Aren/NetBSD
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
2e19323388
NetBSD/sys/kern/init_sysctl.c

1603 lines
45 KiB
C
Raw Normal View History

Add a proper defflag for GPROF, and include opt_gprof.h, otherwise we're not gonna go very far.
2018-02-04 20:31:51 +03:00
/* $NetBSD: init_sysctl.c,v 1.214 2018/02/04 17:31:51 maxv Exp $ */
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
/*-
More changes to improve kern_descrip.c. - Avoid atomics in more places. - Remove the per-descriptor mutex, and just use filedesc_t::fd_lock. It was only being used to synchronize close, and in any case we needed to take fd_lock to free the descriptor slot. - Optimize certain paths for the <NDFDFILE case. - Sprinkle more comments and assertions. - Cache more stuff in filedesc_t. - Fix numerous minor bugs spotted along the way. - Restructure how the open files array is maintained, for clarity and so that we can eliminate the membar_consumer() call in fd_getfile(). This is mostly syntactic sugar; the main functional change is that fd_nfiles now lives alongside the open file array. Some measurements with libmicro: - simple file syscalls are like close() are between 1 to 10% faster. - some nice improvements, e.g. poll(1000) which is ~50% faster.
2009-05-25 01:41:25 +04:00
* Copyright (c) 2003, 2007, 2008, 2009 The NetBSD Foundation, Inc.
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
* All rights reserved.
*
* This code is derived from software contributed to The NetBSD Foundation
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
* by Andrew Brown, and by Andrew Doran.
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
* ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
* TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
* BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*/
RCSid police
2003-12-29 01:12:00 +03:00
#include <sys/cdefs.h>
Add a proper defflag for GPROF, and include opt_gprof.h, otherwise we're not gonna go very far.
2018-02-04 20:31:51 +03:00
__KERNEL_RCSID(0, "$NetBSD: init_sysctl.c,v 1.214 2018/02/04 17:31:51 maxv Exp $");
RCSid police
2003-12-29 01:12:00 +03:00
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
#include "opt_sysv.h"
merge christos-time_t
2009-01-11 05:45:45 +03:00
#include "opt_compat_netbsd.h"
Use "defopt MODULAR" in sys/conf/files, and #include "opt_modular.h" in all kernel sources that use the MODULAR option. Proposed in tech-kern on 18 Jan 2009.
2009-02-14 01:41:00 +03:00
#include "opt_modular.h"
Add a proper defflag for GPROF, and include opt_gprof.h, otherwise we're not gonna go very far.
2018-02-04 20:31:51 +03:00
#include "opt_gprof.h"
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
#include "pty.h"
#include <sys/types.h>
#include <sys/param.h>
#include <sys/sysctl.h>
remove a homegrown definition of CPU_INFO_FOREACH.
2007-08-06 15:51:46 +04:00
#include <sys/cpu.h>
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
#include <sys/errno.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/unistd.h>
#include <sys/disklabel.h>
First step of random number subsystem rework described in <20111022023242.BA26F14A158@mail.netbsd.org>. This change includes the following: An initial cleanup and minor reorganization of the entropy pool code in sys/dev/rnd.c and sys/dev/rndpool.c. Several bugs are fixed. Some effort is made to accumulate entropy more quickly at boot time. A generic interface, "rndsink", is added, for stream generators to request that they be re-keyed with good quality entropy from the pool as soon as it is available. The arc4random()/arc4randbytes() implementation in libkern is adjusted to use the rndsink interface for rekeying, which helps address the problem of low-quality keys at boot time. An implementation of the FIPS 140-2 statistical tests for random number generator quality is provided (libkern/rngtest.c). This is based on Greg Rose's implementation from Qualcomm. A new random stream generator, nist_ctr_drbg, is provided. It is based on an implementation of the NIST SP800-90 CTR_DRBG by Henric Jungheim. This generator users AES in a modified counter mode to generate a backtracking-resistant random stream. An abstraction layer, "cprng", is provided for in-kernel consumers of randomness. The arc4random/arc4randbytes API is deprecated for in-kernel use. It is replaced by "cprng_strong". The current cprng_fast implementation wraps the existing arc4random implementation. The current cprng_strong implementation wraps the new CTR_DRBG implementation. Both interfaces are rekeyed from the entropy pool automatically at intervals justifiable from best current cryptographic practice. In some quick tests, cprng_fast() is about the same speed as the old arc4randbytes(), and cprng_strong() is about 20% faster than rnd_extract_data(). Performance is expected to improve. The AES code in src/crypto/rijndael is no longer an optional kernel component, as it is required by cprng_strong, which is not an optional kernel component. The entropy pool output is subjected to the rngtest tests at startup time; if it fails, the system will reboot. There is approximately a 3/10000 chance of a false positive from these tests. Entropy pool _input_ from hardware random numbers is subjected to the rngtest tests at attach time, as well as the FIPS continuous-output test, to detect bad or stuck hardware RNGs; if any are detected, they are detached, but the system continues to run. A problem with rndctl(8) is fixed -- datastructures with pointers in arrays are no longer passed to userspace (this was not a security problem, but rather a major issue for compat32). A new kernel will require a new rndctl. The sysctl kern.arandom() and kern.urandom() nodes are hooked up to the new generators, but the /dev/*random pseudodevices are not, yet. Manual pages for the new kernel interfaces are forthcoming.
2011-11-20 02:51:18 +04:00
#include <sys/cprng.h>
Remove the "target" argment from vfs_drainvnodes() as it is always equal to "desiredvnodes" and move its definition from sys/vnode.h to sys/vnode_impl.h. Extend vfs_drainvnodes() to also wait for deferred vrele to flush and replace the call to vrele_flush() with a call to vfs_drainvnodes().
2016-12-14 18:46:57 +03:00
#include <sys/vnode_impl.h> /* For vfs_drainvnodes(). */
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
#include <sys/mount.h>
#include <sys/namei.h>
#include <dev/cons.h>
#include <sys/socketvar.h>
#include <sys/file.h>
Add kern.file2. As kern.proc2 is to kern.proc, so is kern.file2 to kern.file, namely a 32/64 bit clean sysctl interface to the same data. It also borrows a few things from struct vnode (if applicable) and from struct proc, just to tie things together a bit more. You can walk this list "by file" or "by pid". The former method is similar to kern.file but omits the filehead, and the latter can give you duplicates if multiple processes have the same struct file open, but tells you which process it is.
2005-03-09 08:02:17 +03:00
#include <sys/filedesc.h>
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
#include <sys/tty.h>
malloc -> kmem_alloc.
2009-01-17 10:02:35 +03:00
#include <sys/kmem.h>
Add a new kern.messages sysctl to allow kernel message verbosity to be altered after boot. Fixes PR kern/46539 using patch submitted by Nat Sloss.
2016-05-31 08:44:19 +03:00
#include <sys/reboot.h>
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
#include <sys/resource.h>
#include <sys/resourcevar.h>
#include <sys/exec.h>
#include <sys/conf.h>
#include <sys/device.h>
implement a security.setid_core node as discussed on tech-kern@ and tech-security@.
2006-02-02 20:48:51 +03:00
#include <sys/stat.h>
integrate kauth.
2006-05-15 01:15:11 +04:00
#include <sys/kauth.h>
Add a -t+S flag to ktrace for tracing activity related to sysctl. MIB names will be displayed, with data readen and written as well.
2006-09-24 02:01:04 +04:00
#include <sys/ktrace.h>
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
machine/{bus,cpu,intr}.h -> sys/{bus,cpu,intr}.h
2007-10-19 15:59:34 +04:00
#include <sys/cpu.h>
Include opt_posix.h for the P1003_1B_SEMAPHORE define. Include <machine/cpu.h> just to be sure.
2003-12-06 23:06:11 +03:00
implement a security.setid_core node as discussed on tech-kern@ and tech-security@.
2006-02-02 20:48:51 +03:00
int security_setidcore_dump;
char security_setidcore_path[MAXPATHLEN] = "/var/crash/%n.core";
uid_t security_setidcore_owner = 0;
gid_t security_setidcore_group = 0;
mode_t security_setidcore_mode = (S_IRUSR|S_IWUSR);
Introduce ``security.curtain'', new node for security features and settings, and new variable for controlling access to objects based on user-id.
2005-09-07 20:26:15 +04:00
More preparation for modularizing the SYSVxxx options. Here we change the kern.ipc.sysvxxx sysctls into dynamic values, so each sub-component of SYSVxxx can declare its own availability.
2015-05-13 04:00:16 +03:00
/*
* Current status of SysV IPC capability. Initially, these are
* 0 if the capability is not built-in to the kernel, but can
* be updated if the appropriate kernel module is (auto)loaded.
*/
int kern_has_sysvmsg = 0;
int kern_has_sysvshm = 0;
int kern_has_sysvsem = 0;
Merge newlock2 to head.
2007-02-10 00:55:00 +03:00
static const u_int sysctl_lwpprflagmap[] = {
Change the process/lwp flags seen by userland via sysctl back to the P_*/L_* naming convention, and rename the in-kernel flags to avoid conflict. (P_ -> PK_, L_ -> LW_ ). Add back the (now unused) LSDEAD constant. Restores source compatibility with pre-newlock2 tools like ps or top. Reviewed by Andrew Doran.
2007-02-18 01:31:36 +03:00
LPR_DETACHED, L_DETACHED,
Merge newlock2 to head.
2007-02-10 00:55:00 +03:00
0
};
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
/*
* try over estimating by 5 procs/lwps
*/
#define KERN_LWPSLOP (5 * sizeof(struct kinfo_lwp))
Changes to make ktrace LKM friendly and reduce ifdef KTRACE. Proposed on tech-kern.
2007-08-15 16:07:23 +04:00
static int dcopyout(struct lwp *, const void *, void *, size_t);
Add a -t+S flag to ktrace for tracing activity related to sysctl. MIB names will be displayed, with data readen and written as well.
2006-09-24 02:01:04 +04:00
Changes to make ktrace LKM friendly and reduce ifdef KTRACE. Proposed on tech-kern.
2007-08-15 16:07:23 +04:00
static int
Convert to ansi definitions from old-style definitons.
2008-02-27 22:57:18 +03:00
dcopyout(struct lwp *l, const void *kaddr, void *uaddr, size_t len)
Add a -t+S flag to ktrace for tracing activity related to sysctl. MIB names will be displayed, with data readen and written as well.
2006-09-24 02:01:04 +04:00
{
int error;
error = copyout(kaddr, uaddr, len);
Changes to make ktrace LKM friendly and reduce ifdef KTRACE. Proposed on tech-kern.
2007-08-15 16:07:23 +04:00
ktrmibio(-1, UIO_READ, uaddr, len, error);
Add a -t+S flag to ktrace for tracing activity related to sysctl. MIB names will be displayed, with data readen and written as well.
2006-09-24 02:01:04 +04:00
return error;
}
Count the number of CPUs at boot and stash in 'ncpu'. Eventually should have each CPU register at attach, so we can figure out the topology for the scheduler.
2007-02-15 23:32:47 +03:00
When DIAGNOSTIC is defined, provide a kern.panic_now sysctl to conviniently and reliably panic the system
2006-03-27 00:07:21 +04:00
#ifdef DIAGNOSTIC
static int sysctl_kern_trigger_panic(SYSCTLFN_PROTO);
#endif
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
static int sysctl_kern_maxvnodes(SYSCTLFN_PROTO);
Add a new kern.messages sysctl to allow kernel message verbosity to be altered after boot. Fixes PR kern/46539 using patch submitted by Nat Sloss.
2016-05-31 08:44:19 +03:00
static int sysctl_kern_messages(SYSCTLFN_PROTO);
Make kern.rtc_offset writable at securelevel <= 0. This allows boot-time adjustment when a machine runs other OSes with RTC == localtime.
2003-12-27 02:49:39 +03:00
static int sysctl_kern_rtc_offset(SYSCTLFN_PROTO);
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
static int sysctl_kern_maxproc(SYSCTLFN_PROTO);
static int sysctl_kern_hostid(SYSCTLFN_PROTO);
static int sysctl_kern_defcorename(SYSCTLFN_PROTO);
static int sysctl_kern_cptime(SYSCTLFN_PROTO);
Also make declaration of sysctl_kern_maxptys() depend on NPTY > 0. Makes the mvme68k RAMDISK kernel compile again.
2003-12-07 13:31:32 +03:00
#if NPTY > 0
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
static int sysctl_kern_maxptys(SYSCTLFN_PROTO);
Also make declaration of sysctl_kern_maxptys() depend on NPTY > 0. Makes the mvme68k RAMDISK kernel compile again.
2003-12-07 13:31:32 +03:00
#endif /* NPTY > 0 */
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
static int sysctl_kern_lwp(SYSCTLFN_PROTO);
static int sysctl_kern_forkfsleep(SYSCTLFN_PROTO);
static int sysctl_kern_root_partition(SYSCTLFN_PROTO);
static int sysctl_kern_drivers(SYSCTLFN_PROTO);
implement a security.setid_core node as discussed on tech-kern@ and tech-security@.
2006-02-02 20:48:51 +03:00
static int sysctl_security_setidcore(SYSCTLFN_PROTO);
static int sysctl_security_setidcorename(SYSCTLFN_PROTO);
Add a new sysctl 'cp_id' that returns the array of cpu id values. Requested by me, implemented by atatat.
2005-06-16 18:55:58 +04:00
static int sysctl_kern_cpid(SYSCTLFN_PROTO);
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
static int sysctl_hw_usermem(SYSCTLFN_PROTO);
static int sysctl_hw_cnmagic(SYSCTLFN_PROTO);
static void fill_lwp(struct lwp *l, struct kinfo_lwp *kl);
/*
* ********************************************************************
* section 1: setup routines
* ********************************************************************
- fix comment sentence capitalization. - whitespace cleanup. No functional changes.
2007-10-16 20:05:50 +04:00
* These functions are stuffed into a link set for sysctl setup
* functions. They're never called or referenced from anywhere else.
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
* ********************************************************************
*/
/*
* this setup routine is a replacement for kern_sysctl()
*/
SYSCTL_SETUP(sysctl_kern_setup, "sysctl kern subtree setup")
{
extern int kern_logsigexit; /* defined in kern/kern_sig.c */
extern fixpt_t ccpu; /* defined in kern/kern_synch.c */
extern int dumponpanic; /* defined in kern/subr_prf.c */
move security.setid_core.* to kern.coredump.setid.*, as requested by yamt@.
2006-07-15 01:55:19 +04:00
const struct sysctlnode *rnode;
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "maxvnodes",
SYSCTL_DESCR("Maximum number of vnodes"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
sysctl_kern_maxvnodes, 0, NULL, 0,
CTL_KERN, KERN_MAXVNODES, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "maxproc",
SYSCTL_DESCR("Maximum number of simultaneous processes"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
sysctl_kern_maxproc, 0, NULL, 0,
CTL_KERN, KERN_MAXPROC, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "maxfiles",
SYSCTL_DESCR("Maximum number of open files"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
NULL, 0, &maxfiles, 0,
CTL_KERN, KERN_MAXFILES, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "argmax",
SYSCTL_DESCR("Maximum number of bytes of arguments to "
"execve(2)"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
NULL, ARG_MAX, NULL, 0,
CTL_KERN, KERN_ARGMAX, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
Prefer that kern.hostid is printed in hex, not as a signed decimal, and avoid accidental sign-extension when setting it.
2004-04-16 17:25:40 +04:00
CTLFLAG_PERMANENT|CTLFLAG_READWRITE|CTLFLAG_HEX,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "hostid",
SYSCTL_DESCR("System host ID number"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
sysctl_kern_hostid, 0, NULL, 0,
CTL_KERN, KERN_HOSTID, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_STRUCT, "vnode",
SYSCTL_DESCR("System vnode table"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
sysctl_kern_vnode, 0, NULL, 0,
CTL_KERN, KERN_VNODE, CTL_EOL);
#ifndef GPROF
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_NODE, "profiling",
SYSCTL_DESCR("Profiling information (not available)"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
sysctl_notavail, 0, NULL, 0,
CTL_KERN, KERN_PROF, CTL_EOL);
#endif
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "posix1version",
SYSCTL_DESCR("Version of ISO/IEC 9945 (POSIX 1003.1) "
"with which the operating system attempts "
"to comply"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
NULL, _POSIX_VERSION, NULL, 0,
CTL_KERN, KERN_POSIX1, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "ngroups",
SYSCTL_DESCR("Maximum number of supplemental groups"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
NULL, NGROUPS_MAX, NULL, 0,
CTL_KERN, KERN_NGROUPS, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "job_control",
SYSCTL_DESCR("Whether job control is available"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
NULL, 1, NULL, 0,
CTL_KERN, KERN_JOB_CONTROL, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "saved_ids",
be explicit in the description for POSIX saved set-id that this is for POSIX-style, not sane-style. (ie, add "POSIX " to the description.)
2005-04-18 14:46:39 +04:00
SYSCTL_DESCR("Whether POSIX saved set-group/user ID is "
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
"available"), NULL,
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
#ifdef _POSIX_SAVED_IDS
1,
#else /* _POSIX_SAVED_IDS */
0,
#endif /* _POSIX_SAVED_IDS */
NULL, 0, CTL_KERN, KERN_SAVED_IDS, CTL_EOL);
Expose the kernel's boothowto(9) variable through the sysctl kern.boothowto variable. Part of the /etc/rc silent changes requested in PR 41946 and proposed in tech-userlevel.
2009-09-11 22:14:58 +04:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_HEX,
CTLTYPE_INT, "boothowto",
SYSCTL_DESCR("Flags from boot loader"),
NULL, 0, &boothowto, sizeof(boothowto),
CTL_KERN, CTL_CREATE, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_STRUCT, "boottime",
SYSCTL_DESCR("System boot time"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
NULL, 0, &boottime, sizeof(boottime),
CTL_KERN, KERN_BOOTTIME, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "maxpartitions",
SYSCTL_DESCR("Maximum number of partitions allowed per "
"disk"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
NULL, MAXPARTITIONS, NULL, 0,
CTL_KERN, KERN_MAXPARTITIONS, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT,
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
CTLTYPE_STRUCT, "timex", NULL,
sysctl_notavail, 0, NULL, 0,
CTL_KERN, KERN_TIMEX, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "rtc_offset",
SYSCTL_DESCR("Offset of real time clock from UTC in "
"minutes"),
Make kern.rtc_offset writable at securelevel <= 0. This allows boot-time adjustment when a machine runs other OSes with RTC == localtime.
2003-12-27 02:49:39 +03:00
sysctl_kern_rtc_offset, 0, &rtc_offset, 0,
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
CTL_KERN, KERN_RTC_OFFSET, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_STRING, "root_device",
SYSCTL_DESCR("Name of the root device"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
sysctl_root_device, 0, NULL, 0,
CTL_KERN, KERN_ROOT_DEVICE, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "fsync",
SYSCTL_DESCR("Whether the POSIX 1003.1b File "
"Synchronization Option is available on "
"this system"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
NULL, 1, NULL, 0,
CTL_KERN, KERN_FSYNC, CTL_EOL);
PR/34837: Mindaguas: Add SysV SHM dynamic reallocation and locking to the physical memory
2006-11-26 00:40:04 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT,
CTLTYPE_NODE, "ipc",
SYSCTL_DESCR("SysV IPC options"),
NULL, 0, NULL, 0,
CTL_KERN, KERN_SYSVIPC, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
More preparation for modularizing the SYSVxxx options. Here we change the kern.ipc.sysvxxx sysctls into dynamic values, so each sub-component of SYSVxxx can declare its own availability.
2015-05-13 04:00:16 +03:00
CTLFLAG_PERMANENT|CTLFLAG_READONLY,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "sysvmsg",
SYSCTL_DESCR("System V style message support available"),
More preparation for modularizing the SYSVxxx options. Here we change the kern.ipc.sysvxxx sysctls into dynamic values, so each sub-component of SYSVxxx can declare its own availability.
2015-05-13 04:00:16 +03:00
NULL, 0, &kern_has_sysvmsg, sizeof(int),
CTL_KERN, KERN_SYSVIPC, KERN_SYSVIPC_MSG, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
More preparation for modularizing the SYSVxxx options. Here we change the kern.ipc.sysvxxx sysctls into dynamic values, so each sub-component of SYSVxxx can declare its own availability.
2015-05-13 04:00:16 +03:00
CTLFLAG_PERMANENT|CTLFLAG_READONLY,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "sysvsem",
SYSCTL_DESCR("System V style semaphore support "
More preparation for modularizing the SYSVxxx options. Here we change the kern.ipc.sysvxxx sysctls into dynamic values, so each sub-component of SYSVxxx can declare its own availability.
2015-05-13 04:00:16 +03:00
"available"),
NULL, 0, &kern_has_sysvsem, sizeof(int),
CTL_KERN, KERN_SYSVIPC, KERN_SYSVIPC_SEM, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
More preparation for modularizing the SYSVxxx options. Here we change the kern.ipc.sysvxxx sysctls into dynamic values, so each sub-component of SYSVxxx can declare its own availability.
2015-05-13 04:00:16 +03:00
CTLFLAG_PERMANENT|CTLFLAG_READONLY,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "sysvshm",
SYSCTL_DESCR("System V style shared memory support "
More preparation for modularizing the SYSVxxx options. Here we change the kern.ipc.sysvxxx sysctls into dynamic values, so each sub-component of SYSVxxx can declare its own availability.
2015-05-13 04:00:16 +03:00
"available"),
NULL, 0, &kern_has_sysvshm, sizeof(int),
CTL_KERN, KERN_SYSVIPC, KERN_SYSVIPC_SHM, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "synchronized_io",
SYSCTL_DESCR("Whether the POSIX 1003.1b Synchronized "
"I/O Option is available on this system"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
NULL, 1, NULL, 0,
CTL_KERN, KERN_SYNCHRONIZED_IO, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "iov_max",
SYSCTL_DESCR("Maximum number of iovec structures per "
"process"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
NULL, IOV_MAX, NULL, 0,
CTL_KERN, KERN_IOV_MAX, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "mapped_files",
SYSCTL_DESCR("Whether the POSIX 1003.1b Memory Mapped "
"Files Option is available on this system"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
NULL, 1, NULL, 0,
CTL_KERN, KERN_MAPPED_FILES, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "memlock",
SYSCTL_DESCR("Whether the POSIX 1003.1b Process Memory "
"Locking Option is available on this "
"system"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
NULL, 1, NULL, 0,
CTL_KERN, KERN_MEMLOCK, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "memlock_range",
SYSCTL_DESCR("Whether the POSIX 1003.1b Range Memory "
"Locking Option is available on this "
"system"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
NULL, 1, NULL, 0,
CTL_KERN, KERN_MEMLOCK_RANGE, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "memory_protection",
SYSCTL_DESCR("Whether the POSIX 1003.1b Memory "
"Protection Option is available on this "
"system"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
NULL, 1, NULL, 0,
CTL_KERN, KERN_MEMORY_PROTECTION, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "login_name_max",
SYSCTL_DESCR("Maximum login name length"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
NULL, LOGIN_NAME_MAX, NULL, 0,
CTL_KERN, KERN_LOGIN_NAME_MAX, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_STRING, "defcorename",
SYSCTL_DESCR("Default core file name"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
sysctl_kern_defcorename, 0, defcorename, MAXPATHLEN,
CTL_KERN, KERN_DEFCORENAME, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "logsigexit",
SYSCTL_DESCR("Log process exit when caused by signals"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
NULL, 0, &kern_logsigexit, 0,
CTL_KERN, KERN_LOGSIGEXIT, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "fscale",
SYSCTL_DESCR("Kernel fixed-point scale factor"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
NULL, FSCALE, NULL, 0,
CTL_KERN, KERN_FSCALE, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "ccpu",
SYSCTL_DESCR("Scheduler exponential decay value"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
NULL, 0, &ccpu, 0,
CTL_KERN, KERN_CCPU, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_STRUCT, "cp_time",
SYSCTL_DESCR("Clock ticks spent in different CPU states"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
sysctl_kern_cptime, 0, NULL, 0,
CTL_KERN, KERN_CP_TIME, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_STRUCT, "consdev",
SYSCTL_DESCR("Console device"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
sysctl_consdev, 0, NULL, sizeof(dev_t),
CTL_KERN, KERN_CONSDEV, CTL_EOL);
#if NPTY > 0
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "maxptys",
SYSCTL_DESCR("Maximum number of pseudo-ttys"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
sysctl_kern_maxptys, 0, NULL, 0,
CTL_KERN, KERN_MAXPTYS, CTL_EOL);
#endif /* NPTY > 0 */
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "maxphys",
SYSCTL_DESCR("Maximum raw I/O transfer size"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
NULL, MAXPHYS, NULL, 0,
CTL_KERN, KERN_MAXPHYS, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "monotonic_clock",
SYSCTL_DESCR("Implementation version of the POSIX "
"1003.1b Monotonic Clock Option"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
/* XXX _POSIX_VERSION */
NULL, _POSIX_MONOTONIC_CLOCK, NULL, 0,
CTL_KERN, KERN_MONOTONIC_CLOCK, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "labelsector",
SYSCTL_DESCR("Sector number containing the disklabel"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
NULL, LABELSECTOR, NULL, 0,
CTL_KERN, KERN_LABELSECTOR, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "labeloffset",
SYSCTL_DESCR("Offset of the disklabel within the "
"sector"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
NULL, LABELOFFSET, NULL, 0,
CTL_KERN, KERN_LABELOFFSET, CTL_EOL);
Add getlabelusesmbr(), as proposed in http://mail-index.netbsd.org/tech-userlevel/2011/08/25/msg005404.html This is used by disk tools such as disklabel(8) to dynamically decide is the undelyling platform uses a disklabel-in-mbr-partition or not (instead of using a compile-time list of ports). getlabelusesmbr() reads the sysctl kern.labelusesmbr, takes its value from the machdep #define LABELUSESMBR. For evbmips, make LABELUSESMBR 1 if the platform uses pmon as bootloader, and 0 (the previous value) otherwise.
2011-08-30 16:39:49 +04:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
CTLTYPE_INT, "labelusesmbr",
SYSCTL_DESCR("disklabel is inside MBR partition"),
NULL, LABELUSESMBR, NULL, 0,
CTL_KERN, CTL_CREATE, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_NODE, "lwp",
SYSCTL_DESCR("System-wide LWP information"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
sysctl_kern_lwp, 0, NULL, 0,
CTL_KERN, KERN_LWP, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "forkfsleep",
SYSCTL_DESCR("Milliseconds to sleep on fork failure due "
"to process limits"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
sysctl_kern_forkfsleep, 0, NULL, 0,
CTL_KERN, KERN_FORKFSLEEP, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "posix_threads",
SYSCTL_DESCR("Version of IEEE Std 1003.1 and its "
"Threads option to which the system "
"attempts to conform"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
/* XXX _POSIX_VERSION */
NULL, _POSIX_THREADS, NULL, 0,
CTL_KERN, KERN_POSIX_THREADS, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
P1003_1B_SEMAPHORE is no longer optional.
2012-03-11 01:51:48 +04:00
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "posix_semaphores",
SYSCTL_DESCR("Version of IEEE Std 1003.1 and its "
"Semaphores option to which the system "
"attempts to conform"), NULL,
P1003_1B_SEMAPHORE is no longer optional.
2012-03-11 01:51:48 +04:00
200112, NULL, 0,
CTL_KERN, KERN_POSIX_SEMAPHORES, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "posix_barriers",
SYSCTL_DESCR("Version of IEEE Std 1003.1 and its "
"Barriers option to which the system "
"attempts to conform"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
/* XXX _POSIX_VERSION */
NULL, _POSIX_BARRIERS, NULL, 0,
CTL_KERN, KERN_POSIX_BARRIERS, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "posix_timers",
SYSCTL_DESCR("Version of IEEE Std 1003.1 and its "
"Timers option to which the system "
"attempts to conform"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
/* XXX _POSIX_VERSION */
NULL, _POSIX_TIMERS, NULL, 0,
CTL_KERN, KERN_POSIX_TIMERS, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "posix_spin_locks",
SYSCTL_DESCR("Version of IEEE Std 1003.1 and its Spin "
"Locks option to which the system attempts "
"to conform"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
/* XXX _POSIX_VERSION */
NULL, _POSIX_SPIN_LOCKS, NULL, 0,
CTL_KERN, KERN_POSIX_SPIN_LOCKS, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_IMMEDIATE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "posix_reader_writer_locks",
SYSCTL_DESCR("Version of IEEE Std 1003.1 and its "
"Read-Write Locks option to which the "
"system attempts to conform"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
/* XXX _POSIX_VERSION */
NULL, _POSIX_READER_WRITER_LOCKS, NULL, 0,
CTL_KERN, KERN_POSIX_READER_WRITER_LOCKS, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "dump_on_panic",
SYSCTL_DESCR("Perform a crash dump on system panic"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
NULL, 0, &dumponpanic, 0,
CTL_KERN, KERN_DUMP_ON_PANIC, CTL_EOL);
When DIAGNOSTIC is defined, provide a kern.panic_now sysctl to conviniently and reliably panic the system
2006-03-27 00:07:21 +04:00
#ifdef DIAGNOSTIC
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
CTLTYPE_INT, "panic_now",
SYSCTL_DESCR("Trigger a panic"),
sysctl_kern_trigger_panic, 0, NULL, 0,
CTL_KERN, CTL_CREATE, CTL_EOL);
#endif
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "root_partition",
SYSCTL_DESCR("Root partition on the root device"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
sysctl_kern_root_partition, 0, NULL, 0,
CTL_KERN, KERN_ROOT_PARTITION, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_STRUCT, "drivers",
SYSCTL_DESCR("List of all drivers with block and "
"character device numbers"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
sysctl_kern_drivers, 0, NULL, 0,
CTL_KERN, KERN_DRIVERS, CTL_EOL);
Add a new sysctl 'cp_id' that returns the array of cpu id values. Requested by me, implemented by atatat.
2005-06-16 18:55:58 +04:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT,
CTLTYPE_STRUCT, "cp_id",
SYSCTL_DESCR("Mapping of CPU number to CPU id"),
sysctl_kern_cpid, 0, NULL, 0,
CTL_KERN, KERN_CP_ID, CTL_EOL);
move security.setid_core.* to kern.coredump.setid.*, as requested by yamt@.
2006-07-15 01:55:19 +04:00
sysctl_createv(clog, 0, NULL, &rnode,
CTLFLAG_PERMANENT,
CTLTYPE_NODE, "coredump",
SYSCTL_DESCR("Coredump settings."),
NULL, 0, NULL, 0,
CTL_KERN, CTL_CREATE, CTL_EOL);
sysctl_createv(clog, 0, &rnode, &rnode,
CTLFLAG_PERMANENT,
CTLTYPE_NODE, "setid",
SYSCTL_DESCR("Set-id processes' coredump settings."),
NULL, 0, NULL, 0,
CTL_CREATE, CTL_EOL);
sysctl_createv(clog, 0, &rnode, NULL,
CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
CTLTYPE_INT, "dump",
SYSCTL_DESCR("Allow set-id processes to dump core."),
sysctl_security_setidcore, 0, &security_setidcore_dump,
sizeof(security_setidcore_dump),
CTL_CREATE, CTL_EOL);
sysctl_createv(clog, 0, &rnode, NULL,
CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
CTLTYPE_STRING, "path",
SYSCTL_DESCR("Path pattern for set-id coredumps."),
sysctl_security_setidcorename, 0,
Add some pre-processor magic to verify that the type of the data item passed to sysctl_createv() actually matches the declared type for the item itself. In the places where the caller specifies a function and a structure address (typically the 'softc') an explicit (void *) cast is now needed. Fixes bugs in sys/dev/acpi/asus_acpi.c sys/dev/bluetooth/bcsp.c sys/kern/vfs_bio.c sys/miscfs/syncfs/sync_subr.c and setting AcpiGbl_EnableAmlDebugObject. (mostly passing the address of a uint64_t when typed as CTLTYPE_INT). I've test built quite a few kernels, but there may be some unfixed MD fallout. Most likely passing &char[] to char *. Also add CTLFLAG_UNSIGNED for unsiged decimals - not set yet.
2012-06-03 01:36:41 +04:00
security_setidcore_path,
move security.setid_core.* to kern.coredump.setid.*, as requested by yamt@.
2006-07-15 01:55:19 +04:00
sizeof(security_setidcore_path),
CTL_CREATE, CTL_EOL);
sysctl_createv(clog, 0, &rnode, NULL,
CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
CTLTYPE_INT, "owner",
SYSCTL_DESCR("Owner id for set-id processes' cores."),
sysctl_security_setidcore, 0, &security_setidcore_owner,
0,
CTL_CREATE, CTL_EOL);
sysctl_createv(clog, 0, &rnode, NULL,
CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
CTLTYPE_INT, "group",
SYSCTL_DESCR("Group id for set-id processes' cores."),
sysctl_security_setidcore, 0, &security_setidcore_group,
0,
CTL_CREATE, CTL_EOL);
sysctl_createv(clog, 0, &rnode, NULL,
CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
CTLTYPE_INT, "mode",
SYSCTL_DESCR("Mode for set-id processes' cores."),
sysctl_security_setidcore, 0, &security_setidcore_mode,
0,
CTL_CREATE, CTL_EOL);
Merge wrstuden-revivesa into HEAD.
2008-10-15 10:51:17 +04:00
sysctl_createv(clog, 0, NULL, NULL,
Remove COMPAT_SA / KERN_SA. Welcome to 6.99.3! Approved by core@.
2012-02-20 01:05:51 +04:00
CTLFLAG_IMMEDIATE|CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
Merge wrstuden-revivesa into HEAD.
2008-10-15 10:51:17 +04:00
CTLTYPE_INT, "no_sa_support",
avoid code duplication.
2011-02-05 04:22:12 +03:00
SYSCTL_DESCR("0 if the kernel supports SA, otherwise "
"it doesn't"),
Remove COMPAT_SA / KERN_SA. Welcome to 6.99.3! Approved by core@.
2012-02-20 01:05:51 +04:00
NULL, 1, NULL, 0,
avoid code duplication.
2011-02-05 04:22:12 +03:00
CTL_KERN, CTL_CREATE, CTL_EOL);
BUILDINFO part 2: expose sysctl kern.buildinfo
2014-08-03 13:15:21 +04:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT,
CTLTYPE_STRING, "configname",
SYSCTL_DESCR("Name of config file"),
NULL, 0, __UNCONST(kernel_ident), 0,
CTL_KERN, CTL_CREATE, CTL_EOL);
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT,
CTLTYPE_STRING, "buildinfo",
SYSCTL_DESCR("Information from build environment"),
NULL, 0, __UNCONST(buildinfo), 0,
CTL_KERN, CTL_CREATE, CTL_EOL);
Add a new kern.messages sysctl to allow kernel message verbosity to be altered after boot. Fixes PR kern/46539 using patch submitted by Nat Sloss.
2016-05-31 08:44:19 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
CTLTYPE_INT, "messages",
SYSCTL_DESCR("Kernel message verbosity"),
sysctl_kern_messages, 0, NULL, 0,
CTL_KERN, CTL_CREATE, CTL_EOL);
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
}
Move a bunch of sysctl nodes from init_sysctl (kitchen sink sysctl file) to init_sysctl_base (only base kernel defs). Main motivation was to fix sysconf(_SC_NPROCESSORS) for Rumprun. As reported by neeraj on irc, it returned -1 before this fix, so we were doing imaginary computing.
2015-08-25 17:52:31 +03:00
SYSCTL_SETUP(sysctl_hw_misc_setup, "sysctl hw subtree misc setup")
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
{
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_INT, "usermem",
SYSCTL_DESCR("Bytes of non-kernel memory"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
sysctl_hw_usermem, 0, NULL, 0,
CTL_HW, HW_USERMEM, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_READWRITE|CTLFLAG_HEX,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_STRING, "cnmagic",
SYSCTL_DESCR("Console magic key sequence"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
sysctl_hw_cnmagic, 0, NULL, CNS_LEN,
CTL_HW, HW_CNMAGIC, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT,
Lots of sysctl descriptions (if someone wants to help out here, that would be good) mostly copied from sysctl(3). This takes care of the top-level, most of kern.* and hw.* (modulo the ath and bge stuff), and all of proc.*. If you don't want the added rodata in your kernel, use "options SYSCTL_NO_DESCR" in your kernel config.
2004-04-08 10:20:29 +04:00
CTLTYPE_QUAD, "usermem64",
SYSCTL_DESCR("Bytes of non-kernel memory"),
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
sysctl_hw_usermem, 0, NULL, 0,
CTL_HW, HW_USERMEM64, CTL_EOL);
}
#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,
};
/*
* this setup routine is a replacement for debug_sysctl()
*
* note that it creates several nodes per defined debug variable
*/
SYSCTL_SETUP(sysctl_debug_setup, "sysctl debug subtree setup")
{
struct ctldebug *cdp;
char nodename[20];
int i;
/*
* two ways here:
*
* the "old" way (debug.name -> value) which was emulated by
* the sysctl(8) binary
*
* the new way, which the sysctl(8) binary was actually using
node debug
node debug.0
- fix comment sentence capitalization. - whitespace cleanup. No functional changes.
2007-10-16 20:05:50 +04:00
string debug.0.name
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
int debug.0.value
int debug.name
*/
for (i = 0; i < CTL_DEBUG_MAXID; i++) {
cdp = debugvars[i];
if (cdp->debugname == NULL || cdp->debugvar == NULL)
continue;
snprintf(nodename, sizeof(nodename), "debug%d", i);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_HIDDEN,
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
CTLTYPE_NODE, nodename, NULL,
NULL, 0, NULL, 0,
CTL_DEBUG, i, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_HIDDEN,
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
CTLTYPE_STRING, "name", NULL,
- add const. - remove unnecessary casts. - add __UNCONST casts and mark them with XXXUNCONST as necessary.
2005-05-30 02:24:14 +04:00
/*XXXUNCONST*/
NULL, 0, __UNCONST(cdp->debugname), 0,
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
CTL_DEBUG, i, CTL_DEBUG_NAME, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT|CTLFLAG_HIDDEN,
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
CTLTYPE_INT, "value", NULL,
NULL, 0, cdp->debugvar, 0,
CTL_DEBUG, i, CTL_DEBUG_VALUE, CTL_EOL);
Tango on sysctl_createv() and flags. The flags have all been renamed, and sysctl_createv() now uses more arguments.
2004-03-24 18:34:46 +03:00
sysctl_createv(clog, 0, NULL, NULL,
CTLFLAG_PERMANENT,
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
CTLTYPE_INT, cdp->debugname, NULL,
NULL, 0, cdp->debugvar, 0,
CTL_DEBUG, CTL_CREATE, CTL_EOL);
}
}
#endif /* DEBUG */
/*
* ********************************************************************
* section 2: private node-specific helper routines.
* ********************************************************************
*/
When DIAGNOSTIC is defined, provide a kern.panic_now sysctl to conviniently and reliably panic the system
2006-03-27 00:07:21 +04:00
#ifdef DIAGNOSTIC
static int
sysctl_kern_trigger_panic(SYSCTLFN_ARGS)
{
int newtrig, error;
struct sysctlnode node;
newtrig = 0;
node = *rnode;
node.sysctl_data = &newtrig;
error = sysctl_lookup(SYSCTLFN_CALL(&node));
if (error || newp == NULL)
return (error);
if (newtrig != 0)
panic("Panic triggered");
return (error);
}
#endif
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
/*
- fix comment sentence capitalization. - whitespace cleanup. No functional changes.
2007-10-16 20:05:50 +04:00
* sysctl helper routine for kern.maxvnodes. Drain vnodes if
update a comment to match with the previous change (rev.1.12).
2003-12-20 10:33:03 +03:00
* new value is lower than desiredvnodes and then calls reinit
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
* routines that needs to adjust to the new value.
*/
static int
sysctl_kern_maxvnodes(SYSCTLFN_ARGS)
{
PR kern/38927: processes getting stuck in uvm_map (cv_timedwait), hanging machine Assume that a vnode (and associated data structures) costs 2kB in the worst imaginable case. Don't allow sysctl to set desiredvnodes to a value that would use more than 75% of KVA or 75% of physical memory.
2008-06-16 15:26:28 +04:00
int error, new_vnodes, old_vnodes, new_max;
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
struct sysctlnode node;
new_vnodes = desiredvnodes;
node = *rnode;
node.sysctl_data = &new_vnodes;
error = sysctl_lookup(SYSCTLFN_CALL(&node));
if (error || newp == NULL)
return (error);
Add sanity check to sysctl_kern_maxvnodes.
2012-10-03 11:22:59 +04:00
/*
* sysctl passes down unsigned values, require them
* to be positive
*/
if (new_vnodes <= 0)
return (EINVAL);
calculate vnode cache size based on the resource it gets allocated from this stops setting kern.maxvnodes to high so it exhausts available space in kmem http://mail-index.netbsd.org/tech-kern/2013/03/08/msg015095.html
2013-03-18 17:36:21 +04:00
/* Limits: 75% of kmem and physical memory. */
new_max = calc_cache_size(vmem_size(kmem_arena, VMEM_FREE|VMEM_ALLOC),
75, 75) / VNODE_COST;
PR kern/38927: processes getting stuck in uvm_map (cv_timedwait), hanging machine Assume that a vnode (and associated data structures) costs 2kB in the worst imaginable case. Don't allow sysctl to set desiredvnodes to a value that would use more than 75% of KVA or 75% of physical memory.
2008-06-16 15:26:28 +04:00
if (new_vnodes > new_max)
new_vnodes = new_max;
restore functionality to decrease kern.maxvnodes which has been backed out during sysctl rework.
2003-12-20 10:26:27 +03:00
old_vnodes = desiredvnodes;
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
desiredvnodes = new_vnodes;
Remove the "target" argment from vfs_drainvnodes() as it is always equal to "desiredvnodes" and move its definition from sys/vnode.h to sys/vnode_impl.h. Extend vfs_drainvnodes() to also wait for deferred vrele to flush and replace the call to vrele_flush() with a call to vfs_drainvnodes().
2016-12-14 18:46:57 +03:00
error = vfs_drainvnodes();
Add a global vnode cache: - vcache_get() retrieves a referenced and initialised vnode / fs node pair. - vcache_remove() removes a vnode / fs node pair from the cache. On cache miss vcache_get() calls new vfs operation vfs_loadvnode() to initialise a vnode / fs node pair. This call is guaranteed exclusive, no other thread will try to load this vnode / fs node pair. Convert ufs/ext2fs, ufs/ffs and ufs/mfs to use this interface. Remove now unused ufs/ufs_ihash Discussed on tech-kern. Welcome to 6.99.41
2014-05-08 12:21:53 +04:00
if (error) {
desiredvnodes = old_vnodes;
return (error);
restore functionality to decrease kern.maxvnodes which has been backed out during sysctl rework.
2003-12-20 10:26:27 +03:00
}
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
vfs_reinit();
nchreinit();
return (0);
}
Add a new kern.messages sysctl to allow kernel message verbosity to be altered after boot. Fixes PR kern/46539 using patch submitted by Nat Sloss.
2016-05-31 08:44:19 +03:00
/*
* sysctl helper routine for kern.messages.
* Alters boothowto to display kernel messages in increasing verbosity
* from 0 to 4.
*/
#define MAXMESSAGES 4
static int
sysctl_kern_messages(SYSCTLFN_ARGS)
{
int error, messageverbose, messagemask, newboothowto;
struct sysctlnode node;
messagemask = (AB_NORMAL|AB_QUIET|AB_SILENT|AB_VERBOSE|AB_DEBUG);
switch (boothowto & messagemask) {
case AB_SILENT:
messageverbose = 0;
break;
case AB_QUIET:
messageverbose = 1;
break;
case AB_VERBOSE:
messageverbose = 3;
break;
case AB_DEBUG:
messageverbose = 4;
break;
case AB_NORMAL:
default:
messageverbose = 2;
}
node = *rnode;
node.sysctl_data = &messageverbose;
error = sysctl_lookup(SYSCTLFN_CALL(&node));
if (error || newp == NULL)
return (error);
if (messageverbose < 0 || messageverbose > MAXMESSAGES)
return EINVAL;
/* Set boothowto */
newboothowto = boothowto & ~messagemask;
switch (messageverbose) {
case 0:
newboothowto |= AB_SILENT;
break;
case 1:
newboothowto |= AB_QUIET;
break;
case 3:
newboothowto |= AB_VERBOSE;
break;
case 4:
newboothowto |= AB_DEBUG;
break;
case 2:
default: /* Messages default to normal. */
break;
}
boothowto = newboothowto;
return (0);
}
Make kern.rtc_offset writable at securelevel <= 0. This allows boot-time adjustment when a machine runs other OSes with RTC == localtime.
2003-12-27 02:49:39 +03:00
/*
* sysctl helper routine for rtc_offset - set time after changes
*/
static int
sysctl_kern_rtc_offset(SYSCTLFN_ARGS)
{
- make settime take timespec. - avoid wrapping of time in settime. - pass struct proc down so that we can log a detailed message.
2005-12-05 03:16:33 +03:00
struct timespec ts, delta;
int error, new_rtc_offset;
Make kern.rtc_offset writable at securelevel <= 0. This allows boot-time adjustment when a machine runs other OSes with RTC == localtime.
2003-12-27 02:49:39 +03:00
struct sysctlnode node;
new_rtc_offset = rtc_offset;
node = *rnode;
node.sysctl_data = &new_rtc_offset;
error = sysctl_lookup(SYSCTLFN_CALL(&node));
if (error || newp == NULL)
return (error);
First take at security model abstraction. - Add a few scopes to the kernel: system, network, and machdep. - Add a few more actions/sub-actions (requests), and start using them as opposed to the KAUTH_GENERIC_ISSUSER place-holders. - Introduce a basic set of listeners that implement our "traditional" security model, called "bsd44". This is the default (and only) model we have at the moment. - Update all relevant documentation. - Add some code and docs to help folks who want to actually use this stuff: * There's a sample overlay model, sitting on-top of "bsd44", for fast experimenting with tweaking just a subset of an existing model. This is pretty cool because it's *really* straightforward to do stuff you had to use ugly hacks for until now... * And of course, documentation describing how to do the above for quick reference, including code samples. All of these changes were tested for regressions using a Python-based testsuite that will be (I hope) available soon via pkgsrc. Information about the tests, and how to write new ones, can be found on: http://kauth.linbsd.org/kauthwiki NOTE FOR DEVELOPERS: *PLEASE* don't add any code that does any of the following: - Uses a KAUTH_GENERIC_ISSUSER kauth(9) request, - Checks 'securelevel' directly, - Checks a uid/gid directly. (or if you feel you have to, contact me first) This is still work in progress; It's far from being done, but now it'll be a lot easier. Relevant mailing list threads: http://mail-index.netbsd.org/tech-security/2006/01/25/0011.html http://mail-index.netbsd.org/tech-security/2006/03/24/0001.html http://mail-index.netbsd.org/tech-security/2006/04/18/0000.html http://mail-index.netbsd.org/tech-security/2006/05/15/0000.html http://mail-index.netbsd.org/tech-security/2006/08/01/0000.html http://mail-index.netbsd.org/tech-security/2006/08/25/0000.html Many thanks to YAMAMOTO Takashi, Matt Thomas, and Christos Zoulas for help stablizing kauth(9). Full credit for the regression tests, making sure these changes didn't break anything, goes to Matt Fleming and Jaime Fournier. Happy birthday Randi! :)
2006-09-09 00:58:56 +04:00
if (kauth_authorize_system(l->l_cred, KAUTH_SYSTEM_TIME,
KAUTH_REQ_SYSTEM_TIME_RTCOFFSET,
- Use KAUTH_ARG() instead of casts, - Don't ignore return value of settime() in sysctl_kern_rtc_offset(), as suggested by yamt@. Note: the kauth(9) call in sysctl_kern_rtc_offset() is bogus, but this will be addressed separately.
2007-12-10 21:58:02 +03:00
KAUTH_ARG(new_rtc_offset), NULL, NULL))
Make kern.rtc_offset writable at securelevel <= 0. This allows boot-time adjustment when a machine runs other OSes with RTC == localtime.
2003-12-27 02:49:39 +03:00
return (EPERM);
if (rtc_offset == new_rtc_offset)
return (0);
/* if we change the offset, adjust the time */
- make settime take timespec. - avoid wrapping of time in settime. - pass struct proc down so that we can log a detailed message.
2005-12-05 03:16:33 +03:00
nanotime(&ts);
delta.tv_sec = 60 * (new_rtc_offset - rtc_offset);
delta.tv_nsec = 0;
timespecadd(&ts, &delta, &ts);
Make kern.rtc_offset writable at securelevel <= 0. This allows boot-time adjustment when a machine runs other OSes with RTC == localtime.
2003-12-27 02:49:39 +03:00
rtc_offset = new_rtc_offset;
- Use KAUTH_ARG() instead of casts, - Don't ignore return value of settime() in sysctl_kern_rtc_offset(), as suggested by yamt@. Note: the kauth(9) call in sysctl_kern_rtc_offset() is bogus, but this will be addressed separately.
2007-12-10 21:58:02 +03:00
return (settime(l->l_proc, &ts));
Make kern.rtc_offset writable at securelevel <= 0. This allows boot-time adjustment when a machine runs other OSes with RTC == localtime.
2003-12-27 02:49:39 +03:00
}
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
/*
- fix comment sentence capitalization. - whitespace cleanup. No functional changes.
2007-10-16 20:05:50 +04:00
* sysctl helper routine for kern.maxproc. Ensures that the new
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
* values are not too low or too high.
*/
static int
sysctl_kern_maxproc(SYSCTLFN_ARGS)
{
int error, nmaxproc;
struct sysctlnode node;
nmaxproc = maxproc;
node = *rnode;
node.sysctl_data = &nmaxproc;
error = sysctl_lookup(SYSCTLFN_CALL(&node));
if (error || newp == NULL)
return (error);
if (nmaxproc < 0 || nmaxproc >= PID_MAX)
return (EINVAL);
#ifdef __HAVE_CPU_MAXPROC
if (nmaxproc > cpu_maxproc())
return (EINVAL);
#endif
maxproc = nmaxproc;
return (0);
}
/*
- fix comment sentence capitalization. - whitespace cleanup. No functional changes.
2007-10-16 20:05:50 +04:00
* sysctl helper function for kern.hostid. The hostid is a long, but
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
* we export it as an int, so we need to give it a little help.
*/
static int
sysctl_kern_hostid(SYSCTLFN_ARGS)
{
int error, inthostid;
struct sysctlnode node;
Fix a comment. Approved by Andrew Brown.
2004-05-03 17:39:50 +04:00
inthostid = hostid; /* XXX assumes sizeof int <= sizeof long */
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
node = *rnode;
node.sysctl_data = &inthostid;
error = sysctl_lookup(SYSCTLFN_CALL(&node));
if (error || newp == NULL)
return (error);
Prefer that kern.hostid is printed in hex, not as a signed decimal, and avoid accidental sign-extension when setting it.
2004-04-16 17:25:40 +04:00
hostid = (unsigned)inthostid;
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
return (0);
}
/*
- fix comment sentence capitalization. - whitespace cleanup. No functional changes.
2007-10-16 20:05:50 +04:00
* sysctl helper routine for kern.defcorename. In the case of a new
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
* string being assigned, check that it's not a zero-length string.
* (XXX the check in -current doesn't work, but do we really care?)
*/
static int
sysctl_kern_defcorename(SYSCTLFN_ARGS)
{
int error;
for some random places, use PNBUF_GET/PUT rather than - on-stack buffer - malloc(MAXPATHLEN)
2006-02-04 15:09:50 +03:00
char *newcorename;
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
struct sysctlnode node;
for some random places, use PNBUF_GET/PUT rather than - on-stack buffer - malloc(MAXPATHLEN)
2006-02-04 15:09:50 +03:00
newcorename = PNBUF_GET();
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
node = *rnode;
node.sysctl_data = &newcorename[0];
memcpy(node.sysctl_data, rnode->sysctl_data, MAXPATHLEN);
error = sysctl_lookup(SYSCTLFN_CALL(&node));
for some random places, use PNBUF_GET/PUT rather than - on-stack buffer - malloc(MAXPATHLEN)
2006-02-04 15:09:50 +03:00
if (error || newp == NULL) {
goto done;
}
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
/*
* when sysctl_lookup() deals with a string, it's guaranteed
- fix comment sentence capitalization. - whitespace cleanup. No functional changes.
2007-10-16 20:05:50 +04:00
* to come back nul terminated. So there. :)
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
*/
for some random places, use PNBUF_GET/PUT rather than - on-stack buffer - malloc(MAXPATHLEN)
2006-02-04 15:09:50 +03:00
if (strlen(newcorename) == 0) {
error = EINVAL;
} else {
memcpy(rnode->sysctl_data, node.sysctl_data, MAXPATHLEN);
error = 0;
}
done:
PNBUF_PUT(newcorename);
return error;
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
}
/*
- fix comment sentence capitalization. - whitespace cleanup. No functional changes.
2007-10-16 20:05:50 +04:00
* sysctl helper routine for kern.cp_time node. Adds up cpu time
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
* across all cpus.
*/
static int
sysctl_kern_cptime(SYSCTLFN_ARGS)
{
struct sysctlnode node = *rnode;
u_intN_t -> uintN_t
2005-12-26 21:41:36 +03:00
uint64_t *cp_time = NULL;
Count the number of CPUs at boot and stash in 'ncpu'. Eventually should have each CPU register at attach, so we can figure out the topology for the scheduler.
2007-02-15 23:32:47 +03:00
int error, n = ncpu, i;
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
struct cpu_info *ci;
CPU_INFO_ITERATOR cii;
/*
* if you specifically pass a buffer that is the size of the
* sum, or if you are probing for the size, you get the "sum"
* of cp_time (and the size thereof) across all processors.
*
* alternately, you can pass an additional mib number and get
* cp_time for that particular processor.
*/
switch (namelen) {
case 0:
- fix comment sentence capitalization. - whitespace cleanup. No functional changes.
2007-10-16 20:05:50 +04:00
if (*oldlenp == sizeof(uint64_t) * CPUSTATES || oldp == NULL) {
u_intN_t -> uintN_t
2005-12-26 21:41:36 +03:00
node.sysctl_size = sizeof(uint64_t) * CPUSTATES;
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
n = -1; /* SUM */
}
else {
u_intN_t -> uintN_t
2005-12-26 21:41:36 +03:00
node.sysctl_size = n * sizeof(uint64_t) * CPUSTATES;
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
n = -2; /* ALL */
}
break;
case 1:
if (name[0] < 0 || name[0] >= n)
Adjust error returns in kern.cp_time when a specific processor is being requested so that (1) the uniprocessor case and the multiprocessor case are more similar and (2) so that we return ENOENT when a non-existent processor is requested (which is both more sensible and follows the general order of things anyway).
2003-12-29 01:24:12 +03:00
return (ENOENT); /* ENOSUCHPROCESSOR */
u_intN_t -> uintN_t
2005-12-26 21:41:36 +03:00
node.sysctl_size = sizeof(uint64_t) * CPUSTATES;
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
n = name[0];
/*
* adjust these so that sysctl_lookup() will be happy
*/
name++;
namelen--;
Add missing break.
2003-12-07 13:33:03 +03:00
break;
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
default:
return (EINVAL);
}
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
cp_time = kmem_alloc(node.sysctl_size, KM_SLEEP);
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
node.sysctl_data = cp_time;
memset(cp_time, 0, node.sysctl_size);
for (CPU_INFO_FOREACH(cii, ci)) {
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
if (n <= 0) {
for (i = 0; i < CPUSTATES; i++) {
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
cp_time[i] += ci->ci_schedstate.spc_cp_time[i];
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
}
}
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
/*
* if a specific processor was requested and we just
* did it, we're done here
*/
if (n == 0)
break;
/*
* if doing "all", skip to next cp_time set for next processor
*/
if (n == -2)
cp_time += CPUSTATES;
/*
* if we're doing a specific processor, we're one
* processor closer
*/
if (n > 0)
n--;
}
error = sysctl_lookup(SYSCTLFN_CALL(&node));
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
kmem_free(node.sysctl_data, node.sysctl_size);
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
return (error);
}
#if NPTY > 0
/*
- fix comment sentence capitalization. - whitespace cleanup. No functional changes.
2007-10-16 20:05:50 +04:00
* sysctl helper routine for kern.maxptys. Ensures that any new value
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
* is acceptable to the pty subsystem.
*/
static int
sysctl_kern_maxptys(SYSCTLFN_ARGS)
{
int pty_maxptys(int, int); /* defined in kern/tty_pty.c */
- add const. - remove unnecessary casts. - add __UNCONST casts and mark them with XXXUNCONST as necessary.
2005-05-30 02:24:14 +04:00
int error, xmax;
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
struct sysctlnode node;
/* get current value of maxptys */
- add const. - remove unnecessary casts. - add __UNCONST casts and mark them with XXXUNCONST as necessary.
2005-05-30 02:24:14 +04:00
xmax = pty_maxptys(0, 0);
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
node = *rnode;
- add const. - remove unnecessary casts. - add __UNCONST casts and mark them with XXXUNCONST as necessary.
2005-05-30 02:24:14 +04:00
node.sysctl_data = &xmax;
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
error = sysctl_lookup(SYSCTLFN_CALL(&node));
if (error || newp == NULL)
return (error);
nuke trailing whitespace
2005-02-27 00:34:55 +03:00
- add const. - remove unnecessary casts. - add __UNCONST casts and mark them with XXXUNCONST as necessary.
2005-05-30 02:24:14 +04:00
if (xmax != pty_maxptys(xmax, 1))
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
return (EINVAL);
return (0);
}
#endif /* NPTY > 0 */
/*
* sysctl helper routine to do kern.lwp.* work.
*/
static int
sysctl_kern_lwp(SYSCTLFN_ARGS)
{
struct kinfo_lwp klwp;
struct proc *p;
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
struct lwp *l2, *l3;
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
char *where, *dp;
int pid, elem_size, elem_count;
int buflen, needed, error;
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
bool gotit;
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
Sysctl functions called for "generic" nodes should forward "query" requests (where possible), rather than returning errors.
2003-12-29 01:36:37 +03:00
if (namelen == 1 && name[0] == CTL_QUERY)
Properly fix the constipated lossage wrt -Wcast-qual and the sysctl code. I know it's not the prettiest code, but it seems to work rather well in spite of itself.
2005-06-09 06:19:59 +04:00
return (sysctl_query(SYSCTLFN_CALL(rnode)));
Sysctl functions called for "generic" nodes should forward "query" requests (where possible), rather than returning errors.
2003-12-29 01:36:37 +03:00
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
dp = where = oldp;
buflen = where != NULL ? *oldlenp : 0;
error = needed = 0;
In sysctl_kern_lwp adjust offsets into the mib entries so that they are now correct. Fixes problems with "ps -s" not working. Also use KERN_LWPSLOP instead of KERN_PROCSLOP. Both changes from Andrew Brown.
2003-12-13 02:21:44 +03:00
if (newp != NULL || namelen != 3)
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
return (EINVAL);
In sysctl_kern_lwp adjust offsets into the mib entries so that they are now correct. Fixes problems with "ps -s" not working. Also use KERN_LWPSLOP instead of KERN_PROCSLOP. Both changes from Andrew Brown.
2003-12-13 02:21:44 +03:00
pid = name[0];
elem_size = name[1];
elem_count = name[2];
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
sysctl_unlock();
merge yamt-idlelwp branch. asked by core@. some ports still needs work. from doc/BRANCHES: idle lwp, and some changes depending on it. 1. separate context switching and thread scheduling. (cf. gmcgarry_ctxsw) 2. implement idle lwp. 3. clean up related MD/MI interfaces. 4. make scheduler(s) modular.
2007-05-17 18:51:11 +04:00
if (pid == -1) {
Network protocol interrupts can now block on locks, so merge the globals proclist_mutex and proclist_lock into a single adaptive mutex (proc_lock). Implications: - Inspecting process state requires thread context, so signals can no longer be sent from a hardware interrupt handler. Signal activity must be deferred to a soft interrupt or kthread. - As the proc state locking is simplified, it's now safe to take exit() and wait() out from under kernel_lock. - The system spends less time at IPL_SCHED, and there is less lock activity.
2008-04-24 19:35:27 +04:00
mutex_enter(proc_lock);
don't forget to skip marker processes.
2009-05-12 15:42:12 +04:00
PROCLIST_FOREACH(p, &allproc) {
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
/* Grab a hold on the process. */
if (!rw_tryenter(&p->p_reflock, RW_READER)) {
continue;
}
Network protocol interrupts can now block on locks, so merge the globals proclist_mutex and proclist_lock into a single adaptive mutex (proc_lock). Implications: - Inspecting process state requires thread context, so signals can no longer be sent from a hardware interrupt handler. Signal activity must be deferred to a soft interrupt or kthread. - As the proc state locking is simplified, it's now safe to take exit() and wait() out from under kernel_lock. - The system spends less time at IPL_SCHED, and there is less lock activity.
2008-04-24 19:35:27 +04:00
mutex_exit(proc_lock);
merge yamt-idlelwp branch. asked by core@. some ports still needs work. from doc/BRANCHES: idle lwp, and some changes depending on it. 1. separate context switching and thread scheduling. (cf. gmcgarry_ctxsw) 2. implement idle lwp. 3. clean up related MD/MI interfaces. 4. make scheduler(s) modular.
2007-05-17 18:51:11 +04:00
Merge proc::p_mutex and proc::p_smutex into a single adaptive mutex, since we no longer need to guard against access from hardware interrupt handlers. Additionally, if cloning a process with CLONE_SIGHAND, arrange to have the child process share the parent's lock so that signal state may be kept in sync. Partially addresses PR kern/37437.
2008-04-24 22:39:20 +04:00
mutex_enter(p->p_lock);
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
LIST_FOREACH(l2, &p->p_lwps, l_sibling) {
if (buflen >= elem_size && elem_count > 0) {
lwp_lock(l2);
fill_lwp(l2, &klwp);
lwp_unlock(l2);
Merge proc::p_mutex and proc::p_smutex into a single adaptive mutex, since we no longer need to guard against access from hardware interrupt handlers. Additionally, if cloning a process with CLONE_SIGHAND, arrange to have the child process share the parent's lock so that signal state may be kept in sync. Partially addresses PR kern/37437.
2008-04-24 22:39:20 +04:00
mutex_exit(p->p_lock);
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
/*
* Copy out elem_size, but not
* larger than the size of a
* struct kinfo_proc2.
*/
error = dcopyout(l, &klwp, dp,
min(sizeof(klwp), elem_size));
if (error) {
rw_exit(&p->p_reflock);
goto cleanup;
}
Merge proc::p_mutex and proc::p_smutex into a single adaptive mutex, since we no longer need to guard against access from hardware interrupt handlers. Additionally, if cloning a process with CLONE_SIGHAND, arrange to have the child process share the parent's lock so that signal state may be kept in sync. Partially addresses PR kern/37437.
2008-04-24 22:39:20 +04:00
mutex_enter(p->p_lock);
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
LIST_FOREACH(l3, &p->p_lwps,
l_sibling) {
if (l2 == l3)
break;
}
if (l3 == NULL) {
Merge proc::p_mutex and proc::p_smutex into a single adaptive mutex, since we no longer need to guard against access from hardware interrupt handlers. Additionally, if cloning a process with CLONE_SIGHAND, arrange to have the child process share the parent's lock so that signal state may be kept in sync. Partially addresses PR kern/37437.
2008-04-24 22:39:20 +04:00
mutex_exit(p->p_lock);
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
rw_exit(&p->p_reflock);
error = EAGAIN;
goto cleanup;
}
dp += elem_size;
buflen -= elem_size;
elem_count--;
}
needed += elem_size;
merge yamt-idlelwp branch. asked by core@. some ports still needs work. from doc/BRANCHES: idle lwp, and some changes depending on it. 1. separate context switching and thread scheduling. (cf. gmcgarry_ctxsw) 2. implement idle lwp. 3. clean up related MD/MI interfaces. 4. make scheduler(s) modular.
2007-05-17 18:51:11 +04:00
}
Merge proc::p_mutex and proc::p_smutex into a single adaptive mutex, since we no longer need to guard against access from hardware interrupt handlers. Additionally, if cloning a process with CLONE_SIGHAND, arrange to have the child process share the parent's lock so that signal state may be kept in sync. Partially addresses PR kern/37437.
2008-04-24 22:39:20 +04:00
mutex_exit(p->p_lock);
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
/* Drop reference to process. */
Network protocol interrupts can now block on locks, so merge the globals proclist_mutex and proclist_lock into a single adaptive mutex (proc_lock). Implications: - Inspecting process state requires thread context, so signals can no longer be sent from a hardware interrupt handler. Signal activity must be deferred to a soft interrupt or kthread. - As the proc state locking is simplified, it's now safe to take exit() and wait() out from under kernel_lock. - The system spends less time at IPL_SCHED, and there is less lock activity.
2008-04-24 19:35:27 +04:00
mutex_enter(proc_lock);
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
rw_exit(&p->p_reflock);
merge yamt-idlelwp branch. asked by core@. some ports still needs work. from doc/BRANCHES: idle lwp, and some changes depending on it. 1. separate context switching and thread scheduling. (cf. gmcgarry_ctxsw) 2. implement idle lwp. 3. clean up related MD/MI interfaces. 4. make scheduler(s) modular.
2007-05-17 18:51:11 +04:00
}
Network protocol interrupts can now block on locks, so merge the globals proclist_mutex and proclist_lock into a single adaptive mutex (proc_lock). Implications: - Inspecting process state requires thread context, so signals can no longer be sent from a hardware interrupt handler. Signal activity must be deferred to a soft interrupt or kthread. - As the proc state locking is simplified, it's now safe to take exit() and wait() out from under kernel_lock. - The system spends less time at IPL_SCHED, and there is less lock activity.
2008-04-24 19:35:27 +04:00
mutex_exit(proc_lock);
merge yamt-idlelwp branch. asked by core@. some ports still needs work. from doc/BRANCHES: idle lwp, and some changes depending on it. 1. separate context switching and thread scheduling. (cf. gmcgarry_ctxsw) 2. implement idle lwp. 3. clean up related MD/MI interfaces. 4. make scheduler(s) modular.
2007-05-17 18:51:11 +04:00
} else {
Network protocol interrupts can now block on locks, so merge the globals proclist_mutex and proclist_lock into a single adaptive mutex (proc_lock). Implications: - Inspecting process state requires thread context, so signals can no longer be sent from a hardware interrupt handler. Signal activity must be deferred to a soft interrupt or kthread. - As the proc state locking is simplified, it's now safe to take exit() and wait() out from under kernel_lock. - The system spends less time at IPL_SCHED, and there is less lock activity.
2008-04-24 19:35:27 +04:00
mutex_enter(proc_lock);
Remove pfind() and pgfind(), fix locking in various broken uses of these. Rename real routines to proc_find() and pgrp_find(), remove PFIND_* flags and have consistent behaviour. Provide proc_find_raw() for special cases. Fix memory leak in sysctl_proc_corename(). COMPAT_LINUX: rework ptrace() locking, minimise differences between different versions per-arch. Note: while this change adds some formal cosmetics for COMPAT_DARWIN and COMPAT_IRIX - locking there is utterly broken (for ages). Fixes PR/43176.
2010-07-01 06:38:26 +04:00
p = proc_find(pid);
merge yamt-idlelwp branch. asked by core@. some ports still needs work. from doc/BRANCHES: idle lwp, and some changes depending on it. 1. separate context switching and thread scheduling. (cf. gmcgarry_ctxsw) 2. implement idle lwp. 3. clean up related MD/MI interfaces. 4. make scheduler(s) modular.
2007-05-17 18:51:11 +04:00
if (p == NULL) {
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
error = ESRCH;
Network protocol interrupts can now block on locks, so merge the globals proclist_mutex and proclist_lock into a single adaptive mutex (proc_lock). Implications: - Inspecting process state requires thread context, so signals can no longer be sent from a hardware interrupt handler. Signal activity must be deferred to a soft interrupt or kthread. - As the proc state locking is simplified, it's now safe to take exit() and wait() out from under kernel_lock. - The system spends less time at IPL_SCHED, and there is less lock activity.
2008-04-24 19:35:27 +04:00
mutex_exit(proc_lock);
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
goto cleanup;
}
/* Grab a hold on the process. */
gotit = rw_tryenter(&p->p_reflock, RW_READER);
Network protocol interrupts can now block on locks, so merge the globals proclist_mutex and proclist_lock into a single adaptive mutex (proc_lock). Implications: - Inspecting process state requires thread context, so signals can no longer be sent from a hardware interrupt handler. Signal activity must be deferred to a soft interrupt or kthread. - As the proc state locking is simplified, it's now safe to take exit() and wait() out from under kernel_lock. - The system spends less time at IPL_SCHED, and there is less lock activity.
2008-04-24 19:35:27 +04:00
mutex_exit(proc_lock);
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
if (!gotit) {
error = ESRCH;
goto cleanup;
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
}
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
Merge proc::p_mutex and proc::p_smutex into a single adaptive mutex, since we no longer need to guard against access from hardware interrupt handlers. Additionally, if cloning a process with CLONE_SIGHAND, arrange to have the child process share the parent's lock so that signal state may be kept in sync. Partially addresses PR kern/37437.
2008-04-24 22:39:20 +04:00
mutex_enter(p->p_lock);
merge yamt-idlelwp branch. asked by core@. some ports still needs work. from doc/BRANCHES: idle lwp, and some changes depending on it. 1. separate context switching and thread scheduling. (cf. gmcgarry_ctxsw) 2. implement idle lwp. 3. clean up related MD/MI interfaces. 4. make scheduler(s) modular.
2007-05-17 18:51:11 +04:00
LIST_FOREACH(l2, &p->p_lwps, l_sibling) {
if (buflen >= elem_size && elem_count > 0) {
lwp_lock(l2);
fill_lwp(l2, &klwp);
lwp_unlock(l2);
Merge proc::p_mutex and proc::p_smutex into a single adaptive mutex, since we no longer need to guard against access from hardware interrupt handlers. Additionally, if cloning a process with CLONE_SIGHAND, arrange to have the child process share the parent's lock so that signal state may be kept in sync. Partially addresses PR kern/37437.
2008-04-24 22:39:20 +04:00
mutex_exit(p->p_lock);
merge yamt-idlelwp branch. asked by core@. some ports still needs work. from doc/BRANCHES: idle lwp, and some changes depending on it. 1. separate context switching and thread scheduling. (cf. gmcgarry_ctxsw) 2. implement idle lwp. 3. clean up related MD/MI interfaces. 4. make scheduler(s) modular.
2007-05-17 18:51:11 +04:00
/*
* Copy out elem_size, but not larger than
* the size of a struct kinfo_proc2.
*/
error = dcopyout(l, &klwp, dp,
min(sizeof(klwp), elem_size));
Merge from vmlocking: don't hold scheduler locks across copyout().
2007-10-08 22:09:37 +04:00
if (error) {
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
rw_exit(&p->p_reflock);
merge yamt-idlelwp branch. asked by core@. some ports still needs work. from doc/BRANCHES: idle lwp, and some changes depending on it. 1. separate context switching and thread scheduling. (cf. gmcgarry_ctxsw) 2. implement idle lwp. 3. clean up related MD/MI interfaces. 4. make scheduler(s) modular.
2007-05-17 18:51:11 +04:00
goto cleanup;
Merge from vmlocking: don't hold scheduler locks across copyout().
2007-10-08 22:09:37 +04:00
}
Merge proc::p_mutex and proc::p_smutex into a single adaptive mutex, since we no longer need to guard against access from hardware interrupt handlers. Additionally, if cloning a process with CLONE_SIGHAND, arrange to have the child process share the parent's lock so that signal state may be kept in sync. Partially addresses PR kern/37437.
2008-04-24 22:39:20 +04:00
mutex_enter(p->p_lock);
Merge from vmlocking: don't hold scheduler locks across copyout().
2007-10-08 22:09:37 +04:00
LIST_FOREACH(l3, &p->p_lwps, l_sibling) {
if (l2 == l3)
break;
}
sysctl_kern_lwp: Use a correct variable when rechecking if LWP still exists after relocking. Found via CID: 4689. OK by <dsl>.
2007-10-13 14:04:08 +04:00
if (l3 == NULL) {
Merge proc::p_mutex and proc::p_smutex into a single adaptive mutex, since we no longer need to guard against access from hardware interrupt handlers. Additionally, if cloning a process with CLONE_SIGHAND, arrange to have the child process share the parent's lock so that signal state may be kept in sync. Partially addresses PR kern/37437.
2008-04-24 22:39:20 +04:00
mutex_exit(p->p_lock);
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
rw_exit(&p->p_reflock);
Merge from vmlocking: don't hold scheduler locks across copyout().
2007-10-08 22:09:37 +04:00
error = EAGAIN;
goto cleanup;
}
merge yamt-idlelwp branch. asked by core@. some ports still needs work. from doc/BRANCHES: idle lwp, and some changes depending on it. 1. separate context switching and thread scheduling. (cf. gmcgarry_ctxsw) 2. implement idle lwp. 3. clean up related MD/MI interfaces. 4. make scheduler(s) modular.
2007-05-17 18:51:11 +04:00
dp += elem_size;
buflen -= elem_size;
elem_count--;
}
needed += elem_size;
}
Merge proc::p_mutex and proc::p_smutex into a single adaptive mutex, since we no longer need to guard against access from hardware interrupt handlers. Additionally, if cloning a process with CLONE_SIGHAND, arrange to have the child process share the parent's lock so that signal state may be kept in sync. Partially addresses PR kern/37437.
2008-04-24 22:39:20 +04:00
mutex_exit(p->p_lock);
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
/* Drop reference to process. */
rw_exit(&p->p_reflock);
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
}
if (where != NULL) {
*oldlenp = dp - where;
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
if (needed > *oldlenp) {
sysctl_relock();
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
return (ENOMEM);
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
}
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
} else {
Use KERN_PROCSLOP for struct kinfo_proc and KERN_LWPSLOP for struct kinfo_lwp, and not vice versa. Should solve the issue with top dying because it's unable to "allocate memory".
2004-02-21 06:27:57 +03:00
needed += KERN_LWPSLOP;
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
*oldlenp = needed;
}
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
error = 0;
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
cleanup:
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
sysctl_relock();
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
return (error);
}
/*
- fix comment sentence capitalization. - whitespace cleanup. No functional changes.
2007-10-16 20:05:50 +04:00
* sysctl helper routine for kern.forkfsleep node. Ensures that the
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
* given value is not too large or two small, and is at least one
* timer tick if not zero.
*/
static int
sysctl_kern_forkfsleep(SYSCTLFN_ARGS)
{
/* userland sees value in ms, internally is in ticks */
extern int forkfsleep; /* defined in kern/kern_fork.c */
int error, timo, lsleep;
struct sysctlnode node;
lsleep = forkfsleep * 1000 / hz;
node = *rnode;
node.sysctl_data = &lsleep;
error = sysctl_lookup(SYSCTLFN_CALL(&node));
if (error || newp == NULL)
return (error);
/* refuse negative values, and overly 'long time' */
if (lsleep < 0 || lsleep > MAXSLP * 1000)
return (EINVAL);
timo = mstohz(lsleep);
/* if the interval is >0 ms && <1 tick, use 1 tick */
if (lsleep != 0 && timo == 0)
forkfsleep = 1;
else
forkfsleep = timo;
return (0);
}
/*
* sysctl helper routine for kern.root_partition
*/
static int
sysctl_kern_root_partition(SYSCTLFN_ARGS)
{
int rootpart = DISKPART(rootdev);
struct sysctlnode node = *rnode;
node.sysctl_data = &rootpart;
return (sysctl_lookup(SYSCTLFN_CALL(&node)));
}
/*
* sysctl helper function for kern.drivers
*/
static int
sysctl_kern_drivers(SYSCTLFN_ARGS)
{
int error;
size_t buflen;
struct kinfo_drivers kd;
char *start, *where;
const char *dname;
int i;
extern struct devsw_conv *devsw_conv;
extern int max_devsw_convs;
start = where = oldp;
buflen = *oldlenp;
if (where == NULL) {
*oldlenp = max_devsw_convs * sizeof kd;
return 0;
}
/*
* An array of kinfo_drivers structures
*/
error = 0;
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
sysctl_unlock();
Rename specfs_lock as device_lock and move it from specfs to devsw. Relaxes kernel dependency on vfs.
2008-12-29 20:41:18 +03:00
mutex_enter(&device_lock);
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
for (i = 0; i < max_devsw_convs; i++) {
dname = devsw_conv[i].d_name;
if (dname == NULL)
continue;
if (buflen < sizeof kd) {
error = ENOMEM;
break;
}
Clear out the struct kinfo_drivers before stuffing things into it. Avoids leaking garbage from the stack (left over from the earlier call to sysctl_locate()).
2004-04-08 07:35:10 +04:00
memset(&kd, 0, sizeof(kd));
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
kd.d_bmajor = devsw_conv[i].d_bmajor;
kd.d_cmajor = devsw_conv[i].d_cmajor;
strlcpy(kd.d_name, dname, sizeof kd.d_name);
Rename specfs_lock as device_lock and move it from specfs to devsw. Relaxes kernel dependency on vfs.
2008-12-29 20:41:18 +03:00
mutex_exit(&device_lock);
Add a -t+S flag to ktrace for tracing activity related to sysctl. MIB names will be displayed, with data readen and written as well.
2006-09-24 02:01:04 +04:00
error = dcopyout(l, &kd, where, sizeof kd);
Rename specfs_lock as device_lock and move it from specfs to devsw. Relaxes kernel dependency on vfs.
2008-12-29 20:41:18 +03:00
mutex_enter(&device_lock);
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
if (error != 0)
break;
buflen -= sizeof kd;
where += sizeof kd;
}
Rename specfs_lock as device_lock and move it from specfs to devsw. Relaxes kernel dependency on vfs.
2008-12-29 20:41:18 +03:00
mutex_exit(&device_lock);
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
sysctl_relock();
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
*oldlenp = where - start;
return error;
}
implement a security.setid_core node as discussed on tech-kern@ and tech-security@.
2006-02-02 20:48:51 +03:00
static int
sysctl_security_setidcore(SYSCTLFN_ARGS)
{
int newsize, error;
struct sysctlnode node;
node = *rnode;
node.sysctl_data = &newsize;
newsize = *(int *)rnode->sysctl_data;
error = sysctl_lookup(SYSCTLFN_CALL(&node));
if (error || newp == NULL)
return error;
First take at security model abstraction. - Add a few scopes to the kernel: system, network, and machdep. - Add a few more actions/sub-actions (requests), and start using them as opposed to the KAUTH_GENERIC_ISSUSER place-holders. - Introduce a basic set of listeners that implement our "traditional" security model, called "bsd44". This is the default (and only) model we have at the moment. - Update all relevant documentation. - Add some code and docs to help folks who want to actually use this stuff: * There's a sample overlay model, sitting on-top of "bsd44", for fast experimenting with tweaking just a subset of an existing model. This is pretty cool because it's *really* straightforward to do stuff you had to use ugly hacks for until now... * And of course, documentation describing how to do the above for quick reference, including code samples. All of these changes were tested for regressions using a Python-based testsuite that will be (I hope) available soon via pkgsrc. Information about the tests, and how to write new ones, can be found on: http://kauth.linbsd.org/kauthwiki NOTE FOR DEVELOPERS: *PLEASE* don't add any code that does any of the following: - Uses a KAUTH_GENERIC_ISSUSER kauth(9) request, - Checks 'securelevel' directly, - Checks a uid/gid directly. (or if you feel you have to, contact me first) This is still work in progress; It's far from being done, but now it'll be a lot easier. Relevant mailing list threads: http://mail-index.netbsd.org/tech-security/2006/01/25/0011.html http://mail-index.netbsd.org/tech-security/2006/03/24/0001.html http://mail-index.netbsd.org/tech-security/2006/04/18/0000.html http://mail-index.netbsd.org/tech-security/2006/05/15/0000.html http://mail-index.netbsd.org/tech-security/2006/08/01/0000.html http://mail-index.netbsd.org/tech-security/2006/08/25/0000.html Many thanks to YAMAMOTO Takashi, Matt Thomas, and Christos Zoulas for help stablizing kauth(9). Full credit for the regression tests, making sure these changes didn't break anything, goes to Matt Fleming and Jaime Fournier. Happy birthday Randi! :)
2006-09-09 00:58:56 +04:00
if (kauth_authorize_system(l->l_cred, KAUTH_SYSTEM_SETIDCORE,
Don't use KAUTH_RESULT_* where it's not applicable. Prompted by yamt@.
2006-09-13 14:07:42 +04:00
0, NULL, NULL, NULL))
implement a security.setid_core node as discussed on tech-kern@ and tech-security@.
2006-02-02 20:48:51 +03:00
return (EPERM);
*(int *)rnode->sysctl_data = newsize;
return 0;
}
static int
sysctl_security_setidcorename(SYSCTLFN_ARGS)
{
int error;
sysctl_security_setidcorename: don't allocate MAXPATHLEN bytes on stack.
2006-06-17 10:54:58 +04:00
char *newsetidcorename;
implement a security.setid_core node as discussed on tech-kern@ and tech-security@.
2006-02-02 20:48:51 +03:00
struct sysctlnode node;
sysctl_security_setidcorename: don't allocate MAXPATHLEN bytes on stack.
2006-06-17 10:54:58 +04:00
newsetidcorename = PNBUF_GET();
implement a security.setid_core node as discussed on tech-kern@ and tech-security@.
2006-02-02 20:48:51 +03:00
node = *rnode;
sysctl_security_setidcorename: don't allocate MAXPATHLEN bytes on stack.
2006-06-17 10:54:58 +04:00
node.sysctl_data = newsetidcorename;
implement a security.setid_core node as discussed on tech-kern@ and tech-security@.
2006-02-02 20:48:51 +03:00
memcpy(node.sysctl_data, rnode->sysctl_data, MAXPATHLEN);
error = sysctl_lookup(SYSCTLFN_CALL(&node));
sysctl_security_setidcorename: don't allocate MAXPATHLEN bytes on stack.
2006-06-17 10:54:58 +04:00
if (error || newp == NULL) {
goto out;
}
First take at security model abstraction. - Add a few scopes to the kernel: system, network, and machdep. - Add a few more actions/sub-actions (requests), and start using them as opposed to the KAUTH_GENERIC_ISSUSER place-holders. - Introduce a basic set of listeners that implement our "traditional" security model, called "bsd44". This is the default (and only) model we have at the moment. - Update all relevant documentation. - Add some code and docs to help folks who want to actually use this stuff: * There's a sample overlay model, sitting on-top of "bsd44", for fast experimenting with tweaking just a subset of an existing model. This is pretty cool because it's *really* straightforward to do stuff you had to use ugly hacks for until now... * And of course, documentation describing how to do the above for quick reference, including code samples. All of these changes were tested for regressions using a Python-based testsuite that will be (I hope) available soon via pkgsrc. Information about the tests, and how to write new ones, can be found on: http://kauth.linbsd.org/kauthwiki NOTE FOR DEVELOPERS: *PLEASE* don't add any code that does any of the following: - Uses a KAUTH_GENERIC_ISSUSER kauth(9) request, - Checks 'securelevel' directly, - Checks a uid/gid directly. (or if you feel you have to, contact me first) This is still work in progress; It's far from being done, but now it'll be a lot easier. Relevant mailing list threads: http://mail-index.netbsd.org/tech-security/2006/01/25/0011.html http://mail-index.netbsd.org/tech-security/2006/03/24/0001.html http://mail-index.netbsd.org/tech-security/2006/04/18/0000.html http://mail-index.netbsd.org/tech-security/2006/05/15/0000.html http://mail-index.netbsd.org/tech-security/2006/08/01/0000.html http://mail-index.netbsd.org/tech-security/2006/08/25/0000.html Many thanks to YAMAMOTO Takashi, Matt Thomas, and Christos Zoulas for help stablizing kauth(9). Full credit for the regression tests, making sure these changes didn't break anything, goes to Matt Fleming and Jaime Fournier. Happy birthday Randi! :)
2006-09-09 00:58:56 +04:00
if (kauth_authorize_system(l->l_cred, KAUTH_SYSTEM_SETIDCORE,
Don't use KAUTH_RESULT_* where it's not applicable. Prompted by yamt@.
2006-09-13 14:07:42 +04:00
0, NULL, NULL, NULL)) {
sysctl_security_setidcorename: don't allocate MAXPATHLEN bytes on stack.
2006-06-17 10:54:58 +04:00
error = EPERM;
goto out;
}
if (strlen(newsetidcorename) == 0) {
error = EINVAL;
goto out;
}
implement a security.setid_core node as discussed on tech-kern@ and tech-security@.
2006-02-02 20:48:51 +03:00
memcpy(rnode->sysctl_data, node.sysctl_data, MAXPATHLEN);
sysctl_security_setidcorename: don't allocate MAXPATHLEN bytes on stack.
2006-06-17 10:54:58 +04:00
out:
PNBUF_PUT(newsetidcorename);
return error;
implement a security.setid_core node as discussed on tech-kern@ and tech-security@.
2006-02-02 20:48:51 +03:00
}
Add a new sysctl 'cp_id' that returns the array of cpu id values. Requested by me, implemented by atatat.
2005-06-16 18:55:58 +04:00
/*
- fix comment sentence capitalization. - whitespace cleanup. No functional changes.
2007-10-16 20:05:50 +04:00
* sysctl helper routine for kern.cp_id node. Maps cpus to their
Add a new sysctl 'cp_id' that returns the array of cpu id values. Requested by me, implemented by atatat.
2005-06-16 18:55:58 +04:00
* cpuids.
*/
static int
sysctl_kern_cpid(SYSCTLFN_ARGS)
{
struct sysctlnode node = *rnode;
u_intN_t -> uintN_t
2005-12-26 21:41:36 +03:00
uint64_t *cp_id = NULL;
Count the number of CPUs at boot and stash in 'ncpu'. Eventually should have each CPU register at attach, so we can figure out the topology for the scheduler.
2007-02-15 23:32:47 +03:00
int error, n = ncpu;
Add a new sysctl 'cp_id' that returns the array of cpu id values. Requested by me, implemented by atatat.
2005-06-16 18:55:58 +04:00
struct cpu_info *ci;
CPU_INFO_ITERATOR cii;
/*
- fix comment sentence capitalization. - whitespace cleanup. No functional changes.
2007-10-16 20:05:50 +04:00
* Here you may either retrieve a single cpu id or the whole
* set. The size you get back when probing depends on what
Comment in new cp_id implementation was wrong since I abandoned rewriting it in favor of some testing and then never got back to it. It's better now.
2005-06-18 03:53:21 +04:00
* you ask for.
Add a new sysctl 'cp_id' that returns the array of cpu id values. Requested by me, implemented by atatat.
2005-06-16 18:55:58 +04:00
*/
switch (namelen) {
case 0:
u_intN_t -> uintN_t
2005-12-26 21:41:36 +03:00
node.sysctl_size = n * sizeof(uint64_t);
Add a new sysctl 'cp_id' that returns the array of cpu id values. Requested by me, implemented by atatat.
2005-06-16 18:55:58 +04:00
n = -2; /* ALL */
break;
case 1:
if (name[0] < 0 || name[0] >= n)
return (ENOENT); /* ENOSUCHPROCESSOR */
u_intN_t -> uintN_t
2005-12-26 21:41:36 +03:00
node.sysctl_size = sizeof(uint64_t);
Add a new sysctl 'cp_id' that returns the array of cpu id values. Requested by me, implemented by atatat.
2005-06-16 18:55:58 +04:00
n = name[0];
/*
* adjust these so that sysctl_lookup() will be happy
*/
name++;
namelen--;
break;
default:
return (EINVAL);
}
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
cp_id = kmem_alloc(node.sysctl_size, KM_SLEEP);
Add a new sysctl 'cp_id' that returns the array of cpu id values. Requested by me, implemented by atatat.
2005-06-16 18:55:58 +04:00
node.sysctl_data = cp_id;
memset(cp_id, 0, node.sysctl_size);
for (CPU_INFO_FOREACH(cii, ci)) {
if (n <= 0)
Use cpu_index(), not ci_cpuid.
2008-05-12 18:28:22 +04:00
cp_id[0] = cpu_index(ci);
Add a new sysctl 'cp_id' that returns the array of cpu id values. Requested by me, implemented by atatat.
2005-06-16 18:55:58 +04:00
/*
* if a specific processor was requested and we just
* did it, we're done here
*/
if (n == 0)
break;
/*
* if doing "all", skip to next cp_id slot for next processor
*/
if (n == -2)
cp_id++;
/*
* if we're doing a specific processor, we're one
* processor closer
*/
if (n > 0)
n--;
}
error = sysctl_lookup(SYSCTLFN_CALL(&node));
Patch up sysctl locking: - Lock processes, credentials, filehead etc correctly. - Acquire a read hold on sysctl_treelock if only doing a query. - Don't wire down the output buffer. It doesn't work correctly and the code regularly does long term sleeps with it held - it's not worth it. - Don't hold locks other than sysctl_lock while doing copyout(). - Drop sysctl_lock while doing copyout / allocating memory in a few places. - Don't take kernel_lock for sysctl. - Fix a number of bugs spotted along the way
2008-01-07 19:12:52 +03:00
kmem_free(node.sysctl_data, node.sysctl_size);
Add a new sysctl 'cp_id' that returns the array of cpu id values. Requested by me, implemented by atatat.
2005-06-16 18:55:58 +04:00
return (error);
}
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
/*
- fix comment sentence capitalization. - whitespace cleanup. No functional changes.
2007-10-16 20:05:50 +04:00
* sysctl helper routine for hw.usermem and hw.usermem64. Values are
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
* calculate on the fly taking into account integer overflow and the
* current wired count.
*/
static int
sysctl_hw_usermem(SYSCTLFN_ARGS)
{
u_int ui;
u_quad_t uq;
struct sysctlnode node;
node = *rnode;
switch (rnode->sysctl_num) {
whitespace nit
2009-01-17 12:00:24 +03:00
case HW_USERMEM:
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
if ((ui = physmem - uvmexp.wired) > (UINT_MAX / PAGE_SIZE))
ui = UINT_MAX;
else
ui *= PAGE_SIZE;
node.sysctl_data = &ui;
break;
case HW_USERMEM64:
uq = (u_quad_t)(physmem - uvmexp.wired) * PAGE_SIZE;
node.sysctl_data = &uq;
break;
default:
return (EINVAL);
}
return (sysctl_lookup(SYSCTLFN_CALL(&node)));
}
/*
- fix comment sentence capitalization. - whitespace cleanup. No functional changes.
2007-10-16 20:05:50 +04:00
* sysctl helper routine for kern.cnmagic node. Pulls the old value
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
* out, encoded, and stuffs the new value in for decoding.
*/
static int
sysctl_hw_cnmagic(SYSCTLFN_ARGS)
{
char magic[CNS_LEN];
int error;
struct sysctlnode node;
if (oldp)
cn_get_magic(magic, CNS_LEN);
node = *rnode;
node.sysctl_data = &magic[0];
error = sysctl_lookup(SYSCTLFN_CALL(&node));
if (error || newp == NULL)
return (error);
nuke trailing whitespace
2005-02-27 00:34:55 +03:00
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
return (cn_set_magic(magic));
}
/*
* ********************************************************************
* section 3: public helper routines that are used for more than one
* node
* ********************************************************************
*/
/*
* sysctl helper routine for the kern.root_device node and some ports'
* machdep.root_device nodes.
*/
int
sysctl_root_device(SYSCTLFN_ARGS)
{
struct sysctlnode node;
node = *rnode;
split device_t/softc for all remaining drivers. replace "struct device *" with "device_t". use device_xname(), device_unit(), etc.
2012-10-27 21:17:22 +04:00
node.sysctl_data = __UNCONST(device_xname(root_device));
use device_xname() where appropriate OK martin
2008-04-05 00:13:18 +04:00
node.sysctl_size = strlen(device_xname(root_device)) + 1;
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
return (sysctl_lookup(SYSCTLFN_CALL(&node)));
}
/*
* sysctl helper routine for kern.consdev, dependent on the current
- fix comment sentence capitalization. - whitespace cleanup. No functional changes.
2007-10-16 20:05:50 +04:00
* state of the console. Also used for machdep.console_device on some
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
* ports.
*/
int
sysctl_consdev(SYSCTLFN_ARGS)
{
dev_t consdev;
provide compatibility for the older variant of kern.consdev, which used a 32 bit dev_t. Reported by mrg.
2009-08-17 00:28:19 +04:00
uint32_t oconsdev;
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
struct sysctlnode node;
if (cn_tab != NULL)
consdev = cn_tab->cn_dev;
else
consdev = NODEV;
node = *rnode;
provide compatibility for the older variant of kern.consdev, which used a 32 bit dev_t. Reported by mrg.
2009-08-17 00:28:19 +04:00
switch (*oldlenp) {
case sizeof(consdev):
node.sysctl_data = &consdev;
node.sysctl_size = sizeof(consdev);
break;
case sizeof(oconsdev):
oconsdev = (uint32_t)consdev;
node.sysctl_data = &oconsdev;
node.sysctl_size = sizeof(oconsdev);
break;
default:
return EINVAL;
}
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
return (sysctl_lookup(SYSCTLFN_CALL(&node)));
}
/*
* ********************************************************************
* section 4: support for some helpers
* ********************************************************************
*/
PR kern/37917 /bin/ps no longer shows zombies
2008-04-29 20:21:27 +04:00
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
/*
* Fill in a kinfo_lwp structure for the specified lwp.
*/
static void
fill_lwp(struct lwp *l, struct kinfo_lwp *kl)
{
merge yamt-idlelwp branch. asked by core@. some ports still needs work. from doc/BRANCHES: idle lwp, and some changes depending on it. 1. separate context switching and thread scheduling. (cf. gmcgarry_ctxsw) 2. implement idle lwp. 3. clean up related MD/MI interfaces. 4. make scheduler(s) modular.
2007-05-17 18:51:11 +04:00
struct proc *p = l->l_proc;
use binuptime for l_stime/l_rtime.
2007-12-22 04:14:53 +03:00
struct timeval tv;
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
- l_wmesg is not always valid. check l_wchan when using l_wmesg. should fix a crash reported by Juan RP on current-users@. - ttyinfo: lock lwp when accessing l_wmesg. - fill_lwp: add an assertion.
2008-04-05 18:03:16 +04:00
KASSERT(lwp_locked(l, NULL));
Set kinfo_lwp to 0 before filling it so that if someone removes variable assignments from here, kernel memory does not leak to userspace. Bug found, a little bit suprisingly, by the atf ps test which failed due to the column width between the -o holdcnt column being too wide due to the contents displayed being garbage.
2010-06-16 22:49:22 +04:00
memset(kl, 0, sizeof(*kl));
merge yamt-idlelwp branch. asked by core@. some ports still needs work. from doc/BRANCHES: idle lwp, and some changes depending on it. 1. separate context switching and thread scheduling. (cf. gmcgarry_ctxsw) 2. implement idle lwp. 3. clean up related MD/MI interfaces. 4. make scheduler(s) modular.
2007-05-17 18:51:11 +04:00
kl->l_forw = 0;
kl->l_back = 0;
Use new PTRTOUINT64() macro instead of local PTRTOINT64() macro.
2004-02-19 06:57:56 +03:00
kl->l_laddr = PTRTOUINT64(l);
kl->l_addr = PTRTOUINT64(l->l_addr);
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
kl->l_stat = l->l_stat;
kl->l_lid = l->l_lid;
Report L_INMEM in the lwp info as well.
2009-12-12 20:29:34 +03:00
kl->l_flag = L_INMEM;
kl->l_flag |= sysctl_map_flags(sysctl_lwpprflagmap, l->l_prflag);
make l_flags contain more stuff. Fixes top thread display where system processes were always displayed.
2008-07-16 02:25:30 +04:00
kl->l_flag |= sysctl_map_flags(sysctl_lwpflagmap, l->l_flag);
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
kl->l_swtime = l->l_swtime;
kl->l_slptime = l->l_slptime;
Merge newlock2 to head.
2007-02-10 00:55:00 +03:00
if (l->l_stat == LSONPROC)
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
kl->l_schedflags = l->l_cpu->ci_schedstate.spc_flags;
Merge newlock2 to head.
2007-02-10 00:55:00 +03:00
else
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
kl->l_schedflags = 0;
Merge scheduler changes from the vmlocking branch. All discussed on tech-kern: - Invert priority space so that zero is the lowest priority. Rearrange number and type of priority levels into bands. Add new bands like 'kernel real time'. - Ignore the priority level passed to tsleep. Compute priority for sleep dynamically. - For SCHED_4BSD, make priority adjustment per-LWP, not per-process.
2007-11-06 03:42:39 +03:00
kl->l_priority = lwp_eprio(l);
kl->l_usrpri = l->l_priority;
- l_wmesg is not always valid. check l_wchan when using l_wmesg. should fix a crash reported by Juan RP on current-users@. - ttyinfo: lock lwp when accessing l_wmesg. - fill_lwp: add an assertion.
2008-04-05 18:03:16 +04:00
if (l->l_wchan)
Revert rev 1.126-1.128. The original code was correct and rmind and I didn't look correctly at them.
2008-04-02 14:53:23 +04:00
strncpy(kl->l_wmesg, l->l_wmesg, sizeof(kl->l_wmesg));
Use new PTRTOUINT64() macro instead of local PTRTOINT64() macro.
2004-02-19 06:57:56 +03:00
kl->l_wchan = PTRTOUINT64(l->l_wchan);
Use cpu_index(), not ci_cpuid.
2008-05-12 18:28:22 +04:00
kl->l_cpuid = cpu_index(l->l_cpu);
use binuptime for l_stime/l_rtime.
2007-12-22 04:14:53 +03:00
bintime2timeval(&l->l_rtime, &tv);
kl->l_rtime_sec = tv.tv_sec;
kl->l_rtime_usec = tv.tv_usec;
merge yamt-idlelwp branch. asked by core@. some ports still needs work. from doc/BRANCHES: idle lwp, and some changes depending on it. 1. separate context switching and thread scheduling. (cf. gmcgarry_ctxsw) 2. implement idle lwp. 3. clean up related MD/MI interfaces. 4. make scheduler(s) modular.
2007-05-17 18:51:11 +04:00
kl->l_cpticks = l->l_cpticks;
kl->l_pctcpu = l->l_pctcpu;
kl->l_pid = p->p_pid;
if (l->l_name == NULL)
kl->l_name[0] = '\0';
When copying l_name and l_wmesg use KI_LNAMELEN and KI_WMESGLEN respectively, so that we don't care if l_name/wmesg is longer than kl_name/wmesg and the KASSERTs added in previous can go away.
2008-04-02 01:05:37 +04:00
else
Revert rev 1.126-1.128. The original code was correct and rmind and I didn't look correctly at them.
2008-04-02 14:53:23 +04:00
strlcpy(kl->l_name, l->l_name, sizeof(kl->l_name));
Dynamic sysctl. Gone are the old kern_sysctl(), cpu_sysctl(), hw_sysctl(), vfs_sysctl(), etc, routines, along with sysctl_int() et al. Now all nodes are registered with the tree, and nodes can be added (or removed) easily, and I/O to and from the tree is handled generically. Since the nodes are registered with the tree, the mapping from name to number (and back again) can now be discovered, instead of having to be hard coded. Adding new nodes to the tree is likewise much simpler -- the new infrastructure handles almost all the work for simple types, and just about anything else can be done with a small helper function. All existing nodes are where they were before (numerically speaking), so all existing consumers of sysctl information should notice no difference. PS - I'm sorry, but there's a distinct lack of documentation at the moment. I'm working on sysctl(3/8/9) right now, and I promise to watch out for buses.
2003-12-04 22:38:21 +03:00
}
Reference in New Issue Copy Permalink