]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
22 years agoReplace /kernel with /boot/kernel/kernel.
Josef Karthauser [Thu, 9 May 2002 11:47:42 +0000 (11:47 +0000)]
Replace /kernel with /boot/kernel/kernel.

PR: docs/37757
Submitted by: Hiten Pandya <hiten@uk.FreeBSD.org>

22 years agoBack out part of previous commit - K&R doesn't have const.
Dag-Erling Smørgrav [Thu, 9 May 2002 11:24:58 +0000 (11:24 +0000)]
Back out part of previous commit - K&R doesn't have const.

22 years agoCleanup the interface to ip_fw_chk, two of the input arguments
Luigi Rizzo [Thu, 9 May 2002 10:34:57 +0000 (10:34 +0000)]
Cleanup the interface to ip_fw_chk, two of the input arguments
were totally useless and have been removed.

ip_input.c, ip_output.c:
    Properly initialize the "ip" pointer in case the firewall does an
    m_pullup() on the packet.

    Remove some debugging code forgotten long ago.

ip_fw.[ch], bridge.c:
    Prepare the grounds for matching MAC header fields in bridged packets,
    so we can have 'etherfw' functionality without a lot of kernel and
    userland bloat.

22 years agoRemove trace_req().
Jonathan Mini [Thu, 9 May 2002 04:13:41 +0000 (04:13 +0000)]
Remove trace_req().

Reviewed by: alfred, jhb, peter

22 years ago o Correct an error made in revision 1.65: In readv(), if uap->iovcnt is
Alan Cox [Thu, 9 May 2002 02:30:41 +0000 (02:30 +0000)]
 o Correct an error made in revision 1.65: In readv(), if uap->iovcnt is
   out-of-range, drop the file reference before returning.  (This error
   also exists in the RELENG_4 branch.)
 o Eliminate the acquisition and release of Giant in readv()
   now that malloc() and free() are callable without Giant.

22 years agoComment two values I was looking at for GDB.
David E. O'Brien [Thu, 9 May 2002 02:07:51 +0000 (02:07 +0000)]
Comment two values I was looking at for GDB.

22 years agoGcc 3.1 now offers both a C99 and a K&R traditional C preprocessor.
David E. O'Brien [Thu, 9 May 2002 00:52:09 +0000 (00:52 +0000)]
Gcc 3.1 now offers both a C99 and a K&R traditional C preprocessor.
This is the traditional one.

22 years ago= We need `-lcrypt' and `-lcom_err' to when building Kerberos
Jacques Vidrine [Thu, 9 May 2002 00:50:49 +0000 (00:50 +0000)]
= We need `-lcrypt' and `-lcom_err' to when building Kerberos
  applications.
= The Heimdal version number is hiding here also.  Correct it.

22 years agoGet rid of the INTERNALSTATICLIB knob and just use plain INTERNALLIB.
David E. O'Brien [Thu, 9 May 2002 00:44:13 +0000 (00:44 +0000)]
Get rid of the INTERNALSTATICLIB knob and just use plain INTERNALLIB.
INTERNALLIB now implies NOPROFILE, NOMAN, and don't install anything.
Add a NOLIB knob.

22 years agosys_nerr should come from standard headers -- not from a coder's guess
David E. O'Brien [Thu, 9 May 2002 00:41:18 +0000 (00:41 +0000)]
sys_nerr should come from standard headers -- not from a coder's guess
of what is right.

22 years agoUnbreak this file after the last drive-by committing. We have to
Bill Paul [Thu, 9 May 2002 00:37:57 +0000 (00:37 +0000)]
Unbreak this file after the last drive-by committing. We have to
#include route.h before iso88025.h, and we have to dereference
the trld_route array correctly. (NOTE: I'm not altogether sure
that this is really the correct way to traverse this array. This
just eliminates the build warning/error. It may not work right at
runtime, and I have no way to test it since I lack the necessary
hardware.)

Broken by: kbyanc, who gets to wear the pointy hat

22 years agoDon't forget to null-terminate string we got from sysctl (f.e. in 'A' format).
Andrey A. Chernov [Wed, 8 May 2002 23:49:19 +0000 (23:49 +0000)]
Don't forget to null-terminate string we got from sysctl (f.e. in 'A' format).
Stack garbadge may be printed otherwise.

22 years agoOnly saying you're launching other CPUs if there are more than one to launch.
Matt Jacob [Wed, 8 May 2002 23:12:00 +0000 (23:12 +0000)]
Only saying you're launching other CPUs if there are more than one to launch.

22 years agoIt is not clear if "#endif FOO" is allowed by ANSI. Gcc 3.1 complains
David E. O'Brien [Wed, 8 May 2002 20:39:10 +0000 (20:39 +0000)]
It is not clear if "#endif FOO" is allowed by ANSI.  Gcc 3.1 complains
loudly, so don't do it.

22 years agoIt is not clear if "#endif FOO" is allowed by ANSI.
David E. O'Brien [Wed, 8 May 2002 20:37:48 +0000 (20:37 +0000)]
It is not clear if "#endif FOO" is allowed by ANSI.
Gcc 3.1 complains loudly, so don't do it.

22 years agoBuild our native CSU bits before the compiler-specific ones.
David E. O'Brien [Wed, 8 May 2002 20:35:12 +0000 (20:35 +0000)]
Build our native CSU bits before the compiler-specific ones.
This helps in the compiler build.

22 years agoDo my bit for the Perl removal team and shoot my own, rather mangy
Paul Richards [Wed, 8 May 2002 18:44:24 +0000 (18:44 +0000)]
Do my bit for the Perl removal team and shoot my own, rather mangy
looking, dog.

22 years agoSome updates to mention accept filters and how
Mike Silbersack [Wed, 8 May 2002 18:19:01 +0000 (18:19 +0000)]
Some updates to mention accept filters and how
listen queues work in a syncache world.

MFC after: 3 days

22 years agoMake compilable without -DINET6.
Hajimu UMEMOTO [Wed, 8 May 2002 17:20:08 +0000 (17:20 +0000)]
Make compilable without -DINET6.

22 years agoResurrect as an empty file to unbreak the build. We have everything we
Dag-Erling Smørgrav [Wed, 8 May 2002 17:19:02 +0000 (17:19 +0000)]
Resurrect as an empty file to unbreak the build.  We have everything we
need in paths.h.

22 years agoPut 'ftp' back.
Maxim Konovalov [Wed, 8 May 2002 17:06:40 +0000 (17:06 +0000)]
Put 'ftp' back.

Requested by: bmah, jhb

22 years agoLog address family of a connection.
Hajimu UMEMOTO [Wed, 8 May 2002 16:39:58 +0000 (16:39 +0000)]
Log address family of a connection.

Requested by: matusita
Reviewed by: matusita

22 years agoUse libutil instead of pw_{copy,util}.c. Clean up a little. Warnsify.
Dag-Erling Smørgrav [Wed, 8 May 2002 15:42:37 +0000 (15:42 +0000)]
Use libutil instead of pw_{copy,util}.c.  Clean up a little.  Warnsify.

Sponsored by: DARPA, NAI Labs

22 years agoFix a boatload of format errors and constness violations.
Dag-Erling Smørgrav [Wed, 8 May 2002 15:41:15 +0000 (15:41 +0000)]
Fix a boatload of format errors and constness violations.

Sponsored by: DARPA, NAI Labs

22 years agoIf no old_pw was passed to pw_copy, compare just the name.
Dag-Erling Smørgrav [Wed, 8 May 2002 14:52:32 +0000 (14:52 +0000)]
If no old_pw was passed to pw_copy, compare just the name.

Sponsored by: DARPA, NAI Labs

22 years agoMake _msgout take a const char * rather than a char *.
Dag-Erling Smørgrav [Wed, 8 May 2002 14:50:28 +0000 (14:50 +0000)]
Make _msgout take a const char * rather than a char *.

Sponsored by: DARPA, NAI Labs

22 years agoIt is dangerous to use globbing like so in startup scripts:
Jacques Vidrine [Wed, 8 May 2002 14:47:44 +0000 (14:47 +0000)]
It is dangerous to use globbing like so in startup scripts:

  rm -f /tmp/.X11-unix/*

If /tmp/.X11-unix didn't already exist, a user could symlink it to a directory
with files that he wants to wipe out, and wait for next reboot.

Reported by: lumpy <lumpy@the.whole.net>

22 years agoDon't emulate INTERNALLIB.
Ruslan Ermilov [Wed, 8 May 2002 13:56:08 +0000 (13:56 +0000)]
Don't emulate INTERNALLIB.

22 years agoChange whtespace indent in format string to fit new output form of
Jun Kuriyama [Wed, 8 May 2002 11:23:45 +0000 (11:23 +0000)]
Change whtespace indent in format string to fit new output form of
timestamp.

Reviewed by: markm

22 years agoexpand_name fixes:
Alfred Perlstein [Wed, 8 May 2002 09:06:47 +0000 (09:06 +0000)]
expand_name fixes:

.) don't use MAXPATHLEN + 1, fix logic to compensate.
.) style(9) function parameters.
.) fix line wrapping.
.) remove duplicated error and string handling code.
.) don't NUL terminate already NUL terminated string.
.) all string length variables changed from int to size_t.
.) constify variables.
.) catch when corename would be truncated.
.) cast pid_t and uid_t args for format string.
.) add parens around return arguments.

Help and suggestions from: bde

22 years agostyle(9): remove an EOF space.
Maxim Konovalov [Wed, 8 May 2002 07:40:22 +0000 (07:40 +0000)]
style(9): remove an EOF space.

22 years agoImplement 'speed' command: set a maximum read speed. At the moment it
Maxim Konovalov [Wed, 8 May 2002 07:32:40 +0000 (07:32 +0000)]
Implement 'speed' command: set a maximum read speed. At the moment it
works on ATAPI drives only.

PR: kern/35512 (a part of)
Submitted by: Philipp Mergenthaler <philipp.mergenthaler@stud.uni-karlsruhe.de>
Reviewed by: -hackers
MFC after: 1 month

22 years agozapnot the signed bits in atomic_cmpset_32. Previously this did not work with
Jeff Roberson [Wed, 8 May 2002 05:19:56 +0000 (05:19 +0000)]
zapnot the signed bits in atomic_cmpset_32.  Previously this did not work with
negative values because the original value was sign extended but the compared
value was not.

22 years agoFix typo.
Gregory Neil Shapiro [Wed, 8 May 2002 05:17:46 +0000 (05:17 +0000)]
Fix typo.

Submitted by: Marius Strom <marius@marius.org>
MFC after: 1 day
and RE approval

22 years agoRemove runq_findproc. This never worked right in the first place and can
Jake Burkholder [Wed, 8 May 2002 04:39:49 +0000 (04:39 +0000)]
Remove runq_findproc.  This never worked right in the first place and can
be prohibitively expensive.

22 years agoRemove unneeded include.
Jake Burkholder [Wed, 8 May 2002 04:21:58 +0000 (04:21 +0000)]
Remove unneeded include.

22 years agoMake a macro for the guts of tl0_immu_miss, like dmmu_miss and prot.
Jake Burkholder [Wed, 8 May 2002 04:14:16 +0000 (04:14 +0000)]
Make a macro for the guts of tl0_immu_miss, like dmmu_miss and prot.
Rearrange things slightly so that the contents of the tag access
register are read and restored outside of the macros.  The intention
is to pass the page size to look up as an argument to the macros.

22 years agoBring doc up-to-date with current buildworld/buildkernel practices.
John W. De Boskey [Wed, 8 May 2002 01:55:08 +0000 (01:55 +0000)]
Bring doc up-to-date with current buildworld/buildkernel practices.

Reviewed by: jhb

22 years agoThese files haven't been used in a while.
Dag-Erling Smørgrav [Wed, 8 May 2002 01:35:22 +0000 (01:35 +0000)]
These files haven't been used in a while.

22 years agoRoll my own min() (named ISO88025_MIN() so as to not cause conflicts) so
Kelly Yancey [Wed, 8 May 2002 01:08:26 +0000 (01:08 +0000)]
Roll my own min() (named ISO88025_MIN() so as to not cause conflicts) so
that this header may be included from userland where min() may not be
declared (or worse, declared differently).  I open to alternative
solutions.

22 years agoUpdate ARP to access token-ring source routing information at it's new
Kelly Yancey [Wed, 8 May 2002 00:55:29 +0000 (00:55 +0000)]
Update ARP to access token-ring source routing information at it's new
location.

Not tested by: me (sorry!)
Sponsored by: NTT Multimedia Communications Labs

22 years agoUse libutil and libypclnt for all passwd manipulation and NIS needs.
Dag-Erling Smørgrav [Wed, 8 May 2002 00:54:29 +0000 (00:54 +0000)]
Use libutil and libypclnt for all passwd manipulation and NIS needs.

Sponsored by: DARPA, NAI Labs

22 years agoDetect and attach to the ICH3 SMBus controller.
John Baldwin [Wed, 8 May 2002 00:51:54 +0000 (00:51 +0000)]
Detect and attach to the ICH3 SMBus controller.

PR: i386/36972
Submitted by: Colin Perkins <csp@isi.edu>
Requested by: Larry Rosenman <ler@lerctr.org>
MFC After: 7 days

22 years agoAdd passwd manipulation code based on parts of vipw and chpass.
Dag-Erling Smørgrav [Wed, 8 May 2002 00:50:07 +0000 (00:50 +0000)]
Add passwd manipulation code based on parts of vipw and chpass.

Sponsored by: DARPA, NAI Labs

22 years agoAdd ypclnt_havepasswd(), and guard against a possible segfault.
Dag-Erling Smørgrav [Wed, 8 May 2002 00:48:39 +0000 (00:48 +0000)]
Add ypclnt_havepasswd(), and guard against a possible segfault.

Sponsored by: DARPA, NAI Labs

22 years agoReally include <paths.h> (it seems it was brought in indirectly)
Dag-Erling Smørgrav [Wed, 8 May 2002 00:47:02 +0000 (00:47 +0000)]
Really include <paths.h> (it seems it was brought in indirectly)

22 years agoUse <paths.h> rather than "pathnames.h", and fix a couple of whitespace nits.
Dag-Erling Smørgrav [Wed, 8 May 2002 00:47:01 +0000 (00:47 +0000)]
Use <paths.h> rather than "pathnames.h", and fix a couple of whitespace nits.

Sponsored by: DARPA, NAI Labs

22 years agoPAMify.
Dag-Erling Smørgrav [Wed, 8 May 2002 00:43:46 +0000 (00:43 +0000)]
PAMify.

Sponsored by: DARPA, NAI Labs

22 years agoAdd some more commonly-used directories and files.
Dag-Erling Smørgrav [Wed, 8 May 2002 00:37:50 +0000 (00:37 +0000)]
Add some more commonly-used directories and files.

Sponsored by: DARPA, NAI Labs

22 years agoUse pam_lastlog(8)'s new no_fail option.
Dag-Erling Smørgrav [Wed, 8 May 2002 00:33:02 +0000 (00:33 +0000)]
Use pam_lastlog(8)'s new no_fail option.

Sponsored by: DARPA, NAI Labs

22 years agoAdd a no_fail option.
Dag-Erling Smørgrav [Wed, 8 May 2002 00:31:45 +0000 (00:31 +0000)]
Add a no_fail option.

Sponsored by: DARPA, NAI Labs

22 years agoAdd pam_ftpusers(8), which enforces /etc/ftpusers.
Dag-Erling Smørgrav [Wed, 8 May 2002 00:30:10 +0000 (00:30 +0000)]
Add pam_ftpusers(8), which enforces /etc/ftpusers.

Sponsored by: DARPA, NAI Labs

22 years agoM_ZERO the temp buffer in expand_name() otherwise if an error occurs
Alfred Perlstein [Tue, 7 May 2002 23:37:07 +0000 (23:37 +0000)]
M_ZERO the temp buffer in expand_name() otherwise if an error occurs
while logging we may pass a non NUL terminated string to log(9) for a
%s format arg.

22 years agoOops, update usage() to have -i.
Juli Mallett [Tue, 7 May 2002 23:33:44 +0000 (23:33 +0000)]
Oops, update usage() to have -i.

22 years agoAdd handling for any nil-length string passed to -i for the backup extension.
Juli Mallett [Tue, 7 May 2002 23:32:26 +0000 (23:32 +0000)]
Add handling for any nil-length string passed to -i for the backup extension.

Add a note that this is kinda-sorta dangerous to the manual page.

22 years agoAlpha MD bits for Gcc 3.1.
David E. O'Brien [Tue, 7 May 2002 23:28:45 +0000 (23:28 +0000)]
Alpha MD bits for Gcc 3.1.

22 years agoFix bug that causes passwd and friends to fail when the user has a '+' in
Dag-Erling Smørgrav [Tue, 7 May 2002 23:26:00 +0000 (23:26 +0000)]
Fix bug that causes passwd and friends to fail when the user has a '+' in
their passwd file for NIS because _PWF_SOURCE is not set.

Submitted by: amigus (perforce change 10969)

22 years agoUn-shadow the `fname' variable.
Juli Mallett [Tue, 7 May 2002 23:06:47 +0000 (23:06 +0000)]
Un-shadow the `fname' variable.

22 years agoMove ISO88025 source routing information into sockaddr_dl's sdl_data
Kelly Yancey [Tue, 7 May 2002 22:14:06 +0000 (22:14 +0000)]
Move ISO88025 source routing information into sockaddr_dl's sdl_data
field.  This returns the sdl_data field to a variable-length field.  More
importantly, this prevents a easily-reproduceable data-corruption bug when
the interface name plus the hardware address exceed the sdl_data field's
original 12 byte limit.  However, token-ring interfaces may still overflow
the new sdl_data field's 46 byte limit if the interface name exceeds 6
characters (since 6 characters for interface name plus 6 for hardware
address plus 34 for source routing = the size of sdl_data).  Further
refinements could overcome this limitation but would break binary
compatibility; this commit only addresses fixing the bug for
commonly-occuring cases without breaking binary compatibility with the
intention that the functionality can be MFC'ed to -stable.

  See message ID's (both send to -arch):
20020421013332.F87395-100000@gateway.posi.net
20020430181359.G11009-300000@gateway.posi.net
  for a more thorough description of the bug addressed and how to
reproduce it.

Approved by: silence on -arch and -net
Sponsored by: NTT Multimedia Communications Labs
MFC after: 1 week

22 years agoRe-remove kern_random.c and svr4_signal.c. Somehow dillon managed to keep
Peter Wemm [Tue, 7 May 2002 21:54:47 +0000 (21:54 +0000)]
Re-remove kern_random.c and svr4_signal.c.  Somehow dillon managed to keep
on committing to these while they were in the Attic after they had been
removed.  I think this was because he had the file checked out and already
'modified' while markm cvs rm'ed them, and cvs screws up when trying to
"merge" the modifications with the "rm".  And after that the client
state was sufficiently hosed to keep it messed up.  Yay CVS!  (CVS is
very fragile for adding and removing files remotely)

The existence of these files was pointed out by: ru

22 years agoRe-remove these files. Somehow obrien managed to revive them with a
Peter Wemm [Tue, 7 May 2002 21:47:57 +0000 (21:47 +0000)]
Re-remove these files.  Somehow obrien managed to revive them with a
Makefile cleanup sweep even after they were removed.  cvs was very
confused because they were in the Attic still.

22 years agoNew release note: syncache/accept_filter bugfix.
Bruce A. Mah [Tue, 7 May 2002 21:01:46 +0000 (21:01 +0000)]
New release note:  syncache/accept_filter bugfix.

22 years ago- Remove the init routine for loop -- it broke something and I don't have
Andrew R. Reiter [Tue, 7 May 2002 20:50:10 +0000 (20:50 +0000)]
- Remove the init routine for loop -- it broke something and I don't have
  the time to fix it properly at the moment.

22 years ago- Use an array of init functions to be used in a for loop in our
Andrew R. Reiter [Tue, 7 May 2002 20:37:40 +0000 (20:37 +0000)]
- Use an array of init functions to be used in a for loop in our
  initialization routine.

22 years ago- Add atm_init_fn
Andrew R. Reiter [Tue, 7 May 2002 20:36:35 +0000 (20:36 +0000)]
- Add atm_init_fn

22 years ago- Add atm_sock_init()
Andrew R. Reiter [Tue, 7 May 2002 20:13:55 +0000 (20:13 +0000)]
- Add atm_sock_init()
- Move the Atm_pcb storage pool (atm_pcb_pool) to be an UMA zone.

22 years ago- Change the atm_cm_init function to take no argument.
Andrew R. Reiter [Tue, 7 May 2002 20:12:29 +0000 (20:12 +0000)]
- Change the atm_cm_init function to take no argument.

22 years agoAdd a -i option to sed(1) to do inplace editing, to give us an alternative to
Juli Mallett [Tue, 7 May 2002 18:32:18 +0000 (18:32 +0000)]
Add a -i option to sed(1) to do inplace editing, to give us an alternative to
Perl for such things.  The key difference to Perl is that a backup extension
*MUST* be specified, because on one hand it isn't recommended to have options
which optionally take a parameter, and on the other hand, it'd be slightly
unpleasent to implement proper handling for that.

The difference between this and the version posted to developers@ is that it
does handle multiple files in argv after the getopt(3) handling "correctly",
in that the inplace editing-specific code has been moved out to a function,
and that function is used beyond the first file in our linked list.

This option has been documented as FreeBSD-specific in the manpage.

Reviewed by: developers@ (got feedback from: des, fanf, sobomax, roberto,
obrien)
MFC after: 1 week

22 years agoMFOpenBSD: ibss and ibss-master.
Warner Losh [Tue, 7 May 2002 18:16:39 +0000 (18:16 +0000)]
MFOpenBSD: ibss and ibss-master.

ibss is the modern ad-hoc mode.  ibss-master is the same, except that
it creates the ibss network.  This distinction is necessary because
some supported cards (symbol) support the former without supporting
the latter.

A seprate commit will introduce a demo-adhoc mode so that we can
disentwingle the multiple, mutually exclusive meandings of adhoc in
the present state of affairs.

Submitted by: jhay

22 years agoMinor style nit
Warner Losh [Tue, 7 May 2002 18:11:55 +0000 (18:11 +0000)]
Minor style nit

22 years agoTry to use spls more correctly
Warner Losh [Tue, 7 May 2002 18:09:19 +0000 (18:09 +0000)]
Try to use spls more correctly

22 years agoFix the copyright dates.
David E. O'Brien [Tue, 7 May 2002 18:07:43 +0000 (18:07 +0000)]
Fix the copyright dates.

22 years agoThis is two new entries for Taiwanese 16bit PCMCIA cards.
Michael C . Wu [Tue, 7 May 2002 16:50:15 +0000 (16:50 +0000)]
This is two new entries for Taiwanese 16bit PCMCIA cards.
The Blue Concentric CF 802.11b card is a compactflash form-factored card
that does 802.11b, including 128bit encryption.
The Zonet modem pccard is a simple FAX/Modem card.

Both are sold in Guang-Hua Market in Taipei, and functions perfectly
with -current and -stable.

22 years agoDo not forget to increase the number of completely connected sockets in
Seigo Tanimura [Tue, 7 May 2002 16:17:44 +0000 (16:17 +0000)]
Do not forget to increase the number of completely connected sockets in
soisconnected_locked().

Forgotten by: tanimura

22 years agoAdded internal, non-recursive (SUBDIR) versions of the `all-man'
Ruslan Ermilov [Tue, 7 May 2002 15:42:56 +0000 (15:42 +0000)]
Added internal, non-recursive (SUBDIR) versions of the `all-man'
and `maninstall' targets.  This fixes the issue where each subdir
was descended into twice during "make all", and also resurrects
the standardization of `maninstall'.

Urged by: bde

22 years agoInstall PROG before SCRIPTS and FILES, but make sure
Ruslan Ermilov [Tue, 7 May 2002 15:30:49 +0000 (15:30 +0000)]
Install PROG before SCRIPTS and FILES, but make sure
LINKS are still processed last.  This backs out part
of the backout in previous revision.

Install the manpages before descending into SUBDIRs.

22 years agoInstall the manpages before descending into SUBDIRs.
Ruslan Ermilov [Tue, 7 May 2002 15:27:56 +0000 (15:27 +0000)]
Install the manpages before descending into SUBDIRs.

22 years agoDocument CVSCMDARGS variable.
Makoto Matsushita [Tue, 7 May 2002 14:57:22 +0000 (14:57 +0000)]
Document CVSCMDARGS variable.

Suggested by: murray
Reviewed by: ru
MFC after: 3 days
(if re@ allows)

22 years agoFix the output when daily_status_mailq_shorten is set to YES
Brian Somers [Tue, 7 May 2002 13:11:05 +0000 (13:11 +0000)]
Fix the output when daily_status_mailq_shorten is set to YES

PR: 23766
Mostly submitted by: lambert@ssabsd.csw.net
MFC after: 3 days

22 years agoDon't corrupt MP fragments when they're put back on the front of our
Brian Somers [Tue, 7 May 2002 12:48:45 +0000 (12:48 +0000)]
Don't corrupt MP fragments when they're put back on the front of our
inbound queue.

Submitted by: "Amit K. Rao" <arao@niksun.com>
PR: 37813
MFC after: 1 week

Also fix a typo while I'm here.

22 years agoSync with master.passwd.
Maxim Konovalov [Tue, 7 May 2002 11:19:22 +0000 (11:19 +0000)]
Sync with master.passwd.

PR: conf/37796
Submitted by: David Siebo"rger <drs@rucus.ru.ac.za>
MFC after: 3 days

22 years agoAdd myself.
Tony Finch [Tue, 7 May 2002 11:09:40 +0000 (11:09 +0000)]
Add myself.

Reviewed by: dwmalone (mentor)

22 years agoInclude 'www' in the list of forbidden ftp users.
Maxim Konovalov [Tue, 7 May 2002 11:07:39 +0000 (11:07 +0000)]
Include 'www' in the list of forbidden ftp users.

Reviewed by: sheldonh
MFC after: 3 days

22 years agoReconnect db_elf.c to the build (now under "options DDB_NOKLDSYM"). It
Bruce Evans [Tue, 7 May 2002 10:59:52 +0000 (10:59 +0000)]
Reconnect db_elf.c to the build (now under "options DDB_NOKLDSYM").  It
doesn't actually build yet.

22 years agoAdd support for vendor specific RADIUS extensions.
Brian Somers [Tue, 7 May 2002 10:47:18 +0000 (10:47 +0000)]
Add support for vendor specific RADIUS extensions.

Only the extensions from rfc2548 are specified for now.

22 years agoUnderstand the Session-Timeout RADIUS attribute
Brian Somers [Tue, 7 May 2002 10:06:54 +0000 (10:06 +0000)]
Understand the Session-Timeout RADIUS attribute
Store the Filter-Id attribute (we don't do anything with it yet)

Submitted mostly by: andrew pavlov <and@kremenchug.net>

22 years agoAdd smmsp, mailnull and www
Brian Somers [Tue, 7 May 2002 09:38:02 +0000 (09:38 +0000)]
Add smmsp, mailnull and www

PR: 37796
Submitted by: drs@rucus.ru.ac.za
MFC after: 1 week

Note, I don't usually look after pppd (just ppp), so I haven't
removed the ftp and ingres users as I guess they're there for
a reason....

22 years agoMake NO_CXX, really mean NO_CXX.
David E. O'Brien [Tue, 7 May 2002 07:49:12 +0000 (07:49 +0000)]
Make NO_CXX, really mean NO_CXX.

22 years agoMake sure the file in Attic is really "dead".
Ruslan Ermilov [Tue, 7 May 2002 06:57:04 +0000 (06:57 +0000)]
Make sure the file in Attic is really "dead".

Problem found by: ken

22 years agoo Add ar.lc to the pcb.
Marcel Moolenaar [Tue, 7 May 2002 05:52:27 +0000 (05:52 +0000)]
o  Add ar.lc to the pcb.
o  Create pcb_save as the backend for savectx and cpu_switch.
o  While here, use explicit bundling for pcb_save and optimize
   for compactness (~87% density).

o  Not part of the commit is a backend pcb_restore. restorectx()
   still jumps halfway into cpu_switch().

22 years agoSwitch from just holding the interlock to holding the standard lock throughout
Jeff Roberson [Tue, 7 May 2002 02:44:06 +0000 (02:44 +0000)]
Switch from just holding the interlock to holding the standard lock throughout
getnewvnode().  This is safer.  In the future, we should investigate requiring
only the interlock to get the vnode object.

22 years agoComment a knob.
David E. O'Brien [Tue, 7 May 2002 02:26:40 +0000 (02:26 +0000)]
Comment a knob.

22 years agoThe default ASM_OUTPUT_ADDR_DIFF_ELT will now handle our mixed a.out/ELF.
David E. O'Brien [Tue, 7 May 2002 02:25:52 +0000 (02:25 +0000)]
The default ASM_OUTPUT_ADDR_DIFF_ELT will now handle our mixed a.out/ELF.

22 years agolibstdc++ version 3 (Gcc 3.1 is a consumer) uses the typical `config.h'.
David E. O'Brien [Tue, 7 May 2002 01:03:51 +0000 (01:03 +0000)]
libstdc++ version 3 (Gcc 3.1 is a consumer) uses the typical `config.h'.

22 years agoFixed missing quotes in the default for MKDEPCMD in the CC != cc case.
Bruce Evans [Tue, 7 May 2002 00:15:53 +0000 (00:15 +0000)]
Fixed missing quotes in the default for MKDEPCMD in the CC != cc case.
"CC='cc -Dfoo' make depend" was very broken.

22 years agoInclude <sys/lockmgr.h> for definitions of deprecated locking interfaces --
Bruce Evans [Mon, 6 May 2002 23:34:13 +0000 (23:34 +0000)]
Include <sys/lockmgr.h> for definitions of deprecated locking interfaces --
don't depend on namespace pollution in other headers.

22 years agoUse __FBSDID
David E. O'Brien [Mon, 6 May 2002 23:24:39 +0000 (23:24 +0000)]
Use __FBSDID

22 years agoWe typically don't add trailing /'s.
David E. O'Brien [Mon, 6 May 2002 23:24:16 +0000 (23:24 +0000)]
We typically don't add trailing /'s.

22 years agoGCC's libstdc++ version 3 has the "libgcc" like bits (start up / terminate
David E. O'Brien [Mon, 6 May 2002 22:07:26 +0000 (22:07 +0000)]
GCC's libstdc++ version 3 has the "libgcc" like bits (start up / terminate
code, `new' handler, exception handling, etc.) split out into libsupc++.

22 years agoMake funsetown() take a 'struct sigio **' so that the locking can
Alfred Perlstein [Mon, 6 May 2002 19:31:28 +0000 (19:31 +0000)]
Make funsetown() take a 'struct sigio **' so that the locking can
be done internally.

Ensure that no one can fsetown() to a dying process/pgrp.  We need
to check the process for P_WEXIT to see if it's exiting.  Process
groups are already safe because there is no such thing as a pgrp
zombie, therefore the proctree lock completely protects the pgrp
from having sigio structures associated with it after it runs
funsetownlst.

Add sigio lock to witness list under proctree and allproc, but over
proc and pgrp.

Seigo Tanimura helped with this.