sysctl: Add indirection of symbols to remove clash with sanitizers

Add indirection and symbol renaming under MKSANITIZER for the linked in
version of sysctlbyname and sysctlgetmibinfo.

branches: 1.21.30; 1.21.32;
use the proper variable

branches: 1.20.6; 1.20.12;
RUMP_ACTION -> RUMPPRG

make sysctl(8) work as a rump client

Comment out CFLAGS+=-g.

branches: 1.17.28; 1.17.30;
PPR/29909: Manuel Bouyer: sysctl dumps core if kern.consdev returns unknown
device. If we cannot determine the device name of the console, print the
console dev_t in hex.

Move sysctlbyname(), sysctlnametomib(), and sysctlgetmibinfo() from
sysctl(8) into libc, making the minor number jump. Add prototypes to
sys/sysctl.h, fix sets, modify man pages, etc. That oughta cover it.

New sysctl(8) binary. Performs auto-discovery and can add/remove
nodes from the tree. Never needs to be recompiled again.

tweaks for fparseln(3) move from libutil to libc:
- remove #include <util.h> if nothing else needed it
- remove LDFLAGS+=-lutil if nothing else needed it

Use ${NETBSDSRCDIR}/some/path instead of ${.CURDIR}/../../some/path

handle new emulation sysctls, and cleanup function duplication.

Add a "-f file" flag to process directives from a file.

don't install /usr/sbin/sysctl (backward compat symlink)

move sysctl to /sbin/sysctl. symlink from /usr/sbin/sysctl
is supplied for backward compatibility.

add /sbin/sysctl to base/mi (does not remove /usr/sbin/sysctl
as we have a symlink)

Moved from usr.sbin/sysctl/Makefile,v

s/CFLAGS/CPPFLAGS/ for -D and -I.

IPv6/IPsec sysctl MIB support.

WARNSify

enable WARNS=1 by default, but disable in unclean 3rd party code

New-style RCS ids.

specify man pages the new way.

branches: 1.1.1;
Initial revision

Move sysctlbyname(), sysctlnametomib(), and sysctlgetmibinfo() from
sysctl(8) into libc, making the minor number jump. Add prototypes to
sys/sysctl.h, fix sets, modify man pages, etc. That oughta cover it.

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22308, verified by myself.

Don't declare 'extern opt*' getopt variables.

Moved from usr.sbin/sysctl/pathconf.c,v

getopt returns -1 not EOF

New-style RCS ids.

branches: 1.1.1;
Initial revision

sysctl: Add indirection of symbols to remove clash with sanitizers

Add indirection and symbol renaming under MKSANITIZER for the linked in
version of sysctlbyname and sysctlgetmibinfo.

branches: 1.2.46; 1.2.48;
use crunchops for crunchables

RUMP_ACTION -> RUMPPRG

allow -q flag to work for reads.

Remove stale references to the "read only at securelevel [12]" flags in
the documentation and code comments.

Don't print an error with sysctl -w name?=value if name does not exist.
Should be pulled up to 5.0

branches: 1.159.2;
Reference secmodel_securelevel(9) manual page.

branches: 1.158.4;
Move TNF licenses to 2 clause form

branches: 1.157.12; 1.157.14;
Move the description of sysctl MIBs from sysctl.3 to a new manual page
sysctl.7. Remove the list of MIBs from sysctl.8 so we don't have to
maintain duplicate information, as proposed by YAMAMOTO Takashi on
tech-userlevel. Also remove references to header files from sysctl.8.

The numeric constants remain documented, they are still needed in some
cases. See the discussion on tech-userlevel. ("mib list in sysctl.8")

OK by YAMAMOTO Takashi.

branches: 1.156.2;
PR/34837: Mindaguas: Add SysV SHM dynamic reallocation and locking to the
physical memory

Initial implementation of PaX Segvguard (this is still work-in-progress,
it's just to get it out of my local tree).

Change the PaX mprotect(2) restrictions' "global_protection" knob to
just "global" -- it's shorter and more readable. Update documentation.

Import of TCP ECN algorithm for congestion control.
Both available for IPv4 and IPv6.
Basic implementation test results are available at
http://netbsd-soc.sourceforge.net/projects/ecn/testresults.html.

Work sponsored by the Google Summer of Code project 2006.
Special thanks to Kentaro Kurahone, Allen Briggs and Matt Thomas for their
help, comments and support during the project.

add net.inet.ip.maxflows. Bump date.

Bump date for previous.

document timecounter sysctls

branches: 1.149.2;
move security.setid_core.* to kern.coredump.setid.*, as requested by yamt@.

okay, since there was no way to divide this to two commits, here it goes..

introduce fileassoc(9), a kernel interface for associating meta-data with
files using in-kernel memory. this is very similar to what we had in
veriexec till now, only abstracted so it can be used more easily by more
consumers.

this also prompted the redesign of the interface, making it work on vnodes
and mounts and not directly on devices and inodes. internally, we still
use file-id but that's gonna change soon... the interface will remain
consistent.

as a result, veriexec went under some heavy changes to conform to the new
interface. since we no longer use device numbers to identify file-systems,
the veriexec sysctl stuff changed too: kern.veriexec.count.dev_N is now
kern.veriexec.tableN.* where 'N' is NOT the device number but rather a
way to distinguish several mounts.

also worth noting is the plugging of unmount/delete operations
wrt/fileassoc and veriexec.

tons of input from yamt@, wrstuden@, martin@, and christos@.

bump date (.Dd)

document Common Address Redundancy Protocol sysctls, aka CARP

ok joerg@

Introduce PaX MPROTECT -- mprotect(2) restrictions used to strengthen
W^X mappings.

Disabled by default.

First proposed in:

http://mail-index.netbsd.org/tech-security/2005/12/18/0000.html

More information in:

http://pax.grsecurity.net/docs/mprotect.txt

Read relevant parts of options(4) and sysctl(3) before using!

Lots of thanks to the PaX author and Matt Thomas.

add the 3 opencrypto sysctls.

Bump date for security.*

- make use of the recently added mode_bits for security.setid_core.mode;
- document setid_core variables.

add some more to kern.

remove dup cnmagic.

Sync and sort ddb, hw, kern, vm.

oops - this should not have been commited. remove sugid_coredump line.

grrr... another space -> tab...

space -> tab

Sync net.{inet,inet6,key}

add vm.inactivepct.

add vm.idlezero. noted by Hubert Feyrer.

bump date for the previous.

add ddb.commandonenter.

Mention "kern.bufq.strategies", bump date.

Add missing comma.

Document security level for sysctl and security.curtain.

Hi Hubert! :)

Document kern.hardclock_ticks. Pointed out by Hubert.

Drop trailing whitespace.

Mention vfs.sync.*, bump date.

Handle net.inet.tcp.debug, net.inet.tcp.debx, net.ns.spp.debug and
net.ns.spp.debx. Bump man page date.

Added net.bpf.peers and net.bpf.stats and bumped the date.

Bump date for previous. <> -> \*[Lt]\*[Gt].

Add man-page bits about the 'count' node.

Remove common code for returning supported fingerprints. This is done now
via sysctl(8) using kern.veriexec.algorithms.

Also add an entry for the 'algorithms' variable in sysctl.8 forgotten in
the last commit.

Some changes in veriexec.

New features:

- Add a veriexec_report() routine to make most reporting consistent and
remove some common code.
- Add 'strict' mode that controls how veriexec behaves.
- Add sysctl knobs:
o kern.veriexec.verbose controls verbosity levels. Value: 0, 1.
o kern.veriexec.strict controls strict level. Values: 0, 1, 2. See
documentation in sysctl(3) for details.
o kern.veriexec.algorithms returns a string with a space separated
list of supported hashing algorithms in veriexec.
- Updated documentation in man pages for sysctl(3) and sysctl(8).

Bug fixes:

- veriexec_removechk(): Code cleanup + handle FINGERPRINT_NOTEVAL
correctly.
- exec_script(): Don't pass 0 as flag when executing a script; use the
defined VERIEXEC_INDIRECT - which is 1. Makes indirect execution
enforcement work.
- Fix some printing formats and types..

branches: 1.118.2;
PR/28782: OBATA Akio: Document that kern.rtc_offset is writable.

add vfs.cd9660.utf8_joliet, and couple other vfs.* entries while here

bump date and add TNF copyright

Add vm.bufcache, vm.bufmem, vm.bufmem_lowater, m.bufmem_hiwater (PR misc/27247, misc/27233).

Bump date for removal of net.key.random_int.

remove net.key.random_int

Bump date for last.

Note net.inet.arp.* entries.

branches: 1.111.2;
Remove duplicate and superfluous words.

Bump date for last.

Bring sysctl man pages up to date (wrt new query interface, the
versioning, and descriptions).

Bump date for previous.

Document net.bpf.maxbufsize in sysctl(8).
NB: bpf isn't a PF_, so where to list it in sysctl(3)?

Used to say "type=", but now says "size=" since that's what it's
supposed to say.

Simplify Oo/Oc to Op, since it has only one simple short argument.

Update sysctl.8 man page to cover all the new stuff.

Slight option ordering change.

Forgot the date....

Add kern.drivers and kern.root_partition to match recent kern_sysctl.c
Add kern.root_device which was absent.

Combine multiple single-letter options.

add -e flag to set the separator to '=' where the default is ' = '.
this allows sysctl output to fed back into itself. inspired by
FreeBSD's sysctl(8).

ok'd by atatat.

put flags in a list with descriptions. add proper description of -w.

Mention /etc/sysctl.conf. Bump date. Closes PR 22213.

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22308, verified by myself.

Bump date for last, and replace some \*[Lt]/\*[Gt] with .Aq.

Add somaxkva.

add tcp.ident.

add checkinterface

Consistent tab usage.

Use "its" instead of "it's" where appropriate.
From Soren Jacobsen in PR 20730.

hw.physpages -> hw.physmem64, hw.userpages -> hw.usermem64

Document net.inet.tcp.init_win_local.

Document hw.physpages and hw.userpages.

Add sysconf(3) knobs for recent additions.

Add several missing items, and be less rageous about memory locking.

Add kern.forkfsleep sysctl - set/get time (in miliseconds) for which
process would be forced to sleep in fork() if it hits either global
or user maxproc limit. Default is zero (no forced sleep).
Maximum is 20 seconds.

Add two sysctls: kern.labelsector and kern.labeloffset.
These are of use to userland code which previously depended on the
hard-coded values of LABELSECTOR and LABELOFFSET to figure out the
location of the disklabel for a particular platform.

With the introduction of umbrella ports such as evbarm, evbmips, etc,
the location of the disklabel may vary between kernels for the same
MACHINE. This sysctl will allow userland programs to remain independent
of the particular flavour of MACHINE in such cases.

Added two sysctl-able flags: proc.curproc.stopfork and proc.curproc.stopexec
that can be used to block a process after fork(2) or exec(2) calls. The
new process is created in the SSTOP state and is never scheduled for running.

This feature is designed so that it is esay to attach the process using gdb
before it has done anything.

It works also with sproc, kthread_create, clone...

New sentence, new line. From Robert Elz.

Replace some \*[Lt]...\*[Gt] with .Aq ...

document vfs.nfs.iothreads

document net.inet6.ip6.maxfrags

branches: 1.77.2;
document net.key.* sysctl. provide sysctl MIB for controlling
proposal payload on ACQUIRE message. sync w/kame

rename: net.inet6.ip6.bindv6only -> net.inet6.ip6.v6only
sync w/kame.

Add -q flag, for use with -w and -f, which suppresses output after a set.

Generate <>& symbolically. I'm avoiding .../dist/... directories for now.

Document the new kern.tkstat.* sysctls.

Prod from lukem - remember to add hw.disk* here too.

update for changed vm knobs.

Sort sections, whitespace nits.

Added descriptions of the new sysctls for controlling the disposition
of IPv4 routes added via redirects, rediraccept and redirtimeout.

Remove duplicate listing of "net.inet.tcp.init_win" system variable as
reported in PR 13760 by Don Yuniskis <auryn@gci-net.com>

Add port of high performance pipe implementation written by John S. Dyson
for FreeBSD project. Besides huge speed boost compared with socketpair-based
pipes, this implementation also uses pagable kernel memory instead of mbufs.

Significant differences to FreeBSD version:
* uses uvm_loan() facility for direct write
* async/SIGIO handling correct also for sync writer, async reader
* limits settable via sysctl, amountpipekva and nbigpipes available via sysctl
* pipes are unidirectional - this is enforced on file descriptor level
for now only, the code would be updated to take advantage of it
eventually
* uses lockmgr(9)-based locks instead of home brew variant
* scatter-gather write is handled correctly for direct write case, data
is transferred by PIPE_DIRECT_CHUNK bytes maximum, to avoid running out of kva

All FreeBSD/NetBSD specific code is within appropriate #ifdef, in preparation
to feed changes back to FreeBSD tree.

This pipe implementation is optional for now, add 'options NEW_PIPE'
to your kernel config to use it.

net.inet.ip.maxfragpackets defines the maximum size of ip reass queue
(prevents fragment flood from chewing up mbuf memory space).
derived from KAME net.inet6.ip6.maxfragpackets.

add UBC memory-usage balancing. we track the number of pages in use for
each of the basic types (anonymous data, executable image, cached files)
and prevent the pagedaemon from reusing a given page if that would reduce
the count of that type of page below a sysctl-setable minimum threshold.
the thresholds are controlled via three new sysctl tunables:
vm.anonmin, vm.vnodemin, and vm.vtextmin. these tunables are the
percentages of pageable memory reserved for each usage, and we do not allow
the sum of the minimums to be more than 95% so that there's always some
memory that can be reused.

during ip6/icmp6 inbound packet processing, do not call log() nor printf() in
normal operation (/var can get filled up by flodding bogus packets).
sysctl net.inet6.icmp6.nd6_debug will turn on diagnostic messages.
(#define ND6_DEBUG will turn it on by default)

improve stats in ND6 code.

lots of synchronziation with kame (including comments and cometic ones).

Document that there's currently no registry for vendor sysctls.
After discussion with John Hawkinson and Frank van der Linden.

* Document the vendor.* sysctl branch
* in sysctl.3, sort the list of CTL_ prefixes and sync with sysctl.h

document net.inet6.icmp6.mtudisc_{lo,hi}wat.

Document new hw.cnmagic sysctl(8) variable.

add couple of missing entries, update kern.maxptys entry - it is not raise only

Ah, so _that_'s how you get a backslash in a macro argument.
Thanks to itojun for pointing it out.

Various formatting cleanups (mostly to use .Ql).
Note that getting a single backslash inside Ql seems to need SIXTEEN
of them in the source!

fix typo (remplaced --> replaced)

document kern.maxptys/KERN_MAXPTYS
note that kern.maxvnodes is raise only

implement net.inet6.ip6.{anon,low}port{min,max} sysctl variable.

document net.inet.ip.lowport{min,max}

nuke net.inet*.ip*.*ratelimit.

add net.inet.tcp.rstppslimit

document sysctl variable "net.inet.icmp.errppslimit".

add description for net.inet6.icmp6.{errppslimit,nd6_maxnudhint}

<vm/vm_param.h> is now <uvm/uvm_param.h>

branches: 1.47.4;
correct FILES section. mention IPv6/IPsec headers.

Update usage with -f.

Add a "-f file" flag to process directives from a file.

add hw.alignbytes sysctl mib. this gives you the value of ALIGNBYTES
at the kernel compilation time (ALIGNBYTES that the kernel uses).

remove net.inet6.ip6.nd6_proxyall sysctl.
support "ndp -s <ip6> <mac> proxy" for proxy NDP.

List vfs.generic.usermount in manpage. Don't try to handle machdep.diskinfo
for the i386, thus avoiding a warning message in 'sysctl -a'.

Note net.inet.icmp.errratelimit and net.inet.tcp.rstratelimit.

Add reference to the include file with TCP sysctl constants. Fixes bin/9378.

Document KERN_LOGSIGEXIT.
sort sysctl variable list in sysctl.8

Moved from usr.sbin/sysctl/sysctl.8,v

make IPV6_BINDV6ONLY setsockopt available. it controls behavior of
AF_INET6 wildcard listening socket. heavily documented in ip6(4).
net.inet6.ip6.bindv6only defines default value. default is 1.

"options INET6_BINDV6ONLY" removes any code fragment that supports
IPV6_BINDV6ONLY == 0 case (not defopt'ed as use of this is rare).

add missing variables under net.inet6.ip6.

add net.inet6.icmp6.nodeinfo sysctl.
this allows you to disable/enable ICMPv6 node information query/reply
processing (which tells remote end the gethostname(3) setting, interface
addresses on the node, and some other things - documented in
draft-ietf-ipngwg-icmp-name-lookup* or something alike).

to test it, try ping6 -w ::1 with nodeinfo=0 and nodeinfo=1.
(sync with kame change)

Fix some minor typos and word usage nits.

Typos (from OpenBSD)

document ddb.fromconsole & DBCTL_FROMCONSOLE, description taken from options(4)

Add handling of the proc hierarchy. Document it, as well as kern.defcorename.

branches: 1.30.2;
IPv6/IPsec sysctl MIB support.

Add new sysctl (net.inet.tcp.log_refused) that when set, causes refused TCP
connections to be logged.

Add support for kern.mbuf.*

Forgot to add these a while ago.

Create tcp.keepidle, tcp.keepintvl, tcp.keepcnt, tcp.slowhz sysctls.

Document shortcorename support and control.

Permit checking the availability of the POSIX File Synchronization Option
(a/k/a fsync(2)), System V style message queues, semaphores and shared memory
at runtime by adding a sysctl variable for each.

Note net.inet.tcp.ack_on_push.

Note the existence of net.inet.ip.mtudisctimeout.

Note the presence of net.inet.tcp.: mssdflt, sack, win_scale, timestamps,
compat_42, cwm, cwm_burstsize.

Fix to ensure that the correct MSS is advertised for loopback
TCP connections by using the MTU of the interface. Also added
a knob, mss_ifmtu, to force all connections to use the MTU of
the interface to calculate the advertised MSS.

macroize BSD, NetBSD, FreeBSD and misc cleanup

document net.inet.ip.anonport{min,max}

Mention net.inet.tcp.init_win.

Fix a paste-o.

branches: 1.15.2;
change sysctl net.inet.icmp.mtudisc to net.inet.ip.mtudisc

Note net.inet.ip.subnetsarelocal.

Commit userland part of pr-1891.

Document new UDP and TCP tunables.

Update the manpage for hw.machine_arch.

Fix broken uses of Dd. Both the mdoc and mdoc.samples pages agree:
.Dd is supposed to be invoked like:
.Dd month day, year
e.g. ".Dd January 25, 1989", rather than:
.Dd "month day, year"
which is what these pages did.

added net.inet.ip.allowsrcrt

Grok and document CTL_DDB.

Document the new autonice sysctls

Add a net.inet.ip.directed-broadcast sysctl as suggested by
Darren Reed <darrenr@vitruvius.arbld.unimelb.edu.au> in PR #1227.
This change is slightly different than the one submitted by Darren in
that the DIRECTED_BROADCAST compile-time option will behave like it used
to so that existing configurations utilizing it won't have to change.

Add net.inet.ip.forwsrcrt: if zero, the system will not forward
source-routed packets. Note this value is protected by kernel security
level; it can only be changed if securelevel < 1.

New-style RCS ids.

Add support for the net.inet.tcp sysctl group and document. From
John Kohl <jtk@kolvir.blrc.ma.us>.

Mention the `kern.maxpartitions' and `kern.rawpartition' variables.

branches: 1.1.1;
Initial revision

commonize case block

don't pass add=1 to print_tree() when displaying by specifying nodes.

For example, "sysctl -w kern.mbuf=" had been displayed extra node names in the MIB.
# sysctl -w kern.mbuf=
kern.mbuf.mbuf.msize = 512
kern.mbuf.mbuf.mclbytes = 2048
:

kern.hashstat takes too long and it is meaningless here.

sysctl: Add indirection of symbols to remove clash with sanitizers

Add indirection and symbol renaming under MKSANITIZER for the linked in
version of sysctlbyname and sysctlgetmibinfo.

branches: 1.161.2;
kern.boottime was changed from a struct timeval to a struct timespec
in January 2009 (the Christos' time merge, when time_t went to 64 bits).

sysctl needs to catch up. (So do other progs, which will happen, eventually,
but most of them are unaffected in any practical way.)

If you are running a system (NetBSD 6 or later) without this change, try
sysctl -nn kern.boottime
and marvel at the result (in theory, seconds.microseconds) most
probably being something like:
jinx$ sysctl -nn kern.boottime
1540801874.999995564
(There is a 1 in 1000 chance your system will have booted
in the interval [0 , 999999] nanoseconds after some second,
in which case this will not be observed. You should get
(almost) the same value after this change - just now it is as
it should be (there should now always be 9 digits after the '.').

On the other hand, if you're on a big-endian 64 bit host (running
64 bit sysctl) you would have always seen 0 for the microseconds field.
That should be fixed by this.

In sysctl(7) also document what we mean by "the time the system booted".

XXX Pullup -8
XXX Pullup -7
XXX Pullup -6 (oops, missed that one...)

branches: 1.160.2; 1.160.4;
updates for GCC 6.4:

identd has aliasing violations, use -fno-strict-aliasing.

newfs_msdos's getbpbinfo() has missing {} issues.

sysctl's kern_cp_id() has missing {} issues.

say ipsec stats are viewable with netstat.

branches: 1.158.2;
Apply patch from PR 43587, mostly from martin and kre. When writing a
sysctl variable using ?= fails with EPERM, don't print an error
message.

Ideally setting a sysctl to the same value it already has should also
not fail regardless of permissions, but this would need to be done in
the kernel.

branches: 1.157.2;
mention ip6addrctl

Add intrctl(8).

remove debugging

add handler for /kern/evcnt and print the actual error when failing.

Get rid of all sysc_init_field uses - initialize fields directly in C99
notation.

Do not loop over children if a node is not marked as CTLTYPE_NODE.

branches: 1.151.2;
use __func__, EXIT_{SUCCESS,FAILURE}

give err() some reason to avoid:
sysctl.c:2713:5: error: zero-length gnu_printf format string [-Werror=format-zero-length]

Add missing free() in error path.

use __BITMAP_TYPE

fix bug found by clang.

switch from fd_set to bitmap macros, by popular demand.

put back line I accidentally deleted.

add a handler for the reserved port sysctl variable.

branches: 1.143.2;
Add support for unsigned decimal values.

Add __printflike attribution to use vprintf and friends with an argument
as format string.

PR/44961: Jukka Ruohonen: for sysctl's with built-in handlers, return OPNOTSUPP
if we don't have handlers instead of using the handler we have and silently
failing on attempts to write a node that cannot be written.

branches: 1.140.2;
Only print errors when -A or required. Fix the error message to match others.

include <netipsec/ipsec.h> rather than <netinet6/ipsec.h> from userland
where possible, for consistency and compatibility to FreeBSD
(exception: KAME specific statistics gathering in netstat(1) and systat(1))

branches: 1.138.2;
Use __dead

Instead of using strncat and re-scanning the strings all the time, use and
append printf function.

no -A or -a with -q

allow -q flag to work for reads.

PR/44864: Paul Ripke: Compile regular expressions on demand and only once.

RUMP_ACTION -> RUMPPRG

make sysctl(8) work as a rump client

implement CTLTYPE_BOOL support. it was entirely missing. HI MATT!

Remove stale references to the "read only at securelevel [12]" flags in
the documentation and code comments.

Don't print an error with sysctl -w name?=value if name does not exist.
Should be pulled up to 5.0

fix sign-compare issues

Format the address/offset of hexdump in hex, not decimal.

Addresses my PR bin/41035

Add missing free() in machdep_diskinfo().

branches: 1.125.2;
fix dev_t format.

branches: 1.124.2;
Remove the \n and tabs from the __COPYRIGHT() strings.
(Tweak some to use a consistent format.)

branches: 1.123.2;
Convert to new 2 clause license

branches: 1.122.12; 1.122.14;
PR/35304: C J Coleman: Off-by-one nul byte overwrite in mode_bits routine
of sbin/sysctl

Exit with non-zero on errors.

branches: 1.120.2;
PR/34965: Nicolas Joly: sysctl(1) small memory leak

prefix name in kern.drivers

support printing kern.drivers

CTLFLAG_READONLY1 and CTLFLAG_READONLY2 were deprecated in earlier commit, remove them here too

move security.setid_core.* to kern.coredump.setid.*, as requested by yamt@.

Coverity CID 2784: Add more checks for value==NULL.

Coverity CID 2763: Add more checks for value==NULL.

Coverity CID 2764: Avoid null reference

Coverity CID 786: Avoid NULL dereference.

detect integer overflow differently. previous change broke negative sysctl
values.

PR/17441: John F. Woods: integer sysctl does not accept numbers > 0x7fffffff
Use unsigned int in the range comparison, and use strerror() instead of
home brewed error strings.

- make use of the recently added mode_bits for security.setid_core.mode;
- document setid_core variables.

add support for parsing file mode bits.

when printed, you'll see something like "0600 (rw-------)", like the
ls output. when reading input you can either specify octal mode (0600)
or chmod-like (u=rw).

ideas from atatat@ and kjk@; okay and lots of help from atatat@.

Handle net.inet.tcp.debug, net.inet.tcp.debx, net.ns.spp.debug and
net.ns.spp.debx. Bump man page date.

net.inet?.*.stats are viewable with netstat(1).

Inform the user that net.bpf.stats and net.bpf.peers are viewable with
netstat(1).

Move WARNS=3 to the Makefile.inc, and add a little const to the remaining
programs that did not compile before.

Add code to handle cp_id. From atatat.

PPR/29909: Manuel Bouyer: sysctl dumps core if kern.consdev returns unknown
device. If we cannot determine the device name of the console, print the
console dev_t in hex.

make sysctl -n print the real console tty name and -nn print the numeric
value.
XXX: -n means don't print name, not numeric. We should stop overloading
it and use a different flag.

If a "create" or "destroy" operation succeeds, mark the cached tree as
"stale" so that the next time we try to read or write to it, we can
purge (and refresh) it.

Addresses PR 29222.

Fix possible segmentation fault when retrieving descriptions. Thought
I committed this a while ago. I guess the fact that no one filed a pr
meant no one else found it. :)

Use regexes instead of static lists of annoying numbers to recognize
sysctl nodes that have "helpers". This is more concise, imho more
easy to understand, and has the added bonus of making it *possible* to
assign helpers to dynamically numbered nodes.

branches: 1.97.2;
Make requestors of kern.file2 be referred to pstat (the same as
kern.file) and add EINVAL to the list of errno values which are
silently ignored when walking the tree.

Pass dynamic buffer pointer to display_string(), not static buffer
pointer. Causes...misfunction if the kernel says the buffer needs to
be too much larger.

constify handlers[]

When converting a string to a number, also make sure that you didn't
convert an empty string to a zero.

Follow on to PR bin/25115 in private email.

Allocate adjusted size, not fixed size.

Some lint cleaning, strip leading (and sometimes trailing) whitespace
from values to be assigned when processing a file. Clean up error
reporting (print the name of the file and the line number), and tidy
some numeric conversions. Continue after most "errors" when
processing a -f argument.

Addresses PR bin/25115.

If the node was specifically requested but has no description, print
the "no description" message instead of skipping (as happens when
recursing into a subtree). This does not apply to non-node-type
nodes.

Don't copy the size that sysctl gave back directly, copy the size of
the description. *thwap*

Don't copy the size that sysctl gave back directly, copy the size of
the description.

Do nothing else for nodes with no children if -d is used (dflag is
set). Otherwise, some nodes (those with "printers") will also print
the "use foo ..." message.

Add -d flag to usage message(s).

branches: 1.86.2;
Make this too gcc2-compileable.

Move sysctlbyname(), sysctlnametomib(), and sysctlgetmibinfo() from
sysctl(8) into libc, making the minor number jump. Add prototypes to
sys/sysctl.h, fix sets, modify man pages, etc. That oughta cover it.

Implement sysctl descriptions. Now all that remains is actually to
write them.

Tango on sysctl_createv() and flags. The flags have all been renamed,
and sysctl_createv() now uses more arguments.

A little syntactic sugar

Make this:

% sysctl -A kern.cp_time

(and also "sysctl -A" without the kern.cp_time argument) print this:

kern.cp_time: user = 851648, nice = 683496, sys = 69383, ...
kern.cp_time.0: user = 85648, nice = 68496, sys = 6983, ...
...

instead of just:

kern.cp_time.0: user = 85648, nice = 68496, sys = 6983, ...
...

More better display of kern.cp_time for MP machines. Now we use one
sysctl() call to query for each of three different display modes:

(1) sum across all cpus

% sysctl kern.cp_time
kern.cp_time: user = 93240, nice = 1507, sys = 17252, ...

(2) data for just cpu 0

% sysctl kern.cp_time.0
kern.cp_time.0: user = 93282, nice = 1507, sys = 17264, ...

(3) each cpu individually up to hw.ncpu

% sysctl -A kern.cp_time
kern.cp_time.0: user = 93349, nice = 1507, sys = 17280, ...
kern.cp_time.1: user = 93403, nice = 1507, sys = 17291, ...
...

Miscellaneous display bugfixes.

Rename sysctlnametomib() to sysctlgetmibinfo() and add FreeBSD
compatible sysctlnametomib() and sysctlbyname() functions.

These are intended to move to libc real soon now.

Homogenize usage messages: make the 'usage' word all lowercase, as this seems
to be the most common practice in our tree.

sysctlnametomib() is a private function for the moment.

New sysctl(8) binary. Performs auto-discovery and can add/remove
nodes from the tree. Never needs to be recompiled again.

Make per-protocol network input queue stats visible to userland via
sysctl. Add a protocol-independent sysctl handler to show the per-protocol
"struct ifq' statistics. Add IP(v4) specific call to the handler.
Other protocols can show their per-protocol input statistics by
allocating a sysclt node and calling sysctl_ifq() with their own struct ifq *.

As posted to tech-kern plus improvements/cleanup suggested by Andrew Brown.

Combine multiple single-letter options.

add -e flag to set the separator to '=' where the default is ' = '.
this allows sysctl output to fed back into itself. inspired by
FreeBSD's sysctl(8).

ok'd by atatat.

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22308, verified by myself.

Remove the KERN_DRIVERS case handler (prototype code?).

Print hw.cnmagic as escaped hex bytes (unless -n given).
Change 'special' to be a numeral not a bit, it isn't ever used as a bit
and I can't see why it should ever be used that way.

Avoid a strict-alias warning in gcc 3.3.

simplify more

simplify by strlcpy/cat

Tweak behaviour introduced in rev 1.50;
* by default, warning messages go to stderr
* if -A or -a is used, send warning messages to stdout instead.

The behaviour in 1.50 meant that
somevar=`sysctl -n machdep.booted_kernel 2>/dev/null`
could end up with $somevar="machdep.booted_kernel: the value is not available"
rather than the more useful $somevar="".

(This way, "sysctl -A | grep ..." is still useful.)

Quell formwat warning for LP64 for the diskinfo printf.

display machdep.diskinfo for i386 and x86_84
(agreed by christos)

On Darwin, mach_init is the system bootstrap process. It is responsible
for forking the traditional UNIX init(8) and it does the Mach port naming
service. We need mach_init for the naming service, but unfortunately, it
will only act as such if its PID is 1. We introduce a sysctl
(emul.darwin.init_pid) to fool a given process into thinking its PID is 1.
That way we can get mach_init into behaving as the name server.

Typical use:
/sbin/sysctl -w emul.darwin.init_pid=$$ ; exec /emul/darwin/sbin/mach_init

tweaks for fparseln(3) move from libutil to libc:
- remove #include <util.h> if nothing else needed it
- remove LDFLAGS+=-lutil if nothing else needed it

Added sysctl to change all IRIX kernel values reported by uname and systeminfo:
OS name, hw name, kernel version, and so forth.

When printing out nodes of type "string", don't print anything if
sysctl() reports that 0 bytes were returned.

Reported by Matt Green with "sysctl hw.disknames" on a system with no
disks.

branches: 1.58.2;
Add -q flag, for use with -w and -f, which suppresses output after a set.

remove debugging line.

handle new emulation sysctls, and cleanup function duplication.

fix inconsistencies with USEAPP; centralize the flags check so that we
don't have to deal with it in the future.

Descend into kern.tkstat.

ANSIfy.

Fix a tyop in a comment.

Teach sysctl(8) that hw.diskstats is a structure.

change a lot of the "error" messages to display to stdout instead of
stderr, so that "sysctl -A" output is actually usable. requested by perry

show net.inet6.tcp6.* as mirror image of net.inet.tcp. better for future
INET-less kernels.

pipe sysctl stuff is now in <sys/pipe.h>

Add port of high performance pipe implementation written by John S. Dyson
for FreeBSD project. Besides huge speed boost compared with socketpair-based
pipes, this implementation also uses pagable kernel memory instead of mbufs.

Significant differences to FreeBSD version:
* uses uvm_loan() facility for direct write
* async/SIGIO handling correct also for sync writer, async reader
* limits settable via sysctl, amountpipekva and nbigpipes available via sysctl
* pipes are unidirectional - this is enforced on file descriptor level
for now only, the code would be updated to take advantage of it
eventually
* uses lockmgr(9)-based locks instead of home brew variant
* scatter-gather write is handled correctly for direct write case, data
is transferred by PIPE_DIRECT_CHUNK bytes maximum, to avoid running out of kva

All FreeBSD/NetBSD specific code is within appropriate #ifdef, in preparation
to feed changes back to FreeBSD tree.

This pipe implementation is optional for now, add 'options NEW_PIPE'
to your kernel config to use it.

add UBC memory-usage balancing. we track the number of pages in use for
each of the basic types (anonymous data, executable image, cached files)
and prevent the pagedaemon from reusing a given page if that would reduce
the count of that type of page below a sysctl-setable minimum threshold.
the thresholds are controlled via three new sysctl tunables:
vm.anonmin, vm.vnodemin, and vm.vtextmin. these tunables are the
percentages of pageable memory reserved for each usage, and we do not allow
the sum of the minimums to be more than 95% so that there's always some
memory that can be reused.

convert to use getprogname()

fix nested extern.

Number of toplevel names grew by one, account for it.

use %ll_ instead of the less standard %q_

Free storage allocated by fparseln.

Handle KERN_MSGBUF and KERN_CONSDEV gracefully.

sync with sys/sysctl.h change (is it the right way to workaround this?)

we now need <uvm/uvm_param.h>

branches: 1.37.2;
s/PIMCTL/PIM6CTL/ to avoid future confusion.

LP64 format problems. u_int64_t on alpha != %llu

cp_time[] is an array of u_int64_t's now.

branches: 1.34.2;
Handle kern.proc_args as well.

s/fprintf(stdout/printf(/

Deal with new structure returning sysctl()s.

Fix minor annoyance - when showing the boottime, the ctime() result
already has a trailing newline, so don't add an extra one.

Don't declare 'extern opt*' getopt variables.

Update usage with -f.

Fix warning.

Add a "-f file" flag to process directives from a file.

List vfs.generic.usermount in manpage. Don't try to handle machdep.diskinfo
for the i386, thus avoiding a warning message in 'sysctl -a'.

Support vm.nkmempages.

fix include file path.

Moved from usr.sbin/sysctl/sysctl.c,v

Add handling of the proc hierarchy. Document it, as well as kern.defcorename.

branches: 1.21.2;
do not include sys/netinet6/in6_systm.h, this has been empty.

IPv6/IPsec sysctl MIB support.

Add support for kern.mbuf.*

Direct the user to use nfsstat to print nfs statistics.

Implement support for the vfs. toplevel.

Fix typos lossage on the Alpha.

Include unistd.h to import declaration of getopt(3)

WARNSify

Merge with Lite2 + local changes

Tell user to use xnptdc to view kern.ntptime, as a workaround until
sysctl and libntp agree about "progname".
Based on a patch suggested by Soren S. Jorvang <soren@t.dk> in PR bin/4433.

getopt returns -1 not EOF

Grok and document CTL_DDB.

New-style RCS ids.

Add support for the net.inet.tcp sysctl group and document. From
John Kohl <jtk@kolvir.blrc.ma.us>.

typeof(timeval.tv_sec) != time_t

Correct argument in error msg. (PR #1087)

be safe with types

Print tickadj.

re-enable net sysctl code

punt networking goo for now

branches: 1.1.1;
Initial revision

sysctl: Add indirection of symbols to remove clash with sanitizers

Add indirection and symbol renaming under MKSANITIZER for the linked in
version of sysctlbyname and sysctlgetmibinfo.

branches: 1.1.46; 1.1.48;
RUMP_ACTION -> RUMPPRG

sysctl: Add indirection of symbols to remove clash with sanitizers

Add indirection and symbol renaming under MKSANITIZER for the linked in
version of sysctlbyname and sysctlgetmibinfo.

branches: 1.1.46; 1.1.48;
RUMP_ACTION -> RUMPPRG

Move sysctlbyname(), sysctlnametomib(), and sysctlgetmibinfo() from
sysctl(8) into libc, making the minor number jump. Add prototypes to
sys/sysctl.h, fix sets, modify man pages, etc. That oughta cover it.

Remove #define of __USE_NEW_SYSCTL. I'm not even sure I remember what
I used it for, but we certainly don't need it now.

This file should have a copyright notice on it.

New node version and layout. This should take care of the netbsd32
emulation problem, formalizes the versioning (should it ever be needed
again), and provides a slot for descriptions.

The new sysctl query interface returns the same information as the old
one, but you must pass in an empty node that indicates the version
you're using.

Tango on sysctl_createv() and flags. The flags have all been renamed,
and sysctl_createv() now uses more arguments.

Rename sysctlnametomib() to sysctlgetmibinfo() and add FreeBSD
compatible sysctlnametomib() and sysctlbyname() functions.

These are intended to move to libc real soon now.

Avoid using things that don't smell like they're the right size (the
size should be a multiple of sizeof(struct sysctlnode)), and avoid
leaking memory in the case of things that don't have children.

sysctlnametomib() is a private function for the moment.

New sysctl(8) binary. Performs auto-discovery and can add/remove
nodes from the tree. Never needs to be recompiled again.