 	    case CTLFLAG_READWRITE:
   1393 		rw = 1;
   1394 		break;
   1395 	}
   1396 
   1397 	/*
   1398 	 * it appears not to be writable at this time, so if someone
   1399 	 * tried to write to it, we must tell them to go away
   1400 	 */
   1401 	if (!rw && newp != NULL)
   1402 		return (EPERM);
   1403 
   1404 	/*
   1405 	 * step one, copy out the stuff we have presently
   1406 	 */
   1407 	if (rnode->sysctl_flags & CTLFLAG_IMMEDIATE) {
   1408 		switch (SYSCTL_TYPE(rnode->sysctl_flags)) {
   1409 		case CTLTYPE_INT:
   1410 			d = &rnode->sysctl_idata;
   1411 			break;
   1412 		case CTLTYPE_QUAD:
   1413 			d = &rnode->sysctl_qdata;
   1414 			break;
   1415 		default:
   1416 			return (EINVAL);
   1417 		}
   1418 	}
   1419 	else
   1420 		d = rnode->sysctl_data;
   1421 	if (SYSCTL_TYPE(rnode->sysctl_flags) == CTLTYPE_STRING)
   1422 		sz = strlen(d) + 1; /* XXX@@@ possible fault here */
   1423 	else
   1424 		sz = rnode->sysctl_size;
   1425 	if (oldp != NULL)
   1426 		error = sysctl_copyout(l, d, oldp, MIN(sz, *oldlenp));
   1427 	if (error)
   1428 		return (error);
   1429 	*oldlenp = sz;
   1430 
   1431 	/*
   1432 	 * are we done?
   1433 	 */
   1434 	if (newp == NULL || newlen == 0)
   1435 		return (0);
   1436 
   1437 	/*
   1438 	 * hmm...not done.  must now "copy in" new value.  re-adjust
   1439 	 * sz to maximum value (strings are "weird").
   1440 	 */
   1441 	sz = rnode->sysctl_size;
   1442 	switch (SYSCTL_TYPE(rnode->sysctl_flags)) {
   1443 	case CTLTYPE_INT:
   1444 	case CTLTYPE_QUAD:
   1445 	case CTLTYPE_STRUCT:
   1446 		/*
   1447 		 * these data must be *exactly* the same size coming
   1448 		 * in.
   1449 		 */
   1450 		if (newlen != sz)
   1451 			return (EINVAL);
   1452 		error = sysctl_copyin(l, newp, d, sz);
   1453 		break;
   1454 	case CTLTYPE_STRING: {
   1455 		/*
   1456 		 * strings, on the other hand, can be shorter, and we
   1457 		 * let userland be sloppy about the trailing nul.
   1458 		 */
   1459 		char *newbuf;
   1460 
   1461 		/*
   1462 		 * too much new string?
   1463 		 */
   1464 		if (newlen > sz)
   1465 			return (EINVAL);
   1466 
   1467 		/*
   1468 		 * temporary copy of new inbound string
   1469 		 */
   1470 		len = MIN(sz, newlen);
   1471 		newbuf = malloc(len, M_SYSCTLDATA, M_WAITOK|M_CANFAIL);
   1472 		if (newbuf == NULL)
   1473 			return (ENOMEM);
   1474 		error = sysctl_copyin(l, newp, newbuf, len);
   1475 		if (error) {
   1476 			FREE(newbuf, M_SYSCTLDATA);
   1477 			return (error);
   1478 		}
   1479 
   1480 		/*
   1481 		 * did they null terminate it, or do we have space
   1482 		 * left to do it ourselves?
   1483 		 */
   1484 		if (newbuf[len - 1] != '\0' && len == sz) {
   1485 			FREE(newbuf, M_SYSCTLDATA);
   1486 			return (EINVAL);
   1487 		}
   1488 
   1489 		/*
   1490 		 * looks good, so pop it into place and zero the rest.
   1491 		 */
   1492 		if (len > 0)
   1493 			memcpy(rnode->sysctl_data, newbuf, len);
   1494 		if (sz != len)
   1495 			memset((char*)rnode->sysctl_data + len, 0, sz - len);
   1496 		FREE(newbuf, M_SYSCTLDATA);
   1497 		break;
   1498 	}
   1499 	default:
   1500 		return (EINVAL);
   1501 	}
   1502 
   1503 	return (error);
   1504 }
   1505 
   1506 /*
   1507  * sysctl_mmap -- Dispatches sysctl mmap requests to those nodes that
   1508  * purport to handle it.  This interface isn't fully fleshed out yet,
   1509  * unfortunately.
   1510  */
   1511 static int
   1512 sysctl_mmap(SYSCTLFN_RWARGS)
   1513 {
   1514 	struct sysctlnode nnode, *node;
   1515 	int error;
   1516 
   1517 	if (SYSCTL_VERS(rnode->sysctl_flags) != SYSCTL_VERSION) {
   1518 		printf("sysctl_mmap: rnode %p wrong version\n", rnode);
   1519 		return (EINVAL);
   1520 	}
   1521 
   1522 	/*
   1523 	 * let's just pretend that didn't happen, m'kay?
   1524 	 */
   1525 	if (l == NULL)
   1526 		return (EPERM);
   1527 
   1528 	/*
   1529 	 * is this a sysctlnode description of an mmap request?
   1530 	 */
   1531 	if (newp == NULL || newlen != sizeof(struct sysctlnode))
   1532 		return (EINVAL);
   1533 	error = sysctl_copyin(l, newp, &nnode, sizeof(nnode));
   1534 	if (error)
   1535 		return (error);
   1536 
   1537 	/*
   1538 	 * does the node they asked for exist?
   1539 	 */
   1540 	if (namelen != 1)
   1541 		return (EOPNOTSUPP);
   1542 	node = rnode;
   1543         error = sysctl_locate(l, &nnode.sysctl_num, 1, &node, NULL);
   1544 	if (error)
   1545 		return (error);
   1546 
   1547 	/*
   1548 	 * does this node that we have found purport to handle mmap?
   1549 	 */
   1550 	if (node->sysctl_func == NULL ||
   1551 	    !(node->sysctl_flags & CTLFLAG_MMAP))
   1552 		return (EOPNOTSUPP);
   1553 
   1554 	/*
   1555 	 * well...okay, they asked for it.
   1556 	 */
   1557 	return ((*node->sysctl_func)(SYSCTLFN_CALL(node)));
   1558 }
   1559 
   1560 /*
   1561  * ********************************************************************
   1562  * Section 3: Create and destroy from inside the kernel
   1563  * ********************************************************************
   1564  * sysctl_createv() and sysctl_destroyv() are simpler-to-use
   1565  * interfaces for the kernel to fling new entries into the mib and rip
   1566  * them out later.  In the case of sysctl_createv(), the returned copy
   1567  * of the node (see sysctl_create()) will be translated back into a
   1568  * pointer to the actual node.
   1569  *
   1570  * Note that sysctl_createv() will return 0 if the create request
   1571  * matches an existing node (ala mkdir -p), and that sysctl_destroyv()
   1572  * will return 0 if the node to be destroyed already does not exist
   1573  * (aka rm -f) or if it is a parent of other nodes.
   1574  *
   1575  * This allows two (or more) different subsystems to assert sub-tree
   1576  * existence before populating their own nodes, and to remove their
   1577  * own nodes without orphaning the others when they are done.
   1578  * ********************************************************************
   1579  */
   1580 int
   1581 sysctl_createv(struct sysctllog **log, int cflags,
   1582 	       struct sysctlnode **rnode, struct sysctlnode **cnode,
   1583 	       int flags, int type, const char *namep, const char *descr,
   1584 	       sysctlfn func, u_quad_t qv, void *newp, size_t newlen,
   1585 	       ...)
   1586 {
   1587 	va_list ap;
   1588 	int error, ni, namelen, name[CTL_MAXNAME];
   1589 	struct sysctlnode *pnode, nnode, onode;
   1590 	size_t sz;
   1591 
   1592 	if (cnode != NULL)
   1593 		*cnode = NULL;
   1594 	if (cflags != 0)
   1595 		return (EINVAL);
   1596 
   1597 	/*
   1598 	 * where are we putting this?
   1599 	 */
   1600 	if (rnode != NULL && *rnode == NULL) {
   1601 		printf("sysctl_createv: rnode NULL\n");
   1602 		return (EINVAL);
   1603 	}
   1604 
   1605 	/*
   1606 	 * what is it?
   1607 	 */
   1608 	flags = SYSCTL_VERSION|SYSCTL_TYPE(type)|SYSCTL_FLAGS(flags);
   1609 
   1610 	/*
   1611 	 * where do we put it?
   1612 	 */
   1613 	va_start(ap, newlen);
   1614 	namelen = 0;
   1615 	ni = -1;
   1616 	do {
   1617 		if (++ni == CTL_MAXNAME)
   1618 			return (ENAMETOOLONG);
   1619 		name[ni] = va_arg(ap, int);
   1620 		/*
   1621 		 * sorry, this is not supported from here
   1622 		 */
   1623 		if (name[ni] == CTL_CREATESYM)
   1624 			return (EINVAL);
   1625 	} while (name[ni] != CTL_EOL && name[ni] != CTL_CREATE);
   1626 	namelen = ni + (name[ni] == CTL_CREATE ? 1 : 0);
   1627 	va_end(ap);
   1628 
   1629 	/*
   1630 	 * what's it called
   1631 	 */
   1632 	if (strlcpy(nnode.sysctl_name, namep, sizeof(nnode.sysctl_name)) >
   1633 	    sizeof(nnode.sysctl_name))
   1634 		return (ENAMETOOLONG);
   1635 
   1636 	/*
   1637 	 * cons up the description of the new node
   1638 	 */
   1639 	nnode.sysctl_num = name[namelen - 1];
   1640 	name[namelen - 1] = CTL_CREATE;
   1641 	nnode.sysctl_size = newlen;
   1642 	nnode.sysctl_flags = flags;
   1643 	if (type == CTLTYPE_NODE) {
   1644 		nnode.sysctl_csize = 0;
   1645 		nnode.sysctl_clen = 0;
   1646 		nnode.sysctl_child = NULL;
   1647 		if (flags & CTLFLAG_ALIAS)
   1648 			nnode.sysctl_alias = qv;
   1649 	}
   1650 	else if (flags & CTLFLAG_IMMEDIATE) {
   1651 		switch (type) {
   1652 		case CTLTYPE_INT:
   1653 			nnode.sysctl_idata = qv;
   1654 			break;
   1655 		case CTLTYPE_QUAD:
   1656 			nnode.sysctl_qdata = qv;
   1657 			break;
   1658 		default:
   1659 			return (EINVAL);
   1660 		}
   1661 	}
   1662 	else {
   1663 		nnode.sysctl_data = newp;
   1664 	}
   1665 	nnode.sysctl_func = func;
   1666 	nnode.sysctl_parent = NULL;
   1667 	nnode.sysctl_ver = 0;
   1668 
   1669 	/*
   1670 	 * initialize lock state -- we need locks if the main tree has
   1671 	 * been marked as complete, but since we could be called from
   1672 	 * either there, or from a device driver (say, at device
   1673 	 * insertion), or from an lkm (at lkm load time, say), we
   1674 	 * don't really want to "wait"...
   1675 	 */
   1676 	error = sysctl_lock(NULL, NULL, 0);
   1677 	if (error)
   1678 		return (error);
   1679 
   1680 	/*
   1681 	 * locate the prospective parent of the new node, and if we
   1682 	 * find it, add the new node.
   1683 	 */
   1684 	sz = sizeof(onode);
   1685 	pnode = rnode ? *rnode : NULL;
   1686 	error = sysctl_locate(NULL, &name[0], namelen - 1, &pnode, &ni);
   1687 	if (error) {
   1688 		sysctl_unlock(NULL);
   1689 		return (error);
   1690 	}
   1691 	error = sysctl_create(&name[ni], namelen - ni, &onode, &sz,
   1692 			      &nnode, sizeof(nnode), &name[0], NULL,
   1693 			      pnode);
   1694 
   1695 	/*
   1696 	 * unfortunately the node we wanted to create is already
   1697 	 * there.  if the node that's already there is a reasonable
   1698 	 * facsimile of the node we wanted to create, just pretend
   1699 	 * (for the caller's benefit) that we managed to create the
   1700 	 * node they wanted.
   1701 	 */
   1702 	if (error == EEXIST) {
   1703 		/* name is the same as requested... */
   1704 		if (strcmp(nnode.sysctl_name, onode.sysctl_name) == 0 &&
   1705 		    /* they want the same function... */
   1706 		    nnode.sysctl_func == onode.sysctl_func &&
   1707 		    /* number is the same as requested, or... */
   1708 		    (nnode.sysctl_num == onode.sysctl_num ||
   1709 		     /* they didn't pick a number... */
   1710 		     nnode.sysctl_num == CTL_CREATE)) {
   1711 			/*
   1712 			 * collision here from trying to create
   1713 			 * something that already existed; let's give
   1714 			 * our customers a hand and tell them they got
   1715 			 * what they wanted.
   1716 			 */
   1717 #ifdef SYSCTL_DEBUG_CREATE
   1718 			printf("cleared\n");
   1719 #endif /* SYSCTL_DEBUG_CREATE */
   1720 			error = 0;
   1721 		}
   1722 	}
   1723 
   1724 	if (error == 0 && cnode != NULL) {
   1725 		/*
   1726 		 * sysctl_create() gave us back a copy of the node,
   1727 		 * but we need to know where it actually is...
   1728 		 */
   1729 		pnode = rnode ? *rnode : NULL;
   1730 		error = sysctl_locate(NULL, &name[0], namelen - 1, &pnode, &ni);
   1731 
   1732 		/*
   1733 		 * manual scan of last layer so that aliased nodes
   1734 		 * aren't followed.
   1735 		 */
   1736 		if (error == 0) {
   1737 			for (ni = 0; ni < pnode->sysctl_clen; ni++)
   1738 				if (pnode->sysctl_child[ni].sysctl_num ==
   1739 				    onode.sysctl_num)
   1740 					break;
   1741 			if (ni < pnode->sysctl_clen)
   1742 				pnode = &pnode->sysctl_child[ni];
   1743 			else
   1744 				error = ENOENT;
   1745 		}
   1746 
   1747 		/*
   1748 		 * not expecting an error here, but...
   1749 		 */
   1750 		if (error == 0) {
   1751 			if (cnode != NULL)
   1752 				*cnode = pnode;
   1753 		}
   1754 		else {
   1755 			printf("sysctl_create succeeded but node not found?!\n");
   1756 			/*
   1757 			 *  confusing, but the create said it
   1758 			 * succeeded, so...
   1759 			 */
   1760 			error = 0;
   1761 		}
   1762 	}
   1763 
   1764 	/*
   1765 	 * now it should be safe to release the lock state.  note that
   1766 	 * the pointer to the newly created node being passed back may
   1767 	 * not be "good" for very long.
   1768 	 */
   1769 	sysctl_unlock(NULL);
   1770 
   1771 	if (error != 0) {
   1772 		printf("sysctl_createv: sysctl_create(%s) returned %d\n",
   1773 		       nnode.sysctl_name, error);
   1774 #if 0
   1775 		if (error != ENOENT)
   1776 			sysctl_dump(&onode);
   1777 #endif
   1778 	}
   1779 
   1780 	return (error);
   1781 }
   1782 
   1783 int
   1784 sysctl_destroyv(struct sysctlnode *rnode, ...)
   1785 {
   1786 	va_list ap;
   1787 	int error, name[CTL_MAXNAME], namelen, ni;
   1788 	struct sysctlnode *pnode, *node, dnode;
   1789 	size_t sz;
   1790 
   1791 	va_start(ap, rnode);
   1792 	namelen = 0;
   1793 	ni = 0;
   1794 	do {
   1795 		if (ni == CTL_MAXNAME)
   1796 			return (ENAMETOOLONG);
   1797 		name[ni] = va_arg(ap, int);
   1798 	} while (name[ni++] != CTL_EOL);
   1799 	namelen = ni - 1;
   1800 	va_end(ap);
   1801 
   1802 	/*
   1803 	 * i can't imagine why we'd be destroying a node when the tree
   1804 	 * wasn't complete, but who knows?
   1805 	 */
   1806 	error = sysctl_lock(NULL, NULL, 0);
   1807 	if (error)
   1808 		return (error);
   1809 
   1810 	/*
   1811 	 * where is it?
   1812 	 */
   1813 	node = rnode;
   1814 	error = sysctl_locate(NULL, &name[0], namelen - 1, &node, &ni);
   1815 	if (error) {
   1816 		/* they want it gone and it's not there, so... */
   1817 		sysctl_unlock(NULL);
   1818 		return (error == ENOENT ? 0 : error);
   1819 	}
   1820 
   1821 	/*
   1822 	 * set up the deletion
   1823 	 */
   1824 	pnode = node;
   1825 	node = &dnode;
   1826 	memset(&dnode, 0, sizeof(dnode));
   1827 	dnode.sysctl_num = name[namelen - 1];
   1828 
   1829 	/*
   1830 	 * we found it, now let's nuke it
   1831 	 */
   1832 	name[namelen - 1] = CTL_DESTROY;
   1833 	sz = 0;
   1834 	error = sysctl_destroy(&name[namelen - 1], 1, NULL, &sz,
   1835 			       node, sizeof(*node), &name[0], NULL,
   1836 			       pnode);
   1837 	if (error == ENOTEMPTY)
   1838 		/*
   1839 		 * think of trying to delete "foo" when "foo.bar"
   1840 		 * (which someone else put there) is still in
   1841 		 * existence
   1842 		 */
   1843 		error = 0;
   1844 
   1845         sysctl_unlock(NULL);
   1846 
   1847 	return (error);
   1848 }
   1849 
   1850 #if 0
   1851 /*
   1852  * ********************************************************************
   1853  * the dump routine.  i haven't yet decided how (if at all) i'll call
   1854  * this from userland when it's in the kernel.
   1855  * ********************************************************************
   1856  */
   1857 static const char *
   1858 sf(int f)
   1859 {
   1860 	static char s[256];
   1861 	char *c;
   1862 
   1863 	s[0] = '\0';
   1864 	c = "";
   1865 
   1866 #define print_flag(_f, _s, _c, _q, _x) \
   1867 	if (((_f) & (__CONCAT(CTLFLAG_,_x))) == (__CONCAT(CTLFLAG_,_q))) { \
   1868 		strlcat((_s), (_c), sizeof(_s)); \
   1869 		strlcat((_s), __STRING(_q), sizeof(_s)); \
   1870 		(_c) = ","; \
   1871 		(_f) &= ~__CONCAT(CTLFLAG_,_x); \
   1872 	}
   1873 
   1874 	print_flag(f, s, c, READONLY,  READWRITE);
   1875 	print_flag(f, s, c, READONLY1, READWRITE);
   1876 	print_flag(f, s, c, READONLY2, READWRITE);
   1877 	print_flag(f, s, c, READWRITE, READWRITE);
   1878 	print_flag(f, s, c, ANYWRITE,  ANYWRITE);
   1879 	print_flag(f, s, c, PRIVATE,   PRIVATE);
   1880 	print_flag(f, s, c, PERMANENT, PERMANENT);
   1881 	print_flag(f, s, c, OWNDATA,   OWNDATA);
   1882 	print_flag(f, s, c, IMMEDIATE, IMMEDIATE);
   1883 	print_flag(f, s, c, HEX,       HEX);
   1884 	print_flag(f, s, c, ROOT,      ROOT);
   1885 	print_flag(f, s, c, ANYNUMBER, ANYNUMBER);
   1886 	print_flag(f, s, c, HIDDEN,    HIDDEN);
   1887 	print_flag(f, s, c, ALIAS,     ALIAS);
   1888 #undef print_flag
   1889 
   1890 	if (f) {
   1891 		char foo[9];
   1892 		snprintf(foo, sizeof(foo), "%x", f);
   1893 		strlcat(s, c, sizeof(s));
   1894 		strlcat(s, foo, sizeof(s));
   1895 	}
   1896 
   1897 	return (s);
   1898 }
   1899 
   1900 static const char *
   1901 st(int t)
   1902 {
   1903 
   1904 	switch (t) {
   1905 	case CTLTYPE_NODE:
   1906 		return "NODE";
   1907 	case CTLTYPE_INT:
   1908 		return "INT";
   1909 	case CTLTYPE_STRING:
   1910 		return "STRING";
   1911 	case CTLTYPE_QUAD:
   1912 		return "QUAD";
   1913 	case CTLTYPE_STRUCT:
   1914 		return "STRUCT";
   1915 	}
   1916 
   1917 	return "???";
   1918 }
   1919 
   1920 void
   1921 sysctl_dump(const struct sysctlnode *d)
   1922 {
   1923 	static char nmib[64], smib[256];
   1924 	static int indent;
   1925 	struct sysctlnode *n;
   1926 	char *np, *sp, tmp[20];
   1927 	int i;
   1928 
   1929 	if (d == NULL)
   1930 		return;
   1931 
   1932 	np = &nmib[strlen(nmib)];
   1933 	sp = &smib[strlen(smib)];
   1934 
   1935 	if (!(d->sysctl_flags & CTLFLAG_ROOT)) {
   1936 		snprintf(tmp, sizeof(tmp), "%d", d->sysctl_num);
   1937 		strcat(nmib, ".");
   1938 		strcat(smib, ".");
   1939 		strcat(nmib, tmp);
   1940 		strcat(smib, d->sysctl_name);
   1941 		printf("%s -> %s (%d)\n", &nmib[1], &smib[1],
   1942 		       SYSCTL_TYPE(d->sysctl_flags));
   1943 	}
   1944 
   1945 	if (1) {
   1946 		printf("%*s%p:\tsysctl_name  [%s]\n", indent, "",
   1947 		       d, d->sysctl_name);
   1948 		printf("%*s\t\tsysctl_num    %d\n",   indent, "",
   1949 		       d->sysctl_num);
   1950 		printf("%*s\t\tsysctl_flags  %x (flags=%x<%s> type=%d<%s> "
   1951 		       "size=%zu)\n",
   1952 		       indent, "", d->sysctl_flags,
   1953 		       SYSCTL_FLAGS(d->sysctl_flags),
   1954 		       sf(SYSCTL_FLAGS(d->sysctl_flags)),
   1955 		       SYSCTL_TYPE(d->sysctl_flags),
   1956 		       st(SYSCTL_TYPE(d->sysctl_flags)),
   1957 		       d->sysctl_size);
   1958 		if (SYSCTL_TYPE(d->sysctl_flags) == CTLTYPE_NODE) {
   1959 			printf("%*s\t\tsysctl_csize  %d\n",   indent, "",
   1960 			       d->sysctl_csize);
   1961 			printf("%*s\t\tsysctl_clen   %d\n",   indent, "",
   1962 			       d->sysctl_clen);
   1963 			printf("%*s\t\tsysctl_child  %p\n",   indent, "",
   1964 			       d->sysctl_child);
   1965 		}
   1966 		else
   1967 			printf("%*s\t\tsysctl_data   %p\n",   indent, "",
   1968 			       d->sysctl_data);
   1969 		printf("%*s\t\tsysctl_func   %p\n",   indent, "",
   1970 		       d->sysctl_func);
   1971 		printf("%*s\t\tsysctl_parent %p\n",   indent, "",
   1972 		       d->sysctl_parent);
   1973 		printf("%*s\t\tsysctl_ver    %d\n",   indent, "",
   1974 		       d->sysctl_ver);
   1975 	}
   1976 
   1977 	if (SYSCTL_TYPE(d->sysctl_flags) == CTLTYPE_NODE) {
   1978 		indent += 8;
   1979 		n = d->sysctl_child;
   1980 		for (i = 0; i < d->sysctl_clen; i++) {
   1981 			sysctl_dump(&n[i]);
   1982 		}
   1983 		indent -= 8;
   1984 	}
   1985 
   1986 	np[0] = '\0';
   1987 	sp[0] = '\0';
   1988 }
   1989 #endif /* 0 */
   1990 
   1991 /*
   1992  * ********************************************************************
   1993  * Deletes an entire n-ary tree.  Not recommended unless you know why
   1994  * you're doing it.  Personally, I don't know why you'd even think
   1995  * about it.
   1996  * ********************************************************************
   1997  */
   1998 void
   1999 sysctl_free(struct sysctlnode *rnode)
   2000 {
   2001 	struct sysctlnode *node, *pnode;
   2002 
   2003 	if (SYSCTL_VERS(rnode->sysctl_flags) != SYSCTL_VERSION) {
   2004 		printf("sysctl_free: rnode %p wrong version\n", rnode);
   2005 		return;
   2006 	}
   2007 
   2008 	if (rnode == NULL)
   2009 		rnode = &sysctl_root;
   2010 	pnode = rnode;
   2011 
   2012 	node = pnode->sysctl_child;
   2013 	do {
   2014 		while (node != NULL && pnode->sysctl_csize > 0) {
   2015 			while (node <
   2016 			       &pnode->sysctl_child[pnode->sysctl_clen] &&
   2017 			       (SYSCTL_TYPE(node->sysctl_flags) !=
   2018 				CTLTYPE_NODE ||
   2019 				node->sysctl_csize == 0)) {
   2020 				if (SYSCTL_FLAGS(node->sysctl_flags) &
   2021 				    CTLFLAG_OWNDATA) {
   2022 					if (node->sysctl_data != NULL) {
   2023 						FREE(node->sysctl_data,
   2024 						     M_SYSCTLDATA);
   2025 						node->sysctl_data = NULL;
   2026 					}
   2027 				}
   2028 				node++;
   2029 			}
   2030 			if (node < &pnode->sysctl_child[pnode->sysctl_clen]) {
   2031 				pnode = node;
   2032 				node = node->sysctl_child;
   2033 			}
   2034 			else
   2035 				break;
   2036 		}
   2037 		if (pnode->sysctl_child != NULL)
   2038 			FREE(pnode->sysctl_child, M_SYSCTLNODE);
   2039 		pnode->sysctl_clen = 0;
   2040 		pnode->sysctl_csize = 0;
   2041 		pnode->sysctl_child = NULL;
   2042 		node = pnode;
   2043 		pnode = node->sysctl_parent;
   2044 	} while (pnode != NULL && node != rnode);
   2045 }
   2046 
   2047 /*
   2048  * ********************************************************************
   2049  * old_sysctl -- A routine to bridge old-style internal calls to the
   2050  * new infrastructure.
   2051  * ********************************************************************
   2052  */
   2053 int
   2054 old_sysctl(int *name, u_int namelen, void *oldp, size_t *oldlenp,
   2055 	   void *newp, size_t newlen, struct lwp *l)
   2056 {
   2057 	int error;
   2058 	size_t savelen = *oldlenp;
   2059 
   2060 	error = sysctl_lock(l, oldp, savelen);
   2061 	if (error)
   2062 		return (error);
   2063 	error = sysctl_dispatch(name, namelen, oldp, oldlenp,
   2064 				newp, newlen, name, l, NULL);
   2065 	sysctl_unlock(l);
   2066 	if (error == 0 && oldp != NULL && savelen < *oldlenp)
   2067 		error = ENOMEM;
   2068 
   2069 	return (error);
   2070 }
   2071 
   2072 /*
   2073  * ********************************************************************
   2074  * Section 4: Generic helper routines
   2075  * ********************************************************************
   2076  * "helper" routines that can do more finely grained access control,
   2077  * construct structures from disparate information, create the
   2078  * appearance of more nodes and sub-trees, etc.  for example, if
   2079  * CTL_PROC wanted a helper function, it could respond to a CTL_QUERY
   2080  * with a dynamically created list of nodes that represented the
   2081  * currently running processes at that instant.
   2082  * ********************************************************************
   2083  */
   2084 
   2085 /*
   2086  * first, a few generic helpers that provide:
   2087  *
   2088  * sysctl_needfunc()		a readonly interface that emits a warning
   2089  * sysctl_notavail()		returns EOPNOTSUPP (generic error)
   2090  * sysctl_null()		an empty return buffer with no error
   2091  */
   2092 int
   2093 sysctl_needfunc(SYSCTLFN_ARGS)
   2094 {
   2095 	int error;
   2096 
   2097 	printf("!!SYSCTL_NEEDFUNC!!\n");
   2098 
   2099 	if (newp != NULL || namelen != 0)
   2100 		return (EOPNOTSUPP);
   2101 
   2102 	error = 0;
   2103 	if (oldp != NULL)
   2104 		error = sysctl_copyout(l, rnode->sysctl_data, oldp,
   2105 				       MIN(rnode->sysctl_size, *oldlenp));
   2106 	*oldlenp = rnode->sysctl_size;
   2107 
   2108 	return (error);
   2109 }
   2110 
   2111 int
   2112 sysctl_notavail(SYSCTLFN_ARGS)
   2113 {
   2114 
   2115 	if (namelen == 1 && name[0] == CTL_QUERY)
   2116 		return (sysctl_query(SYSCTLFN_CALL(rnode)));
   2117 
   2118 	return (EOPNOTSUPP);
   2119 }
   2120 
   2121 int
   2122 sysctl_null(SYSCTLFN_ARGS)
   2123 {
   2124 
   2125 	*oldlenp = 0;
   2126 
   2127 	return (0);
   2128 }
   2129 
   2130 /*
   2131  * ********************************************************************
   2132  * Section 5: The machinery that makes it all go
   2133  * ********************************************************************
   2134  * Memory "manglement" routines.  Not much to this, eh?
   2135  * ********************************************************************
   2136  */
   2137 static int
   2138 sysctl_alloc(struct sysctlnode *p, int x)
   2139 {
   2140 	int i;
   2141 	struct sysctlnode *n;
   2142 
   2143 	assert(p->sysctl_child == NULL);
   2144 
   2145 	if (x == 1)
   2146 		MALLOC(n, struct sysctlnode *,
   2147 		       sizeof(struct sysctlnode),
   2148 		       M_SYSCTLNODE, M_WAITOK|M_CANFAIL);
   2149 	else
   2150 		MALLOC(n, struct sysctlnode *,
   2151 		       SYSCTL_DEFSIZE * sizeof(struct sysctlnode),
   2152 		       M_SYSCTLNODE, M_WAITOK|M_CANFAIL);
   2153 	if (n == NULL)
   2154 		return (ENOMEM);
   2155 
   2156 	if (x == 1) {
   2157 		memset(n, 0, sizeof(struct sysctlnode));
   2158 		p->sysctl_csize = 1;
   2159 	}
   2160 	else {
   2161 		memset(n, 0, SYSCTL_DEFSIZE * sizeof(struct sysctlnode));
   2162 		p->sysctl_csize = SYSCTL_DEFSIZE;
   2163 	}
   2164 	p->sysctl_clen = 0;
   2165 
   2166 	for (i = 0; i < p->sysctl_csize; i++)
   2167 		n[i].sysctl_parent = p;
   2168 
   2169 	p->sysctl_child = n;
   2170 	return (0);
   2171 }
   2172 
   2173 static int
   2174 sysctl_realloc(struct sysctlnode *p)
   2175 {
   2176 	int i, j;
   2177 	struct sysctlnode *n;
   2178 
   2179 	assert(p->sysctl_csize == p->sysctl_clen);
   2180 
   2181 	/*
   2182 	 * how many do we have...how many should we make?
   2183 	 */
   2184 	i = p->sysctl_clen;
   2185 	n = malloc(2 * i * sizeof(struct sysctlnode), M_SYSCTLNODE,
   2186 		   M_WAITOK|M_CANFAIL);
   2187 	if (n == NULL)
   2188 		return (ENOMEM);
   2189 
   2190 	/*
   2191 	 * move old children over...initialize new children
   2192 	 */
   2193 	memcpy(n, p->sysctl_child, i * sizeof(struct sysctlnode));
   2194 	memset(&n[i], 0, i * sizeof(struct sysctlnode));
   2195 	p->sysctl_csize = 2 * i;
   2196 
   2197 	/*
   2198 	 * reattach moved (and new) children to parent; if a moved
   2199 	 * child node has children, reattach the parent pointers of
   2200 	 * grandchildren
   2201 	 */
   2202         for (i = 0; i < p->sysctl_csize; i++) {
   2203                 n[i].sysctl_parent = p;
   2204 		if (n[i].sysctl_child != NULL) {
   2205 			for (j = 0; j < n[i].sysctl_csize; j++)
   2206 				n[i].sysctl_child[j].sysctl_parent = &n[i];
   2207 		}
   2208 	}
   2209 
   2210 	/*
   2211 	 * get out with the old and in with the new
   2212 	 */
   2213 	FREE(p->sysctl_child, M_SYSCTLNODE);
   2214 	p->sysctl_child = n;
   2215 
   2216 	return (0);
   2217 }
   2218