]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
19 years ago- add wall_cmos_clock and adjkerntz variables, required by msdosfs
Peter Grehan [Fri, 4 Feb 2005 01:41:38 +0000 (01:41 +0000)]
- add wall_cmos_clock and adjkerntz variables, required by msdosfs
- support adjkerntz sysctl to silence NTP, though it's a null
  implementation at the moment.

19 years agoConvert bus_space_barrier() into a null inline function rather than an
Peter Grehan [Fri, 4 Feb 2005 01:27:21 +0000 (01:27 +0000)]
Convert bus_space_barrier() into a null inline function rather than an
empty macro to avoid many compile warnings in the USB code.

19 years ago- add definitions for MPC7447A/7448 (i.e. miniMac)
Peter Grehan [Fri, 4 Feb 2005 01:17:04 +0000 (01:17 +0000)]
- add definitions for MPC7447A/7448 (i.e. miniMac)
- expand MPC745X_P macro to include these

Obtained from: NetBSD

19 years agoHID0 updates:
Peter Grehan [Fri, 4 Feb 2005 01:14:38 +0000 (01:14 +0000)]
HID0 updates:
 - updated relevant models for High BAT enable bit
 - fixed bug in BHTCLR/XAEN constants
 - added LRSTK and FOLD bits

19 years agoSort PANASONIC products numerically
Warner Losh [Thu, 3 Feb 2005 23:45:20 +0000 (23:45 +0000)]
Sort PANASONIC products numerically

19 years agotakawata tells me that KXLC0003 also works with this driver, so add it
Warner Losh [Thu, 3 Feb 2005 23:39:29 +0000 (23:39 +0000)]
takawata tells me that KXLC0003 also works with this driver, so add it
to the list.  Note it is untested.

Also kill some whitespace at the end of the line...

19 years agoRemove the entries for isp and ispfw instead of leaving them in an inconsistent
Scott Long [Thu, 3 Feb 2005 23:10:17 +0000 (23:10 +0000)]
Remove the entries for isp and ispfw instead of leaving them in an inconsistent
state.

19 years ago- Split out PCI support.
Matthew N. Dodd [Thu, 3 Feb 2005 23:01:01 +0000 (23:01 +0000)]
- Split out PCI support.
- Add previously removed ISA support.

Submitted by: David S. Madole <david AT madole.net>

19 years agos/worman/woman/
Jens Schweikhardt [Thu, 3 Feb 2005 22:48:29 +0000 (22:48 +0000)]
s/worman/woman/

19 years ago- Use svc_getrpccaller() rather than svc_getcaller() for using
Jun Kuriyama [Thu, 3 Feb 2005 22:21:19 +0000 (22:21 +0000)]
- Use svc_getrpccaller() rather than svc_getcaller() for using
  xt_rtaddr member of SVCXPRT structure.  This allows to use IPv6
  address stored in "struct sockaddr_storage" in "struct netbuf".
- Output the reason of getnameinfo() error.

Reviewed by: alfred

19 years ago- Sort options.
Pawel Jakub Dawidek [Thu, 3 Feb 2005 21:34:39 +0000 (21:34 +0000)]
- Sort options.
- Put 'break' into separate line.

19 years agoFix whitespace.
Pawel Jakub Dawidek [Thu, 3 Feb 2005 21:25:35 +0000 (21:25 +0000)]
Fix whitespace.

19 years agoClarify a bit the BUGS section.
Giorgos Keramidas [Thu, 3 Feb 2005 21:16:23 +0000 (21:16 +0000)]
Clarify a bit the BUGS section.

PR: docs/77024
Submitted by: Adrian Filipi <adrian@ubergeeks.com>
Reviewed by: ru

19 years agoConnect ng_device(4) to the build.
Ruslan Ermilov [Thu, 3 Feb 2005 19:54:58 +0000 (19:54 +0000)]
Connect ng_device(4) to the build.

19 years agoAnytime we write to the RTC's status B register to possibly enable
John Baldwin [Thu, 3 Feb 2005 19:06:03 +0000 (19:06 +0000)]
Anytime we write to the RTC's status B register to possibly enable
interrupts, read from the interrupt status register to clear any pending
interrupts.  Otherwise in some rare cases the RTC would never fire any
interrupts as it constantly thinks it has an interrupt pending.

PR: i386/17800
PR: kern/76776
Submitted by: Jose M. Alcaide jose at we dot lc dot ehu dot es
MFC after: 2 weeks

19 years agocorrect check for unicast key being setup; wpa_supplicant in particular
Sam Leffler [Thu, 3 Feb 2005 16:21:07 +0000 (16:21 +0000)]
correct check for unicast key being setup; wpa_supplicant in particular
uses only the global key state so looking in the per-node key slot is wrong

Submitted by: Tai-hwa Liang

19 years agoMFlibc: Remove extra closing parenthesis.
Ruslan Ermilov [Thu, 3 Feb 2005 15:51:39 +0000 (15:51 +0000)]
MFlibc: Remove extra closing parenthesis.

19 years agoWe do want to print '\n'/'\r'.
Pawel Jakub Dawidek [Thu, 3 Feb 2005 15:47:30 +0000 (15:47 +0000)]
We do want to print '\n'/'\r'.

Pointed out by: stefanf

19 years ago'c' cannot be -1.
Pawel Jakub Dawidek [Thu, 3 Feb 2005 15:23:27 +0000 (15:23 +0000)]
'c' cannot be -1.

Submitted by: stefanf

19 years ago- Move gets() function to libkern (I want to use it outside vfs_mount.c).
Pawel Jakub Dawidek [Thu, 3 Feb 2005 15:10:58 +0000 (15:10 +0000)]
- Move gets() function to libkern (I want to use it outside vfs_mount.c).
- Add buffer size limitations (overflow will not be possible anymore).
- Add 'visible' option, which will allow for passphrase reading in the
  future.
- Remove special treatment of '@' and '#', those two are only confusing.

Discussed with: rwatson
MFC after: 2 weeks

19 years agoReflect recent code changes.
Ruslan Ermilov [Thu, 3 Feb 2005 13:33:34 +0000 (13:33 +0000)]
Reflect recent code changes.

19 years ago- Remove bogus O_CREAT flag. We really don't want to create a file here.
Pawel Jakub Dawidek [Thu, 3 Feb 2005 13:12:17 +0000 (13:12 +0000)]
- Remove bogus O_CREAT flag. We really don't want to create a file here.

  PR: bin/67793
  Submitted by: Amir Shalem <amir@boom.org.il>

- Sync usage with manual page.

Approved by: phk
MFC after: 1 week

19 years agoParse "getifname" using the standard parse string type.
Ruslan Ermilov [Thu, 3 Feb 2005 13:03:31 +0000 (13:03 +0000)]
Parse "getifname" using the standard parse string type.

Fixed an off-by-one error when dealing with interface name
(if_xname is NUL-terminated).

Don't waste time making a copy of if_xname in constructor.

19 years agoFixed an off-by-one error when dealing with interface name
Ruslan Ermilov [Thu, 3 Feb 2005 12:54:18 +0000 (12:54 +0000)]
Fixed an off-by-one error when dealing with interface name
(if_xname is NUL-terminated).

Don't waste time making a copy of if_xname in attach().

19 years agoFixed an off-by-one error when dealing with interface name
Ruslan Ermilov [Thu, 3 Feb 2005 12:50:10 +0000 (12:50 +0000)]
Fixed an off-by-one error when dealing with interface name
(if_xname is IFNAMSIZ-sized and NUL-terminated).

19 years agoAdded ASCII version of the NGM_EIFACE_GET_IFNAME message, "getifname".
Ruslan Ermilov [Thu, 3 Feb 2005 11:52:42 +0000 (11:52 +0000)]
Added ASCII version of the NGM_EIFACE_GET_IFNAME message, "getifname".

19 years agoConstify an argument to a function.
Hartmut Brandt [Thu, 3 Feb 2005 11:51:25 +0000 (11:51 +0000)]
Constify an argument to a function.

Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoGet rid of a number of unneccessary memory allocations.
Hartmut Brandt [Thu, 3 Feb 2005 11:36:19 +0000 (11:36 +0000)]
Get rid of a number of unneccessary memory allocations.

Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoRemoved unused includes.
Ruslan Ermilov [Thu, 3 Feb 2005 11:28:53 +0000 (11:28 +0000)]
Removed unused includes.

19 years agoIgnore consonants, and remove more duplicates.
Ruslan Ermilov [Thu, 3 Feb 2005 10:54:07 +0000 (10:54 +0000)]
Ignore consonants, and remove more duplicates.

19 years agoWhen entering siocnputc() with (kdb_active), don't acquire (or
Robert Watson [Thu, 3 Feb 2005 10:35:05 +0000 (10:35 +0000)]
When entering siocnputc() with (kdb_active), don't acquire (or
release) the sio spin mutex, as use of synchronization primitives in
the debugger can result in substantial problems.  With this patch in
place entering the debugger via a serial console is made
substantially more reliable.

MFC after: 1 week
Tested by: kris
Discussed with: bde

19 years agoRemoved more duplicates (forgot to ignore underscore).
Ruslan Ermilov [Thu, 3 Feb 2005 09:44:30 +0000 (09:44 +0000)]
Removed more duplicates (forgot to ignore underscore).

19 years agoKeep the fortune above this.
Ruslan Ermilov [Thu, 3 Feb 2005 09:28:45 +0000 (09:28 +0000)]
Keep the fortune above this.

Reviewed by: sheldonh

19 years agoRemove more duplicates.
Ruslan Ermilov [Thu, 3 Feb 2005 09:27:01 +0000 (09:27 +0000)]
Remove more duplicates.

19 years agoIgnore vowels, and remove more duplicates, mostly due to typos.
Ruslan Ermilov [Thu, 3 Feb 2005 09:13:02 +0000 (09:13 +0000)]
Ignore vowels, and remove more duplicates, mostly due to typos.

19 years agoTrim some obnoxious warlording.
Dag-Erling Smørgrav [Thu, 3 Feb 2005 08:47:14 +0000 (08:47 +0000)]
Trim some obnoxious warlording.

19 years agoRemoved more duplicates, by ignoring "i", "you", "we",
Ruslan Ermilov [Thu, 3 Feb 2005 08:40:08 +0000 (08:40 +0000)]
Removed more duplicates, by ignoring "i", "you", "we",
"an", "is", "are", "will", "'ll", "is", "'s" to find
them.  More to follow...

19 years agoFix alignment in the last commit.
Ruslan Ermilov [Thu, 3 Feb 2005 08:07:22 +0000 (08:07 +0000)]
Fix alignment in the last commit.

19 years ago- remove NO_MODULES since they've been working for some time
Peter Grehan [Thu, 3 Feb 2005 06:28:17 +0000 (06:28 +0000)]
- remove NO_MODULES since they've been working for some time
- add -fno-omit-framepointer to CFLAGS if DDB is enabled so
  backtraces will still work when optimization is enabled.
  Stolen from Makefile.amd64.

19 years agoDon't build syscons, uart or vpo on PPC.
Peter Grehan [Thu, 3 Feb 2005 06:12:43 +0000 (06:12 +0000)]
Don't build syscons, uart or vpo on PPC.

19 years agoFix up assignment of negative number to char. Char's are unsigned by
Peter Grehan [Thu, 3 Feb 2005 02:35:28 +0000 (02:35 +0000)]
Fix up assignment of negative number to char. Char's are unsigned by
default on PowerPC.

Approved by: mdodd

19 years agoMF dev/wi: merge the pccard lists.
Warner Losh [Thu, 3 Feb 2005 00:40:35 +0000 (00:40 +0000)]
MF dev/wi: merge the pccard lists.

# this hasn't compiled for two weeks and no one noticed!

MFC After: 7 days or so

19 years agoMerged fortunes with fortunes2.
Ruslan Ermilov [Thu, 3 Feb 2005 00:20:37 +0000 (00:20 +0000)]
Merged fortunes with fortunes2.
Sort fortunes (except for the first entry).

19 years agoFix bugs.
Ruslan Ermilov [Thu, 3 Feb 2005 00:18:06 +0000 (00:18 +0000)]
Fix bugs.

Due to one of the bugs (^C was the first character of each line
when sorting), the -b option of sort(1) didn't take any effect,
so stay bug-compatible, and remove the -b option.

19 years agoDon't lose the frame pointer for PPC modules: backtrace doesn't work.
Peter Grehan [Wed, 2 Feb 2005 23:44:24 +0000 (23:44 +0000)]
Don't lose the frame pointer for PPC modules: backtrace doesn't work.

19 years agoModify rpcgen -t output to something that's a lot more useful in
Alfred Perlstein [Wed, 2 Feb 2005 22:14:10 +0000 (22:14 +0000)]
Modify rpcgen -t output to something that's a lot more useful in
today's modern "no sir, not today" compilers.

Basically, when building the table:

 use NULL instead of 0,
 use braces around array initializers,
 case the function pointers to xdrproc_t,
 don't cast function pointers that do not need casting.

MFC After: 1 week

19 years ago - Use a seperate malloc tag for saved inode contents to help in debugging
Jeff Roberson [Wed, 2 Feb 2005 20:30:47 +0000 (20:30 +0000)]
 - Use a seperate malloc tag for saved inode contents to help in debugging
   memory modified after free errors.

Sponsored by: Isilon Systems, Inc.

19 years ago - Correct a typo in kern_rename. tvfslocked should be initialized from
Jeff Roberson [Wed, 2 Feb 2005 17:17:15 +0000 (17:17 +0000)]
 - Correct a typo in kern_rename.  tvfslocked should be initialized from
   tond and not fromnd.  This could lead us to leak Giant, or unlock it
   twice, depending on the filesystems involved.  renames within a single
   filesystem would not have caused any problems.

Sponsored by: Isilon Systems, Inc.

19 years agoDefine SI_SUB_AUDIT, the system boot event to initialize the audit
Robert Watson [Wed, 2 Feb 2005 14:41:03 +0000 (14:41 +0000)]
Define SI_SUB_AUDIT, the system boot event to initialize the audit
subsystem.

Obtained from: TrustedBSD Project
Submitted by: Wayne Salamon <wsalamon at computer dot org>

19 years agoAdd a stub audit_kevents.h, which defines exactly one audit event:
Robert Watson [Wed, 2 Feb 2005 14:27:36 +0000 (14:27 +0000)]
Add a stub audit_kevents.h, which defines exactly one audit event:
AUE_NULL.  This is a place-holder to allow other audit infrastructure
to be introduced, such as an updated syscalls.master file format,
while the license on the real audit_kevents.h is fixed.

19 years agoBack out previous commit, bde@ provided an example of something this
Ken Smith [Wed, 2 Feb 2005 14:21:01 +0000 (14:21 +0000)]
Back out previous commit, bde@ provided an example of something this
breaks.

19 years agoFix the comment.
Ruslan Ermilov [Wed, 2 Feb 2005 14:02:40 +0000 (14:02 +0000)]
Fix the comment.

19 years agoWhitespace and "const" changes to reduce diffs to RELENG_4.
Ruslan Ermilov [Wed, 2 Feb 2005 13:27:03 +0000 (13:27 +0000)]
Whitespace and "const" changes to reduce diffs to RELENG_4.
(Gives the same object when compiled without NETGRAPH_DEBUG.)

19 years agoMake enomem static - it's referenced only in util.c.
Hartmut Brandt [Wed, 2 Feb 2005 11:25:05 +0000 (11:25 +0000)]
Make enomem static - it's referenced only in util.c.

Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoRemove a couple of unused buffer functions.
Hartmut Brandt [Wed, 2 Feb 2005 11:21:26 +0000 (11:21 +0000)]
Remove a couple of unused buffer functions.

Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoAdd a place-holder f_label void * for a future struct label pointer
Robert Watson [Wed, 2 Feb 2005 10:55:32 +0000 (10:55 +0000)]
Add a place-holder f_label void * for a future struct label pointer
required for the port of SELinux FLASK/TE to FreeBSD using the MAC
Framework.

19 years agoRemoved 1695 fortunes also found in fortunes2.
Ruslan Ermilov [Wed, 2 Feb 2005 10:24:11 +0000 (10:24 +0000)]
Removed 1695 fortunes also found in fortunes2.

19 years agoFix "Duplicate mbuf free panic".
Pyun YongHyeon [Wed, 2 Feb 2005 08:35:11 +0000 (08:35 +0000)]
Fix "Duplicate mbuf free panic".
The cause of "Duplicate mbuf free panic" is in the programming
error of hme_load_txmbuf(). The code path of the panic is the
following.

1. Due to unknown reason DMA engine was freezed. So TX descritors
   of HME become full and the last failed attempt to transmit a
   packet had set its associated mbuf address to hme_txdesc
   structure. Also the failed packet is requeued into interface
   queue structure in order to retrasmit it when there are more
   available TX descritors.

2. Since DMA engine was freezed, if_timer starts to decrement its
   counter. When if_timer expires it tries to reset HME. During
   the reset phase, hme_meminit() is called and it frees all
   associated mbuf with descriptors. The last failed mbuf is also
   freed here.

3. After HME reset completed, HME starts to retransmit packets
   by dequeing the first packet in interface queue.(Note! the
   packet was already freed in hme_meminit()!)

4. When a TX completion interrupt is posted by the HME, driver
   tries to free the successfylly transmitted mbuf. Since the
   mbuf was freed in step2, now we get "Duplicate mbuf free panic".

However, the real cause is in DMA engine freeze. Since no fatal
errors reported via interrupts, there might be other cause of
the freeze. I tried hard to understand the cause of DMA engine
freeze but couldn't find any clues. It seems that the freeze
happens under very high network loads(e.g. 7.5-8.0 MB/s TX speed).

Though this fix is not enough to eliminate DMA engine freeze it's
better than panic.

Reported by: jhb via sparc64 ML

19 years agoConvert several typedefs from beeing pointers to structs to be the structs
Hartmut Brandt [Wed, 2 Feb 2005 07:36:18 +0000 (07:36 +0000)]
Convert several typedefs from beeing pointers to structs to be the structs
itself. This will ease constification (think of what 'const Ptr foo'
means if Ptr is a pointer to a struct).

Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoIt was noticed that we do not change a file's access time when it gets
Ken Smith [Wed, 2 Feb 2005 00:21:38 +0000 (00:21 +0000)]
It was noticed that we do not change a file's access time when it gets
executed.  This appears to violate most of the UNIX-ish standards.
One example quote from:

  http://www.opengroup.org/onlinepubs/009695399/functions/exec.html

    Upon successful completion, the exec functions shall mark for update
    the st_atime field of the file. If an exec function failed but was
    able to locate the process image file, whether the st_atime field is
    marked for update is unspecified. Should the exec function succeed,
    the process image file shall be considered to have been opened with
    open().

This appears to take care of it for ufs filesystems, doing the necessary
sanity checks (read-only filesystem, etc) without violating any other
standards (setting atime for any open appears to be allowed in any standards
I could find).

Noticed by: cperciva
Reviewed by: kan, rwatson

19 years ago - Or MPSAFE with the correct set of flags in stat(). This affected only
Jeff Roberson [Tue, 1 Feb 2005 23:43:46 +0000 (23:43 +0000)]
 - Or MPSAFE with the correct set of flags in stat().  This affected only
   the LOOKUP_SHARED case.

Spotted by: jhb

19 years agoForgot to mark the IRQ as MPSAFE.
Poul-Henning Kamp [Tue, 1 Feb 2005 20:34:47 +0000 (20:34 +0000)]
Forgot to mark the IRQ as MPSAFE.

19 years agoThe two PLANEX cards listed in pccarddevs, identified only by their
Warner Losh [Tue, 1 Feb 2005 18:28:09 +0000 (18:28 +0000)]
The two PLANEX cards listed in pccarddevs, identified only by their
CIS, weren't actually used anywhere (other than the generic PC Card
code when certain variables are defined).  They aren't used in NetBSD
either.  Make things simpler by removing them.  Change PLANEX_2 to
PLANEX and tweak wi and owi to use that instead.  The PLANEX id seems
to actually be pci ID assigned to planex, not its pcmcia id.  Ooops.
I don't know if this is a reporting error from where this entry came
from, or if it is a mistake on PLANEX's part.  I suspect the latter,
as ACTIONTEC and NEWMEDIA made the same mistake (although new media
may be because it uses an advansys chip inside).  Make a note of this
in the file.  The 0xc entires may be JEITA assigned, so note that as
well.

# This leaves just 3 entries that are totally unknown: airvast, archos
# and edimax although the arivast number is the same assigned to
# avertec in usb...

19 years agoAdd a IEEE488 driver for PCIIA compatible cards.
Poul-Henning Kamp [Tue, 1 Feb 2005 16:59:23 +0000 (16:59 +0000)]
Add a IEEE488 driver for PCIIA compatible cards.

This driver implements "unaddressed listen only mode", which is what
printers and plotters commonly do on GP-IB busses.

This means that you can capture print/plot like output from your
instruments by configuring them as necessary (good luck!) and

cat -u /dev/gpib0l > /tmp/somefile

Since there is no way to know when no more output is comming you
will have to ctrl-C the cat process when it is done (that is why
the -u is important).

19 years agoIgnore case and all non-word characters, and remove more duplicates.
Ruslan Ermilov [Tue, 1 Feb 2005 16:34:38 +0000 (16:34 +0000)]
Ignore case and all non-word characters, and remove more duplicates.

19 years ago- Add few lines from my personal signature file
Pav Lucistnik [Tue, 1 Feb 2005 16:22:44 +0000 (16:22 +0000)]
- Add few lines from my personal signature file

Suggested by: Florent Thoumie
Approved by: mux

19 years agoCleanup prototypes and remove dead code
Alexey Zelkin [Tue, 1 Feb 2005 16:04:55 +0000 (16:04 +0000)]
Cleanup prototypes and remove dead code

MFC After:      3 days

19 years agoAvoid unconditional increasing of variable leading to SIGSEGV, if req'd
Alexey Zelkin [Tue, 1 Feb 2005 16:01:55 +0000 (16:01 +0000)]
Avoid unconditional increasing of variable leading to SIGSEGV, if req'd
parameter for -lang is missing.

19 years agoRemove another duplicate fortune.
Sean Kelly [Tue, 1 Feb 2005 15:48:48 +0000 (15:48 +0000)]
Remove another duplicate fortune.

19 years agoMove all the remaining list functions into one C-file and remove
Hartmut Brandt [Tue, 1 Feb 2005 15:28:41 +0000 (15:28 +0000)]
Move all the remaining list functions into one C-file and remove
double documentation comments. Remove the 3rd clause (from 4) of the
BSD license because these files have only the UCB copyright.

Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoRemoved full duplicates that only differ in amount of whitespace
Ruslan Ermilov [Tue, 1 Feb 2005 15:15:33 +0000 (15:15 +0000)]
Removed full duplicates that only differ in amount of whitespace
and line breaks.

19 years agoRemove a duplicate.
Dag-Erling Smørgrav [Tue, 1 Feb 2005 14:10:41 +0000 (14:10 +0000)]
Remove a duplicate.

19 years agoIn case of various tunneling protocols, mbuf may pass several interfaces
Gleb Smirnoff [Tue, 1 Feb 2005 14:07:05 +0000 (14:07 +0000)]
In case of various tunneling protocols, mbuf may pass several interfaces
before entering ng_netflow. In this case it will have not NULL m_pkthdr.rcvif.
However, it will enter ng_iface soon with another index. So let in_ifIndex
value configured by user override m_pkthdr.rcvif.

Reported by: Damir Bikmuhametov
MFC after: 1 week

19 years agoRemove debugging stuff that crept in in the previous commit.
Hartmut Brandt [Tue, 1 Feb 2005 11:39:45 +0000 (11:39 +0000)]
Remove debugging stuff that crept in in the previous commit.

19 years ago- change all u_int_XX to uint_XX
Peter Grehan [Tue, 1 Feb 2005 11:17:24 +0000 (11:17 +0000)]
- change all u_int_XX to uint_XX
- cast param for atomic_subtract_long, since Netgraph uses it.

19 years agoClean up include files and file including. Split nonints.h into pieces
Hartmut Brandt [Tue, 1 Feb 2005 10:50:37 +0000 (10:50 +0000)]
Clean up include files and file including. Split nonints.h into pieces
that get included just where they are needed. All headers include the
headers that they need to compile (just with an empty .c file). Sort
includes alphabetically where apropriate and fix some duplicate commenting
for struct Job, struct GNode and struct Shell by removing one version and
inlining the comments into the structure declaration (the comments have been
somewhat outdated).

This patch does not contain functional changes (checked with md5).

Submitted by: Max Okumoto <okumoto@ucsd.edu>

19 years agoIn addition to the PAM environment, export a handful of useful PAM items.
Dag-Erling Smørgrav [Tue, 1 Feb 2005 10:37:07 +0000 (10:37 +0000)]
In addition to the PAM environment, export a handful of useful PAM items.

Suggested by: Ed Maste <emaste@phaedrus.sandvine.ca>

19 years agoAdd openpam_free_envlist(3).
Dag-Erling Smørgrav [Tue, 1 Feb 2005 10:21:07 +0000 (10:21 +0000)]
Add openpam_free_envlist(3).

19 years agoThis commit was generated by cvs2svn to compensate for changes in r141098,
Dag-Erling Smørgrav [Tue, 1 Feb 2005 10:16:17 +0000 (10:16 +0000)]
This commit was generated by cvs2svn to compensate for changes in r141098,
which included commits to RCS files with non-trunk default branches.

19 years agoVendor import of OpenPAM Feterita.
Dag-Erling Smørgrav [Tue, 1 Feb 2005 10:16:17 +0000 (10:16 +0000)]
Vendor import of OpenPAM Feterita.

19 years agoAdd to CFLAGS, rather than overriding it. Use LDFLAGS for the -L option,
Nik Clayton [Tue, 1 Feb 2005 09:26:34 +0000 (09:26 +0000)]
Add to CFLAGS, rather than overriding it.  Use LDFLAGS for the -L option,
and LDLIBS to specify -ltap.

Submitted by: ru

19 years agoUse LC_MESSAGES locale category instead of value of LANG environment variable
Alexey Zelkin [Tue, 1 Feb 2005 08:48:15 +0000 (08:48 +0000)]
Use LC_MESSAGES locale category instead of value of LANG environment variable
while opening of tcsh's message catalog.  It gives us more flexibility, and
do not change current behaviour (since we call setlocale() here)

Submitted by: Rostislav Krasny <rosti_bsd@yahoo.com>
PR: bin/76773
MFC After: 3 days

19 years agoBUS_SPACE_UNRESTRICTED shouldn't be used with the bus_alloc_resource
Warner Losh [Tue, 1 Feb 2005 07:43:34 +0000 (07:43 +0000)]
BUS_SPACE_UNRESTRICTED shouldn't be used with the bus_alloc_resource
interface.  Instead, move to the convenience _any interface.

19 years agoSort functions.
Nate Lawson [Tue, 1 Feb 2005 06:36:27 +0000 (06:36 +0000)]
Sort functions.

19 years agoUpdate copyright, remove "all rights reserved" (since they are not
Bosko Milekic [Tue, 1 Feb 2005 03:17:52 +0000 (03:17 +0000)]
Update copyright, remove "all rights reserved" (since they are not
all reserved, as the lisence makes clear), and strike the third clause
(now this is a 2-clause liberal BSDL as are the rest of files I hold
copyright over).

19 years agoLink against libtap (ports/devel/libtap). Replace the calls to assert()
Nik Clayton [Tue, 1 Feb 2005 01:53:35 +0000 (01:53 +0000)]
Link against libtap (ports/devel/libtap).  Replace the calls to assert()
with calls to libtap::ok1(), and make sure the correct number of tests is
planned for, and that the exit code is correct.

19 years agoadd cpu_idle support for 21066A based lca systems
Bernd Walter [Mon, 31 Jan 2005 23:07:42 +0000 (23:07 +0000)]
add cpu_idle support for 21066A based lca systems

19 years agoUse a taskqueue to handle port status changes.
Shunsuke Akiyama [Mon, 31 Jan 2005 13:58:10 +0000 (13:58 +0000)]
Use a taskqueue to handle port status changes.
Calling ucom layer directly from interrupt context make a panic.

MFC after: 1 week

19 years agonit in /*-
Warner Losh [Mon, 31 Jan 2005 08:16:45 +0000 (08:16 +0000)]
nit in /*-

19 years agoYikes! Fix a typo in a function name that managed to occur twice.
Scott Long [Mon, 31 Jan 2005 07:28:04 +0000 (07:28 +0000)]
Yikes!  Fix a typo in a function name that managed to occur twice.

Submitted by: yongari

19 years agoRemove unsigned casts.
Suleiman Souhlal [Mon, 31 Jan 2005 04:45:45 +0000 (04:45 +0000)]
Remove unsigned casts.

Discussed with: delphij, stefanf
Approved by: grehan (mentor)

19 years agoAdd the keyboard system device before we probe for the keyboard.
Marcel Moolenaar [Mon, 31 Jan 2005 04:31:22 +0000 (04:31 +0000)]
Add the keyboard system device before we probe for the keyboard.
The presence or absence of a keyboard does not change whether an
UART is designed as a keyboard port or not and thus whether we
can use the port as a TTY or not.
We now call sunkbd_attach() even when we didn't previously find
a keyboard. Emit a useful message stating that no keyboard was
found, but don't do anything else.

MFC after: 5 days

19 years agoRevert rev. 1.5.
Marcel Moolenaar [Mon, 31 Jan 2005 04:25:48 +0000 (04:25 +0000)]
Revert rev. 1.5.
It should be safe to attach to all Z8530 controllers again. In fact,
the keyboard works on Ultra 2 machines.

MFC after: 5 days

19 years agoUpdate an additional reference to the rate of ISN tick callouts that was
Robert Watson [Mon, 31 Jan 2005 01:35:01 +0000 (01:35 +0000)]
Update an additional reference to the rate of ISN tick callouts that was
missed in tcp_subr.c:1.216: projected_offset must also reflect how often
the tcp_isn_tick() callout will fire.

MFC after: 2 weeks
Submitted by: silby

19 years agoRemove a dead vtophys() reference.
Scott Long [Mon, 31 Jan 2005 01:26:22 +0000 (01:26 +0000)]
Remove a dead vtophys() reference.

19 years agoChange the state allocator from using regular malloc to using
Christian S.J. Peron [Mon, 31 Jan 2005 00:48:39 +0000 (00:48 +0000)]
Change the state allocator from using regular malloc to using
a UMA zone instead. This should eliminate a bit of the locking
overhead associated with with malloc and reduce the memory
consumption associated with each new state.

Reviewed by: rwatson, andre
Silence on: ipfw@
MFC after: 1 week

19 years agoHave tcp_isn_tick() fire 100 times a second, rather than HZ times a
Robert Watson [Sun, 30 Jan 2005 23:30:28 +0000 (23:30 +0000)]
Have tcp_isn_tick() fire 100 times a second, rather than HZ times a
second; since the default hz has changed to 1000 times a second,
this resulted in unecessary work being performed.

MFC after: 2 weeks
Discussed with: phk, cperciva
General head nod: silby

19 years agoFix a logic bug that caused DSR to never be deasserted.
Marcel Moolenaar [Sun, 30 Jan 2005 22:14:30 +0000 (22:14 +0000)]
Fix a logic bug that caused DSR to never be deasserted.

MFC after: 5 days

19 years agoUpdate the text of an assertion to reflect changes made in revision 1.148.
Alan Cox [Sun, 30 Jan 2005 21:29:47 +0000 (21:29 +0000)]
Update the text of an assertion to reflect changes made in revision 1.148.
Submitted by: tegge

Eliminate an unnecessary, temporary increment of the backing object's
reference count in vm_object_qcollapse().  Reviewed by: tegge

19 years agoThe current version of mkisofs is warning us that the -P option will
Ken Smith [Sun, 30 Jan 2005 21:10:52 +0000 (21:10 +0000)]
The current version of mkisofs is warning us that the -P option will
be changing with the next version, and we should be using -publisher
instead.

Also warned by: trhodes
MFC after: 1 week