]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
23 years agoFix logic error which caused a page fault in kernel mode immediately
jdp [Sat, 19 Aug 2000 23:50:41 +0000 (23:50 +0000)]
Fix logic error which caused a page fault in kernel mode immediately
on boot up.

Submitted by: Bosko Milekic <bmilekic@technokratis.com>

23 years agoFixed a missing splx() in if_addmulti(). Was broken in rev.1.28.
bde [Sat, 19 Aug 2000 22:10:10 +0000 (22:10 +0000)]
Fixed a missing splx() in if_addmulti().  Was broken in rev.1.28.

23 years agoDon't build pccard for now until I can resolve the linker interaction
imp [Sat, 19 Aug 2000 19:26:03 +0000 (19:26 +0000)]
Don't build pccard for now until I can resolve the linker interaction
issue that I'm seeing.

23 years agoWork around a possible kernel linker bug. It appears that the data
imp [Sat, 19 Aug 2000 19:25:23 +0000 (19:25 +0000)]
Work around a possible kernel linker bug.  It appears that the data
elements defined by foo_if.c aren't sharable amoung modules (I'm
working on a bug report for it now), or else I don't understand
something.  It showed up as kobj functions not being called.

In any event, link in the pcic and pccard parts of the modules to work
around the problem for now.

23 years agoo Move to using PCCARD_SOFTC(dev)
imp [Sat, 19 Aug 2000 19:22:04 +0000 (19:22 +0000)]
o Move to using PCCARD_SOFTC(dev)
o fill in the size of the actual softc, rather than 1 in data structure
o minor debugging improvements.

23 years agoo move to using PCIC_SOFTC(dev) to get the softc
imp [Sat, 19 Aug 2000 19:20:25 +0000 (19:20 +0000)]
o move to using PCIC_SOFTC(dev) to get the softc
o move pcic_deactivate into pcic_detach
o Better debug messages

23 years agoRemove obsolete entry about wdc hangs
jkh [Sat, 19 Aug 2000 18:54:37 +0000 (18:54 +0000)]
Remove obsolete entry about wdc hangs

23 years agofix SNDCTL_DSP_GETODELAY
cg [Sat, 19 Aug 2000 18:29:46 +0000 (18:29 +0000)]
fix SNDCTL_DSP_GETODELAY

PR: kern/20340
Submitted by: Takashi SHIRAI <shirai@nintendo.co.jp>

23 years agodon't complain about bad irqs if using a shared irq
cg [Sat, 19 Aug 2000 18:22:20 +0000 (18:22 +0000)]
don't complain about bad irqs if using a shared irq

23 years agoRecognize ESS 1888 in Dell Latitude XPi
cg [Sat, 19 Aug 2000 18:17:15 +0000 (18:17 +0000)]
Recognize ESS 1888 in Dell Latitude XPi

PR: kern/18714
Submitted by: Sean O'Connell <sean@stat.Duke.EDU>

23 years agoFix the arguments to [(1) (AKA test(1)). If the string being tested
markm [Sat, 19 Aug 2000 12:28:23 +0000 (12:28 +0000)]
Fix the arguments to [(1) (AKA test(1)). If the string being tested
by -n is nonexistant, then the following -d was misinterpreted with
a strange error. By putting double quotes (") around the argument,
we can be sure there is _something_ there that we can check a zero
length against.

23 years agoSilence -Wall; main()'s second argument was "unsigned char **",
markm [Sat, 19 Aug 2000 11:45:18 +0000 (11:45 +0000)]
Silence -Wall; main()'s second argument was "unsigned char **",
and this had filtered down into too many other places.

23 years agoDisabled serial console.
nyan [Sat, 19 Aug 2000 10:19:43 +0000 (10:19 +0000)]
Disabled serial console.

23 years agoReplace the mbuf external reference counting code with something
dwmalone [Sat, 19 Aug 2000 08:32:59 +0000 (08:32 +0000)]
Replace the mbuf external reference counting code with something
that should be better.

The old code counted references to mbuf clusters by using the offset
of the cluster from the start of memory allocated for mbufs and
clusters as an index into an array of chars, which did the reference
counting. If the external storage was not a cluster then reference
counting had to be done by the code using that external storage.

NetBSD's system of linked lists of mbufs was cosidered, but Alfred
felt it would have locking issues when the kernel was made more
SMP friendly.

The system implimented uses a pool of unions to track external
storage. The union contains an int for counting the references and
a pointer for forming a free list. The reference counts are
incremented and decremented atomically and so should be SMP friendly.
This system can track reference counts for any sort of external
storage.

Access to the reference counting stuff is now through macros defined
in mbuf.h, so it should be easier to make changes to the system in
the future.

The possibility of storing the reference count in one of the
referencing mbufs was considered, but was rejected 'cos it would
often leave extra mbufs allocated. Storing the reference count in
the cluster was also considered, but because the external storage
may not be a cluster this isn't an option.

The size of the pool of reference counters is available in the
stats provided by "netstat -m".

PR: 19866
Submitted by: Bosko Milekic <bmilekic@dsuper.net>
Reviewed by: alfred (glanced at by others on -net)

23 years agoCall bus_generic_detach on pcic before deleting the children. This
imp [Sat, 19 Aug 2000 05:01:07 +0000 (05:01 +0000)]
Call bus_generic_detach on pcic before deleting the children.  This
has the side effect of detaching the children before I delete them.
When I put the last commits in a loop for loop it died after 5 or so
iterations.  After this change, I lasted 50 before I stopped the test.

23 years agoRemove pcic_isa_identify. I don't think it is needed.
imp [Sat, 19 Aug 2000 04:47:09 +0000 (04:47 +0000)]
Remove pcic_isa_identify.  I don't think it is needed.
Add pcic_detach which removes all of the pcic's children.

This lets me load/unload pcic multiple times w/o having multiple
instances of each slot attached to pcic.

23 years agoIncrease the default NAPIC from 1 to 2 as a bandaid until we allocate
msmith [Fri, 18 Aug 2000 20:09:15 +0000 (20:09 +0000)]
Increase the default NAPIC from 1 to 2 as a bandaid until we allocate
these dynamically (ie. typically you shouldn't have to set NAPIC at all)

23 years agoRemove gratuitous free() call when we use special .SHELL target.
ru [Fri, 18 Aug 2000 10:41:37 +0000 (10:41 +0000)]
Remove gratuitous free() call when we use special .SHELL target.

23 years agoRemove an unwanted space from the "Shutting down daemon processes"
sheldonh [Fri, 18 Aug 2000 10:34:11 +0000 (10:34 +0000)]
Remove an unwanted space from the "Shutting down daemon processes"
message.

PR: 20692
Submitted by: Andre Albsmeier <andre.albsmeier@mchp.siemens.de>

23 years agoIntroduce vop_stdinactive() and make it the default if no vop_inactive
phk [Fri, 18 Aug 2000 10:01:02 +0000 (10:01 +0000)]
Introduce vop_stdinactive() and make it the default if no vop_inactive
is declared.

Sort and prune a few vop_op[].

23 years agoFix bug in rev 1.54; the terminating `;;' for the ibcs2_enable=YES
sheldonh [Fri, 18 Aug 2000 09:43:18 +0000 (09:43 +0000)]
Fix bug in rev 1.54; the terminating `;;' for the ibcs2_enable=YES
case got lost.

23 years agoDocument ibcs2_loaders and provide a default for it in
sheldonh [Fri, 18 Aug 2000 09:37:50 +0000 (09:37 +0000)]
Document ibcs2_loaders and provide a default for it in
etc/defaults/rc.conf .

23 years agoMerged from sys/i386/i386/machdep.c revision 1.402.
kato [Fri, 18 Aug 2000 09:22:01 +0000 (09:22 +0000)]
Merged from sys/i386/i386/machdep.c revision 1.402.

23 years ago(forced commit)
sheldonh [Fri, 18 Aug 2000 08:53:26 +0000 (08:53 +0000)]
(forced commit)

The previous change should have been attributed to:

PR: 20692
Submitted by: Andre Albsmeier <andre.albsmeier@mchp.siemens.de>

23 years agoThe output of this script got a little broken in rev 1.225, with
sheldonh [Fri, 18 Aug 2000 08:11:23 +0000 (08:11 +0000)]
The output of this script got a little broken in rev 1.225, with
the output for the linux_enable and svr4_enable stuff "floating
in the middle of nowhere".

Give them their own section, called "additional ABI support".

23 years agoDon't rely on a shell scripts to do the simple job of loading
sheldonh [Fri, 18 Aug 2000 08:07:02 +0000 (08:07 +0000)]
Don't rely on a shell scripts to do the simple job of loading
kernel modules for ibcs2_enable and svr4_enable.

Don't rely on a shell script to do the neglibly less simple
job of loading a kernel module and running one command for
linux_enable.

These shell scripts are going away.

23 years agoThe slip driver used to allocate a mbuf cluster without attaching
dwmalone [Fri, 18 Aug 2000 08:02:31 +0000 (08:02 +0000)]
The slip driver used to allocate a mbuf cluster without attaching
it to a mbuf. This patch makes it attach it to mbuf.  This patch
is in preperation for Bosko Milekic's mbuf external reference
counting patches.

PR: 19866 (first stage)
Submitted by: Ian Dowse <iedowse@maths.tcd.ie>
Reviewed by: alfred

23 years agoFix the urio and ugen devices (I meant unit2minor, not minor2unit)
brian [Fri, 18 Aug 2000 07:01:19 +0000 (07:01 +0000)]
Fix the urio and ugen devices (I meant unit2minor, not minor2unit)

Problem pointed out by: markm

23 years agoAdd a PPP_NOSUID hook
brian [Fri, 18 Aug 2000 00:09:46 +0000 (00:09 +0000)]
Add a PPP_NOSUID hook

23 years agoMake -DNOSUID (or -DPPP_NOSUID) possible to build ppp without SUID
brian [Fri, 18 Aug 2000 00:01:44 +0000 (00:01 +0000)]
Make -DNOSUID (or -DPPP_NOSUID) possible to build ppp without SUID
capabilities.

23 years agoWarn that the ``alias'' command is depricated.
brian [Fri, 18 Aug 2000 00:01:30 +0000 (00:01 +0000)]
Warn that the ``alias'' command is depricated.
We still process it for now though.

23 years ago(forced commit: previous commit message was truncated)
archie [Thu, 17 Aug 2000 23:51:47 +0000 (23:51 +0000)]
(forced commit: previous commit message was truncated)

Fix two bugs:

- The ftpPassive() function seemed to think that the PASV command
  is a toggle. This is not true (however, the ftp(1) "pass" command
  is indeed a toggle). So no need to emit a "PASV" command each time
  this function is called; that's handled by ftp_file_op().

- check_passive() is supposed to check if FTP_PASSIVE_MODE is defined,
  and if so, override ftpPassive(). However, it was overriding
  ftpPassive() even if FTP_PASSIVE_MODE was not defined, rendering
  calls to the ftpPassive() function completely ineffectual.

Also, clarify the relationship between ftpPassive() and the
FTP_PASSIVE_MODE environment variable in the man page.

23 years agoFix two bugs:
archie [Thu, 17 Aug 2000 23:46:13 +0000 (23:46 +0000)]
Fix two bugs:
- The ftpPassive()

23 years agoAllow people to set the default boot slice with boot0cfg.
dwmalone [Thu, 17 Aug 2000 18:42:13 +0000 (18:42 +0000)]
Allow people to set the default boot slice with boot0cfg.

PR: 18923
Submitted by: Ian Dowse <iedowse@maths.tcd.ie>
Reviewed by: jhb
Approved by: rnordier

23 years agoRemove the warning about recording on the Solo. The workarounds appear
nsayer [Thu, 17 Aug 2000 18:30:59 +0000 (18:30 +0000)]
Remove the warning about recording on the Solo. The workarounds appear
to clear up the hwptr problems.

23 years agoRemove extraneous arguments to the Os (operating system) and Nm
sheldonh [Thu, 17 Aug 2000 16:57:31 +0000 (16:57 +0000)]
Remove extraneous arguments to the Os (operating system) and Nm
(name) macros.

Do not terminate the cross-reference list in the SEE ALSO section
with a period.

23 years agoImplement the GDB counterpart to use hardware watchpoints in connection
joerg [Thu, 17 Aug 2000 16:27:26 +0000 (16:27 +0000)]
Implement the GDB counterpart to use hardware watchpoints in connection
with Brian's kernel support for i386 debug registers.  This makes
watchpoints actually usable for real-life problems.  Note: you can
only set watchpoints on 1-, 2- or 4-byte locations, gdb automatically
falls back to [sloooow] software watchpoints when attempting to use
them on variables which don't fit into this category.  To circumvent
this, one can use the following hack:

watch *(int *)0x<some address>

David O'Brien is IMHO considering to get this fully integrated into the
official GDB, but as long as we've got the i386/* files sitting around
in our private FreeBSD tree here, the feature can now be tested more
extensively, so i'm committing this for the time being.

This work has been done in order to debug a tix toolkit problem, thus
it has been sponsored by teh Deutsche Post AG.

Reviewed by: bsd (not the operating system, but Brian :-)

23 years agoFix style bugs and poor wording introduced in rev 1.12.
sheldonh [Thu, 17 Aug 2000 16:09:27 +0000 (16:09 +0000)]
Fix style bugs and poor wording introduced in rev 1.12.

Submitted by: bde

23 years agoFix style bugs and inconsistencies introduced in rev 1.16.
sheldonh [Thu, 17 Aug 2000 16:08:06 +0000 (16:08 +0000)]
Fix style bugs and inconsistencies introduced in rev 1.16.

Submitted by: bde

23 years agoThe previous delta was wrong; an empty MAKEOBJDIRPREFIX should
sheldonh [Thu, 17 Aug 2000 14:41:48 +0000 (14:41 +0000)]
The previous delta was wrong; an empty MAKEOBJDIRPREFIX should
cause the working directory to be used.  Make it so.

When we're more convinced that it'll work, we might try this
to avoid a shell invocation:

.if defined(MAKEOBJDIRPREFIX) && !empty(MAKEOBJDIRPREFIX) &&
    exists(${CANONICALOBJDIR}/)

Reported by: bde

23 years agoUse an MRU of 1504 with 1500 byte MRRUs
brian [Thu, 17 Aug 2000 14:40:42 +0000 (14:40 +0000)]
Use an MRU of 1504 with 1500 byte MRRUs

23 years agoMake this file look prettier
brian [Thu, 17 Aug 2000 14:39:24 +0000 (14:39 +0000)]
Make this file look prettier
Bump the MRU by 4 bytes to make room for the MP header
Down the autoload threshold to a practical value
Don't specify the ISDN bandwidth as 65536 (ahem!)
Don't specifiy a carrier period (the default of 6 seconds is fine)

23 years agoDon't let an illegal value for dr7 get set, which can lead to an
bsd [Thu, 17 Aug 2000 14:35:23 +0000 (14:35 +0000)]
Don't let an illegal value for dr7 get set, which can lead to an
unexpected TRCTRAP.

Reported by: John W. De Boskey <jwd@FreeBSD.org>

23 years agoIf we're in MP mode with a single open link, MP link level compression
brian [Thu, 17 Aug 2000 14:14:54 +0000 (14:14 +0000)]
If we're in MP mode with a single open link, MP link level compression
isn't open and the links MRU >= our MRRU, send outbound traffic as
PROTO_IP rather than PROTO_MP.  This shaves some bytes off the front
of each packet 'till the second link is brought up.

Idea obtained from: Cisco

23 years agoDon't set an arbitrary limit on username lengths; use MAXLOGNAME
sheldonh [Thu, 17 Aug 2000 12:31:17 +0000 (12:31 +0000)]
Don't set an arbitrary limit on username lengths; use MAXLOGNAME
instead.

PR: 20675
Submitted by: Vladimir B Grebenschikov <vova@express.ru>

23 years ago- Fixed the conversion to bus_space interface.
nyan [Thu, 17 Aug 2000 12:15:45 +0000 (12:15 +0000)]
- Fixed the conversion to bus_space interface.
- Added PC-98 Cbus devices support.
  The original patch is submitted by chi@bd.mbn.or.jp (Chiharu Shibata)
- Removed old ed driver.

23 years agoAllow a /etc/finger.conf file that contains finger aliases
brian [Thu, 17 Aug 2000 10:59:17 +0000 (10:59 +0000)]
Allow a /etc/finger.conf file that contains finger aliases
This allows people who's email address differs from their account name
to be fingerable.

Submitted by: Mark Knight <markk@knigma.org>

23 years agoClear extraneous arguments to the name (Nm) macro.
sheldonh [Thu, 17 Aug 2000 08:17:16 +0000 (08:17 +0000)]
Clear extraneous arguments to the name (Nm) macro.
Mark PRISON_ROOT and EPERM as defined values (Dv).

23 years agoRevert gratuitous whitespace changes from revisions 1.111 and 1.112.
jhb [Thu, 17 Aug 2000 07:51:48 +0000 (07:51 +0000)]
Revert gratuitous whitespace changes from revisions 1.111 and 1.112.

23 years agoMention that basic network options now include firewall/security options as
jhb [Thu, 17 Aug 2000 06:04:13 +0000 (06:04 +0000)]
Mention that basic network options now include firewall/security options as
well.

23 years agoAdd support for loading this as a module (before there were some
imp [Thu, 17 Aug 2000 05:52:18 +0000 (05:52 +0000)]
Add support for loading this as a module (before there were some
problems).  Also pick an irq if one isn't specified.  Ditto for
memory.

Add additional printfs in the error paths.  These will disappear in time.

There are still some panics, but this is a good checkpoint.

23 years agoAdd in a hack that installs 'src/sys/${MACHINE_ARCH}/GENERIC.hints' as
jhb [Thu, 17 Aug 2000 05:43:41 +0000 (05:43 +0000)]
Add in a hack that installs 'src/sys/${MACHINE_ARCH}/GENERIC.hints' as
/boot/device.hints in the bin dist during releases so that current snapshots
have a chance of booting up ok after installing.  The real fix for this
problem is to rewrite userconfig in Forth, stick it in the loader, axe
userconfig from the kernel, and extract the hints from the booted kernel in
sysinstall similar to the way we generate /boot/kernel.conf right now.  For
now, however, this will have to do.

23 years agoFix style bugs (including ones introduced from OpenBSD).
green [Wed, 16 Aug 2000 23:37:16 +0000 (23:37 +0000)]
Fix style bugs (including ones introduced from OpenBSD).

23 years agoAllow use of the ${MAKE_SHELL} variable to specify alternate shells for
green [Wed, 16 Aug 2000 23:31:43 +0000 (23:31 +0000)]
Allow use of the ${MAKE_SHELL} variable to specify alternate shells for
make(1) to use.  Setting it to "sh" and "ksh" are the only values which
work right ATM; I wouldn't expect "csh" to get you far ;)

23 years agoFix a couple cases where p_trespass wasn't transitioned into place.
green [Wed, 16 Aug 2000 23:28:54 +0000 (23:28 +0000)]
Fix a couple cases where p_trespass wasn't transitioned into place.

Make RTP_SET (rtprio) only accessible to real root, not root in jails.

23 years agoFix a whitespace bogon.
jhb [Wed, 16 Aug 2000 23:08:28 +0000 (23:08 +0000)]
Fix a whitespace bogon.

23 years agoPrepare for a cleanup of pmap module API pollution introduced by the
tegge [Wed, 16 Aug 2000 21:24:44 +0000 (21:24 +0000)]
Prepare for a cleanup of pmap module API pollution introduced by the
suggested fix in PR 12378.

Keep track of all existing pmaps independent of existing processes.

This allows for a process to temporarily connect to a different address
space without the risk of missing an update of the original address space if
the kernel grows.

pmap_pinit2() is no longer needed on the i386 platform but is left as a
stub until the alpha pmap code is updated.

PR: 12378

23 years agoMention more functions in the headers so whatis(1) can find them, and
markm [Wed, 16 Aug 2000 19:53:38 +0000 (19:53 +0000)]
Mention more functions in the headers so whatis(1) can find them, and
create links to functions so added found.

23 years agoGrammar police; make the English style more consistent throughout,
markm [Wed, 16 Aug 2000 19:52:03 +0000 (19:52 +0000)]
Grammar police; make the English style more consistent throughout,
and give suser_xxx some more airtime.

23 years agoAdd in the linux, tcl83, and tk83 ports categories.
jhb [Wed, 16 Aug 2000 18:44:40 +0000 (18:44 +0000)]
Add in the linux, tcl83, and tk83 ports categories.

PR: bin/20328
Submitted by: obrien

23 years agoAdd new 'ruby' ports category.
jhb [Wed, 16 Aug 2000 17:44:35 +0000 (17:44 +0000)]
Add new 'ruby' ports category.

PR: conf/20629
Submitted by: asami

23 years agoAdd unit2minor() and use it where appropriate so that we handle
brian [Wed, 16 Aug 2000 16:42:50 +0000 (16:42 +0000)]
Add unit2minor() and use it where appropriate so that we handle
units > 255 properly.

Fix usio while I'm there

23 years agoRemove unnecessary extern definition of strrchr. It is defined in
imp [Wed, 16 Aug 2000 16:39:42 +0000 (16:39 +0000)]
Remove unnecessary extern definition of strrchr.  It is defined in
string.h, which already was included.

23 years agoAdd cross reference to the ast device.
imp [Wed, 16 Aug 2000 16:38:47 +0000 (16:38 +0000)]
Add cross reference to the ast device.
Remove reference to the now defunct wt device.

23 years agooptreset is declared in unistd.h, so we don't need to declare it here again.
imp [Wed, 16 Aug 2000 15:36:14 +0000 (15:36 +0000)]
optreset is declared in unistd.h, so we don't need to declare it here again.

Reviews by: ken

23 years agoDo not allow the cleandir target to blow away the entire source
sheldonh [Wed, 16 Aug 2000 14:58:42 +0000 (14:58 +0000)]
Do not allow the cleandir target to blow away the entire source
directory when MAKEOBJDIRPREFIX is defined but empty.

Reported by: Vernon Schryver <vjs@calcite.rhyolite.com>

23 years agobackout previous change for now
darrenr [Wed, 16 Aug 2000 13:59:50 +0000 (13:59 +0000)]
backout previous change for now

23 years agoadd extern for inetsw
darrenr [Wed, 16 Aug 2000 13:54:56 +0000 (13:54 +0000)]
add extern for inetsw

23 years agoISDN B channels have a bandwidth of 64000, not 65536
brian [Wed, 16 Aug 2000 13:54:04 +0000 (13:54 +0000)]
ISDN B channels have a bandwidth of 64000, not 65536

Pointed out by: Christian Weisgerber <naddy@mips.inka.de>

23 years agoDisable part of my 8-bits fixes from December 1999.
cracauer [Wed, 16 Aug 2000 12:23:57 +0000 (12:23 +0000)]
Disable part of my 8-bits fixes from December 1999.

Serious fix still needed, see discussion on -current
(Subject: /bin/sh dumps core with here-document of 8bit text)

Problem in this code originally spotted by
Jun Kuriyama <kuriyama@FreeBSD.org>

23 years agoFix type builtin for absolute paths and relative paths with directory
cracauer [Wed, 16 Aug 2000 12:08:02 +0000 (12:08 +0000)]
Fix type builtin for absolute paths and relative paths with directory
names in them.

Also use a colon in the answer of `type` everytime the questioned item
is not usable.

PR: bin/20567

23 years agoIf the -p option is used, use setproctitle() to hide it's argument
brian [Wed, 16 Aug 2000 12:01:17 +0000 (12:01 +0000)]
If the -p option is used, use setproctitle() to hide it's argument
(a password) asap.

23 years agoFrom submitter:
cracauer [Wed, 16 Aug 2000 10:39:43 +0000 (10:39 +0000)]
From submitter:
growstackblock() sometimes relocates a stack_block considered empty
without properly relocating stack marks referencing that block.
The first call to popstackmark() with the unrelocated stack mark
as argument then causes sh to abort.

Relocating the relevant stack marks seems to solve this problem.

The patch changes the semantics of popstackmark() somewhat.  It can
only be called once after a call to setstackmark(), thus cmdloop() in
main.c needs an extra call to setstackmark().

PR: bin/19983
Submitted by: Tor.Egge@fast.no
Reviewed by: Gerald Pfeifer <pfeifer@dbai.tuwien.ac.at>

23 years agoadd missing splx(), when outgoing interface queue is full on tunnelled
itojun [Wed, 16 Aug 2000 09:56:45 +0000 (09:56 +0000)]
add missing splx(), when outgoing interface queue is full on tunnelled
IPsec packet output.  KAME PR 280.

23 years agoDecrease the size of mfsroot.flp for PC-98 to 1.2MB.
nyan [Wed, 16 Aug 2000 09:30:47 +0000 (09:30 +0000)]
Decrease the size of mfsroot.flp for PC-98 to 1.2MB.

23 years agoFix `control socket: Protocol not supported' failure in
ru [Wed, 16 Aug 2000 09:12:33 +0000 (09:12 +0000)]
Fix `control socket: Protocol not supported' failure in
standalone -D mode when neither -4 nor -6 is specified.

23 years agosetproctitle() doesn't need to be called with root privs, so move
brian [Wed, 16 Aug 2000 09:07:23 +0000 (09:07 +0000)]
setproctitle() doesn't need to be called with root privs, so move
it from id.c into defs.c

23 years agoDefault the bandwidth of an ISDN B channel to 65536bps
brian [Wed, 16 Aug 2000 09:06:59 +0000 (09:06 +0000)]
Default the bandwidth of an ISDN B channel to 65536bps

23 years agooptreset is declared in unistd.h now.
imp [Wed, 16 Aug 2000 07:36:30 +0000 (07:36 +0000)]
optreset is declared in unistd.h now.

23 years agoRearrange the probe a bit, hopefully this will help to eleminate
sos [Wed, 16 Aug 2000 07:09:04 +0000 (07:09 +0000)]
Rearrange the probe a bit, hopefully this will help to eleminate
some of the fake devices sometimes seen on single device ATA
channels.

Proberly fail on failures in ata-disk.c, retry instead of hang.

Cleanup the VIA probe/init code a bit.

Fix a couble of missing free's in atapi-cd.c in the changer code.

23 years agoDon't explicitly declare optarg and optind. These are declared in
imp [Wed, 16 Aug 2000 05:14:49 +0000 (05:14 +0000)]
Don't explicitly declare optarg and optind.  These are declared in
unistd.h, which is already included.

23 years agoopen_drive:
grog [Wed, 16 Aug 2000 04:31:37 +0000 (04:31 +0000)]
open_drive:
  Add support for AMD RAID controllers as "disks".

Requested-by: Marius Bendiksen <mbendiks@eunet.no>
  Remove potential panic when attempting to open non-existent drivers.

init_drive: Return error codes correctly.  Previously it would
            occasionally return 0.  The error was redetected
            elsewhere, but this was causing a number of confusing
            error messages.

23 years agoAdd another USR modem.
peter [Tue, 15 Aug 2000 21:03:28 +0000 (21:03 +0000)]
Add another USR modem.

PR: 20621
Submitted by: Masanori Taira <mtaira@logicaleffect.com>

23 years agoMention the gawk-3.0.6 upgrade.
sheldonh [Tue, 15 Aug 2000 18:05:05 +0000 (18:05 +0000)]
Mention the gawk-3.0.6 upgrade.

23 years agoRemove redundant code left over on removal of /dev/rXXX handling
rnordier [Tue, 15 Aug 2000 18:04:16 +0000 (18:04 +0000)]
Remove redundant code left over on removal of /dev/rXXX handling
in r1.12.

23 years agoMove the running of the Linux and SVR4 compat bits to the arch independent
obrien [Tue, 15 Aug 2000 18:02:49 +0000 (18:02 +0000)]
Move the running of the Linux and SVR4 compat bits to the arch independent
location.

Sponsored by:   LinuxWorld frustration

23 years agoPartially revert r1.12, in which a warnx() if the device is not a
rnordier [Tue, 15 Aug 2000 17:42:24 +0000 (17:42 +0000)]
Partially revert r1.12, in which a warnx() if the device is not a
character device is changed to an errx().  This unnecessarily
compromises device independence.

23 years agoGrammar police; "has been written by" --> "was written by".
markm [Tue, 15 Aug 2000 15:23:46 +0000 (15:23 +0000)]
Grammar police; "has been written by" --> "was written by".

23 years agoImprove the cross-references.
markm [Tue, 15 Aug 2000 15:14:13 +0000 (15:14 +0000)]
Improve the cross-references.

23 years agoMake the pppoe-in config entry a bit more useful (and verbose)
brian [Tue, 15 Aug 2000 15:13:16 +0000 (15:13 +0000)]
Make the pppoe-in config entry a bit more useful (and verbose)

23 years agoAdd a sample ifconfig entry for an IPX address family address, to give
sheldonh [Tue, 15 Aug 2000 15:09:34 +0000 (15:09 +0000)]
Add a sample ifconfig entry for an IPX address family address, to give
IPX folks a fighting chance of figuring this out themselves.  I can't
work out how to document this carefully in rc.conf(5), but this ought
to close the PR.

PR: 17904
Reported by: John Gelnaw <jeg@hawk.circa.ufl.edu>

23 years agoFix grammar and a function name.
markm [Tue, 15 Aug 2000 15:09:13 +0000 (15:09 +0000)]
Fix grammar and a function name.

23 years agoDescribe the process used for the gawk-3.0.6 update in
sheldonh [Tue, 15 Aug 2000 14:42:50 +0000 (14:42 +0000)]
Describe the process used for the gawk-3.0.6 update in
FREEBSD-upgrade and remove the config.h which should never
have existed in the contrib area.  That file has been moved
into gnu/usr.bin/awk where it belongs.

23 years agoInstall mode 644, not mode 664 during kernel compile. Yes, Peter, I am
kris [Tue, 15 Aug 2000 11:06:56 +0000 (11:06 +0000)]
Install mode 644, not mode 664 during kernel compile. Yes, Peter, I am
psycho :-)

23 years agoUpdate for the newly-updated gawk-3.0.6.
sheldonh [Tue, 15 Aug 2000 10:47:38 +0000 (10:47 +0000)]
Update for the newly-updated gawk-3.0.6.

23 years agoUpdate for the newly-updated gawk-3.0.6.
sheldonh [Tue, 15 Aug 2000 10:31:12 +0000 (10:31 +0000)]
Update for the newly-updated gawk-3.0.6.

23 years agoMaintain input and output throughput averages and choose the highest
brian [Tue, 15 Aug 2000 10:25:42 +0000 (10:25 +0000)]
Maintain input and output throughput averages and choose the highest
of the two when calculating the MP throughput average for the ``set
autoload'' implementation.

This makes more sense as all links I know of are full-duplex.  This
also means that people may need to adjust their autoload settings
as 100% bandwidth is now the theoretical maximum rather than 200%
(but of course, halfing the current settings is probably not the
correct answer either!).

This involves a ppp version bump as we need to pass an extra
throughput array through the MP local domain socket.

23 years agoMerge the vendor release of gawk-3.0.6 onto HEAD.
sheldonh [Tue, 15 Aug 2000 10:20:01 +0000 (10:20 +0000)]
Merge the vendor release of gawk-3.0.6 onto HEAD.

This addresses one known problem in the PR database.

PR: 13921
Reported by: Jin Guojun <j_guojun@lbl.gov>

23 years agoThis commit was generated by cvs2svn to compensate for changes in r64666,
sheldonh [Tue, 15 Aug 2000 10:02:07 +0000 (10:02 +0000)]
This commit was generated by cvs2svn to compensate for changes in r64666,
which included commits to RCS files with non-trunk default branches.

23 years agoUpdate vendor branch to gawk-3.0.6.
sheldonh [Tue, 15 Aug 2000 10:02:07 +0000 (10:02 +0000)]
Update vendor branch to gawk-3.0.6.