1/*-
2 * SPDX-License-Identifier: BSD-3-Clause
3 *
4 * Copyright (c) 1980, 1986, 1989, 1993
5 *	The Regents of the University of California.  All rights reserved.
6 * (c) UNIX System Laboratories, Inc.
7 * All or some portions of this file are derived from material licensed
8 * to the University of California by American Telephone and Telegraph
9 * Co. or Unix System Laboratories, Inc. and are reproduced herein with
10 * the permission of UNIX System Laboratories, Inc.
11 *
12 * Redistribution and use in source and binary forms, with or without
13 * modification, are permitted provided that the following conditions
14 * are met:
15 * 1. Redistributions of source code must retain the above copyright
16 *    notice, this list of conditions and the following disclaimer.
17 * 2. Redistributions in binary form must reproduce the above copyright
18 *    notice, this list of conditions and the following disclaimer in the
19 *    documentation and/or other materials provided with the distribution.
20 * 3. Neither the name of the University nor the names of its contributors
21 *    may be used to endorse or promote products derived from this software
22 *    without specific prior written permission.
23 *
24 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
25 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
26 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
27 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
28 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
29 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
30 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
31 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
32 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
33 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34 * SUCH DAMAGE.
35 *
36 *	@(#)param.c	8.3 (Berkeley) 8/20/94
37 */
38
39#include <sys/cdefs.h>
40__FBSDID("$FreeBSD$");
41
42#include "opt_param.h"
43#include "opt_msgbuf.h"
44#include "opt_maxusers.h"
45
46#include <sys/param.h>
47#include <sys/systm.h>
48#include <sys/kernel.h>
49#include <sys/limits.h>
50#include <sys/msgbuf.h>
51#include <sys/sysctl.h>
52#include <sys/proc.h>
53
54#include <vm/vm.h>
55#include <vm/vm_param.h>
56#include <vm/pmap.h>
57
58/*
59 * System parameter formulae.
60 */
61
62#ifndef HZ
63#  if defined(__mips__) || defined(__arm__)
64#    define	HZ 100
65#  else
66#    define	HZ 1000
67#  endif
68#  ifndef HZ_VM
69#    define	HZ_VM 100
70#  endif
71#else
72#  ifndef HZ_VM
73#    define	HZ_VM HZ
74#  endif
75#endif
76#define	NPROC (20 + 16 * maxusers)
77#ifndef NBUF
78#define NBUF 0
79#endif
80#ifndef MAXFILES
81#define	MAXFILES (40 + 32 * maxusers)
82#endif
83
84static int sysctl_kern_vm_guest(SYSCTL_HANDLER_ARGS);
85
86int	hz;				/* system clock's frequency */
87int	tick;				/* usec per tick (1000000 / hz) */
88struct bintime tick_bt;			/* bintime per tick (1s / hz) */
89sbintime_t tick_sbt;
90int	maxusers;			/* base tunable */
91int	maxproc;			/* maximum # of processes */
92int	maxprocperuid;			/* max # of procs per user */
93int	maxfiles;			/* sys. wide open files limit */
94int	maxfilesperproc;		/* per-proc open files limit */
95int	msgbufsize;			/* size of kernel message buffer */
96int	nbuf;
97int	bio_transient_maxcnt;
98int	ngroups_max;			/* max # groups per process */
99int	nswbuf;
100pid_t	pid_max = PID_MAX;
101long	maxswzone;			/* max swmeta KVA storage */
102long	maxbcache;			/* max buffer cache KVA storage */
103long	maxpipekva;			/* Limit on pipe KVA */
104int	vm_guest = VM_GUEST_NO;		/* Running as virtual machine guest? */
105u_long	maxtsiz;			/* max text size */
106u_long	dfldsiz;			/* initial data size limit */
107u_long	maxdsiz;			/* max data size */
108u_long	dflssiz;			/* initial stack size limit */
109u_long	maxssiz;			/* max stack size */
110u_long	sgrowsiz;			/* amount to grow stack */
111
112SYSCTL_INT(_kern, OID_AUTO, hz, CTLFLAG_RDTUN | CTLFLAG_NOFETCH, &hz, 0,
113    "Number of clock ticks per second");
114SYSCTL_INT(_kern, OID_AUTO, nbuf, CTLFLAG_RDTUN | CTLFLAG_NOFETCH, &nbuf, 0,
115    "Number of buffers in the buffer cache");
116SYSCTL_INT(_kern, OID_AUTO, nswbuf, CTLFLAG_RDTUN | CTLFLAG_NOFETCH, &nswbuf, 0,
117    "Number of swap buffers");
118SYSCTL_INT(_kern, OID_AUTO, msgbufsize, CTLFLAG_RDTUN | CTLFLAG_NOFETCH, &msgbufsize, 0,
119    "Size of the kernel message buffer");
120SYSCTL_LONG(_kern, OID_AUTO, maxswzone, CTLFLAG_RDTUN | CTLFLAG_NOFETCH, &maxswzone, 0,
121    "Maximum memory for swap metadata");
122SYSCTL_LONG(_kern, OID_AUTO, maxbcache, CTLFLAG_RDTUN | CTLFLAG_NOFETCH, &maxbcache, 0,
123    "Maximum value of vfs.maxbufspace");
124SYSCTL_INT(_kern, OID_AUTO, bio_transient_maxcnt, CTLFLAG_RDTUN | CTLFLAG_NOFETCH,
125    &bio_transient_maxcnt, 0,
126    "Maximum number of transient BIOs mappings");
127SYSCTL_ULONG(_kern, OID_AUTO, maxtsiz, CTLFLAG_RWTUN | CTLFLAG_NOFETCH, &maxtsiz, 0,
128    "Maximum text size");
129SYSCTL_ULONG(_kern, OID_AUTO, dfldsiz, CTLFLAG_RWTUN | CTLFLAG_NOFETCH, &dfldsiz, 0,
130    "Initial data size limit");
131SYSCTL_ULONG(_kern, OID_AUTO, maxdsiz, CTLFLAG_RWTUN | CTLFLAG_NOFETCH, &maxdsiz, 0,
132    "Maximum data size");
133SYSCTL_ULONG(_kern, OID_AUTO, dflssiz, CTLFLAG_RWTUN | CTLFLAG_NOFETCH, &dflssiz, 0,
134    "Initial stack size limit");
135SYSCTL_ULONG(_kern, OID_AUTO, maxssiz, CTLFLAG_RWTUN | CTLFLAG_NOFETCH, &maxssiz, 0,
136    "Maximum stack size");
137SYSCTL_ULONG(_kern, OID_AUTO, sgrowsiz, CTLFLAG_RWTUN | CTLFLAG_NOFETCH, &sgrowsiz, 0,
138    "Amount to grow stack on a stack fault");
139SYSCTL_PROC(_kern, OID_AUTO, vm_guest, CTLFLAG_RD | CTLTYPE_STRING,
140    NULL, 0, sysctl_kern_vm_guest, "A",
141    "Virtual machine guest detected?");
142
143/*
144 * The elements of this array are ordered based upon the values of the
145 * corresponding enum VM_GUEST members.
146 */
147static const char *const vm_guest_sysctl_names[] = {
148	"none",
149	"generic",
150	"xen",
151	"hv",
152	"vmware",
153	"kvm",
154	"bhyve",
155	NULL
156};
157CTASSERT(nitems(vm_guest_sysctl_names) - 1 == VM_LAST);
158
159/*
160 * Boot time overrides that are not scaled against main memory
161 */
162void
163init_param1(void)
164{
165
166#if !defined(__mips__) && !defined(__arm64__) && !defined(__sparc64__)
167	TUNABLE_INT_FETCH("kern.kstack_pages", &kstack_pages);
168#endif
169	hz = -1;
170	TUNABLE_INT_FETCH("kern.hz", &hz);
171	if (hz == -1)
172		hz = vm_guest > VM_GUEST_NO ? HZ_VM : HZ;
173	tick = 1000000 / hz;
174	tick_sbt = SBT_1S / hz;
175	tick_bt = sbttobt(tick_sbt);
176
177	/*
178	 * Arrange for ticks to wrap 10 minutes after boot to help catch
179	 * sign problems sooner.
180	 */
181	ticks = INT_MAX - (hz * 10 * 60);
182
183#ifdef VM_SWZONE_SIZE_MAX
184	maxswzone = VM_SWZONE_SIZE_MAX;
185#endif
186	TUNABLE_LONG_FETCH("kern.maxswzone", &maxswzone);
187#ifdef VM_BCACHE_SIZE_MAX
188	maxbcache = VM_BCACHE_SIZE_MAX;
189#endif
190	TUNABLE_LONG_FETCH("kern.maxbcache", &maxbcache);
191	msgbufsize = MSGBUF_SIZE;
192	TUNABLE_INT_FETCH("kern.msgbufsize", &msgbufsize);
193
194	maxtsiz = MAXTSIZ;
195	TUNABLE_ULONG_FETCH("kern.maxtsiz", &maxtsiz);
196	dfldsiz = DFLDSIZ;
197	TUNABLE_ULONG_FETCH("kern.dfldsiz", &dfldsiz);
198	maxdsiz = MAXDSIZ;
199	TUNABLE_ULONG_FETCH("kern.maxdsiz", &maxdsiz);
200	dflssiz = DFLSSIZ;
201	TUNABLE_ULONG_FETCH("kern.dflssiz", &dflssiz);
202	maxssiz = MAXSSIZ;
203	TUNABLE_ULONG_FETCH("kern.maxssiz", &maxssiz);
204	sgrowsiz = SGROWSIZ;
205	TUNABLE_ULONG_FETCH("kern.sgrowsiz", &sgrowsiz);
206
207	/*
208	 * Let the administrator set {NGROUPS_MAX}, but disallow values
209	 * less than NGROUPS_MAX which would violate POSIX.1-2008 or
210	 * greater than INT_MAX-1 which would result in overflow.
211	 */
212	ngroups_max = NGROUPS_MAX;
213	TUNABLE_INT_FETCH("kern.ngroups", &ngroups_max);
214	if (ngroups_max < NGROUPS_MAX)
215		ngroups_max = NGROUPS_MAX;
216
217	/*
218	 * Only allow to lower the maximal pid.
219	 * Prevent setting up a non-bootable system if pid_max is too low.
220	 */
221	TUNABLE_INT_FETCH("kern.pid_max", &pid_max);
222	if (pid_max > PID_MAX)
223		pid_max = PID_MAX;
224	else if (pid_max < 300)
225		pid_max = 300;
226
227	TUNABLE_INT_FETCH("vfs.unmapped_buf_allowed", &unmapped_buf_allowed);
228}
229
230/*
231 * Boot time overrides that are scaled against main memory
232 */
233void
234init_param2(long physpages)
235{
236
237	/* Base parameters */
238	maxusers = MAXUSERS;
239	TUNABLE_INT_FETCH("kern.maxusers", &maxusers);
240	if (maxusers == 0) {
241		maxusers = physpages / (2 * 1024 * 1024 / PAGE_SIZE);
242		if (maxusers < 32)
243			maxusers = 32;
244#ifdef VM_MAX_AUTOTUNE_MAXUSERS
245                if (maxusers > VM_MAX_AUTOTUNE_MAXUSERS)
246                        maxusers = VM_MAX_AUTOTUNE_MAXUSERS;
247#endif
248                /*
249                 * Scales down the function in which maxusers grows once
250                 * we hit 384.
251                 */
252                if (maxusers > 384)
253                        maxusers = 384 + ((maxusers - 384) / 8);
254        }
255
256	/*
257	 * The following can be overridden after boot via sysctl.  Note:
258	 * unless overriden, these macros are ultimately based on maxusers.
259	 * Limit maxproc so that kmap entries cannot be exhausted by
260	 * processes.
261	 */
262	maxproc = NPROC;
263	TUNABLE_INT_FETCH("kern.maxproc", &maxproc);
264	if (maxproc > (physpages / 12))
265		maxproc = physpages / 12;
266	if (maxproc > pid_max)
267		maxproc = pid_max;
268	maxprocperuid = (maxproc * 9) / 10;
269
270	/*
271	 * The default limit for maxfiles is 1/12 of the number of
272	 * physical page but not less than 16 times maxusers.
273	 * At most it can be 1/6 the number of physical pages.
274	 */
275	maxfiles = imax(MAXFILES, physpages / 8);
276	TUNABLE_INT_FETCH("kern.maxfiles", &maxfiles);
277	if (maxfiles > (physpages / 4))
278		maxfiles = physpages / 4;
279	maxfilesperproc = (maxfiles / 10) * 9;
280	TUNABLE_INT_FETCH("kern.maxfilesperproc", &maxfilesperproc);
281
282	/*
283	 * Cannot be changed after boot.
284	 */
285	nbuf = NBUF;
286	TUNABLE_INT_FETCH("kern.nbuf", &nbuf);
287	TUNABLE_INT_FETCH("kern.bio_transient_maxcnt", &bio_transient_maxcnt);
288
289	/*
290	 * The default for maxpipekva is min(1/64 of the kernel address space,
291	 * max(1/64 of main memory, 512KB)).  See sys_pipe.c for more details.
292	 */
293	maxpipekva = (physpages / 64) * PAGE_SIZE;
294	TUNABLE_LONG_FETCH("kern.ipc.maxpipekva", &maxpipekva);
295	if (maxpipekva < 512 * 1024)
296		maxpipekva = 512 * 1024;
297	if (maxpipekva > (VM_MAX_KERNEL_ADDRESS - VM_MIN_KERNEL_ADDRESS) / 64)
298		maxpipekva = (VM_MAX_KERNEL_ADDRESS - VM_MIN_KERNEL_ADDRESS) /
299		    64;
300}
301
302/*
303 * Sysctl stringifying handler for kern.vm_guest.
304 */
305static int
306sysctl_kern_vm_guest(SYSCTL_HANDLER_ARGS)
307{
308	return (SYSCTL_OUT_STR(req, vm_guest_sysctl_names[vm_guest]));
309}
310