]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
23 years agoDocument the PANIC_REBOOT_WAIT_TIME option.
dd [Sun, 10 Jun 2001 00:30:49 +0000 (00:30 +0000)]
Document the PANIC_REBOOT_WAIT_TIME option.

PR: 22228
Submitted by: Keith Jones <keith@mithy.demon.co.uk>

23 years agofdc doesn't support talking to QIC40/QIC80 tape drives, and hasn't since
imp [Sun, 10 Jun 2001 00:27:18 +0000 (00:27 +0000)]
fdc doesn't support talking to QIC40/QIC80 tape drives, and hasn't since
ft was removed from the tree prior to branching releng_3!

23 years agoAdd a comma between the two names in the CONTRIBUTERS section.
dd [Sun, 10 Jun 2001 00:23:05 +0000 (00:23 +0000)]
Add a comma between the two names in the CONTRIBUTERS section.

23 years agoTwo fixes to the out-of-swap process termination code. First, start killing
dillon [Sat, 9 Jun 2001 18:06:58 +0000 (18:06 +0000)]
Two fixes to the out-of-swap process termination code.   First, start killing
processes a little earlier to avoid a deadlock.  Second, when calculating
the 'largest process' do not just count RSS.  Instead count the RSS + SWAP
used by the process.  Without this the code tended to kill small
inconsequential processes like, oh, sshd, rather then one of the many
'eatmem 200MB' I run on a whim :-).  This fix has been extensively tested on
-stable and somewhat tested on -current and will be MFCd in a few days.

Shamed into fixing this by: ps

23 years agofix RCS ID style nit
obrien [Sat, 9 Jun 2001 16:54:11 +0000 (16:54 +0000)]
fix RCS ID style nit

23 years agoFirst approximation of an ia64 EFI loader. Not functional.
dfr [Sat, 9 Jun 2001 16:49:51 +0000 (16:49 +0000)]
First approximation of an ia64 EFI loader. Not functional.

23 years agoSimplify, with the power of new mdoc(7).
ru [Sat, 9 Jun 2001 11:20:31 +0000 (11:20 +0000)]
Simplify, with the power of new mdoc(7).

23 years agoFix /etc/rc so that all remote filesystems, not just NFS, are mounted
sheldonh [Sat, 9 Jun 2001 09:47:05 +0000 (09:47 +0000)]
Fix /etc/rc so that all remote filesystems, not just NFS, are mounted
after the network is configured.

Future remote filesystem types should be added to remote_fstypes, which
is a whitespace-delimited list of type:description pairs.

The specific problem this fixes right now is that smbfs mounts described
in /etc/ftsab were failing.

23 years agoBump __FreeBSD_version to reflect machine/ioctl_fd.c moving to sys/fdcio.h
brian [Sat, 9 Jun 2001 09:09:45 +0000 (09:09 +0000)]
Bump __FreeBSD_version to reflect machine/ioctl_fd.c moving to sys/fdcio.h

23 years agoFirst stab at adding back in CL-PD6729 support.
imp [Sat, 9 Jun 2001 07:34:17 +0000 (07:34 +0000)]
First stab at adding back in CL-PD6729 support.

23 years agoAdd PC9801-102 CBUS card to the list of plug and play devices. Some
imp [Sat, 9 Jun 2001 06:43:54 +0000 (06:43 +0000)]
Add PC9801-102 CBUS card to the list of plug and play devices.  Some
mapping of irq 6 may be required to use that irq, but if so,
additional commits will follow.

Submitted by: Hiroshi TSUKADA-san

23 years agoAdd -j to the usage string.
dd [Sat, 9 Jun 2001 06:14:05 +0000 (06:14 +0000)]
Add -j to the usage string.

PR: 27986
Submitted by: Cyrille Lefevre <clefevre@redirect.to>

23 years agoID style nit.
obrien [Sat, 9 Jun 2001 05:24:14 +0000 (05:24 +0000)]
ID style nit.

23 years agoUpdate copyright.
obrien [Sat, 9 Jun 2001 05:22:31 +0000 (05:22 +0000)]
Update copyright.

23 years agoFix style of defines.
obrien [Sat, 9 Jun 2001 05:21:17 +0000 (05:21 +0000)]
Fix style of defines.

23 years agoStyle fix FreeBSD ID, and change continuation style slightly.
obrien [Sat, 9 Jun 2001 05:01:24 +0000 (05:01 +0000)]
Style fix FreeBSD ID, and change continuation style slightly.

23 years agoSince we use getopt (in rev. 1.12), there's no need for the case's for
dd [Sat, 9 Jun 2001 04:32:46 +0000 (04:32 +0000)]
Since we use getopt (in rev. 1.12), there's no need for the case's for
the individual options to increment argv and decrement argc.  This
caused the -T option to swallow an extra argument.

PR: 27982
Submitted by: Samuel Greear <sgreear@vsni.com>

23 years agoNormalize the use of sizeof according to style(9).
dd [Sat, 9 Jun 2001 03:56:16 +0000 (03:56 +0000)]
Normalize the use of sizeof according to style(9).

Reviewed by: md5(1)

23 years agodumpmag is 'u_long dumpmag' in the kernel, not 'int'. If this worked on
peter [Sat, 9 Jun 2001 01:41:03 +0000 (01:41 +0000)]
dumpmag is 'u_long dumpmag' in the kernel, not 'int'.  If this worked on
the Alpha, it was because of luck that it was little endian.

Tidy up the dumpsize/dumppages confusion and the out of date comments.

23 years agoXref dladdr(3).
dd [Fri, 8 Jun 2001 22:50:33 +0000 (22:50 +0000)]
Xref dladdr(3).

Requested by: -hackers

23 years agoSwitch to the user's uid before attempting to unlink the auth forwarding
green [Fri, 8 Jun 2001 22:22:09 +0000 (22:22 +0000)]
Switch to the user's uid before attempting to unlink the auth forwarding
file, nullifying the effects of a race.

Obtained from: OpenBSD

23 years agoFix a "this sentence no verb" grammar bogon.
schweikh [Fri, 8 Jun 2001 21:20:14 +0000 (21:20 +0000)]
Fix a "this sentence no verb" grammar bogon.
MFC after: 1 week

23 years agoNew release notes: OLDCARD PCI support, savecore(8) 2GB RAM fix,
bmah [Fri, 8 Jun 2001 19:18:50 +0000 (19:18 +0000)]
New release notes:  OLDCARD PCI support, savecore(8) 2GB RAM fix,
inetd(8) faith syntax, ident support in inetd(8).

MFCs noted:  Linux-PAM 0.75

Move PAM authenticators to be in the right place and update.

23 years agoMFS: Apply edits from 4-STABLE errata file to -CURRENT. Also add some
bmah [Fri, 8 Jun 2001 18:35:09 +0000 (18:35 +0000)]
MFS:  Apply edits from 4-STABLE errata file to -CURRENT.  Also add some
comments and notes to the effect that this file won't have any real
content until after 5.0-RELEASE.

23 years agoGo ahead and request 0x44000000 through 0xfffffff instead of just
imp [Fri, 8 Jun 2001 18:31:51 +0000 (18:31 +0000)]
Go ahead and request 0x44000000 through 0xfffffff instead of just
0xefffffff

# Note, this is bogus, but less bogus than before.

23 years agoAlways print at least 2 bytes for IN_CLASSB_NET networks.
ru [Fri, 8 Jun 2001 15:44:17 +0000 (15:44 +0000)]
Always print at least 2 bytes for IN_CLASSB_NET networks.
Always print at least 3 bytes for IN_CLASSC_NET networks.

The standard 193.0.0 class C network for example, will now
be displayed as "193.0.0" as opposed to the confusing 193.

PR: bin/21546
MFC after: 1 week

23 years agoForced commit:
bmah [Fri, 8 Jun 2001 15:31:54 +0000 (15:31 +0000)]
Forced commit:

The last commit to this file introduced some text that was "Copyright (C)
2000 BSDI".  It was asserted (correctly) that permission should have been
sought from someone at BSDi / WindRiver first.

In response to a query on this point, jkh wrote:

> Hmmm.  I'm not sure we ever asserted any particular copyright, but in
> any case it's certainly not clear who one would ask.  Me?  If so,
> I give my permission. :)

Astutely pointed out by: murray
Approved by: jkh

23 years agoSync up to v 1.9 of NetBSD's db_trace.c to get access to the
gallatin [Fri, 8 Jun 2001 13:38:02 +0000 (13:38 +0000)]
Sync up to v 1.9 of NetBSD's db_trace.c to get access to the
greatly improved traceback code from Ross Harvey.  This code
requires the use of more traceback friendly temporary labels
at kernel entry points, hence the changes to exception.s and
asm.h

Reviewed by: jhb, dfr
Obtained from: NetBSD
MFC after: 1 week

23 years agoWhen changing an indirect route, kernel routing code allocates
ru [Fri, 8 Jun 2001 12:44:25 +0000 (12:44 +0000)]
When changing an indirect route, kernel routing code allocates
a route to the gateway and caches it in the route structure.
It may happen (if the routing table is screwed) that the gateway
route is the same route as the one being modified, in which case
a kernel reports EDQUOT.  Be more verbose about this:

# route add -net 10 192.168.4.65
add net 10: gateway 192.168.4.65
# netstat -rn -finet
Routing tables

Internet:
Destination        Gateway            Flags     Refs     Use     Netif Expire
default            192.168.4.65       UGSc        1        7      rl0
10                 192.168.4.65       UGSc        0        0      rl0
127.0.0.1          127.0.0.1          UH          0      178      lo0
192.168.4          link#1             UC          2        0      rl0 =>
192.168.4.65       0:d0:b7:16:9c:c6   UHLW        2        0      rl0   1123

Before:

# route change -net 10 10.0.0.1
route: writing to routing socket: Disc quota exceeded
change net 10: gateway 10.0.0.1: Disc quota exceeded

After:

# ./route change -net 10 10.0.0.1
route: writing to routing socket: Disc quota exceeded
change net 10: gateway 10.0.0.1: gateway uses the same route

PR: bin/1093, misc/26833

23 years agoCatch up to csareg.h rev 1.3.
tanimura [Fri, 8 Jun 2001 11:57:39 +0000 (11:57 +0000)]
Catch up to csareg.h rev 1.3.

Forgotten by: cg

23 years agoThe date for Australia day in
phk [Fri, 8 Jun 2001 11:53:58 +0000 (11:53 +0000)]
The date for Australia day in
src/usr.bin/caldendar/calenders/calendar.holiday
is incorrect.
From looking through webcvs it seems like the error is in Open/NetBSD also.

PR: 27960
Submitted by: Harley Anderson <Warragul@selcon.com.au>

23 years agoAdd support for yet another Promise ATA100 variant
sos [Fri, 8 Jun 2001 09:51:33 +0000 (09:51 +0000)]
Add support for yet another Promise ATA100 variant

Minor fix to the VIA setup code.

23 years agoChange the host/bits syntax introduced in route.c,v 1.24 to the
ru [Fri, 8 Jun 2001 09:07:34 +0000 (09:07 +0000)]
Change the host/bits syntax introduced in route.c,v 1.24 to the
net/bits syntax, for consistency with netstat(1) in particular.

OK'ed by: phk
MFC after: 1 week

23 years agoThe TI-1031 is more like the TI-113x chips rather than the 12xx or
imp [Fri, 8 Jun 2001 07:16:56 +0000 (07:16 +0000)]
The TI-1031 is more like the TI-113x chips rather than the 12xx or
higher chips.  Treat it as if it were a 113x.  This is correct as far
as 16-bit cards go, at least how we're using it.

# It appears that my TI-1031 based pci card that YAMAMOTO shigeru-san gave
# me on my trip to Japan now works.

23 years ago"Fix" the previous initial attempt at fixing TUNABLE_INT(). This time
peter [Fri, 8 Jun 2001 05:24:21 +0000 (05:24 +0000)]
"Fix" the previous initial attempt at fixing TUNABLE_INT().  This time
around, use a common function for looking up and extracting the tunables
from the kernel environment.  This saves duplicating the same function
over and over again.  This way typically has an overhead of 8 bytes + the
path string, versus about 26 bytes + the path string.

23 years agoCorrect a typo.
dd [Fri, 8 Jun 2001 04:41:21 +0000 (04:41 +0000)]
Correct a typo.

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

23 years agoRemove xref to blackhole(4). sysctl has nothing to do with
dd [Fri, 8 Jun 2001 03:12:32 +0000 (03:12 +0000)]
Remove xref to blackhole(4).  sysctl has nothing to do with
blackhole(4), except that blackhole(4) uses sysctl's.  This xref
obviously isn't appropriate unless we want to xref all the other man
pages which mention sysctls, which we obviously don't (we may want to
list those sysctls, but that's another story).

PR: 27937
Submitted by: yar

23 years ago'int dumpsize; /* memory size in bytes */' is not good for machines
peter [Fri, 8 Jun 2001 03:06:24 +0000 (03:06 +0000)]
'int dumpsize; /* memory size in bytes */' is not good for machines
with 2GB or more of ram.

23 years agoAdd Alpha User's Q&A section to the installation guide.
bmah [Thu, 7 Jun 2001 22:38:46 +0000 (22:38 +0000)]
Add Alpha User's Q&A section to the installation guide.

Obtained from: 4.3-RELEASE CD-ROM liner notes
MFC after: 3 days

23 years agocleanup to get rid of most warnings on alpha
tomsoft [Thu, 7 Jun 2001 20:43:00 +0000 (20:43 +0000)]
cleanup to get rid of most warnings on alpha
and yes now it also works on alpha

Reviewed by: chm
MFC after: 3 weeks

23 years agoenable vchan building
cg [Thu, 7 Jun 2001 20:16:17 +0000 (20:16 +0000)]
enable vchan building

23 years agoenable vchan compilation
cg [Thu, 7 Jun 2001 20:12:11 +0000 (20:12 +0000)]
enable vchan compilation

23 years agolock sound device when adding/removing channels
cg [Thu, 7 Jun 2001 20:06:22 +0000 (20:06 +0000)]
lock sound device when adding/removing channels
implement setblocksize for vchans
don't panic when doing certain ioctls or aborting on a vchan
xmms now works with vchans

23 years agoFix the number of bytes allocated by realloc when more space is needed
ghelmer [Thu, 7 Jun 2001 20:05:18 +0000 (20:05 +0000)]
Fix the number of bytes allocated by realloc when more space is needed
for the vector of arguments.

MFC after: 1 week

23 years agoCreate curs_termcap.3 --> termcap.3 MLINK, so many manpages in our tree that
sobomax [Thu, 7 Jun 2001 17:27:44 +0000 (17:27 +0000)]
Create curs_termcap.3 --> termcap.3 MLINK, so many manpages in our tree that
reference to termcap.3 will not be pointing to nowhere.

MFC after: 1 week

23 years agoCorrect cross-reference:
sobomax [Thu, 7 Jun 2001 16:59:19 +0000 (16:59 +0000)]
Correct cross-reference:
portmap.8 --> rpcbind.8

Submitted by: .Xr testing script

23 years agos/format/byte order/ (for clarity).
ru [Thu, 7 Jun 2001 15:15:38 +0000 (15:15 +0000)]
s/format/byte order/ (for clarity).

23 years agoadd new directories for XFree86-4.1.0.
sf [Thu, 7 Jun 2001 14:46:55 +0000 (14:46 +0000)]
add new directories for XFree86-4.1.0.

Reviewed by: ru

23 years agoAdd 'df' string to gettytab - the strftime(3) format for %d
yar [Thu, 7 Jun 2001 13:53:23 +0000 (13:53 +0000)]
Add 'df' string to gettytab - the strftime(3) format for %d
in the banner messages (of course, defaults to "%+").

Submitted by: Gleb Smirnoff <glebius@tak.estra.ru>

23 years ago- Exit 1 if "add", "change", or "delete" operation fails.
ru [Thu, 7 Jun 2001 13:50:24 +0000 (13:50 +0000)]
- Exit 1 if "add", "change", or "delete" operation fails.

PR: bin/12489

- Use inet_ntoa(3) where it should have been used.  This
  part of code simply wasn't converted to the "new" style
  after the routename() function was converted from the
  protocol-generic version to protocol-specific version
  in CSRG revision 5.6.

MFC after: 1 week

23 years ago- Renamed the ``proxy'' modifier of the -d flag to ``pub'',
ru [Thu, 7 Jun 2001 12:35:53 +0000 (12:35 +0000)]
- Renamed the ``proxy'' modifier of the -d flag to ``pub'',
  to be consistent with the -s flag.  Updated documentation
  on what this modifier does.

- Added the ``only'' keyword to the -s and -S flags, that
  could be used to created "proxy-only" published entries.
  Previously, arp(8) created an entry of this type only
  in the absence of the route to a destination.

PR: bin/12357
MFC after: 1 week

23 years agoFix bug introduced by myself that often resulted in a session having
markm [Thu, 7 Jun 2001 08:45:23 +0000 (08:45 +0000)]
Fix bug introduced by myself that often resulted in a session having
SIGINTR (^C) and SIGSTP (^Z) masked.

Reported by: bde, sobomax
Submitted by: sobomax

23 years agoClean up the PAM lib lists. Particularly relevant to the KRB5 case.
markm [Thu, 7 Jun 2001 08:31:30 +0000 (08:31 +0000)]
Clean up the PAM lib lists. Particularly relevant to the KRB5 case.

23 years agomaster.passwd(5) -> passwd(5).
ru [Thu, 7 Jun 2001 07:27:52 +0000 (07:27 +0000)]
master.passwd(5) -> passwd(5).

23 years agoQuiet the READ_TOC errors that can appear on empty CDR's.
sos [Thu, 7 Jun 2001 07:21:20 +0000 (07:21 +0000)]
Quiet the READ_TOC errors that can appear on empty CDR's.

Assure cur_write_speed can be no less than 177 to avoid divide by 0 error.

23 years agoIn the "Message from Talk_Daemon" announcement, print the date as well
dd [Thu, 7 Jun 2001 05:26:57 +0000 (05:26 +0000)]
In the "Message from Talk_Daemon" announcement, print the date as well
as the current time.  It's nice to know whether the talk request you
see was sent just a few minutes ago (assuming you didn't hear the
bell), or if it's been decaying for days (weeks?).

23 years agoNull terminate buffer.
jlemon [Thu, 7 Jun 2001 05:26:26 +0000 (05:26 +0000)]
Null terminate buffer.

PR: 23150
Submitted by: Dan Nelson <dnelson@emsphone.com>
MFC in: 1 week

23 years agoMove IPFilter into contrib.
jlemon [Thu, 7 Jun 2001 05:13:35 +0000 (05:13 +0000)]
Move IPFilter into contrib.

23 years agoFix previous commit which inadverdently deleted a section.
jlemon [Thu, 7 Jun 2001 05:04:53 +0000 (05:04 +0000)]
Fix previous commit which inadverdently deleted a section.

23 years agoRelocate IPFilter from sys/netinet to sys/contrib/ipfilter.
jlemon [Thu, 7 Jun 2001 04:06:21 +0000 (04:06 +0000)]
Relocate IPFilter from sys/netinet to sys/contrib/ipfilter.

23 years agoBack out part of my previous commit. This was a last minute change
peter [Thu, 7 Jun 2001 03:17:26 +0000 (03:17 +0000)]
Back out part of my previous commit.  This was a last minute change
and I botched testing.  This is a perfect example of how NOT to do
this sort of thing. :-(

23 years agoSilence the remaining warnings and clamp down with WARNS=2.
dd [Thu, 7 Jun 2001 02:57:58 +0000 (02:57 +0000)]
Silence the remaining warnings and clamp down with WARNS=2.

Not objected to by: -audit

23 years agourandom(4) -> random(4) in comments.
dd [Thu, 7 Jun 2001 02:32:18 +0000 (02:32 +0000)]
urandom(4) -> random(4) in comments.

PR: 27858
Submitted by: Yoshihiro Koya <Yoshihiro.Koya@math.yokohama-cu.ac.jp>
Reviewed by: md5(1)
Approved by: markm

23 years agourandom(4) -> random(4)
dd [Thu, 7 Jun 2001 02:31:03 +0000 (02:31 +0000)]
urandom(4) -> random(4)

PR: 27858
Submitted by: Yoshihiro Koya <Yoshihiro.Koya@math.yokohama-cu.ac.jp>
Approved by: markm

23 years agoCall vn_close on the backing file vnode if ufs_extattr_enable failed to
tmm [Thu, 7 Jun 2001 00:11:32 +0000 (00:11 +0000)]
Call vn_close on the backing file vnode if ufs_extattr_enable failed to
avoid leaking it.

Reviewed by: rwatson

23 years agoDon't allow the user to change the subject; instead, always generate
dd [Thu, 7 Jun 2001 00:05:19 +0000 (00:05 +0000)]
Don't allow the user to change the subject; instead, always generate
it from the Synopsis field.  There's no reason for the subject to be
different, since all that does is cause confusion.  Users may get
confused because they may think the subject and synopsis are supposed
to be different, and developers may get confused because it may look
like there are two different problems.

Requested by: ru

23 years agoFix an instance of NDINIT in the extattrctl syscall: LOCKLEAF was or'ed
tmm [Wed, 6 Jun 2001 23:34:38 +0000 (23:34 +0000)]
Fix an instance of NDINIT in the extattrctl syscall: LOCKLEAF was or'ed
to the operation parameter, not to the flags as it should be.

Reviewed by: rwatson

23 years agoAdd PnP IDs for AHA-1530 and AHA-1520 cards.
imp [Wed, 6 Jun 2001 22:32:29 +0000 (22:32 +0000)]
Add PnP IDs for AHA-1530 and AHA-1520 cards.

PR: 19497, 18378
Submitted by: Martijn Plak <martijn@be3.com>

23 years agoMake the TUNABLE_*() macros look and behave more consistantly like the
peter [Wed, 6 Jun 2001 22:17:08 +0000 (22:17 +0000)]
Make the TUNABLE_*() macros look and behave more consistantly like the
SYSCTL_*() macros.  TUNABLE_INT_DECL() was an odd name because it didn't
actually declare the int, which is what the name suggests it would do.

23 years ago- Remember to set the 'extsts enable' bit in the CFG register to enable the
wpaul [Wed, 6 Jun 2001 22:16:23 +0000 (22:16 +0000)]
- Remember to set the 'extsts enable' bit in the CFG register to enable the
  use of the extsts field in DMA descriptors. We need this to tell the chip
  to calculate TCP/IP checksums in hardware on a per-packet basis.

- Fix the unions in DMA descriptor structures. Breakage on alpha led
  me to realize I'd done it wrong the first time.

23 years agoNew release notes: SYN flood resistance, fmt(1) rewrite, df(1) -l,
bmah [Wed, 6 Jun 2001 22:08:26 +0000 (22:08 +0000)]
New release notes:  SYN flood resistance, fmt(1) rewrite, df(1) -l,
traceroute(8) default maximum TTL.

Modified release notes:  reverse sense of ipfw(8) -d and add ipfw(8) -e.

23 years agoWe don't need to hold a lock just to test a flag.
jhb [Wed, 6 Jun 2001 22:05:48 +0000 (22:05 +0000)]
We don't need to hold a lock just to test a flag.

23 years agoActivate the kthread mechanism for doing usb bus discovery. This means
peter [Wed, 6 Jun 2001 22:00:03 +0000 (22:00 +0000)]
Activate the kthread mechanism for doing usb bus discovery.  This means
that device add/remove will work without usbd running.  usbd is still
used for execing stuff, but that is all now.  Ideally it could be replaced
by a devd some day.  Until now, usbd had to be running so that the
USB_DISCOVER ioctl could be called to walk the tree when an attachment
status change was noticed.

Among the changes:
- when a detach happens, remove any pending 'attach' messages or the system
suffers from whiplash from exec moused / kill moused loops if you do lots
of attach/detach and later start usbd.
- tweaks related to kthread differences
- disable the select handler for the old interface (never return success).
I have not removed it yet or old usbd's will abort.  That can get removed
later once usbd is cleaned up and things have stabilized for a few weeks.
- get Giant in the kthread.
- a couple of minor potential bug fixes (usb_nevents vs malloc failure etc)

Pre-approved by: n_hibma (ages and ages ago)

23 years agoInitialize the sa_flags member of an auto struct sigaction variable.
jdp [Wed, 6 Jun 2001 21:08:15 +0000 (21:08 +0000)]
Initialize the sa_flags member of an auto struct sigaction variable.
It contained whatever garbage happened to be on the stack.

PR: bin/27457 (not confirmed yet, but almost certain)
MFC after: 1 day

23 years agoStyle fixes from Sascha
phk [Wed, 6 Jun 2001 21:00:01 +0000 (21:00 +0000)]
Style fixes from Sascha

PR: 16551
Submitted by: Sascha Schumann <sascha@schumann.cx>

23 years agoMention Alexandre Peixoto's share/examples/ipfw/change_rules.sh in the
chris [Wed, 6 Jun 2001 20:56:56 +0000 (20:56 +0000)]
Mention Alexandre Peixoto's share/examples/ipfw/change_rules.sh in the
checklist.

MFC after: 1 week

23 years agoFix a byteorder problem in a diagnostic message.
phk [Wed, 6 Jun 2001 20:52:30 +0000 (20:52 +0000)]
Fix a byteorder problem in a diagnostic message.

PR: 16625
Submitted by: Peter Jeremy <peter.jeremy@ALCATEL.COM.AU>

23 years agoGet rid of se_ctladdrinitok, which doesn't do anything and seemes
dwmalone [Wed, 6 Jun 2001 20:00:42 +0000 (20:00 +0000)]
Get rid of se_ctladdrinitok, which doesn't do anything and seemes
to have been accidently imported when ipv6 support was added to
inetd.

Approved by: ume

23 years agomvprintw is varadic - it doesn't take a va_list. Use vw_printw
dwmalone [Wed, 6 Jun 2001 19:44:58 +0000 (19:44 +0000)]
mvprintw is varadic - it doesn't take a va_list. Use vw_printw
instead.

Submitted by: Andrew Gordon <arg@arg1.demon.co.uk>
MFC after: 1 week

23 years agoSilby's take one on increasing FreeBSD's resistance to SYN floods:
jesper [Wed, 6 Jun 2001 19:41:51 +0000 (19:41 +0000)]
Silby's take one on increasing FreeBSD's resistance to SYN floods:

One way we can reduce the amount of traffic we send in response to a SYN
flood is to eliminate the RST we send when removing a connection from
the listen queue.  Since we are being flooded, we can assume that the
majority of connections in the queue are bogus.  Our RST is unwanted
by these hosts, just as our SYN-ACK was.  Genuine connection attempts
will result in hosts responding to our SYN-ACK with an ACK packet.  We
will automatically return a RST response to their ACK when it gets to us
if the connection has been dropped, so the early RST doesn't serve the
genuine class of connections much.  In summary, we can reduce the number
of packets we send by a factor of two without any loss in functionality
by ensuring that RST packets are not sent when dropping a connection
from the listen queue.

Submitted by: Mike Silbersack <silby@silby.com>
Reviewed by: jesper
MFC after: 2 weeks

23 years agoUse ${EDITOR} as claimed.
dcs [Wed, 6 Jun 2001 19:34:02 +0000 (19:34 +0000)]
Use ${EDITOR} as claimed.

Source rc.conf and use ${firewall_script} instead of rc.firewall.

Textwidth formatting of comments and text.

PR:
Submitted by: Maxim <maxim@news1.macomnet.ru>

(sorry if I got name/email/committer status wrong :)

23 years agoDisable extra TCP/UCP checksum checking in nge_rxeof() for now.
wpaul [Wed, 6 Jun 2001 19:17:10 +0000 (19:17 +0000)]
Disable extra TCP/UCP checksum checking in nge_rxeof() for now.

23 years agoUse LGE_INC() macro to increment tx producer index in lge_encap().
wpaul [Wed, 6 Jun 2001 19:16:02 +0000 (19:16 +0000)]
Use LGE_INC() macro to increment tx producer index in lge_encap().
Disable the extra TCP/UCP checksum checking in lge_rxeof() since it
doesn't appear to actually work as advertised.

23 years agoUnbreak world by adding the new example directiory.
dcs [Wed, 6 Jun 2001 19:11:32 +0000 (19:11 +0000)]
Unbreak world by adding the new example directiory.

Reminded by: Niels Chr. Bank-Pedersen <ncbp@bank-pedersen.dk>
MFC after: 1 week

23 years agoSpelling fix ("independantly" -> "independently")
chris [Wed, 6 Jun 2001 19:07:03 +0000 (19:07 +0000)]
Spelling fix ("independantly" -> "independently")

PR: 25648
Submitted by: Jens Schweikhardt <schweikh@schweikhardt.net>

23 years ago o If the file being tailed is moved back to the original location,
jlemon [Wed, 6 Jun 2001 17:47:13 +0000 (17:47 +0000)]
  o If the file being tailed is moved back to the original location,
    resume using kqueue to monitor it, if possible.
  o auto-clear the read filter on EOF; this is needed when tailing a FIFO.

23 years agoAdd a wrapper for the fifo kqfilter which falls through to the ufs routine.
jlemon [Wed, 6 Jun 2001 17:40:57 +0000 (17:40 +0000)]
Add a wrapper for the fifo kqfilter which falls through to the ufs routine.
This permits the fifo to inherit the ufs VNODE kqfilter.

23 years agoThe kq write filter was hooked up to the wrong socket, and thus was
jlemon [Wed, 6 Jun 2001 17:38:36 +0000 (17:38 +0000)]
The kq write filter was hooked up to the wrong socket, and thus was
not behaving correctly.  Fix by attaching to the correct socket.

Also call so{rw}wakeup in addition to the fifo wakeup, so that any
kqfilters attached to the socket buffer get poked.

23 years agoProperly key RELNOTESng language selection from RELNOTES_LANG instead
bmah [Wed, 6 Jun 2001 16:58:09 +0000 (16:58 +0000)]
Properly key RELNOTESng language selection from RELNOTES_LANG instead
of DOC_LANG.

Submitted by: Thomas Seck <tmseck@web.de>

23 years agoAdd the single suffix rules for FORTRAN.
obrien [Wed, 6 Jun 2001 16:58:08 +0000 (16:58 +0000)]
Add the single suffix rules for FORTRAN.

23 years agoAdd the single suffix rules for C++.
obrien [Wed, 6 Jun 2001 16:51:38 +0000 (16:51 +0000)]
Add the single suffix rules for C++.
(While there, I also moved the single suffix C rules beside the double
suffix ones so they are easier to find)

PR: 24438
Submitted by: Georg-W. Koltermann <gwk@sgi.com>

23 years agoFetch the default maximum TTL value from the net.inet.ip.ttl MIB.
ru [Wed, 6 Jun 2001 16:12:59 +0000 (16:12 +0000)]
Fetch the default maximum TTL value from the net.inet.ip.ttl MIB.

PR: bin/19598
MFC after: 1 week

23 years agoIntroduce an example script of safe changing of ipfw(8) rules.
dcs [Wed, 6 Jun 2001 15:45:04 +0000 (15:45 +0000)]
Introduce an example script of safe changing of ipfw(8) rules.

I'm sure this can be much improved, but this works fairly well.

PR: 27887
Submitted by: Alexandre Peixoto <alexandref@tcoip.com.br>
MFC after: 1 week

23 years agohold the vm_mtx around vm_map_lookup_entry() and vm_map_findspace()
gallatin [Wed, 6 Jun 2001 14:07:52 +0000 (14:07 +0000)]
hold the vm_mtx around vm_map_lookup_entry() and vm_map_findspace()

23 years agoUnbreak setregid(2).
ru [Wed, 6 Jun 2001 13:58:03 +0000 (13:58 +0000)]
Unbreak setregid(2).

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

23 years agoApparently, the vm_mtx must be held around vm_map_find(). I'm assuming
gallatin [Wed, 6 Jun 2001 13:46:16 +0000 (13:46 +0000)]
Apparently, the vm_mtx must be held around vm_map_find().  I'm assuming
that it needn't be held around vm_mmap() since vm_mmap() itself aquires
it.

Sleuthing by: obrien

23 years agoSorry, an "ARCHIVE Python 06408" does not need SA_QUIRK_NOCOMP.
non [Wed, 6 Jun 2001 13:01:44 +0000 (13:01 +0000)]
Sorry, an "ARCHIVE Python 06408" does not need SA_QUIRK_NOCOMP.

23 years agoNote that -a is ignored if -s is also specified.
ru [Wed, 6 Jun 2001 12:48:22 +0000 (12:48 +0000)]
Note that -a is ignored if -s is also specified.

23 years agofmt(1) rewrite by Gareth McCaughan.
ru [Wed, 6 Jun 2001 10:17:05 +0000 (10:17 +0000)]
fmt(1) rewrite by Gareth McCaughan.

PR: bin/5901, bin/12242, bin/19071
Submitted by: Gareth McCaughan <gjm11@dpmms.cam.ac.uk>
Obtained from: OpenBSD + ctype(3), WARNS=2, and manpage cleanup
MFC after: 2 weeks

23 years agoFixed missing parentheses in the definition of KTR_COMPILE. KTR_COMPILE
bde [Wed, 6 Jun 2001 06:58:13 +0000 (06:58 +0000)]
Fixed missing parentheses in the definition of KTR_COMPILE.  KTR_COMPILE
is usually (always?) used in expressions like (KTR_COMPILE & KTR_FOO).
Defining it as KTR_INTR|KTR_PROC gave the wrong value in approximately
8497 places according to error output for compiling LINT.