path: root/sys/kern/kern_sysctl.c

/*	$OpenBSD: kern_sysctl.c,v 1.51 2001/06/24 16:00:47 art Exp $	*/
/*	$NetBSD: kern_sysctl.c,v 1.17 1996/05/20 17:49:05 mrg Exp $	*/

/*-
 * Copyright (c) 1982, 1986, 1989, 1993
 *	The Regents of the University of California.  All rights reserved.
 *
 * This code is derived from software contributed to Berkeley by
 * Mike Karels at Berkeley Software Design, Inc.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. All advertising materials mentioning features or use of this software
 *    must display the following acknowledgement:
 *	This product includes software developed by the University of
 *	California, Berkeley and its contributors.
 * 4. Neither the name of the University nor the names of its contributors
 *    may be used to endorse or promote products derived from this software
 *    without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 *
 *	@(#)kern_sysctl.c	8.4 (Berkeley) 4/14/94
 */

/*
 * sysctl system call.
 */

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/malloc.h>
#include <sys/proc.h>
#include <sys/resourcevar.h>
#include <sys/file.h>
#include <sys/vnode.h>
#include <sys/unistd.h>
#include <sys/buf.h>
#include <sys/ioctl.h>
#include <sys/tty.h>
#include <sys/disklabel.h>
#include <sys/disk.h>
#include <vm/vm.h>
#include <sys/sysctl.h>
#include <sys/msgbuf.h>
#include <sys/dkstat.h>
#include <sys/vmmeter.h>
#include <sys/namei.h>

#if defined(UVM)
#include <uvm/uvm_extern.h>
#endif

#include <sys/mount.h>
#include <sys/syscallargs.h>
#include <dev/rndvar.h>

#ifdef DDB
#include <ddb/db_var.h>
#endif

extern struct forkstat forkstat;
extern struct nchstats nchstats;
extern int nselcoll, fscale;
extern struct disklist_head disklist;
extern fixpt_t ccpu;

int sysctl_diskinit(int, struct proc *);

/*
 * Lock to avoid too many processes vslocking a large amount of memory
 * at the same time.
 */
struct lock sysctl_lock, sysctl_disklock;

#if defined(KMEMSTATS) || defined(DIAGNOSTIC) || defined(FFS_SOFTUPDATES)
struct lock sysctl_kmemlock;
#endif

void
sysctl_init()
{
	lockinit(&sysctl_lock, PLOCK|PCATCH, "sysctl", 0, 0);
	lockinit(&sysctl_disklock, PLOCK|PCATCH, "sysctl_disklock", 0, 0);

#if defined(KMEMSTATS) || defined(DIAGNOSTIC) || defined(FFS_SOFTUPDATES)
	lockinit(&sysctl_kmemlock, PLOCK|PCATCH, "sysctl_kmemlock", 0, 0);
#endif
}

int
sys___sysctl(p, v, retval)
	struct proc *p;
	void *v;
	register_t *retval;
{
	register struct sys___sysctl_args /* {
		syscallarg(int *) name;
		syscallarg(u_int) namelen;
		syscallarg(void *) old;
		syscallarg(size_t *) oldlenp;
		syscallarg(void *) new;
		syscallarg(size_t) newlen;
	} */ *uap = v;
	int error, dolock = 1;
	size_t savelen = 0, oldlen = 0;
	sysctlfn *fn;
	int name[CTL_MAXNAME];

	if (SCARG(uap, new) != NULL &&
	    (error = suser(p->p_ucred, &p->p_acflag)))
		return (error);
	/*
	 * all top-level sysctl names are non-terminal
	 */
	if (SCARG(uap, namelen) > CTL_MAXNAME || SCARG(uap, namelen) < 2)
		return (EINVAL);
	error = copyin(SCARG(uap, name), &name,
		       SCARG(uap, namelen) * sizeof(int));
	if (error)
		return (error);

	switch (name[0]) {
	case CTL_KERN:
		fn = kern_sysctl;
		if (name[2] == KERN_VNODE)	/* XXX */
			dolock = 0;
		break;
	case CTL_HW:
		fn = hw_sysctl;
		break;
	case CTL_VM:
#if defined(UVM)
		fn = uvm_sysctl;
#else
		fn = vm_sysctl;
#endif
		break;
	case CTL_NET:
		fn = net_sysctl;
		break;
	case CTL_FS:
		fn = fs_sysctl;
		break;
	case CTL_VFS:
		fn = vfs_sysctl;
		break;
	case CTL_MACHDEP:
		fn = cpu_sysctl;
		break;
#ifdef DEBUG
	case CTL_DEBUG:
		fn = debug_sysctl;
		break;
#endif
#ifdef DDB
	case CTL_DDB:
		fn = ddb_sysctl;
		break;
#endif
	default:
		return (EOPNOTSUPP);
	}

	if (SCARG(uap, oldlenp) &&
	    (error = copyin(SCARG(uap, oldlenp), &oldlen, sizeof(oldlen))))
		return (error);
	if (SCARG(uap, old) != NULL) {
#if defined(UVM)
		/* XXX - obsolete now that vslock returns an error? */
		if (!uvm_useracc(SCARG(uap, old), oldlen, B_WRITE))
#else
		if (!useracc(SCARG(uap, old), oldlen, B_WRITE))
#endif
			return (EFAULT);
		if ((error = lockmgr(&sysctl_lock, LK_EXCLUSIVE, NULL, p)) != 0)
			return (error);
		if (dolock)
#if defined(UVM)
			if (uvm_vslock(p, SCARG(uap, old), oldlen,
			    VM_PROT_READ|VM_PROT_WRITE) != KERN_SUCCESS) {
				lockmgr(&sysctl_lock, LK_RELEASE, NULL, p);
				return EFAULT;
			}
#else
			vslock(SCARG(uap, old), oldlen);
#endif
		savelen = oldlen;
	}
	error = (*fn)(name + 1, SCARG(uap, namelen) - 1, SCARG(uap, old),
	    &oldlen, SCARG(uap, new), SCARG(uap, newlen), p);
	if (SCARG(uap, old) != NULL) {
		if (dolock)
#if defined(UVM)
			uvm_vsunlock(p, SCARG(uap, old), savelen);
#else
			vsunlock(SCARG(uap, old), savelen);
#endif
		lockmgr(&sysctl_lock, LK_RELEASE, NULL, p);
	}
	if (error)
		return (error);
	if (SCARG(uap, oldlenp))
		error = copyout(&oldlen, SCARG(uap, oldlenp), sizeof(oldlen));
	return (error);
}

/*
 * Attributes stored in the kernel.
 */
char hostname[MAXHOSTNAMELEN];
int hostnamelen;
char domainname[MAXHOSTNAMELEN];
int domainnamelen;
long hostid;
char *disknames = NULL;
struct diskstats *diskstats = NULL;
#ifdef INSECURE
int securelevel = -1;
#else
int securelevel;
#endif

/*
 * kernel related system variables.
 */
int
kern_sysctl(name, namelen, oldp, oldlenp, newp, newlen, p)
	int *name;
	u_int namelen;
	void *oldp;
	size_t *oldlenp;
	void *newp;
	size_t newlen;
	struct proc *p;
{
	int error, level, inthostid;
	extern char ostype[], osrelease[], osversion[], version[];
	extern int somaxconn, sominconn;
	extern int usermount, nosuidcoredump;
	extern long cp_time[CPUSTATES];

	/* all sysctl names at this level are terminal */
	if (namelen != 1 && !(name[0] == KERN_PROC || name[0] == KERN_PROF ||
	    name[0] == KERN_MALLOCSTATS || name[0] == KERN_TTY ||
	    name[0] == KERN_POOL))
		return (ENOTDIR);		/* overloaded */

	switch (name[0]) {
	case KERN_OSTYPE:
		return (sysctl_rdstring(oldp, oldlenp, newp, ostype));
	case KERN_OSRELEASE:
		return (sysctl_rdstring(oldp, oldlenp, newp, osrelease));
	case KERN_OSREV:
		return (sysctl_rdint(oldp, oldlenp, newp, OpenBSD));
	case KERN_OSVERSION:
		return (sysctl_rdstring(oldp, oldlenp, newp, osversion));
	case KERN_VERSION:
		return (sysctl_rdstring(oldp, oldlenp, newp, version));
	case KERN_MAXVNODES:
		return(sysctl_int(oldp, oldlenp, newp, newlen, &desiredvnodes));
	case KERN_MAXPROC:
		return (sysctl_int(oldp, oldlenp, newp, newlen, &maxproc));
	case KERN_MAXFILES:
		return (sysctl_int(oldp, oldlenp, newp, newlen, &maxfiles));
	case KERN_ARGMAX:
		return (sysctl_rdint(oldp, oldlenp, newp, ARG_MAX));
	case KERN_NSELCOLL:
		return (sysctl_rdint(oldp, oldlenp, newp, nselcoll));
	case KERN_SECURELVL:
		level = securelevel;
		if ((error = sysctl_int(oldp, oldlenp, newp, newlen, &level)) ||
		    newp == NULL)
			return (error);
		if ((securelevel > 0 || level < -1)
		    && level < securelevel && p->p_pid != 1)
			return (EPERM);
		securelevel = level;
		return (0);
	case KERN_HOSTNAME:
		error = sysctl_tstring(oldp, oldlenp, newp, newlen,
		    hostname, sizeof(hostname));
		if (newp && !error)
			hostnamelen = newlen;
		return (error);
	case KERN_DOMAINNAME:
		error = sysctl_tstring(oldp, oldlenp, newp, newlen,
		    domainname, sizeof(domainname));
		if (newp && !error)
			domainnamelen = newlen;
		return (error);
	case KERN_HOSTID:
		inthostid = hostid;  /* XXX assumes sizeof long <= sizeof int */
		error =  sysctl_int(oldp, oldlenp, newp, newlen, &inthostid);
		hostid = inthostid;
		return (error);
	case KERN_CLOCKRATE:
		return (sysctl_clockrate(oldp, oldlenp));
	case KERN_BOOTTIME:
		return (sysctl_rdstruct(oldp, oldlenp, newp, &boottime,
		    sizeof(struct timeval)));
	case KERN_VNODE:
		return (sysctl_vnode(oldp, oldlenp, p));
	case KERN_PROC:
		return (sysctl_doproc(name + 1, namelen - 1, oldp, oldlenp));
	case KERN_FILE:
		return (sysctl_file(oldp, oldlenp));
#ifdef GPROF
	case KERN_PROF:
		return (sysctl_doprof(name + 1, namelen - 1, oldp, oldlenp,
		    newp, newlen));
#endif
	case KERN_POSIX1:
		return (sysctl_rdint(oldp, oldlenp, newp, _POSIX_VERSION));
	case KERN_NGROUPS:
		return (sysctl_rdint(oldp, oldlenp, newp, NGROUPS_MAX));
	case KERN_JOB_CONTROL:
		return (sysctl_rdint(oldp, oldlenp, newp, 1));
	case KERN_SAVED_IDS:
#ifdef _POSIX_SAVED_IDS
		return (sysctl_rdint(oldp, oldlenp, newp, 1));
#else
		return (sysctl_rdint(oldp, oldlenp, newp, 0));
#endif
	case KERN_MAXPARTITIONS:
		return (sysctl_rdint(oldp, oldlenp, newp, MAXPARTITIONS));
	case KERN_RAWPARTITION:
		return (sysctl_rdint(oldp, oldlenp, newp, RAW_PART));
	case KERN_NTPTIME:
		return (sysctl_ntptime(oldp, oldlenp));
	case KERN_SOMAXCONN:
		return (sysctl_int(oldp, oldlenp, newp, newlen, &somaxconn));
	case KERN_SOMINCONN:
		return (sysctl_int(oldp, oldlenp, newp, newlen, &sominconn));
	case KERN_USERMOUNT:
		return (sysctl_int(oldp, oldlenp, newp, newlen, &usermount));
	case KERN_RND:
		return (sysctl_rdstruct(oldp, oldlenp, newp, &rndstats,
		    sizeof(rndstats)));
	case KERN_ARND:
		return (sysctl_rdint(oldp, oldlenp, newp, arc4random()));
	case KERN_NOSUIDCOREDUMP:
		return (sysctl_int(oldp, oldlenp, newp, newlen, &nosuidcoredump));
	case KERN_FSYNC:
		return (sysctl_rdint(oldp, oldlenp, newp, 1));
	case KERN_SYSVMSG:
#ifdef SYSVMSG
		return (sysctl_rdint(oldp, oldlenp, newp, 1));
#else
		return (sysctl_rdint(oldp, oldlenp, newp, 0));
#endif
	case KERN_SYSVSEM:
#ifdef SYSVSEM
		return (sysctl_rdint(oldp, oldlenp, newp, 1));
#else
		return (sysctl_rdint(oldp, oldlenp, newp, 0));
#endif
	case KERN_SYSVSHM:
#ifdef SYSVSHM
		return (sysctl_rdint(oldp, oldlenp, newp, 1));
#else
		return (sysctl_rdint(oldp, oldlenp, newp, 0));
#endif
	case KERN_MSGBUFSIZE:
		/*
		 * deal with cases where the message buffer has
		 * become corrupted.
		 */
		if (!msgbufp || msgbufp->msg_magic != MSG_MAGIC)
			return (ENXIO);
		return (sysctl_rdint(oldp, oldlenp, newp, msgbufp->msg_bufs));
	case KERN_MSGBUF:
		/* see note above */
		if (!msgbufp || msgbufp->msg_magic != MSG_MAGIC)
			return (ENXIO);
		return (sysctl_rdstruct(oldp, oldlenp, newp, msgbufp,
		    msgbufp->msg_bufs + sizeof(*msgbufp) - 1));
	case KERN_MALLOCSTATS:
		return (sysctl_malloc(name + 1, namelen - 1, oldp, oldlenp,
		    newp, newlen, p));
	case KERN_CPTIME:
		return (sysctl_rdstruct(oldp, oldlenp, newp, &cp_time,
		    sizeof(cp_time)));
	case KERN_NCHSTATS:
		return (sysctl_rdstruct(oldp, oldlenp, newp, &nchstats,
		    sizeof(struct nchstats)));
	case KERN_FORKSTAT:
		return (sysctl_rdstruct(oldp, oldlenp, newp, &forkstat,
		    sizeof(struct forkstat)));
	case KERN_TTY:
		return (sysctl_tty(name + 1, namelen - 1, oldp, oldlenp,
		    newp, newlen));
	case KERN_FSCALE:
		return (sysctl_rdint(oldp, oldlenp, newp, fscale));
	case KERN_CCPU:
		return (sysctl_rdint(oldp, oldlenp, newp, ccpu));
	case KERN_NPROCS:
		return (sysctl_rdint(oldp, oldlenp, newp, nprocs));
	case KERN_POOL:
		return (sysctl_dopool(name + 1, namelen - 1, oldp, oldlenp));
	default:
		return (EOPNOTSUPP);
	}
	/* NOTREACHED */
}

/*
 * hardware related system variables.
 */
int
hw_sysctl(name, namelen, oldp, oldlenp, newp, newlen, p)
	int *name;
	u_int namelen;
	void *oldp;
	size_t *oldlenp;
	void *newp;
	size_t newlen;
	struct proc *p;
{
	extern char machine[], cpu_model[];
	int err;

	/* all sysctl names at this level are terminal */
	if (namelen != 1)
		return (ENOTDIR);		/* overloaded */

	switch (name[0]) {
	case HW_MACHINE:
		return (sysctl_rdstring(oldp, oldlenp, newp, machine));
	case HW_MODEL:
		return (sysctl_rdstring(oldp, oldlenp, newp, cpu_model));
	case HW_NCPU:
		return (sysctl_rdint(oldp, oldlenp, newp, 1));	/* XXX */
	case HW_BYTEORDER:
		return (sysctl_rdint(oldp, oldlenp, newp, BYTE_ORDER));
	case HW_PHYSMEM:
		return (sysctl_rdint(oldp, oldlenp, newp, ctob(physmem)));
	case HW_USERMEM:
#if defined(UVM)
		return (sysctl_rdint(oldp, oldlenp, newp,
		    ctob(physmem - uvmexp.wired)));
#else
		return (sysctl_rdint(oldp, oldlenp, newp,
		    ctob(physmem - cnt.v_wire_count)));
#endif
	case HW_PAGESIZE:
		return (sysctl_rdint(oldp, oldlenp, newp, PAGE_SIZE));
	case HW_DISKNAMES:
		err = sysctl_diskinit(0, p);
		if (err)
			return err;
		return (sysctl_rdstring(oldp, oldlenp, newp, disknames));
	case HW_DISKSTATS:
		err = sysctl_diskinit(1, p);
		if (err)
			return err;
		return (sysctl_rdstruct(oldp, oldlenp, newp, diskstats,
		    disk_count * sizeof(struct diskstats)));
	case HW_DISKCOUNT:
		return (sysctl_rdint(oldp, oldlenp, newp, disk_count));
	default:
		return (EOPNOTSUPP);
	}
	/* NOTREACHED */
}

#ifdef DEBUG
/*
 * Debugging related system variables.
 */
struct ctldebug debug0, debug1, debug2, debug3, debug4;
struct ctldebug debug5, debug6, debug7, debug8, debug9;
struct ctldebug debug10, debug11, debug12, debug13, debug14;
struct ctldebug debug15, debug16, debug17, debug18, debug19;
static struct ctldebug *debugvars[CTL_DEBUG_MAXID] = {
	&debug0, &debug1, &debug2, &debug3, &debug4,
	&debug5, &debug6, &debug7, &debug8, &debug9,
	&debug10, &debug11, &debug12, &debug13, &debug14,
	&debug15, &debug16, &debug17, &debug18, &debug19,
};
int
debug_sysctl(name, namelen, oldp, oldlenp, newp, newlen, p)
	int *name;
	u_int namelen;
	void *oldp;
	size_t *oldlenp;
	void *newp;
	size_t newlen;
	struct proc *p;
{
	struct ctldebug *cdp;

	/* all sysctl names at this level are name and field */
	if (namelen != 2)
		return (ENOTDIR);		/* overloaded */
	cdp = debugvars[name[0]];
	if (cdp->debugname == 0)
		return (EOPNOTSUPP);
	switch (name[1]) {
	case CTL_DEBUG_NAME:
		return (sysctl_rdstring(oldp, oldlenp, newp, cdp->debugname));
	case CTL_DEBUG_VALUE:
		return (sysctl_int(oldp, oldlenp, newp, newlen, cdp->debugvar));
	default:
		return (EOPNOTSUPP);
	}
	/* NOTREACHED */
}
#endif /* DEBUG */

/*
 * Validate parameters and get old / set new parameters
 * for an integer-valued sysctl function.
 */
int
sysctl_int(oldp, oldlenp, newp, newlen, valp)
	void *oldp;
	size_t *oldlenp;
	void *newp;
	size_t newlen;
	int *valp;
{
	int error = 0;

	if (oldp && *oldlenp < sizeof(int))
		return (ENOMEM);
	if (newp && newlen != sizeof(int))
		return (EINVAL);
	*oldlenp = sizeof(int);
	if (oldp)
		error = copyout(valp, oldp, sizeof(int));
	if (error == 0 && newp)
		error = copyin(newp, valp, sizeof(int));
	return (error);
}

/*
 * As above, but read-only.
 */
int
sysctl_rdint(oldp, oldlenp, newp, val)
	void *oldp;
	size_t *oldlenp;
	void *newp;
	int val;
{
	int error = 0;

	if (oldp && *oldlenp < sizeof(int))
		return (ENOMEM);
	if (newp)
		return (EPERM);
	*oldlenp = sizeof(int);
	if (oldp)
		error = copyout((caddr_t)&val, oldp, sizeof(int));
	return (error);
}

/*
 * Validate parameters and get old / set new parameters
 * for an integer-valued sysctl function.
 */
int
sysctl_quad(oldp, oldlenp, newp, newlen, valp)
	void *oldp;
	size_t *oldlenp;
	void *newp;
	size_t newlen;
	int64_t *valp;
{
	int error = 0;

	if (oldp && *oldlenp < sizeof(int64_t))
		return (ENOMEM);
	if (newp && newlen != sizeof(int64_t))
		return (EINVAL);
	*oldlenp = sizeof(int64_t);
	if (oldp)
		error = copyout(valp, oldp, sizeof(int64_t));
	if (error == 0 && newp)
		error = copyin(newp, valp, sizeof(int64_t));
	return (error);
}

/*
 * As above, but read-only.
 */
int
sysctl_rdquad(oldp, oldlenp, newp, val)
	void *oldp;
	size_t *oldlenp;
	void *newp;
	int64_t val;
{
	int error = 0;

	if (oldp && *oldlenp < sizeof(int64_t))
		return (ENOMEM);
	if (newp)
		return (EPERM);
	*oldlenp = sizeof(int64_t);
	if (oldp)
		error = copyout((caddr_t)&val, oldp, sizeof(int64_t));
	return (error);
}

/*
 * Validate parameters and get old / set new parameters
 * for a string-valued sysctl function.
 */
int
sysctl_string(oldp, oldlenp, newp, newlen, str, maxlen)
	void *oldp;
	size_t *oldlenp;
	void *newp;
	size_t newlen;
	char *str;
	int maxlen;
{
	return sysctl__string(oldp, oldlenp, newp, newlen, str, maxlen, 0);
}

int
sysctl_tstring(oldp, oldlenp, newp, newlen, str, maxlen)
	void *oldp;
	size_t *oldlenp;
	void *newp;
	size_t newlen;
	char *str;
	int maxlen;
{
	return sysctl__string(oldp, oldlenp, newp, newlen, str, maxlen, 1);
}

int
sysctl__string(oldp, oldlenp, newp, newlen, str, maxlen, trunc)
	void *oldp;
	size_t *oldlenp;
	void *newp;
	size_t newlen;
	char *str;
	int maxlen;
	int trunc;
{
	int len, error = 0;
	char c;

	len = strlen(str) + 1;
	if (oldp && *oldlenp < len) {
		if (trunc == 0 || *oldlenp == 0)
			return (ENOMEM);
	}
	if (newp && newlen >= maxlen)
		return (EINVAL);
	if (oldp) {
		if (trunc && *oldlenp < len) {
			/* save & zap NUL terminator while copying */
			c = str[*oldlenp-1];
			str[*oldlenp-1] = '\0';
			error = copyout(str, oldp, *oldlenp);
			str[*oldlenp-1] = c;
		} else {
			*oldlenp = len;
			error = copyout(str, oldp, len);
		}
	}
	if (error == 0 && newp) {
		error = copyin(newp, str, newlen);
		str[newlen] = 0;
	}
	return (error);
}

/*
 * As above, but read-only.
 */
int
sysctl_rdstring(oldp, oldlenp, newp, str)
	void *oldp;
	size_t *oldlenp;
	void *newp;
	char *str;
{
	int len, error = 0;

	len = strlen(str) + 1;
	if (oldp && *oldlenp < len)
		return (ENOMEM);
	if (newp)
		return (EPERM);
	*oldlenp = len;
	if (oldp)
		error = copyout(str, oldp, len);
	return (error);
}

/*
 * Validate parameters and get old / set new parameters
 * for a structure oriented sysctl function.
 */
int
sysctl_struct(oldp, oldlenp, newp, newlen, sp, len)
	void *oldp;
	size_t *oldlenp;
	void *newp;
	size_t newlen;
	void *sp;
	int len;
{
	int error = 0;

	if (oldp && *oldlenp < len)
		return (ENOMEM);
	if (newp && newlen > len)
		return (EINVAL);
	if (oldp) {
		*oldlenp = len;
		error = copyout(sp, oldp, len);
	}
	if (error == 0 && newp)
		error = copyin(newp, sp, len);
	return (error);
}

/*
 * Validate parameters and get old parameters
 * for a structure oriented sysctl function.
 */
int
sysctl_rdstruct(oldp, oldlenp, newp, sp, len)
	void *oldp;
	size_t *oldlenp;
	void *newp, *sp;
	int len;
{
	int error = 0;

	if (oldp && *oldlenp < len)
		return (ENOMEM);
	if (newp)
		return (EPERM);
	*oldlenp = len;
	if (oldp)
		error = copyout(sp, oldp, len);
	return (error);
}

/*
 * Get file structures.
 */
int
sysctl_file(where, sizep)
	char *where;
	size_t *sizep;
{
	int buflen, error;
	struct file *fp;
	char *start = where;

	buflen = *sizep;
	if (where == NULL) {
		/*
		 * overestimate by 10 files
		 */
		*sizep = sizeof(filehead) + (nfiles + 10) * sizeof(struct file);
		return (0);
	}

	/*
	 * first copyout filehead
	 */
	if (buflen < sizeof(filehead)) {
		*sizep = 0;
		return (0);
	}
	error = copyout((caddr_t)&filehead, where, sizeof(filehead));
	if (error)
		return (error);
	buflen -= sizeof(filehead);
	where += sizeof(filehead);

	/*
	 * followed by an array of file structures
	 */
	for (fp = filehead.lh_first; fp != 0; fp = fp->f_list.le_next) {
		if (buflen < sizeof(struct file)) {
			*sizep = where - start;
			return (ENOMEM);
		}
		error = copyout((caddr_t)fp, where, sizeof (struct file));
		if (error)
			return (error);
		buflen -= sizeof(struct file);
		where += sizeof(struct file);
	}
	*sizep = where - start;
	return (0);
}

/*
 * try over estimating by 5 procs
 */
#define KERN_PROCSLOP	(5 * sizeof (struct kinfo_proc))

int
sysctl_doproc(name, namelen, where, sizep)
	int *name;
	u_int namelen;
	char *where;
	size_t *sizep;
{
	register struct proc *p;
	register struct kinfo_proc *dp = (struct kinfo_proc *)where;
	register int needed = 0;
	int buflen = where != NULL ? *sizep : 0;
	int doingzomb;
	struct eproc eproc;
	int error = 0;

	if (namelen != 2 && !(namelen == 1 &&
	    (name[0] == KERN_PROC_ALL || name[0] == KERN_PROC_KTHREAD)))
		return (EINVAL);
	p = LIST_FIRST(&allproc);
	doingzomb = 0;
again:
	for (; p != 0; p = LIST_NEXT(p, p_list)) {
		/*
		 * Skip embryonic processes.
		 */
		if (p->p_stat == SIDL)
			continue;
		/*
		 * TODO - make more efficient (see notes below).
		 * do by session.
		 */
		switch (name[0]) {

		case KERN_PROC_PID:
			/* could do this with just a lookup */
			if (p->p_pid != (pid_t)name[1])
				continue;
			break;

		case KERN_PROC_PGRP:
			/* could do this by traversing pgrp */
			if (p->p_pgrp->pg_id != (pid_t)name[1])
				continue;
			break;

		case KERN_PROC_TTY:
			if ((p->p_flag & P_CONTROLT) == 0 ||
			    p->p_session->s_ttyp == NULL ||
			    p->p_session->s_ttyp->t_dev != (dev_t)name[1])
				continue;
			break;

		case KERN_PROC_UID:
			if (p->p_ucred->cr_uid != (uid_t)name[1])
				continue;
			break;

		case KERN_PROC_RUID:
			if (p->p_cred->p_ruid != (uid_t)name[1])
				continue;
			break;

		case KERN_PROC_ALL:
			if (p->p_flag & P_SYSTEM)
				continue;
			break;
		}
		if (buflen >= sizeof(struct kinfo_proc)) {
			fill_eproc(p, &eproc);
			error = copyout((caddr_t)p, &dp->kp_proc,
					sizeof(struct proc));
			if (error)
				return (error);
			error = copyout((caddr_t)&eproc, &dp->kp_eproc,
					sizeof(eproc));
			if (error)
				return (error);
			dp++;
			buflen -= sizeof(struct kinfo_proc);
		}
		needed += sizeof(struct kinfo_proc);
	}
	if (doingzomb == 0) {
		p = LIST_FIRST(&zombproc);
		doingzomb++;
		goto again;
	}
	if (where != NULL) {
		*sizep = (caddr_t)dp - where;
		if (needed > *sizep)
			return (ENOMEM);
	} else {
		needed += KERN_PROCSLOP;
		*sizep = needed;
	}
	return (0);
}

/*
 * Fill in an eproc structure for the specified process.
 */
void
fill_eproc(p, ep)
	register struct proc *p;
	register struct eproc *ep;
{
	register struct tty *tp;

	ep->e_paddr = p;
	ep->e_sess = p->p_pgrp->pg_session;
	ep->e_pcred = *p->p_cred;
	ep->e_ucred = *p->p_ucred;
	if (p->p_stat == SIDL || P_ZOMBIE(p)) {
		ep->e_vm.vm_rssize = 0;
		ep->e_vm.vm_tsize = 0;
		ep->e_vm.vm_dsize = 0;
		ep->e_vm.vm_ssize = 0;
	} else {
		register struct vmspace *vm = p->p_vmspace;

		ep->e_vm.vm_rssize = vm_resident_count(vm);
		ep->e_vm.vm_tsize = vm->vm_tsize;
		ep->e_vm.vm_dsize = vm->vm_dsize;
		ep->e_vm.vm_ssize = vm->vm_ssize;
	}
	if (p->p_pptr)
		ep->e_ppid = p->p_pptr->p_pid;
	else
		ep->e_ppid = 0;
	ep->e_pgid = p->p_pgrp->pg_id;
	ep->e_jobc = p->p_pgrp->pg_jobc;
	if ((p->p_flag & P_CONTROLT) &&
	     (tp = ep->e_sess->s_ttyp)) {
		ep->e_tdev = tp->t_dev;
		ep->e_tpgid = tp->t_pgrp ? tp->t_pgrp->pg_id : NO_PID;
		ep->e_tsess = tp->t_session;
	} else
		ep->e_tdev = NODEV;
	ep->e_flag = ep->e_sess->s_ttyvp ? EPROC_CTTY : 0;
	if (SESS_LEADER(p))
		ep->e_flag |= EPROC_SLEADER;
	strncpy(ep->e_wmesg, p->p_wmesg ? p->p_wmesg : "", WMESGLEN);
	ep->e_wmesg[WMESGLEN] = '\0';
	ep->e_xsize = ep->e_xrssize = 0;
	ep->e_xccount = ep->e_xswrss = 0;
	strncpy(ep->e_login, ep->e_sess->s_login, MAXLOGNAME-1);
	ep->e_login[MAXLOGNAME-1] = '\0';
	strncpy(ep->e_emul, p->p_emul->e_name, EMULNAMELEN);
	ep->e_emul[EMULNAMELEN] = '\0';
	ep->e_maxrss = p->p_rlimit ? p->p_rlimit[RLIMIT_RSS].rlim_cur : 0;
}

/*
 * Initialize disknames/diskstats for export by sysctl. If update is set,
 * then we simply update the disk statistics information.
 */
int
sysctl_diskinit(update, p)
	int update;
	struct proc *p;
{
	struct diskstats *sdk;
	struct disk *dk;
	int i, tlen, l;

	if ((i = lockmgr(&sysctl_disklock, LK_EXCLUSIVE, NULL, p)) != 0)
		return i;

	if (disk_change) {
		for (dk = TAILQ_FIRST(&disklist), tlen = 0; dk;
		    dk = TAILQ_NEXT(dk, dk_link))
			tlen += strlen(dk->dk_name) + 1;

		if (disknames)
			free(disknames, M_SYSCTL);
		if (diskstats)
			free(diskstats, M_SYSCTL);
		diskstats = NULL;
		disknames = NULL;
		diskstats = malloc(disk_count * sizeof(struct diskstats),
		    M_SYSCTL, M_WAITOK);
		disknames = malloc(tlen, M_SYSCTL, M_WAITOK);

		for (dk = TAILQ_FIRST(&disklist), i = 0, l = 0; dk;
		    dk = TAILQ_NEXT(dk, dk_link), i++) {
			l += sprintf(disknames + l, "%s,",
			    dk->dk_name ? dk->dk_name : "");
			sdk = diskstats + i;
			sdk->ds_busy = dk->dk_busy;
			sdk->ds_xfer = dk->dk_xfer;
			sdk->ds_seek = dk->dk_seek;
			sdk->ds_bytes = dk->dk_bytes;
			sdk->ds_attachtime = dk->dk_attachtime;
			sdk->ds_timestamp = dk->dk_timestamp;
			sdk->ds_time = dk->dk_time;
		}

		/* Eliminate trailing comma */
		if (l != 0)
			disknames[l - 1] = '\0';
		disk_change = 0;
	} else if (update) {
		/* Just update, number of drives hasn't changed */
		for (dk = TAILQ_FIRST(&disklist), i = 0; dk;
		    dk = TAILQ_NEXT(dk, dk_link), i++) {
			sdk = diskstats + i;
			sdk->ds_busy = dk->dk_busy;
			sdk->ds_xfer = dk->dk_xfer;
			sdk->ds_seek = dk->dk_seek;
			sdk->ds_bytes = dk->dk_bytes;
			sdk->ds_attachtime = dk->dk_attachtime;
			sdk->ds_timestamp = dk->dk_timestamp;
			sdk->ds_time = dk->dk_time;
		}
	}
	lockmgr(&sysctl_disklock, LK_RELEASE, NULL, p);
	return 0;
}