]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
23 years agofix typos.
ben [Fri, 29 Dec 2000 01:58:47 +0000 (01:58 +0000)]
fix typos.

PR: 23934
Submitted by: Alex Kapranoff <alex@kapran.bitmcnit.bryansk.su>

23 years agoAdd some text saying that this page may occasionally be out of date with the
ben [Fri, 29 Dec 2000 01:54:39 +0000 (01:54 +0000)]
Add some text saying that this page may occasionally be out of date with the
real make.conf.

23 years agoAdd $FreeBSD$
peter [Fri, 29 Dec 2000 01:23:57 +0000 (01:23 +0000)]
Add $FreeBSD$

23 years agoMap FreeBSD character device hard disks to Linux block device hard disks.
paul [Fri, 29 Dec 2000 00:44:42 +0000 (00:44 +0000)]
Map FreeBSD character device hard disks to Linux block device hard disks.

This fixes the problem with VMWARE not being able to use raw disks.

23 years agoadd a couple off offset defines for ATIO2s
mjacob [Thu, 28 Dec 2000 23:27:54 +0000 (23:27 +0000)]
add a couple off offset defines for ATIO2s

23 years agoroll to 2.01.26 level
mjacob [Thu, 28 Dec 2000 23:23:28 +0000 (23:23 +0000)]
roll to 2.01.26 level

23 years agoremove MAINTAINER file
mjacob [Thu, 28 Dec 2000 23:23:11 +0000 (23:23 +0000)]
remove MAINTAINER file

23 years agoNew release notes: vinum(4) updates, ICMP ECHO and TSTAMP reply
bmah [Thu, 28 Dec 2000 22:11:22 +0000 (22:11 +0000)]
New release notes:  vinum(4) updates, ICMP ECHO and TSTAMP reply
rate limiting, ICMP admin prohibited handling, PECOFF support,
retired kernfs(5), killall(1) behavior restored, rm(1) -v output
change, lpr(1)/lpq(1)/lpd(8) enhancements, pkg_update(1), libdisk
install-time configuration of boot0 for i386.

Also note MFC of an(4) changes for WEP.

23 years agoThis is repo-copied from vnioctl.h
phk [Thu, 28 Dec 2000 20:59:55 +0000 (20:59 +0000)]
This is repo-copied from vnioctl.h

23 years agoAdd mdconfig into the build.
phk [Thu, 28 Dec 2000 20:59:35 +0000 (20:59 +0000)]
Add mdconfig into the build.

23 years agoAdd the "mdctl" device under the "std" stanza.
phk [Thu, 28 Dec 2000 20:59:19 +0000 (20:59 +0000)]
Add the "mdctl" device under the "std" stanza.

23 years agoPreliminary scaffolding for the new integrated vn+md device driver.
phk [Thu, 28 Dec 2000 20:57:57 +0000 (20:57 +0000)]
Preliminary scaffolding for the new integrated vn+md device driver.

I decided to work on the md(4) driver and integrate the vn(4)
functionality into it mainly based on the name being more suitable.
Ideally 'vd' as in "virtual disk" would probably be the most logical
but our sound-master pointed out that this would cause uncontrollable
fits of giggles in the brits.  Another complication would the needed
changes to the ramdisk boot/root functionality.

The vn driver will stay around for some time after I complete this
merge for transition reasons, and I'll make it whine to people that
they should migrate to the md(4) driver for some time before it
dies.

The kernel part of the new md(4) driver will be committed after more
testing.

23 years agoFix the broken options that were in the development version I committed by mistake.
paul [Thu, 28 Dec 2000 19:53:44 +0000 (19:53 +0000)]
Fix the broken options that were in the development version I committed by mistake.

23 years ago* Fix a segfault when timed(8) receives a packet with a bad tsp_type.
ben [Thu, 28 Dec 2000 19:50:07 +0000 (19:50 +0000)]
* Fix a segfault when timed(8) receives a packet with a bad tsp_type.
* Check that received packets aren't too short, as this could cause other
  problems.

Reviewed by: imp, markm

23 years agoBring the man page up to date with the current version of the script.
paul [Thu, 28 Dec 2000 19:42:06 +0000 (19:42 +0000)]
Bring the man page up to date with the current version of the script.

23 years agoAdd pkg_update, which can replace currently installed ports with new versions
paul [Thu, 28 Dec 2000 18:47:23 +0000 (18:47 +0000)]
Add pkg_update, which can replace currently installed ports with new versions
while taking of updating all the dependencies.

23 years agoRemove prototypes for pthread_attr_[gs]etfloatstate(), which we don't
jasone [Thu, 28 Dec 2000 18:46:21 +0000 (18:46 +0000)]
Remove prototypes for pthread_attr_[gs]etfloatstate(), which we don't
implement.

23 years agoMove all VCS ids to a consistent location.
obrien [Thu, 28 Dec 2000 18:21:32 +0000 (18:21 +0000)]
Move all VCS ids to a consistent location.

23 years agoAdd multiple inclusion protection.
obrien [Thu, 28 Dec 2000 18:18:23 +0000 (18:18 +0000)]
Add multiple inclusion protection.

PR: 23902
Submitted by: Christian Weisgerber <naddy@mips.inka.de>

23 years agogrammar gnit: "relationships (plural) _are_ complex"
hoek [Thu, 28 Dec 2000 17:17:49 +0000 (17:17 +0000)]
grammar gnit: "relationships (plural) _are_ complex"

23 years agotypo: modifes -> modifies
hoek [Thu, 28 Dec 2000 16:44:47 +0000 (16:44 +0000)]
typo: modifes -> modifies

23 years agoMdoc(7)ify.
ru [Thu, 28 Dec 2000 16:18:43 +0000 (16:18 +0000)]
Mdoc(7)ify.

23 years agoRetire kernfs (userland part).
des [Thu, 28 Dec 2000 12:59:57 +0000 (12:59 +0000)]
Retire kernfs (userland part).

23 years agoRetire kernfs (kernel part).
des [Thu, 28 Dec 2000 12:17:35 +0000 (12:17 +0000)]
Retire kernfs (kernel part).

23 years agoReflect rev 1.18 in crypt.c. Note that this section is somewhat
peter [Thu, 28 Dec 2000 11:56:45 +0000 (11:56 +0000)]
Reflect rev 1.18 in crypt.c.  Note that this section is somewhat
mangled and could do with some word-smithing.

23 years agoHindsight is wonderful, but I got cold feet over the crypt(3) default
peter [Thu, 28 Dec 2000 11:23:01 +0000 (11:23 +0000)]
Hindsight is wonderful, but I got cold feet over the crypt(3) default
so I am backing it out for now.  The problem is that some random program
calling crypt() could be passing a DES salt and the crypt(3) library
would encrypt it in md5 mode and there would be a password mismatch as a
result.  I wrote a validater function for the DES code to verify that
a salt is valid for DES, but I realized there were too many strange things
to go wrong.  passwd(1), pw(8) etc still generate md5 passwords by default
for /etc/master.passwd, so this is almost academic.  It is a big deal for
things that have their own crypt(3)-ed password strings (.htaccess,
etc etc).  Those are the things I do not want to break.

My DES salt recognizer basically checked if the salt was either 2 or
13 characters long, or began with '_' (_PASSWORD_EFMT1).  I think it
would have worked but I have seen way too much crypt() mishandling
in the past.

23 years agoMerge into a single US-exportable libcrypt, which only provides
peter [Thu, 28 Dec 2000 10:32:02 +0000 (10:32 +0000)]
Merge into a single US-exportable libcrypt, which only provides
one-way hash functions for authentication purposes.  There is no more
"set the libcrypt->libXXXcrypt" nightmare.
- Undo the libmd.so hack, use -D to hide the md5c.c internals.
- Remove the symlink hacks in release/Makefile
- the algorthm is set by set_crypt_format() as before.  If this is
  not called, it tries to heuristically figure out the hash format, and
  if all else fails, it uses the optional auth.conf entry to chose the
  overall default hash.
- Since source has non-hidden crypto in it there may be some issues with
  having the source it in some countries, so preserve the "secure/*"
  division.  You can still build a des-free libcrypt library if you want
  to badly enough.  This should not be a problem in the US or exporting
  from the US as freebsd.org had notified BXA some time ago.  That makes
  this stuff re-exportable by anyone.
- For consistancy, the default in absence of any other clues is md5.  This
  is to try and minimize POLA across buildworld where folk may suddenly
  be activating des-crypt()-hash support.  Since the des hash may not
  always be present, it seemed sensible to make the stronger md5 algorithm
  the default.
All things being equal, no functionality is lost.

Reviewed-by: jkh
(flame-proof suit on)

23 years agoSend a SIGCONT when detaching or continuing the excution of a traced
ps [Thu, 28 Dec 2000 08:34:21 +0000 (08:34 +0000)]
Send a SIGCONT when detaching or continuing the excution of a traced
process.  This fixes a problem when attaching to a process in gdb
and the process staying in the STOP'd state after quiting gdb.
This whole process seems a bit suspect, but this seems to work.

Reviewed by: peter

23 years agoPull out the module path from the loader. ie: if you boot from
peter [Thu, 28 Dec 2000 08:14:58 +0000 (08:14 +0000)]
Pull out the module path from the loader.  ie: if you boot from
/boot/kernel.foobar/* then that had better be in the path ahead of the
others.

Submitted by:  Daniel J. O'Connor <darius@dons.net.au>
PR: 23662

23 years agoWhen trying to deduce the diskname from the name so we can run
phk [Wed, 27 Dec 2000 22:28:40 +0000 (22:28 +0000)]
When trying to deduce the diskname from the name so we can run
parallel fsck's one per drive, use the shortest prefix ending in
a digit rather than the longest prefix ending in a digit.

This makes "/dev/ad0s1a" and "/dev/ad0s2a" appear to both reside
on the disk "/dev/ad0" and consequently they will be fsck'ed
sequentially rather than in parallel as now.

In general this heuristic is rather soft and errorprone.  For
instance ccd may often reside on two or more physical disks.  A
good solution would be to look for passes larger than 1 until no
disks are found in a particular pass, that way people could put
ccd stripes in pass 3... and have them fsck'ed sequentially.

Reviewed by: mjacob

23 years agoSmall fix for bpf compat:
bmilekic [Wed, 27 Dec 2000 22:20:13 +0000 (22:20 +0000)]
Small fix for bpf compat:
Make malloc() use M_NOWAIT istead of M_DONTWAIT and in the
bpf_compat case, define M_NOWAIT to be M_DONTWAIT.

23 years agoUse official API to <sys/queue.h> instead of groping around inside the
phk [Wed, 27 Dec 2000 21:38:06 +0000 (21:38 +0000)]
Use official API to <sys/queue.h> instead of groping around inside the
data structures.

Reviewed by: imp

23 years agoLink stringlist.3 to sl_{add,find,free,init}.3
ben [Wed, 27 Dec 2000 20:00:01 +0000 (20:00 +0000)]
Link stringlist.3 to sl_{add,find,free,init}.3

23 years agoPrepare for mdoc(7)NG.
ru [Wed, 27 Dec 2000 16:52:31 +0000 (16:52 +0000)]
Prepare for mdoc(7)NG.

23 years agoPrepare for mdoc(7)NG.
ru [Wed, 27 Dec 2000 16:37:32 +0000 (16:37 +0000)]
Prepare for mdoc(7)NG.

23 years agoPrepare for mdoc(7)NG.
ru [Wed, 27 Dec 2000 15:30:30 +0000 (15:30 +0000)]
Prepare for mdoc(7)NG.

23 years agoPrepare for mdoc(7)NG.
ru [Wed, 27 Dec 2000 14:40:52 +0000 (14:40 +0000)]
Prepare for mdoc(7)NG.

23 years agoPrepare for mdoc(7)NG.
ru [Wed, 27 Dec 2000 14:22:05 +0000 (14:22 +0000)]
Prepare for mdoc(7)NG.

23 years agoPrepare for mdoc(7)NG.
ru [Wed, 27 Dec 2000 14:15:06 +0000 (14:15 +0000)]
Prepare for mdoc(7)NG.

23 years agoPrepare for mdoc(7)NG.
ru [Wed, 27 Dec 2000 13:59:17 +0000 (13:59 +0000)]
Prepare for mdoc(7)NG.

23 years agoMajor bugfix and minor update. This should resolve the current issues
msmith [Wed, 27 Dec 2000 13:14:56 +0000 (13:14 +0000)]
Major bugfix and minor update.  This should resolve the current issues
with the driver locking up under load.

 - Restructure so that we use a static pool of commands/FIBs, rather than
   allocating them in clusters.  The cluster allocation just made things
   more complicated, and allowed us to waste more memory in peak load
   situations.
 - Make queueing macros more like my other drivers.  This adds queue stats
   for free.  Add some debugging to take advantage of this.
 - Reimplement the periodic timeout scan.  Kick the interrupt handler
   and the start routine every scan as well, just to be safe.  Track busy
   commands properly.
 - Bring resource cleanup into line with resource allocation.  We should
   now clean up correctly after a failed probe/unload/etc.
 - Try to start new commands when old ones are completed.  We weren't doing
   this before, which could lead to deadlock when the controller was full.
 - Don't try to build a new command if we have found a deferred command.
   This could cause us to lose the deferred command.
 - Use diskerr() to report I/O errors.
 - Don't bail if the AdapterInfo structure is the wrong size.  Some variation
   seems to be normal.  We need to improve our handing of 2.x firmware sets.
 - Improve some comments in an attempt to try to make things clearer.
 - Restructure to avoid some warnings.

23 years agochange irq handler slightly, get rid of superflous messages
cg [Wed, 27 Dec 2000 04:04:36 +0000 (04:04 +0000)]
change irq handler slightly, get rid of superflous messages

23 years agoinclude tcp header files to get the prototype for tcp_seq_vs_sess
assar [Wed, 27 Dec 2000 03:02:29 +0000 (03:02 +0000)]
include tcp header files to get the prototype for tcp_seq_vs_sess

23 years agoMake zalloc and zfree non-inline functions. This avoids having to
assar [Wed, 27 Dec 2000 02:54:37 +0000 (02:54 +0000)]
Make zalloc and zfree non-inline functions.  This avoids having to
have the code calling these be compiled with the same setting for
INVARIANTS and SMP.

Reviewed by: dillon

23 years agoFix a quoting problem I introduced.
dougb [Tue, 26 Dec 2000 20:55:18 +0000 (20:55 +0000)]
Fix a quoting problem I introduced.
Fix a tiny style problem while I'm here.

Submitted by: knu

23 years agoFix an annoying message ``gdb: ptrace(PT_GETDBREGS) failed: No such process''
joerg [Tue, 26 Dec 2000 20:38:46 +0000 (20:38 +0000)]
Fix an annoying message ``gdb: ptrace(PT_GETDBREGS) failed: No such process''
when using gdb on a remote target.  The fix is to restrict PT_GETDBREGS
calls to `child' and `freebsd-uthreads' targets solely.

I've been in some conversation with Brian about this, and this solution
seems to be the most appropriate one.

PR: gnu/21685
Submitted by: bsd

23 years agoThis implements a better launder limiting solution. There was a solution
dillon [Tue, 26 Dec 2000 19:41:38 +0000 (19:41 +0000)]
This implements a better launder limiting solution.  There was a solution
in 4.2-REL which I ripped out in -stable and -current when implementing the
low-memory handling solution.  However, maxlaunder turns out to be the saving
grace in certain very heavily loaded systems (e.g. newsreader box).  The new
algorithm limits the number of pages laundered in the first pageout daemon
pass.  If that is not sufficient then suceessive will be run without any
limit.

Write I/O is now pipelined using two sysctls, vfs.lorunningspace and
vfs.hirunningspace.  This prevents excessive buffered writes in the
disk queues which cause long (multi-second) delays for reads.  It leads
to more stable (less jerky) and generally faster I/O streaming to disk
by allowing required read ops (e.g. for indirect blocks and such) to occur
without interrupting the write stream, amoung other things.

NOTE: eventually, filesystem write I/O pipelining needs to be done on a
per-device basis.  At the moment it is globalized.

23 years agoAdd CDR progress ioctl.
sos [Tue, 26 Dec 2000 15:24:53 +0000 (15:24 +0000)]
Add CDR progress ioctl.

23 years agoUse M_ZERO flag on malloc when approbiate.
sos [Tue, 26 Dec 2000 12:05:49 +0000 (12:05 +0000)]
Use M_ZERO flag on malloc when approbiate.

23 years agoUpdate to match the progress interface on blank command.
sos [Tue, 26 Dec 2000 11:56:15 +0000 (11:56 +0000)]
Update to match the progress interface on blank command.

23 years agoUpdate the burncd interface a bit, dont block the ATA channel on
sos [Tue, 26 Dec 2000 11:55:44 +0000 (11:55 +0000)]
Update the burncd interface a bit, dont block the ATA channel on
blank & fixate commands and provide a progress interface for the
blank command (for now)

23 years agoUse diskerr for printing disk hard/soft errors.
sos [Tue, 26 Dec 2000 11:49:57 +0000 (11:49 +0000)]
Use diskerr for printing disk hard/soft errors.

23 years agoAdd these files that were used from contrib/gcc/. They are taken from
obrien [Tue, 26 Dec 2000 10:20:22 +0000 (10:20 +0000)]
Add these files that were used from contrib/gcc/.  They are taken from
GCC 2.7.2.3 as that was the version of GCC in active use before the switch
to ELF.
The GCC 2.9[67] versions of these files carry more baggage and I'm not sure
the are appropriate for this linker.

23 years agoFix for the case where the first two entries returned by readdir() are
will [Tue, 26 Dec 2000 07:36:07 +0000 (07:36 +0000)]
Fix for the case where the first two entries returned by readdir() are
actually NOT '.' and '..'.  Apparently this isn't the case when accessing
a directory via XFS over NFS on SGI systems.  Since I don't have access to
an environment like that, this will sit out in -current for at least six
weeks.  However, the patch proposed by the submitter seems acceptable, so
I've decided to commit it to the tree, in the hope that it will solve some
problems without bringing up others.

PR: 23300
Submitted by: Jim Pirzyk <Jim.Pirzyk@disney.com>

23 years agoAdd OZO8008 - Zoom (33.6k Modem).
tanimura [Tue, 26 Dec 2000 06:52:57 +0000 (06:52 +0000)]
Add OZO8008 - Zoom (33.6k Modem).

PR: kern/23336
Submitted by: Paulo Menezes <root@samurai.dee.uc.pt>

23 years agoMinor newbus/style(9) cleanups.
imp [Tue, 26 Dec 2000 06:38:04 +0000 (06:38 +0000)]
Minor newbus/style(9) cleanups.
o Move the ax88190 code to its own function.
o Move all device_method_t, driver_t and DRIVER_MODULE definitions to the
  end of files.
o Wrap a few lines > 80 characters.
o Use the same devclass for all ed drivers.  This allows machines with
  multiple types of cards to have their cards numbered correctly.  Before,
  you could wind up with two ed0's.
o Protect if_edvar.h from multiple includes because I was there.

23 years agoadd PC-Card entry, ELECOM Laneed LD-CDL/TX
sanpei [Tue, 26 Dec 2000 00:59:34 +0000 (00:59 +0000)]
add PC-Card entry, ELECOM Laneed LD-CDL/TX

ed1 at port 0x240-0x25f irq 3 slot 0 on pccard0
ed1: address 00:90:fe:xx:xx:xx, type Linksys (16 bit)

Submitted by:   taguchi@tohoku.iij.ad.jp

23 years agoadd Melco LPC2-TX
sanpei [Tue, 26 Dec 2000 00:47:41 +0000 (00:47 +0000)]
add Melco LPC2-TX

23 years agoThe change in rev 1.8 makes the C version of killall behave the same as the
obrien [Mon, 25 Dec 2000 20:44:05 +0000 (20:44 +0000)]
The change in rev 1.8 makes the C version of killall behave the same as the
old perl one.

23 years agoUse the real user ID so non-root users can kill setuid processes
ps [Mon, 25 Dec 2000 19:26:07 +0000 (19:26 +0000)]
Use the real user ID so non-root users can kill setuid processes
they started.

PR: 23356
Submitted by: Maxim Konovalov <maxim@macomnet.ru>

23 years agoAdd support for PC-card router configuration. (MFPAO)
toshi [Mon, 25 Dec 2000 09:21:18 +0000 (09:21 +0000)]
Add support for PC-card router configuration. (MFPAO)

o Add the removable_interfaces variable for list of removable network
  interfaces (PC-card ethernet, wireless network and USB ethernet etc).
o ifconfig_<ifn>_alias0, static_routes_<ifn>, removable_route_flush,
  /etc/start_if.<ifn> and /etc/stop_if.<ifn> are support.
o removable_route_flush variable is set to "NO" if you want to use the
  machine as gateway using two or more removable network cards. If
  static routing is needed use static_routes_<ifn> instead of
  static_routes or defaultrouter.
o The optional static_routes_<ifn> variable is likely static_routes.
o /etc/start_if.<ifn> and /etc/stop_if.<ifn> are shell script to be
  specified that are called when a card is inserted or removed.

23 years agomdoc(7) police: formatting fixes.
ru [Mon, 25 Dec 2000 09:08:44 +0000 (09:08 +0000)]
mdoc(7) police: formatting fixes.

23 years agomove parametric defines and structure definitions from .h to .c
cg [Mon, 25 Dec 2000 02:49:28 +0000 (02:49 +0000)]
move parametric defines and structure definitions from .h to .c

23 years agoindent debugging printfs, change a couple of messages slightly
cg [Mon, 25 Dec 2000 02:21:16 +0000 (02:21 +0000)]
indent debugging printfs, change a couple of messages slightly

23 years agoupdate to return correct blocksize- these chips have a fixed irq rate, so
cg [Mon, 25 Dec 2000 01:42:13 +0000 (01:42 +0000)]
update to return correct blocksize- these chips have a fixed irq rate, so
block size varies only with format and rate.

23 years agoDon't abuse the SCCS `@(#)' for RCS.
obrien [Mon, 25 Dec 2000 01:16:12 +0000 (01:16 +0000)]
Don't abuse the SCCS `@(#)' for RCS.

Requested by: bde

23 years agoAdd ioctls to acpi_cmbat and acpi_acad. These use mike's acpi_register_ioctl().
iwasaki [Sun, 24 Dec 2000 19:12:10 +0000 (19:12 +0000)]
Add ioctls to acpi_cmbat and acpi_acad.  These use mike's acpi_register_ioctl().
Fix wrong AML method calling in acpi_cmbat.

23 years agoAs far as I can tell, you have never been able to use 'ab' as a unit
dwmalone [Sun, 24 Dec 2000 16:15:28 +0000 (16:15 +0000)]
As far as I can tell, you have never been able to use 'ab' as a unit
name if 'a' and 'b' are unit names. You can use 'a b' though.

23 years agoo Almost all the config line changed to `auto', except for
toshi [Sun, 24 Dec 2000 15:38:38 +0000 (15:38 +0000)]
o Almost all the config line changed to `auto', except for
  multifunction cards and I recieved reports that the card does not
  workd by `config auto'. (MFPAO)
o Remove static assign of the IRQ number.
o Remove two duplicated entries.
o Join some entries using regex and fixed matching order problem.

These changes for boot.flp.

Suggested by: sanpei

23 years agoFix output of -v option.
toshi [Sun, 24 Dec 2000 15:30:36 +0000 (15:30 +0000)]
Fix output of -v option.

23 years agoRemoved pc98-hack of aout tools.
kato [Sun, 24 Dec 2000 14:58:34 +0000 (14:58 +0000)]
Removed pc98-hack of aout tools.

23 years agoDisabled boot2 for the moment, because it needs aout tools.
kato [Sun, 24 Dec 2000 14:54:34 +0000 (14:54 +0000)]
Disabled boot2 for the moment, because it needs aout tools.

23 years agoBack out the modules `make obj' speedup.
obrien [Sun, 24 Dec 2000 13:29:58 +0000 (13:29 +0000)]
Back out the modules `make obj' speedup.
I cannot find a way to do this with the more esoteric source dir schemes
I thought of.

23 years agoClean up the source directory finding code, and in the process fix a
joe [Sun, 24 Dec 2000 12:21:15 +0000 (12:21 +0000)]
Clean up the source directory finding code, and in the process fix a
problem that occurs when a program's source isn't found.

23 years agoUpdate the "icmp_admin_prohib_like_rst" code to check the tcp-window and
phk [Sun, 24 Dec 2000 10:57:21 +0000 (10:57 +0000)]
Update the "icmp_admin_prohib_like_rst" code to check the tcp-window and
to be configurable with respect to acting only in SYN or in all TCP states.

PR: 23665
Submitted by: Jesper Skriver <jesper@skriver.dk>

23 years agoClean up the -j/-y/--bzip entry in usage().
obrien [Sun, 24 Dec 2000 10:52:23 +0000 (10:52 +0000)]
Clean up the -j/-y/--bzip entry in usage().

Add the -U and --unlink-first options which are the offical verions of our
--unlink localism.
Add support for the "TAR_OPTIONS" environmental variable.

Obtained from: GNU tar 1.13.18

23 years agoAdd -j as the offical "pipe thru bzip2 compression program" option.
obrien [Sun, 24 Dec 2000 10:19:55 +0000 (10:19 +0000)]
Add -j as the offical "pipe thru bzip2 compression program" option.
Make -I equivalent to -T for compatiblity with Solaris.

Obtained from: GNU tar 1.13.18

23 years agofairly substantial rewrite- seperate out play/record code, implement
cg [Sun, 24 Dec 2000 03:56:41 +0000 (03:56 +0000)]
fairly substantial rewrite- seperate out play/record code, implement
setblocksize, simplify resource allocation

23 years agocompensate for broken codecs that return 0xffff on unimplemented registers
cg [Sun, 24 Dec 2000 03:33:21 +0000 (03:33 +0000)]
compensate for broken codecs that return 0xffff on unimplemented registers
instead of 0.

23 years agoHandle multiple words in $daily_clean_disks_files
brian [Sun, 24 Dec 2000 03:22:45 +0000 (03:22 +0000)]
Handle multiple words in $daily_clean_disks_files

PR: 23805
Submitted mostly by: Norbert Papke <npapke@telus.net>

23 years agoChange M_WAIT to M_TRYWAIT, the new name of the flag.
bmilekic [Sun, 24 Dec 2000 03:08:17 +0000 (03:08 +0000)]
Change M_WAIT to M_TRYWAIT, the new name of the flag.

23 years agoimplement setblocksize properly
cg [Sat, 23 Dec 2000 22:24:43 +0000 (22:24 +0000)]
implement setblocksize properly

23 years agoAdd `gcov' to the mix.
obrien [Sat, 23 Dec 2000 19:49:28 +0000 (19:49 +0000)]
Add `gcov' to the mix.

Requested by: David Malone <dwmalone@maths.tcd.ie>
PR: 18574

23 years agoProtect proc.p_pptr and proc.p_children/p_sibling with the
jake [Sat, 23 Dec 2000 19:43:10 +0000 (19:43 +0000)]
Protect proc.p_pptr and proc.p_children/p_sibling with the
proctree_lock.

linprocfs not locked pending response from informal maintainer.

Reviewed by: jhb, -smp@

23 years agoRemove floppy.tree from the build directory.
joe [Sat, 23 Dec 2000 19:33:13 +0000 (19:33 +0000)]
Remove floppy.tree from the build directory.

23 years agoMF4: Update the bridge kernel config file to reflect the structure and
joe [Sat, 23 Dec 2000 19:27:42 +0000 (19:27 +0000)]
MF4: Update the bridge kernel config file to reflect the structure and
     order that Luigi has just committed to RELENG_4.

23 years agoMF4: revision 1.3.2.1.
joe [Sat, 23 Dec 2000 19:11:53 +0000 (19:11 +0000)]
MF4: revision 1.3.2.1.
     inetd is not always present, check before invoking it.

23 years agoDetermine the names of the picobsd releases automatically.
joe [Sat, 23 Dec 2000 16:53:35 +0000 (16:53 +0000)]
Determine the names of the picobsd releases automatically.

Use the 'realclean' target in the crunch.mk file for cleaning the object
build tree.

23 years agouse ``config auto'' for ACCTON EN2216-PCMCIA-ETHERNET
sanpei [Sat, 23 Dec 2000 16:11:50 +0000 (16:11 +0000)]
use ``config auto'' for ACCTON EN2216-PCMCIA-ETHERNET

Submitted by: taguchi@tohoku.iij.ad.jp

23 years agouse ``config auto'' for IBM PCMCIA Ethernet I/II
sanpei [Sat, 23 Dec 2000 16:01:16 +0000 (16:01 +0000)]
use ``config  auto'' for IBM PCMCIA Ethernet I/II

Submitted by: WAKABAYAI Susumu <susumu@wakabaya.net>

23 years agofix MELCO LPC3-TX config
sanpei [Sat, 23 Dec 2000 15:27:54 +0000 (15:27 +0000)]
fix MELCO LPC3-TX config
use ``config auto''
add ether line

Submitted by:   bsd-nomads mailing list
14954: Seigo Tanimura <tanimura@r.dl.itc.u-tokyo.ac.jp>
14967: Yoichi Nakayama <yoichi@eken.phys.nagoya-u.ac.jp>
14971: WAKABAYAI Susumu <susumu@wakabaya.net>

23 years agouse ``config auto "sio" ?'' for Xircom CreditCard Modem CM-56T
sanpei [Sat, 23 Dec 2000 13:07:48 +0000 (13:07 +0000)]
use ``config  auto "sio" ?'' for Xircom CreditCard Modem CM-56T

PR: conf/23068
Submitted by: Matt <kienow@infinet.com>

23 years agoMake sure we have a non-null proc pointer before referring to fields
mjacob [Sat, 23 Dec 2000 07:33:32 +0000 (07:33 +0000)]
Make sure we have a non-null proc pointer before referring to fields
off of it.

23 years agofix a panic if the requested blocksize was greater than the data rate,
cg [Sat, 23 Dec 2000 06:09:43 +0000 (06:09 +0000)]
fix a panic if the requested blocksize was greater than the data rate,
resulting in a divide by 0.

23 years agokill a warning
cg [Sat, 23 Dec 2000 05:19:32 +0000 (05:19 +0000)]
kill a warning

23 years agofix a typo
cg [Sat, 23 Dec 2000 05:16:39 +0000 (05:16 +0000)]
fix a typo

23 years agoadd sys/dev/sound/pcm/buffer.c, lefo out of last commit
cg [Sat, 23 Dec 2000 03:27:09 +0000 (03:27 +0000)]
add sys/dev/sound/pcm/buffer.c, lefo out of last commit

23 years agoupdate code dealing with snd_dbuf objects to do so using a functional interface
cg [Sat, 23 Dec 2000 03:16:13 +0000 (03:16 +0000)]
update code dealing with snd_dbuf objects to do so using a functional interface

modify chn_setblocksize() to pick a default soft-blocksize appropriate to the
sample rate and format in use.  it will aim for a power of two size small
enough to generate block sizes of at most 20ms.  it will also set the
hard-blocksize taking into account rate/format conversions in use.

update drivers to implement setblocksize  correctly:
updated, tested:  sb16, emu10k1, maestro, solo
updated, untested:  ad1816, ess, mss, sb8, csa
not updated:  ds1, es137x, fm801, neomagic, t4dwave, via82c686

i lack hardware to test: ad1816, csa, fm801, neomagic
others will be updated/tested in the next few days.

23 years ago*sigh* The `pecoff' module was added to all arches w/out testing a compile
obrien [Fri, 22 Dec 2000 22:33:15 +0000 (22:33 +0000)]
*sigh* The `pecoff' module was added to all arches w/out testing a compile
on all arches.  This fixes the Alpha kernel build.

23 years agoProbe generic i960 devices for the old AMI signature as well as the current
msmith [Fri, 22 Dec 2000 22:23:56 +0000 (22:23 +0000)]
Probe generic i960 devices for the old AMI signature as well as the current
one, in order to support adapters with older firmware.

Submitted by: "Prasanth S." <prasanths@ami.co.in>