]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
25 years agoAdd JKH's auth.conf-file parser to turn Kerberos on/off in userland.
markm [Fri, 9 Oct 1998 06:31:45 +0000 (06:31 +0000)]
Add JKH's auth.conf-file parser to turn Kerberos on/off in userland.

25 years agoInitialize boot configuration files on floppy correctly. We now
jkh [Fri, 9 Oct 1998 06:28:28 +0000 (06:28 +0000)]
Initialize boot configuration files on floppy correctly.  We now
support 3 kinds of floppies:  boot+mfs (standard), boot, mfs.

25 years agoAdd some evil temporary phys-to-kern translation for mfs.
jkh [Fri, 9 Oct 1998 06:21:12 +0000 (06:21 +0000)]
Add some evil temporary phys-to-kern translation for mfs.

25 years agoAdd support for yet another "cyclone" board, with PCI device id 0x905A.
wpaul [Fri, 9 Oct 1998 03:59:24 +0000 (03:59 +0000)]
Add support for yet another "cyclone" board, with PCI device id 0x905A.
This is a 100BaseFX board with SC fiber media connectors. I don't actually
have one of these but I've been told it works with the xl driver.

Submitted by: Jason Wright from the openbsd group

25 years agoMissing defines for the kernel environment and module metadata lookup
msmith [Fri, 9 Oct 1998 01:44:09 +0000 (01:44 +0000)]
Missing defines for the kernel environment and module metadata lookup
functions

25 years agoDon't try to print out the signal number that caused core to be
jdp [Fri, 9 Oct 1998 01:43:08 +0000 (01:43 +0000)]
Don't try to print out the signal number that caused core to be
dumped.  It is not in a fixed location in the file, so at best,
garbage comes out.  I don't think this code was ever valid.

25 years agoinclude proper header for Mike's new stuff.
jkh [Fri, 9 Oct 1998 01:40:56 +0000 (01:40 +0000)]
include proper header for Mike's new stuff.

25 years agoAdd new variable NO_FILTER_SHLIBS -- it will disable ELF <-> a.out shlib
asami [Fri, 9 Oct 1998 01:27:21 +0000 (01:27 +0000)]
Add new variable NO_FILTER_SHLIBS -- it will disable ELF <-> a.out shlib
name conversion.  Use it for binary ports that come with its own private
shlib dirs, ports that install linux compatibility libraries (thus following
their naming conventions and not ours), etc.

Reviewed by: Martin Cracauer <cracauer@FreeBSD.ORG>

25 years agoAll these have to include stdio.h now.
jkh [Fri, 9 Oct 1998 00:39:09 +0000 (00:39 +0000)]
All these have to include stdio.h now.

25 years agoKernel environment access, preloaded module lookup.
msmith [Fri, 9 Oct 1998 00:31:29 +0000 (00:31 +0000)]
Kernel environment access, preloaded module lookup.

25 years agoInitialise kernel environment and module metadata pointers.
msmith [Fri, 9 Oct 1998 00:31:06 +0000 (00:31 +0000)]
Initialise kernel environment and module metadata pointers.

25 years agoNew files for kernel environment and module metadata interfaces.
msmith [Fri, 9 Oct 1998 00:30:50 +0000 (00:30 +0000)]
New files for kernel environment and module metadata interfaces.

25 years agoFix a panic on SMP systems, caused by sleeping while holding a
jdp [Fri, 9 Oct 1998 00:24:49 +0000 (00:24 +0000)]
Fix a panic on SMP systems, caused by sleeping while holding a
simple-lock.

The reviewer raises the following caveat: "I believe these changes
open a non-critical race condition when adding memory to the pool
for the zone. I think what will happen is that you could have two
threads that are simultaneously adding additional memory when the
pool runs out. This appears to not be a problem, however, since
the re-aquisition of the lock will protect the list pointers."
The submitter agrees that the race is non-critical, and points out
that it already existed for the non-SMP case.  He suggests that
perhaps a sleep lock (using the lock manager) should be used to
close that race.  This might be worth revisiting after 3.0 is
released.

Reviewed by: dg (David Greenman)
Submitted by: tegge (Tor Egge)

25 years agoUse lstat() rather than stat in determining whether a file exists,
jkh [Fri, 9 Oct 1998 00:01:16 +0000 (00:01 +0000)]
Use lstat() rather than stat in determining whether a file exists,
otherwise we'll miss the "symlink exists but points nowhere" case.
Submitted by: asami and/or possibly <ru@ucb.crimea.ua>

25 years agoAllow the module area to be used in order to find the MFS image
jkh [Thu, 8 Oct 1998 23:34:44 +0000 (23:34 +0000)]
Allow the module area to be used in order to find the MFS image
(in addition to allowing it to be compiled in) and stop overloading
the MFS_ROOT variable to store size information.

25 years agoThis thing has its own puts function, so use it.
jkh [Thu, 8 Oct 1998 23:14:02 +0000 (23:14 +0000)]
This thing has its own puts function, so use it.
Submitted by: Matthew Jacob <mjacob@nas.nasa.gov>

25 years agoremove stdio.h include; I forgot Bruce's cardinal rule that header files
jkh [Thu, 8 Oct 1998 23:10:41 +0000 (23:10 +0000)]
remove stdio.h include; I forgot Bruce's cardinal rule that header files
shouldn't include other ones (which, unfortunately, is also a hellish
rule since he broke interfaces like sysctl this way by requiring undocumented
header files to be included just in order to be able to use them now - SIGH!).

25 years agoFix up the kernel environment and module data pointers in the bootinfo if
msmith [Thu, 8 Oct 1998 21:03:41 +0000 (21:03 +0000)]
Fix up the kernel environment and module data pointers in the bootinfo if
they are present.
If we are told where the end of the loaded kernel image is, believe it.

25 years agoMake that annoying "device busy" message dependent on DEBUG (using the
des [Thu, 8 Oct 1998 20:55:54 +0000 (20:55 +0000)]
Make that annoying "device busy" message dependent on DEBUG (using the
DEB macro). There are probably quite a few other messages that warrant
a similar treatment, and many more that should be converted to plain
log messages (e.g. "WARNING: wrintr but write DMA inactive!"). Now
that I think of it, same goes for the CAM code (e.g. the famed "tagged
openings" message)

25 years agoFix a memory leak in rpc.rstatd that shows up when it's run in standalone
ken [Thu, 8 Oct 1998 19:59:40 +0000 (19:59 +0000)]
Fix a memory leak in rpc.rstatd that shows up when it's run in standalone
mode.  (i.e., not from inetd)

PR: bin/8212

25 years agoWhile I'm at it, remove the tickadj manpage from src/share/man0/man8,
des [Thu, 8 Oct 1998 18:52:17 +0000 (18:52 +0000)]
While I'm at it, remove the tickadj manpage from src/share/man0/man8,
though I'm afraid there's a lot more that needs fixing in this file,
judging by 'find /usr/src -name "*.8" -print'.

Spotted-by: glimpse -H /usr/src tickadj

25 years agoRemove all references to tickadj(8) from rc, rc.conf and rc.conf.5.
des [Thu, 8 Oct 1998 18:47:30 +0000 (18:47 +0000)]
Remove all references to tickadj(8) from rc, rc.conf and rc.conf.5.

Disable building tickadj(8) by removing util from SUBDIR in the xntpd
Makefile. Note that the sources are still there and tickadj can still
be built and installed by doing:

# cd /usr/src/usr.sbin/xntpd/util
# make all install

There are enough references to tickadj in e.g. the xntpd documentation
(not to mention the sysctl variables it uses etc.) that I don't feel
up to implementing the final solution right now.

Kinda-approved-by: phk

25 years agoDiscard previous commit.
kato [Thu, 8 Oct 1998 17:20:58 +0000 (17:20 +0000)]
Discard previous commit.

25 years agoOops, discard my previous commits becase sumitted code is for RELENG_2_2.
kato [Thu, 8 Oct 1998 17:16:41 +0000 (17:16 +0000)]
Oops, discard my previous commits becase sumitted code is for RELENG_2_2.

25 years agoAdd entry for EtherEZ98 and Access/PC N98C+ cards.
kato [Thu, 8 Oct 1998 17:07:44 +0000 (17:07 +0000)]
Add entry for EtherEZ98 and Access/PC N98C+ cards.

Reviewed by: kato
Submitted by: chi@bd.mbn.or.jp (Chiharu Shibata)

25 years agoAdd ungermann-Bass Access/PC N98C+ support to fe driver (PC-98).
kato [Thu, 8 Oct 1998 17:06:32 +0000 (17:06 +0000)]
Add ungermann-Bass Access/PC N98C+ support to fe driver (PC-98).

Reviewed by: kato
Submitted by: chi@bd.mbn.or.jp (Chiharu Shibata)

25 years agoAdd SMC EtherEZ98 support to ed driver (PC-98).
kato [Thu, 8 Oct 1998 17:04:59 +0000 (17:04 +0000)]
Add SMC EtherEZ98 support to ed driver (PC-98).

Reviewed by: kato
Submitted by: chi@bd.mbn.or.jp (Chiharu Shibata)

25 years agoBIOS ROM base address is 0xe8000 on PC-98.
kato [Thu, 8 Oct 1998 16:15:22 +0000 (16:15 +0000)]
BIOS ROM base address is 0xe8000 on PC-98.

25 years agoCorrectly update the tail pointer of the transmit queue in tl_start()
wpaul [Thu, 8 Oct 1998 15:45:36 +0000 (15:45 +0000)]
Correctly update the tail pointer of the transmit queue in tl_start()
(one-liner). I have yet to actually encounter any problems due to this
bug, but why take chances.

25 years agoSync with sys/i386/conf/GENERIC revision up to 1.123.
kato [Thu, 8 Oct 1998 12:10:07 +0000 (12:10 +0000)]
Sync with sys/i386/conf/GENERIC revision up to 1.123.

25 years agoSync with sys/i386/i386/userconfig.c revision 1.111.
kato [Thu, 8 Oct 1998 12:09:38 +0000 (12:09 +0000)]
Sync with sys/i386/i386/userconfig.c revision 1.111.

25 years agoFinally settle on a name for the mfsroot image.
jkh [Thu, 8 Oct 1998 10:18:55 +0000 (10:18 +0000)]
Finally settle on a name for the mfsroot image.

25 years agoOnly install boot blocks if !MFSKERN. Too big otherwise!
jkh [Thu, 8 Oct 1998 10:02:28 +0000 (10:02 +0000)]
Only install boot blocks if !MFSKERN.  Too big otherwise!

25 years agoQuiet many compiler warnings.
obrien [Thu, 8 Oct 1998 09:56:10 +0000 (09:56 +0000)]
Quiet many compiler warnings.

25 years agoAdd boot bits to boot floppy.
jkh [Thu, 8 Oct 1998 09:53:32 +0000 (09:53 +0000)]
Add boot bits to boot floppy.

25 years agoMake both types of boot targets.
jkh [Thu, 8 Oct 1998 09:40:17 +0000 (09:40 +0000)]
Make both types of boot targets.

25 years agoFinal cleanup pass over boot floppy changes.
jkh [Thu, 8 Oct 1998 09:06:16 +0000 (09:06 +0000)]
Final cleanup pass over boot floppy changes.

25 years agoAvoid using dmesg to find devices, the buffer may not be big enough.
phk [Thu, 8 Oct 1998 08:56:01 +0000 (08:56 +0000)]
Avoid using dmesg to find devices, the buffer may not be big enough.

Reviewed by: phk
Submitted by: Mike Spengler <mks@networkcs.com>

25 years agoRemove a gratuitous copy.
jkh [Thu, 8 Oct 1998 07:12:37 +0000 (07:12 +0000)]
Remove a gratuitous copy.

25 years agoAdd some rudimentary documentation for my new functions.
jkh [Thu, 8 Oct 1998 06:53:32 +0000 (06:53 +0000)]
Add some rudimentary documentation for my new functions.

25 years agoRemove debug output on disk change.
sos [Thu, 8 Oct 1998 06:41:44 +0000 (06:41 +0000)]
Remove debug output on disk change.

25 years agoAdd auth.conf. JKH Added the code to understand this to libutil, and
markm [Thu, 8 Oct 1998 06:12:04 +0000 (06:12 +0000)]
Add auth.conf. JKH Added the code to understand this to libutil, and
I will be following up with commits to use it in KerberosIV userland.

25 years agoAdd the quirk entry framework to handle disabling the synchronize cache
ken [Thu, 8 Oct 1998 05:46:38 +0000 (05:46 +0000)]
Add the quirk entry framework to handle disabling the synchronize cache
command on drives that don't like it.  Right now, there's just a bogus
quirk entry in the table that doesn't do anything, but that should be
changed once we get actual inquiry data for drives that don't like the
synchronize cache command.

Also, add a shutdown hook that runs through all direct access peripherals
and runs a synchronize cache on them if they're still open, and if
synchronize cache isn't disabled via a quirk entry.

Add a synchronize cache call at the end of dadump() (again, conditionalized
on the quirk entry), so we can insure that the disk cache contents get
flushed to physical media after a dump.

Check the new quirk entry in daclose() to decide whether or not to
synchronize the cache for a disk at final close.

Reviewed by: gibbs

25 years agoCorrect a build error that got past my build test somehow.
jkh [Thu, 8 Oct 1998 01:56:49 +0000 (01:56 +0000)]
Correct a build error that got past my build test somehow.

25 years agoCheck the timeval passed to BIOCSRTIMEOUT with itimerfix. Use tvtohz()
alex [Thu, 8 Oct 1998 00:32:08 +0000 (00:32 +0000)]
Check the timeval passed to BIOCSRTIMEOUT with itimerfix.  Use tvtohz()
to convert the timeval into a tick count.

Suggested by: bde
Reviewed by: bde

Handle hz > 1000 in BIOCGRTIMEOUT.

Pointed out by: bde
Reviewed by: bde
Obtained from: OpenBSD

25 years agoUnreverse a conditional and gzip the mfs image by default.
jkh [Wed, 7 Oct 1998 22:58:22 +0000 (22:58 +0000)]
Unreverse a conditional and gzip the mfs image by default.

25 years agoSupport 3 floppy images as standard now: boot kernel, mfs image and
jkh [Wed, 7 Oct 1998 22:54:04 +0000 (22:54 +0000)]
Support 3 floppy images as standard now:  boot kernel, mfs image and
kernel+mfs image.

25 years agoFix typo.
jkh [Wed, 7 Oct 1998 22:36:48 +0000 (22:36 +0000)]
Fix typo.

25 years agoFix typos.
jkh [Wed, 7 Oct 1998 22:15:24 +0000 (22:15 +0000)]
Fix typos.

25 years agoAdd MFSKERN flag to enable the combining of kernel and MFS fs.
jkh [Wed, 7 Oct 1998 21:53:14 +0000 (21:53 +0000)]
Add MFSKERN flag to enable the combining of kernel and MFS fs.
At some point, it will always be off from the user's POV and used
only internally to build (optionall) both mfskern floppies and
ones where mfs root and kernel go in separate places.

25 years agoBy popular request, mention that mailq responds to the same options as
des [Wed, 7 Oct 1998 19:45:36 +0000 (19:45 +0000)]
By popular request, mention that mailq responds to the same options as
sendmail(8).

PR: i386/8149

25 years agoRead in all the nlist entries properly.
jkh [Wed, 7 Oct 1998 19:42:46 +0000 (19:42 +0000)]
Read in all the nlist entries properly.

25 years agoAdd isa_devtab_cam
jkh [Wed, 7 Oct 1998 19:40:51 +0000 (19:40 +0000)]
Add isa_devtab_cam

25 years agoAdd a simple mechanism for reading property lists from files (which
jkh [Wed, 7 Oct 1998 17:32:49 +0000 (17:32 +0000)]
Add a simple mechanism for reading property lists from files (which
I'll convert sysinstall to use shortly) and a simple call which uses
this mechanism to implement an /etc/auth.conf file.  I'll let Mark Murray
handle the format and checkin of the sample auth.conf file.
Reviewed by: markm

25 years agoAdd missing newline in probe message.
des [Wed, 7 Oct 1998 14:42:24 +0000 (14:42 +0000)]
Add missing newline in probe message.

PR: kern/8168
Submitted by: Stefan Eggers <seggers@semyam.dinoco.de>

25 years agoUse vm_page_xxx() inline functions to manipulate vm_page::flags, vm_page::busy.
luoqi [Wed, 7 Oct 1998 13:59:26 +0000 (13:59 +0000)]
Use vm_page_xxx() inline functions to manipulate vm_page::flags, vm_page::busy.
As a side effect, a few wakeup() calls are added, which might fix some of the
missing vm_page wakeups people have been seeing.

Reviewed by: Doug Rabson <dfr@nlsystems.com>

25 years agoFix syntax errors I introduced.
obrien [Wed, 7 Oct 1998 13:51:46 +0000 (13:51 +0000)]
Fix syntax errors I introduced.

25 years agoAdd strcasecmp source file required for dosfs operation.
msmith [Wed, 7 Oct 1998 13:50:17 +0000 (13:50 +0000)]
Add strcasecmp source file required for dosfs operation.

25 years agoUse MACHINE_ARCH instead of MACHINE to detect x86 arch.
kato [Wed, 7 Oct 1998 13:46:09 +0000 (13:46 +0000)]
Use MACHINE_ARCH instead of MACHINE to detect x86 arch.

Pointed out by: Akio Morita <amorita@meadow.scphys.kyoto-u.ac.jp>

25 years agoFix that pesky boot aguments parsing bug. (I think :-)
peter [Wed, 7 Oct 1998 10:55:46 +0000 (10:55 +0000)]
Fix that pesky boot aguments parsing bug.  (I think :-)

25 years agoRequest the flags back when checking for keyboard status.
msmith [Wed, 7 Oct 1998 07:34:31 +0000 (07:34 +0000)]
Request the flags back when checking for keyboard status.
Submitted by: rnordier

25 years agoAdd entries for the adw device driver.
gibbs [Wed, 7 Oct 1998 03:42:44 +0000 (03:42 +0000)]
Add entries for the adw device driver.

25 years agoahc_pci.c:
gibbs [Wed, 7 Oct 1998 03:40:51 +0000 (03:40 +0000)]
ahc_pci.c:
Disable DPARCKEN in the DSCOMMAND0 register on the aic7890/91/96/97.
Parity checking is broken for some chip/MB combinations and this
is the work around recommended by Adaptec.

dpt_pci.c:
Remove a superflous '{' that prevented DPT_ALLOW_MEMIO from working.

pcireg.h:
Add a definition for Parity Error Reponse bit in the PCI Space
command register.

25 years agoFix a parent tag reference count bug during tag teardown.
gibbs [Wed, 7 Oct 1998 03:38:14 +0000 (03:38 +0000)]
Fix a parent tag reference count bug during tag teardown.

Enable optimization for nobounce_dmamap clients by setting the map
held by the client to NULL.  This allows the macros in bus.h to check
against a constant to avoid function calls.

Don't attempt to 'free()' contigmalloced pages in bus_dmamem_free().
We currently leak these pages, which is not ideal, but is better than
a panic.  The leak will be fixed when contigmalloc is merged into the
bus dma framework after 3.0R.

25 years agoAdd a missing newline.
gibbs [Wed, 7 Oct 1998 03:34:13 +0000 (03:34 +0000)]
Add a missing newline.

25 years agoDon't set the active flag on a transaction resource until it has been
gibbs [Wed, 7 Oct 1998 03:32:57 +0000 (03:32 +0000)]
Don't set the active flag on a transaction resource until it has been
queued.

Perform dma segment setup outside of splcam protection as this can take
some time and the protection is not necessary.

Inline a function.

Clean up some whitespace.

25 years agoAdd ADW_ALLOW_MEMIO and DPT_ALLOW_MEMIO. Until our configuration
gibbs [Wed, 7 Oct 1998 03:30:55 +0000 (03:30 +0000)]
Add ADW_ALLOW_MEMIO and DPT_ALLOW_MEMIO.  Until our configuration
framework properly detects broken PCI-PCI bridges, users will have to
manually enable memory mapped I/O for these drivers. <sigh>

25 years agoAdd a quirk entry for the CFP2107, another drive with broken
gibbs [Wed, 7 Oct 1998 03:25:21 +0000 (03:25 +0000)]
Add a quirk entry for the CFP2107, another drive with broken
tagged queuing support.

Ensure that we report that a device supports tagged queuing even if
the system is waiting a "command count delay" before starting to use
them.

If a user disables disconnects on a device ensure that tagged queuing
is also turned off.  We did the right thing during initial configuration,
but could be confused by manual changes.

25 years agoAdd support for the ASC3550 AdvanSys SCSI Host Controller (aka 940UW).
gibbs [Wed, 7 Oct 1998 03:20:52 +0000 (03:20 +0000)]
Add support for the ASC3550 AdvanSys SCSI Host Controller (aka 940UW).

25 years agoDon't use boot2 on alpha.
jkh [Wed, 7 Oct 1998 03:15:08 +0000 (03:15 +0000)]
Don't use boot2 on alpha.

25 years agoRemove hard-coded constants being passed to doFS.sh and add in a 386/alpha
jkh [Wed, 7 Oct 1998 03:13:36 +0000 (03:13 +0000)]
Remove hard-coded constants being passed to doFS.sh and add in a 386/alpha
configuration section.

25 years agoUp the read capacity timeout from 20 seconds to 60 seconds to keep my
imp [Wed, 7 Oct 1998 03:09:19 +0000 (03:09 +0000)]
Up the read capacity timeout from 20 seconds to 60 seconds to keep my
JAZ drive happy.  This shouldn't impact fast drives, and will keep cam
from failing on very slow ones (that are spinning up, say).  20
seconds was almost long enough, but not in all cases.

Suggested by: gibbs

25 years agoSome fixes for the CD and DA drivers from bde. (and some of my own as
ken [Wed, 7 Oct 1998 02:57:57 +0000 (02:57 +0000)]
Some fixes for the CD and DA drivers from bde.  (and some of my own as
well)  Among them:

[ cd driver ]
1. Old labeling code was still there.
2. Error handling for dsopen() was broken (no test for the `error'
   returned by dsopen(); bogus test of an `error' that is known to be 0).
3. cdopen() closed the physical device after certain errors although there
   may still be open partitions on it.
4. cdclose() closed the physical device although there may still be open
   partitions on it.
5. Some printf format fixes was incomplete or missing.
6. cdioctl() truncated unit numbers mod 256.
7. cdioctl() was missing locking.

[ da driver ]
1. daclose() closed the physical device although there may still be open
   partitions on it.  This was fixed many years ago in sd.c rev.1.57.
2. A minor optimization (the dk_slices != NULL test) in sdopen() became
   uglier in daopen().  It is not worth doing.  da only regressed compared
   with od and my version of sd, since I never committed the change to sd.
   daopen() should probably do less if some partition is already open.
   This is not addressed by the diffs.
[ ... ]
5. "opt_hw_wdog.h" was not included, so the HW_WDOG code was unreachable.

- Added a getdev CCB call in the cdopen() and daopen() calls so that the
  vendor name and device name are available for the disklabel.  (suggested
  by bde)

- Removed vestigal devfs support in both drivers, since we can't properly
  work with devfs yet.  (ask bde for details on this)

- Cleaned up the probe code in both drivers in the failure cases.  There
  were a number of things wrong here.  The peripheral driver instances
  weren't getting properly cleaned up.  Sometimes the wrong probe message
  would get printed out (with the failure message appended), so it wasn't
  very clear that we failed to attach.  SCSI sense information was printed,
  even when the error in question wasn't a SCSI error.  I put similar fixes
  into the changer driver in revision 1.2 of scsi_ch.c.

Reviewed by: gibbs
Submitted by: bde (partially)

25 years agoTurn ALLLANG back on.
jkh [Wed, 7 Oct 1998 02:57:16 +0000 (02:57 +0000)]
Turn ALLLANG back on.

25 years agoAdd spanish doc dirs.
jkh [Wed, 7 Oct 1998 02:55:59 +0000 (02:55 +0000)]
Add spanish doc dirs.
Submitted by: Motoyuki Konno <motoyuki@snipe.rim.or.jp>

25 years ago Enable the DOS filesystem. NOTE: you will have to rebuild libstand in
msmith [Wed, 7 Oct 1998 02:39:32 +0000 (02:39 +0000)]
 Enable the DOS filesystem.  NOTE: you will have to rebuild libstand in
 order to be able to build again.

25 years ago - Drain the keyboard buffer when initialising.
msmith [Wed, 7 Oct 1998 02:39:06 +0000 (02:39 +0000)]
 - Drain the keyboard buffer when initialising.
 - Be pedantic about the return from int 16 fn 01.

25 years ago - VERBOSE_LS is obsolete, as the heap is much better behaved now.
msmith [Wed, 7 Oct 1998 02:38:26 +0000 (02:38 +0000)]
 - VERBOSE_LS is obsolete, as the heap is much better behaved now.
 - Don't whine about nodes we can't stat(); these are usually
   symlinks that lead out of the filesystem.
 - Autoboot is now controlled by $autoboot_delay, which is a value
   in seconds or NO to disable autoboot.
 - Don't autoboot at the end of boot.conf if we have already tried.
 - Add a 'read' command to complement 'echo'.  Both are still hidden.
 - Improve the 'source' command/function so that it is possible to
   source scripts off removable media.  The entire script is read and
   saved before beginning execution.  Script lines beginning with
   '@' will not be echoed when being executed.  Script execution will
   normally terminate at the first error, however if the script line
   begins with '-' this behaviour is overriden for that command.

25 years agoVERBOSE_LS is obsolete
msmith [Wed, 7 Oct 1998 02:33:01 +0000 (02:33 +0000)]
VERBOSE_LS is obsolete

25 years agoFix a memory leak
julian [Wed, 7 Oct 1998 01:30:02 +0000 (01:30 +0000)]
Fix a memory leak
PR: 7923
Submitted by: Archie Cobbs <archie@whistle.com>

        The scandir() function returns -1 if it fails.
In many cases when this happens, it does not free
the memory that it allocated, resulting in a memory
leak, or close the directory opened with opendir().
BAD DOG, BAD!

25 years agoCast the return value of tvtohz() from a long to an int to satisfy the
alex [Tue, 6 Oct 1998 23:17:44 +0000 (23:17 +0000)]
Cast the return value of tvtohz() from a long to an int to satisfy the
compiler that we know what we're doing (the value returned has already
been restricted to int ranges).

Reviewed by: bde

25 years agoMinor cleanup: kill a couple of unused variables, and a couple of
joerg [Tue, 6 Oct 1998 21:12:45 +0000 (21:12 +0000)]
Minor cleanup: kill a couple of unused variables, and a couple of
uninitialized variables.

Obtained from: The isdn4bsd project (partially)

25 years ago(1) Add new variable ${PERL} which is equal to ${PERL5} but without the
asami [Tue, 6 Oct 1998 21:12:14 +0000 (21:12 +0000)]
(1) Add new variable ${PERL} which is equal to ${PERL5} but without the
    version number part (i.e., "<directory>/perl").  Use this to
    substitute #! lines in your perl5 scripts.

Requested and reviewed by: ache

(2) Add new variable WRKDIRPREFIX (defaults to "").  The "work"
    directories are now in ${WRKDIRPREFIX}${.CURDIR}/work by default.
    You can have a read-only ports tree (modulo any broken ports that
    write something to places other than ${WRKDIR}) by setting this to
    a writable location.

    Ports that set WRKDIR explicitly should append this to front so
    they will work when the user has WRKDIRPREFIX set.

Reviewed by: Toshihiko Kodama <kodama@ayame.mfd.cs.fujitsu.co.jp>

25 years agoIn an attempt to reduce the huge number of differences between the
joerg [Tue, 6 Oct 1998 20:47:53 +0000 (20:47 +0000)]
In an attempt to reduce the huge number of differences between the
FreeBSD repository version of this file and the isdn4bsd version,
adopt those changes from the i4b version that make this file
BSD-version independent.  I attempted to avoid uglifying this file too
much, thus deviated a little from the i4b version (and hope they will
adopt the changes, too).

The diffs mostly concentrate on:

. #include differences between the systems
. different callout handling between FreeBSD vs. Net/OpenBSD
. interface naming (Net/OpenBSD store the ASCII name including the
  unit # in struct ifnet, FreeBSD only the name)
. use of random() in FreeBSD vs. time-based pseudo-randomization in
  Net/OpenBSD (for loopback detection ad CHAP challenges -- i
  assume at least OpenBSD could also benefit from random(), but that's
  the way i've got this file)
. interface address list elements are named a little differently
  between FreeBSD and Net/OpenBSD

I attempted to segregate those compat fixes from other code fixes and
enhancements.

Obtained from: The isdn4bsd project

25 years agoDisable multi-lun probing and serial number inquiries for the Exabyte 8200.
ken [Tue, 6 Oct 1998 19:27:19 +0000 (19:27 +0000)]
Disable multi-lun probing and serial number inquiries for the Exabyte 8200.

25 years agoHere are some scripts and man pages for configuring HARP ATM
phk [Tue, 6 Oct 1998 19:24:31 +0000 (19:24 +0000)]
Here are some scripts and man pages for configuring HARP ATM
interfaces.

Reviewed by: phk
Submitted by: Mike Spengler <mks@networkcs.com>

25 years agoEnable the DOS filesystem. This allows reading from various DOS filesystems
msmith [Tue, 6 Oct 1998 19:23:57 +0000 (19:23 +0000)]
Enable the DOS filesystem.  This allows reading from various DOS filesystems
(FAT12/16/32, VFAT).

Make a private copy of strcasecmp, as the "real" one uses the system ctype
header, which introduces locale poisoning.

25 years agoAdd functions for accessing dense and bwx memory for pci devices. These
dfr [Tue, 6 Oct 1998 14:18:40 +0000 (14:18 +0000)]
Add functions for accessing dense and bwx memory for pci devices.  These
routines are necessary to allow the use of certain types of hardware on
the alpha, particularly a Myrinet card.

Submitted by: Andrew Gallatin <gallatin@cs.duke.edu>

25 years ago- Implement enabling write allocate on AMD K5/K6/K6-2 cpus.
kato [Tue, 6 Oct 1998 13:16:29 +0000 (13:16 +0000)]
- Implement enabling write allocate on AMD K5/K6/K6-2 cpus.
  The code was originaly contributed by Kelly Yancey
  <kbyanc@freedomnet.com> in PR i386/6269 and revised by Akio Morita
  <amorita@meadow.scphys.kyoto-u.ac.jp> and me.  Test was performed by
  Akio Morita and Toshiomi Moriki <moriki@db.is.kyushu-u.ac.jp>.
- Fix stylistic bug in identcpu.c.
- Update copyright in initcpu.c
- Fix typo in LINT.

PR: 6269 and 6270

25 years agoDoh! Change the correct version of disktab and backout the change to
dfr [Tue, 6 Oct 1998 12:42:55 +0000 (12:42 +0000)]
Doh!  Change the correct version of disktab and backout the change to
the i386 disktab.

25 years agoAdd a 2880k disk prototype for building MFS install images.
dfr [Tue, 6 Oct 1998 12:06:43 +0000 (12:06 +0000)]
Add a 2880k disk prototype for building MFS install images.

25 years agoTeach libdisk about alpha boot blocks.
dfr [Tue, 6 Oct 1998 11:57:08 +0000 (11:57 +0000)]
Teach libdisk about alpha boot blocks.

25 years agoMake sure that the argv pointers for init are aligned to the correct
dfr [Tue, 6 Oct 1998 11:55:40 +0000 (11:55 +0000)]
Make sure that the argv pointers for init are aligned to the correct
boundary on the alpha.

25 years agoReplace use of non-standard ld -O with a ld -o / mv combination as used
peter [Tue, 6 Oct 1998 11:28:29 +0000 (11:28 +0000)]
Replace use of non-standard ld -O with a ld -o / mv combination as used
elsewhere in the tree.  Binutils doesn't support the -O hack^H^H^H^H
extension.  (actually, it ignores it for option compatability with some
other OS).

25 years agoAvoid the ``ruptime: no hosts in /var/rwho.'' message by not calling
joerg [Tue, 6 Oct 1998 09:46:46 +0000 (09:46 +0000)]
Avoid the ``ruptime: no hosts in /var/rwho.'' message by not calling
rwho iff /var/rwho is empty.  Call `uptime' instead.  This doesn't
belong under `network' right away, but at least reports the same
informaton about the local system.  rwhod is not turned on by default
(for good reason), and i've already seen too many of the above
messages...

25 years agoAdd support for adjkerntz (largely untested).
dfr [Tue, 6 Oct 1998 08:40:18 +0000 (08:40 +0000)]
Add support for adjkerntz (largely untested).

25 years agoFix unaligned accesses when checking DOS MBRs.
dfr [Tue, 6 Oct 1998 08:38:58 +0000 (08:38 +0000)]
Fix unaligned accesses when checking DOS MBRs.

25 years agoAdd MFS_ROOT support.
dfr [Tue, 6 Oct 1998 08:38:04 +0000 (08:38 +0000)]
Add MFS_ROOT support.

25 years agoTeach sysinstall about isa_devtab_cam
msmith [Tue, 6 Oct 1998 07:41:49 +0000 (07:41 +0000)]
Teach sysinstall about isa_devtab_cam

25 years agoteach userconfig about isa_devtab_cam
msmith [Tue, 6 Oct 1998 07:34:26 +0000 (07:34 +0000)]
teach userconfig about isa_devtab_cam