]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
23 years agofix merges from 0.3f
assar [Thu, 21 Jun 2001 02:21:57 +0000 (02:21 +0000)]
fix merges from 0.3f

23 years agoThis commit was generated by cvs2svn to compensate for changes in r78527,
assar [Thu, 21 Jun 2001 02:12:07 +0000 (02:12 +0000)]
This commit was generated by cvs2svn to compensate for changes in r78527,
which included commits to RCS files with non-trunk default branches.

23 years agoimport of heimdal 0.3f
assar [Thu, 21 Jun 2001 02:12:07 +0000 (02:12 +0000)]
import of heimdal 0.3f

23 years agoConstify mdmaybeload(), add a prototype for usage(), and get rid of a
dd [Thu, 21 Jun 2001 02:10:26 +0000 (02:10 +0000)]
Constify mdmaybeload(), add a prototype for usage(), and get rid of a
sizeof(int) == sizeof(long) assumption; clamp down with WARNS=2.

23 years agoDon't assume the length of MD_NAME is 2.
dd [Thu, 21 Jun 2001 01:39:01 +0000 (01:39 +0000)]
Don't assume the length of MD_NAME is 2.

23 years agoFail if -s isn't specified for an MD_MALLOC or MD_SWAP disk; the
dd [Thu, 21 Jun 2001 01:36:09 +0000 (01:36 +0000)]
Fail if -s isn't specified for an MD_MALLOC or MD_SWAP disk; the
driver itself obviously won't configure such a disk, but the error
returned (EDOM) is more cryptic to the average user than it should be.

Also assert that the argument to -u is in fact a valid unit; don't
just accept any string to mean 0.

Approved by: phk

23 years agoAdd etc/rc.d (looking to the future)
obrien [Thu, 21 Jun 2001 01:10:30 +0000 (01:10 +0000)]
Add etc/rc.d (looking to the future)

23 years agoRemove -DNEW_LINKERSET, it is not used here anymore. This is now native.
peter [Thu, 21 Jun 2001 00:08:23 +0000 (00:08 +0000)]
Remove -DNEW_LINKERSET, it is not used here anymore.  This is now native.

23 years agoDon't lock around swap_pager_swap_init() that is only called once during
jhb [Wed, 20 Jun 2001 23:34:06 +0000 (23:34 +0000)]
Don't lock around swap_pager_swap_init() that is only called once during
the pagedaemon's startup code since it calls malloc which results in lock
order reversals.

23 years agoUse 'LIBS+= ...' instead of 'LIBS= ...' in the generated makefile
joe [Wed, 20 Jun 2001 23:21:02 +0000 (23:21 +0000)]
Use 'LIBS+= ...' instead of 'LIBS= ...' in the generated makefile
so that the user can override libraries at build time.  This is
makes it easier to cross-compile crunch builds.

Suggested by: luigi
MFC after: 3 days

23 years agoFix some lock order reversals where we called free() while holding a proc
jhb [Wed, 20 Jun 2001 23:10:06 +0000 (23:10 +0000)]
Fix some lock order reversals where we called free() while holding a proc
lock.  We now use temporary variables to save the process argument pointer
and just update the pointer while holding the lock.  We then perform the
free on the cached pointer after releasing the lock.

23 years agoAdd a missing word.
dd [Wed, 20 Jun 2001 22:10:31 +0000 (22:10 +0000)]
Add a missing word.

Submitted by: Alexander Leidinger <Alexander@Leidinger.net>

23 years agoFix a grammar bogon.
dd [Wed, 20 Jun 2001 22:03:10 +0000 (22:03 +0000)]
Fix a grammar bogon.

23 years agoDocument the mysterious Peter Wemm Murphy Field (PWMF).
jhb [Wed, 20 Jun 2001 20:58:04 +0000 (20:58 +0000)]
Document the mysterious Peter Wemm Murphy Field (PWMF).

Reviewed by: phk, eivind, nra

23 years agoRemove some more duplicate fortunes.
jhb [Wed, 20 Jun 2001 20:25:56 +0000 (20:25 +0000)]
Remove some more duplicate fortunes.

23 years agoCosmetics:
joerg [Wed, 20 Jun 2001 20:21:55 +0000 (20:21 +0000)]
Cosmetics:

. staticize out_fdc(), there's no longer an ft(4) driver sharing its use
. remove in_fdc(), has been used by ft(4) last time, long since obsoleted
  by fd_in()
. move the declaration of fd_clone() to where most of the other function
  declarations are
. de-__P()ify fd_clone(), it's been the only _P()ed function in the
  entire file

23 years ago"Unmount" is spelled "umount", not "unmount".
mph [Wed, 20 Jun 2001 20:14:36 +0000 (20:14 +0000)]
"Unmount" is spelled "umount", not "unmount".

23 years agoAvoid using too many args for ``set m[tr]u''
brian [Wed, 20 Jun 2001 19:55:58 +0000 (19:55 +0000)]
Avoid using too many args for ``set m[tr]u''

This needs to work with OpenBSD's groff too :(

23 years agoResort the last entry to its proper location. D definitely comes before Z.
jhb [Wed, 20 Jun 2001 19:55:41 +0000 (19:55 +0000)]
Resort the last entry to its proper location.  D definitely comes before Z.

23 years agoChange m_devget()'s outdated and unused `offset' argument to actually mean
bmilekic [Wed, 20 Jun 2001 19:48:35 +0000 (19:48 +0000)]
Change m_devget()'s outdated and unused `offset' argument to actually mean
something: offset into the first mbuf of the target chain before copying
the source data over.

Make drivers using m_devget() with a first argument "data - ETHER_ALIGN"
to use the offset argument to pass ETHER_ALIGN in. The way it was previously
done is potentially dangerous if the source data was at the top of a page
and the offset caused the previous page to be copied (if the
previous page has not yet been appropriately mapped).

The old `offset' argument in m_devget() is not used anywhere (it's always
0) and dates back to ~1995 (and earlier?) when support for ethernet trailers
existed. With that support gone, it was merely collecting dust.

Tested on alpha by: jlemon
Partially submitted by: jlemon
Reviewed by: jlemon
MFC after: 3 weeks

23 years agoRemove a duplicate entry.
jhb [Wed, 20 Jun 2001 19:45:15 +0000 (19:45 +0000)]
Remove a duplicate entry.

23 years agoPreemption by an interrupt thread is an involuntary switch, not a voluntary
jhb [Wed, 20 Jun 2001 18:26:41 +0000 (18:26 +0000)]
Preemption by an interrupt thread is an involuntary switch, not a voluntary
one.

Pointy-hat to: me

23 years agoApplied the patch to fix HARP ATM that was broken when KAME IPv6 came
pirzyk [Wed, 20 Jun 2001 17:09:14 +0000 (17:09 +0000)]
Applied the patch to fix HARP ATM that was broken when KAME IPv6 came
into FreeBSD.

PR: kern/23620
Submitted by: rh@matriplex.com
MFC after: 1 month

23 years agoThe serial console break-to-debugger support only functioned while
iedowse [Wed, 20 Jun 2001 16:47:23 +0000 (16:47 +0000)]
The serial console break-to-debugger support only functioned while
the console device was open. At other times, the interrupts that
are used to detect the break signal or ~^B sequence were disabled,
so these events would not be noticed until the next open (e.g. the
next kernel printf). This was mainly a problem while there was no
getty running on the console, such as during bootup or shutdown.

For serial consoles with break-to-debugger support, we now enable
the generation of interrupts at attach time, and we leave them
enabled while the device is closed.

Reviewed by: bde (I've since made chages as per his suggestions)

23 years agoConstify (silence warnings introduced by last commit to sys/module.h)
des [Wed, 20 Jun 2001 16:08:45 +0000 (16:08 +0000)]
Constify (silence warnings introduced by last commit to sys/module.h)

23 years agoUpdate the manpage to reflect reality (and what was already in -stable).
luigi [Wed, 20 Jun 2001 15:01:17 +0000 (15:01 +0000)]
Update the manpage to reflect reality (and what was already in -stable).

23 years agoo Add a -a flag for changing/getting the ALTPIN setting for a digi port.
brian [Wed, 20 Jun 2001 14:52:20 +0000 (14:52 +0000)]
o Add a -a flag for changing/getting the ALTPIN setting for a digi port.
o For the -i switch, only show the device if more than one is given on
  the command line.

23 years agoAllow individual ports to use alternate pin settings (swap dsr & cd)
brian [Wed, 20 Jun 2001 14:52:08 +0000 (14:52 +0000)]
Allow individual ports to use alternate pin settings (swap dsr & cd)
via the new DIGIIO_SETALTPIN ioctl, and allow the port's ALTPIN setting
to be queried via DIGIIO_GETALTPIN.

The initial state and lock devices are normally used to set and/or
lock ALTPIN settings although the device itself may also be used.

ALTPIN settings are applied per-device and apply to both the callin
and callout device at the same time.

23 years agoAdd DIGIIO_SETALTPIN and DIGIIO_GETALTPIN ioctl values
brian [Wed, 20 Jun 2001 14:51:58 +0000 (14:51 +0000)]
Add DIGIIO_SETALTPIN and DIGIIO_GETALTPIN ioctl values

23 years agoMove variable assignement after argument parsing, to improve handling
luigi [Wed, 20 Jun 2001 14:16:48 +0000 (14:16 +0000)]
Move variable assignement after argument parsing, to improve handling
of alternate source trees.

Also improve verbose behaviour.

23 years agoChange default value of rtadvd_enable to NO to be compatible with
ume [Wed, 20 Jun 2001 13:17:07 +0000 (13:17 +0000)]
Change default value of rtadvd_enable to NO to be compatible with
the following description in RFC2461:

    AdvSendAdvertisements
    A flag indicating whether or not the router sends
    periodic Router Advertisements and responds to
    Router Solicitations.

    Default: FALSE

    Note that AdvSendAdvertisements MUST be FALSE by
    default so that a node will not accidentally start
    acting as a router unless it is explicitly
    configured by system management to send Router
    Advertisements.

Submitted by: JINMEI Tatuya <jinmei@isl.rdc.toshiba.co.jp>
MFC after: 1 week

23 years agomade sure to use the correct sa_len for rtalloc().
ume [Wed, 20 Jun 2001 12:32:48 +0000 (12:32 +0000)]
made sure to use the correct sa_len for rtalloc().
sizeof(ro_dst) is not necessarily the correct one.
this change would also fix the recent path MTU discovery problem for the
destination of an incoming TCP connection.

Submitted by: JINMEI Tatuya <jinmei@kame.net>
Obtained from: KAME
MFC after: 2 weeks

23 years agoClose a race where we were releasing the unit resource at the start
brian [Wed, 20 Jun 2001 10:06:28 +0000 (10:06 +0000)]
Close a race where we were releasing the unit resource at the start
of tunclose() rather than the end, and tunopen() grabbed that unit
before tunclose() finished (one process is allocating it while another
is freeing it!).

It may be worth hanging some sort of rw mutex around all specinfo
calls where d_close and the detach handler get a write lock and all
other functions get a read lock.  This would guarantee certain levels
of ``atomicity'' (is that a word?) that people may expect (I believe
Solaris does something like this).

23 years agoWork around the problem in RELENG_4 where the file doesn't actually
dougb [Wed, 20 Jun 2001 07:15:38 +0000 (07:15 +0000)]
Work around the problem in RELENG_4 where the file doesn't actually
exist, and therefore mm_install is returning the "fail" value of
the test instead of the "success" value for install.

This change is a no-op on HEAD, but since the only harm on RELENG_4
ATM is a spurious warning it can follow the usual MFC practice.

Submitted by: A cast of thousands :-/

23 years agoRevert last change: it was unintended at this time.
imp [Wed, 20 Jun 2001 06:24:28 +0000 (06:24 +0000)]
Revert last change: it was unintended at this time.

23 years agoAdd new entries for:
imp [Wed, 20 Jun 2001 06:21:17 +0000 (06:21 +0000)]
Add new entries for:
TI1210, TI4410 and TI4450
and fix what looks like a typo in the OZ6860 entry.

Obtained from: NetBSD

23 years agoDocument the interactive command `what'.
mikeh [Wed, 20 Jun 2001 04:00:43 +0000 (04:00 +0000)]
Document the interactive command `what'.

Reviewed by: ru
MFC after: 2 weeks

23 years agoAdd B921600 (yes, some serial ports can do this, but generic sio not support
ache [Wed, 20 Jun 2001 03:26:41 +0000 (03:26 +0000)]
Add B921600 (yes, some serial ports can do this, but generic sio not support
them yet)

23 years agoFix typo which could lead to memory leak.
mikeh [Wed, 20 Jun 2001 01:56:09 +0000 (01:56 +0000)]
Fix typo which could lead to memory leak.

PR: misc/28283
MFC after: 2 weeks

23 years agoThis compiles cleanly with WARNS=2; make sure it stays that way.
dd [Wed, 20 Jun 2001 01:12:46 +0000 (01:12 +0000)]
This compiles cleanly with WARNS=2; make sure it stays that way.

23 years agoDon't arbitrarily limit tabstop expansions to 256 characters.
dd [Wed, 20 Jun 2001 01:08:58 +0000 (01:08 +0000)]
Don't arbitrarily limit tabstop expansions to 256 characters.

Submitted by: Oliver Fromme <oliver.fromme@heim3.tu-clausthal.de>

23 years agoPut the scheduler, vmdaemon, and pagedaemon kthreads back under Giant for
jhb [Wed, 20 Jun 2001 00:48:20 +0000 (00:48 +0000)]
Put the scheduler, vmdaemon, and pagedaemon kthreads back under Giant for
now.  The proc locking isn't actually safe yet and won't be until the proc
locking is finished.

23 years agoAdd a SEE ALSO section.
dd [Wed, 20 Jun 2001 00:37:34 +0000 (00:37 +0000)]
Add a SEE ALSO section.

23 years agoMiscellaneous mdoc(7) fixes; also expand contractions.
dd [Wed, 20 Jun 2001 00:34:20 +0000 (00:34 +0000)]
Miscellaneous mdoc(7) fixes; also expand contractions.

Submitted by: ru

23 years agoConstify the module name. This silences a few warnings ("initialization
des [Tue, 19 Jun 2001 21:40:14 +0000 (21:40 +0000)]
Constify the module name.  This silences a few warnings ("initialization
discards qualifier"), and probably adds a few where module names are
compared to or passed as non-const strings.

Not-objected-to-by: bde

23 years agoAdd configuration for a FAITH IPv6-to-IPv4 TCP translator.
ume [Tue, 19 Jun 2001 19:27:51 +0000 (19:27 +0000)]
Add configuration for a FAITH IPv6-to-IPv4 TCP translator.
To use a FAITH actually, you also need faithd(8) setup.
Please consult faithd(8) manpage.

23 years agoMade all fields in default output be space separated. Run-together
schweikh [Tue, 19 Jun 2001 17:59:40 +0000 (17:59 +0000)]
Made all fields in default output be space separated. Run-together
columns confuse the heck out of other apps trying to parse vmstat output
(eg sscope). I made sure we're still <= 80 cols per line.
Fixed warnings about unused vars and printf %format mismatches.

Requested by: Eugene Aleynikov <eugenea@infospace.com>
Reviewed by: joerg (implicitly)
MFC after: 2 weeks

23 years agoActually document TCPDEBUG.
wollman [Tue, 19 Jun 2001 17:07:15 +0000 (17:07 +0000)]
Actually document TCPDEBUG.

23 years agoFix punctuation in comment.
wollman [Tue, 19 Jun 2001 17:00:55 +0000 (17:00 +0000)]
Fix punctuation in comment.

23 years agoFix a couple of tips, and add a new one.
nik [Tue, 19 Jun 2001 16:46:58 +0000 (16:46 +0000)]
Fix a couple of tips, and add a new one.

Submitted by: Alexey V. Neyman <avn@any.ru>

23 years agoSuppress update ifnet.iflastchange when processing packets for SNMP
sumikawa [Tue, 19 Jun 2001 15:53:51 +0000 (15:53 +0000)]
Suppress update ifnet.iflastchange when processing packets for SNMP
requirements(RFC1573, interface MIB). This change for 4.4BSD was
first introduced in if_ethersubr.c:1.17->1.18.

BTW, iflastchange on all of IFs are inconsistent. e.g.
     ether, tun: update
     fddi, tokenring, ppp: not update
I'll make patch later.

Obtained from: KAME
MFC after: 2 weeks

23 years agoAdd more headers that are required with -fno-builtin (stdlib and strings)
des [Tue, 19 Jun 2001 15:41:57 +0000 (15:41 +0000)]
Add more headers that are required with -fno-builtin (stdlib and strings)

23 years agoAdd IFT_L2VLAN for supported NDP type. IPv6 over VLAN works now.
sumikawa [Tue, 19 Jun 2001 14:48:02 +0000 (14:48 +0000)]
Add IFT_L2VLAN for supported NDP type. IPv6 over VLAN works now.

Obtained from: KAME
MFC after: 2 weeks

23 years agoMore verbose version of identifycpu() which also contains many more CPU
benno [Tue, 19 Jun 2001 13:27:33 +0000 (13:27 +0000)]
More verbose version of identifycpu() which also contains many more CPU
versions/revisions.

Modified from the original patch to mark G3 and G4 processors as such.

Submitted by: Jeff Schottmiller <jeff@neoscale.com>

23 years ago<stdlib.h> is needed for exit(3) when building with -fno-builtin.
des [Tue, 19 Jun 2001 12:03:14 +0000 (12:03 +0000)]
<stdlib.h> is needed for exit(3) when building with -fno-builtin.

23 years agoConvert the elf loader to the new linker set layout for elf files.
peter [Tue, 19 Jun 2001 07:41:07 +0000 (07:41 +0000)]
Convert the elf loader to the new linker set layout for elf files.
This should make dependencies at load time work like before.  Oops.

Noticed by: markm

23 years agoFix some of the worst formatting bug (seperate commit)
peter [Tue, 19 Jun 2001 06:48:58 +0000 (06:48 +0000)]
Fix some of the worst formatting bug (seperate commit)

23 years agoThis commit was generated by cvs2svn to compensate for changes in r78460,
ru [Tue, 19 Jun 2001 06:43:49 +0000 (06:43 +0000)]
This commit was generated by cvs2svn to compensate for changes in r78460,
which included commits to RCS files with non-trunk default branches.

23 years agoBring in FSF revision 1.17:
ru [Tue, 19 Jun 2001 06:43:49 +0000 (06:43 +0000)]
Bring in FSF revision 1.17:

* src/roff/troff/node.h (node::node): Initialize `last'.

This fixes coredump on \b'\(sq' sequence.

Urged by: eivind

23 years agoCorrect some typos.
dd [Tue, 19 Jun 2001 05:46:30 +0000 (05:46 +0000)]
Correct some typos.

PR: 28251
Submitted by: Alex Kapranoff <alex@kapran.bitmcnit.bryansk.su>

23 years agostat(2) filesystem quotas act on, not filesystem quotas file resides on.
mikeh [Tue, 19 Jun 2001 04:45:50 +0000 (04:45 +0000)]
stat(2) filesystem quotas act on, not filesystem quotas file resides on.

PR: bin/4949
MFC after: 2 weeks

23 years agoCall clnt_destroy() to prevent exhausting resources.
mikeh [Tue, 19 Jun 2001 03:48:26 +0000 (03:48 +0000)]
Call clnt_destroy() to prevent exhausting resources.

PR: bin/14255
Reviewed by: Kenji Tomita <tommy@ti.com>
MFC after: 2 weeks

23 years agoCleanup the mix of styles in this file:
mikeh [Tue, 19 Jun 2001 03:37:21 +0000 (03:37 +0000)]
Cleanup the mix of styles in this file:
  - spaces -> tabs
  - return(0) -> return (0)
  - wrap longlines

Reviewed by: md5
MFC after: 2 weeks

23 years agoThis command has more than four options.
dd [Tue, 19 Jun 2001 03:34:03 +0000 (03:34 +0000)]
This command has more than four options.

23 years agoAdd IBM "High Rate Wireless LAN PC Card", a rebadged Lucent WaveLAN/IEEE.
grog [Tue, 19 Jun 2001 01:21:22 +0000 (01:21 +0000)]
Add IBM "High Rate Wireless LAN PC Card", a rebadged Lucent WaveLAN/IEEE.

MFC after: 2 weeks

23 years agoPrint "-" instead of "" in the grace field if there is no grace
dd [Tue, 19 Jun 2001 00:47:24 +0000 (00:47 +0000)]
Print "-" instead of "" in the grace field if there is no grace
period; this keeps the number of whitespace-separated fields constant.

PR: 15739, 23908
Submitted by: Valentin Nechayev <netch@netch.kiev.ua>

23 years agoCND has been unused for two years, and RND has been unused for five
dd [Tue, 19 Jun 2001 00:39:29 +0000 (00:39 +0000)]
CND has been unused for two years, and RND has been unused for five
years; remove them from CFLAGS.

PR: 23712
Submitted by: Volker Stolz <stolz@i2.informatik.rwth-aachen.de>

23 years agoHook mdmfs into the build.
dd [Mon, 18 Jun 2001 23:47:18 +0000 (23:47 +0000)]
Hook mdmfs into the build.

23 years agoIntroduce mdmfs(8), a wrapper around mdconfig(8), disklabel(8),
dd [Mon, 18 Jun 2001 23:46:58 +0000 (23:46 +0000)]
Introduce mdmfs(8), a wrapper around mdconfig(8), disklabel(8),
newfs(8), and mount(8) that mimics the command line option set of the
deprecated mount_mfs(8).

Approved by: jkh, phk, -hackers

23 years agoadd ircs, per IANA
dbaker [Mon, 18 Jun 2001 23:33:09 +0000 (23:33 +0000)]
add ircs, per IANA

MFC after: 3 days

23 years agoCall va_start/va_end for each call to vfprintf
brian [Mon, 18 Jun 2001 23:24:26 +0000 (23:24 +0000)]
Call va_start/va_end for each call to vfprintf

Obtained from: OpenBSD

23 years agoFix spelling of `errno' and add cross-references.
wollman [Mon, 18 Jun 2001 22:26:31 +0000 (22:26 +0000)]
Fix spelling of `errno' and add cross-references.

23 years agoDocument _SC_IOV_MAX. Add cross-references to pathconf(2), confstr(3), and
wollman [Mon, 18 Jun 2001 22:18:08 +0000 (22:18 +0000)]
Document _SC_IOV_MAX.  Add cross-references to pathconf(2), confstr(3), and
getconf(1).  This document still needs to be sorted and to have other missing
parameters described.

23 years agoFix some memory bugs with regard to jumbo buffers. I made a mistake when
wpaul [Mon, 18 Jun 2001 22:04:40 +0000 (22:04 +0000)]
Fix some memory bugs with regard to jumbo buffers. I made a mistake when
converting from the old external mbuf buffer code to the new (with the
MEXTADD() macro). Also free free list memory correctly in
foo_free_jumbo_mem() routines: grab the head of the list, then
remove it, _then_ free() it.

This fixes the memory corruption problem I've been chasing in the level 1
driver.

23 years agoWhen processing -a, strip off the path of the argument before copying
dd [Mon, 18 Jun 2001 21:57:24 +0000 (21:57 +0000)]
When processing -a, strip off the path of the argument before copying
it to the output.

Reviewed by: alex

23 years agoAdded CTLTYPE_{UINT|LONG|ULONG} types.
pirzyk [Mon, 18 Jun 2001 21:07:04 +0000 (21:07 +0000)]
Added CTLTYPE_{UINT|LONG|ULONG} types.

PR: kern/21132
Reviewed by: no objections to by -arch
MFC after: 1 month

23 years agomodfied sysctl command to allow setting values > 2GB if the mib supports it.
pirzyk [Mon, 18 Jun 2001 21:06:24 +0000 (21:06 +0000)]
modfied sysctl command to allow setting values > 2GB if the mib supports it.

PR: kern/21132
Reviewed by: no objections to by -arch
MFC after: 1 month

23 years agoTypo/mind-o fixes.
wpaul [Mon, 18 Jun 2001 20:57:03 +0000 (20:57 +0000)]
Typo/mind-o fixes.

23 years agoCorrect getopt usage and a syntax error (period used instead of a semicolon).
dd [Mon, 18 Jun 2001 20:36:06 +0000 (20:36 +0000)]
Correct getopt usage and a syntax error (period used instead of a semicolon).

23 years agoAfter one too many PRs on the subject, bite the bullet and define IOV_MAX
wollman [Mon, 18 Jun 2001 20:24:54 +0000 (20:24 +0000)]
After one too many PRs on the subject, bite the bullet and define IOV_MAX
and its associated constants.  Implement _SC_IOV_MAX in the usual way.
Be a bit sloppy about the namespace question; this should get cleared up
in time for 5.0.

MFC after: 1 month

23 years agoXref zcat(1).
dd [Mon, 18 Jun 2001 20:00:39 +0000 (20:00 +0000)]
Xref zcat(1).

23 years agostop csa from panicing in clkrun_hack() - we were using free'd memory
cg [Mon, 18 Jun 2001 19:58:03 +0000 (19:58 +0000)]
stop csa from panicing in clkrun_hack() - we were using free'd memory

don't leak memory in clkrun_hack()

Submitted by: grog (partially)

23 years agoLock Giant in postsig() for the KTRACE case as ktrpsig() needs Giant when
jhb [Mon, 18 Jun 2001 19:23:43 +0000 (19:23 +0000)]
Lock Giant in postsig() for the KTRACE case as ktrpsig() needs Giant when
it writes out to the trace file.

Reported by: peter, gallatin, and others

23 years agoInitialize mutexes needed early on all in the same place so that the
jhb [Mon, 18 Jun 2001 19:19:38 +0000 (19:19 +0000)]
Initialize mutexes needed early on all in the same place so that the
startup routine more closely matches that of alpha and ia64.  At some
point the common mutexes shared across all platforms probably should move
into sys/kern_mutex.c.

23 years ago- Add support for decoding syscall names. (Brought over from the new alpha
jhb [Mon, 18 Jun 2001 19:17:30 +0000 (19:17 +0000)]
- Add support for decoding syscall names.  (Brought over from the new alpha
  trace code that was brought over from NetBSD.)
- Check for "syscall_with_err_pushed" as the label prior to a syscall trap
  frame rather than "Xlcall_syscall" and "Xint0x80_syscall".  We don't
  have a valid trapframe during the short range of code that those two
  symbols now cover.
- Simplify db_next_frame() to avoid duplicating the code for the different
  trap frame types.
- Don't try to trace a swapped-out process.  (Brought over from NetBSD via
  the new alpha trace code.)

23 years agoInclude sys/pcpu.h to get the prototype for globaldata_register() to quiet
jhb [Mon, 18 Jun 2001 19:06:14 +0000 (19:06 +0000)]
Include sys/pcpu.h to get the prototype for globaldata_register() to quiet
a warning.

23 years agoCall bus_teardown_intr when csa_attach fails after the interrupt has
tmm [Mon, 18 Jun 2001 18:36:34 +0000 (18:36 +0000)]
Call bus_teardown_intr when csa_attach fails after the interrupt has
been set up.

PR: kern/28178
Reviewed by: cg
MFC after: 2 weeks

23 years agoIf the username we are trying to look up in the db files won't fit into
dwmalone [Mon, 18 Jun 2001 16:09:24 +0000 (16:09 +0000)]
If the username we are trying to look up in the db files won't fit into
the buffer then act like it doesn't exist. The buffer is always big enough
for any valid username.

PR: 27860
Reviewed by: nectar

23 years agoPick up the correct headers from sys/contrib/ipfilter/netinet.
ru [Mon, 18 Jun 2001 15:55:28 +0000 (15:55 +0000)]
Pick up the correct headers from sys/contrib/ipfilter/netinet.

23 years agoRemoved duplicate copies of files already present in sys/contrib/ipfilter.
ru [Mon, 18 Jun 2001 15:54:04 +0000 (15:54 +0000)]
Removed duplicate copies of files already present in sys/contrib/ipfilter.

Not objected to by: -arch

23 years agoUse linker_reference_module() instead of hard-coding the digi_*
brian [Mon, 18 Jun 2001 15:10:20 +0000 (15:10 +0000)]
Use linker_reference_module() instead of hard-coding the digi_*
module path.

23 years agoAdd linker_reference_module().
brian [Mon, 18 Jun 2001 15:09:33 +0000 (15:09 +0000)]
Add linker_reference_module().

This function loads a module if required, otherwise bumps the reference
count -- the opposite of linker_file_unload().

23 years agoCredit John Prince and Eric Hernes for their work.
brian [Mon, 18 Jun 2001 15:02:33 +0000 (15:02 +0000)]
Credit John Prince and Eric Hernes for their work.

23 years agoAdd support for stateful MPPE (microsoft encryption) providing
brian [Mon, 18 Jun 2001 15:00:22 +0000 (15:00 +0000)]
Add support for stateful MPPE (microsoft encryption) providing
encryption compatibility with Windows 2000.  Stateful encryption
uses less CPU but is bad on lossy transports.

The ``set mppe'' command has been expanded.  If it's used with any
arguments, ppp will insist on encryption, closing LCP if the other
end refuses.

Unfortunately, Microsoft have abused the CCP reset request so that
receiving a reset request does not result in a reset ack when using
MPPE...

Sponsored by: Monzoon Networks AG and FreeBSD Services Limited

23 years agoHandle hardware-imposed MTU/MRU limitations. PPPoE will no longer
brian [Mon, 18 Jun 2001 14:59:36 +0000 (14:59 +0000)]
Handle hardware-imposed MTU/MRU limitations.  PPPoE will no longer
allow MRU/MTU negotiations to exceed 1492.

Add an optional ``max'' specifier to ``set m[rt]u'', ie.

set mtu max 1480

Bump the ppp version number.

Sponsored by: Monzoon Networks AG and FreeBSD Services Limited

23 years agoMake this really work.
ru [Mon, 18 Jun 2001 12:35:14 +0000 (12:35 +0000)]
Make this really work.

23 years agocall pfxlist_onlink_check() at the end of in6_tmpifadd(), to make sure
ume [Mon, 18 Jun 2001 11:37:06 +0000 (11:37 +0000)]
call pfxlist_onlink_check() at the end of in6_tmpifadd(), to make sure
a temporary address generated from a detached public one also detached.

Submitted by: JINMEI Tatuya <jinmei@isl.rdc.toshiba.co.jp>
Obtained from: KAME

23 years agoDon't remove the SI_CHEAPCLONE for unsupported minors
brian [Mon, 18 Jun 2001 09:22:30 +0000 (09:22 +0000)]
Don't remove the SI_CHEAPCLONE for unsupported minors

23 years agoRemove the SI_CHEAPCLONE flag when hanging resources off the dev_t
brian [Mon, 18 Jun 2001 09:21:59 +0000 (09:21 +0000)]
Remove the SI_CHEAPCLONE flag when hanging resources off the dev_t

23 years agoAdd a -d command-line option; when used in conjunction with -f, rarpd
roam [Mon, 18 Jun 2001 06:48:33 +0000 (06:48 +0000)]
Add a -d command-line option; when used in conjunction with -f, rarpd
sends error messages to stderr, normal output to stdout, instead of
logging everything via syslog.

Turn off the FORMAT_AUDIT in the Makefile, until I can figure out how
to disable the check for one single line in the source :(

Reviewed by: dd, silence on -audit
MFC after: 1 month