]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
25 years agoAdd the EGCS g77 FORTRAN compiler.
David E. O'Brien [Sun, 18 Apr 1999 09:33:25 +0000 (09:33 +0000)]
Add the EGCS g77 FORTRAN compiler.

25 years agoMinimum change to convert from home-grown f77 to EGCS's g77.
David E. O'Brien [Sun, 18 Apr 1999 09:22:56 +0000 (09:22 +0000)]
Minimum change to convert from home-grown f77 to EGCS's g77.
(Pre-bdelinted)

25 years agoInstalls a default loader.rc if none exists. Document loader.conf.
Daniel C. Sobral [Sun, 18 Apr 1999 09:21:12 +0000 (09:21 +0000)]
Installs a default loader.rc if none exists. Document loader.conf.

25 years agoA default loader.rc to be installed.
Daniel C. Sobral [Sun, 18 Apr 1999 09:20:08 +0000 (09:20 +0000)]
A default loader.rc to be installed.

25 years agoMake room for g77.
David E. O'Brien [Sun, 18 Apr 1999 09:17:31 +0000 (09:17 +0000)]
Make room for g77.

25 years agoI missed this PnP change for new-bus..
Peter Wemm [Sun, 18 Apr 1999 06:13:33 +0000 (06:13 +0000)]
I missed this PnP change for new-bus..

Submitted by: Blaz Zupan <blaz@gold.amis.net>

25 years agoActually build the new fxp module.
Bruce Evans [Sun, 18 Apr 1999 03:18:33 +0000 (03:18 +0000)]
Actually build the new fxp module.

Updated comments about rotten modules.

Fixed disorder and English errors.

25 years agoSet ifq_maxlen to IFQ_MAXLEN
Paul Richards [Sun, 18 Apr 1999 01:49:16 +0000 (01:49 +0000)]
Set ifq_maxlen to IFQ_MAXLEN

25 years agoSet ifq_maxlen to number of transmit descriptors.
Paul Richards [Sun, 18 Apr 1999 01:37:19 +0000 (01:37 +0000)]
Set ifq_maxlen to number of transmit descriptors.

25 years agoMake WangTek 51000 a fixed block device.
Matt Jacob [Sun, 18 Apr 1999 01:05:03 +0000 (01:05 +0000)]
Make WangTek 51000 a fixed block device.
PR: kern/11194
Obtained from:naddy@mips.rhein-neckar.de

25 years agoThe changes to vm_page.c have broken the assumption that mallocing
Paul Richards [Sun, 18 Apr 1999 01:03:46 +0000 (01:03 +0000)]
The changes to vm_page.c have broken the assumption that mallocing
early will result in memory below 16M so now we need to explicitly use
contigmalloc to get low physical memory.

25 years agoUHCI_DEBUG should of course be OHCI_DEBUG
Nick Hibma [Sat, 17 Apr 1999 21:48:03 +0000 (21:48 +0000)]
UHCI_DEBUG should of course be OHCI_DEBUG

Reported by: MAEKAWA Masahide <bishop@rr.iij4u.or.jp>

25 years agoProbe the SiS 85c503 as a PCI-ISA bridge.
Doug Rabson [Sat, 17 Apr 1999 19:48:45 +0000 (19:48 +0000)]
Probe the SiS 85c503 as a PCI-ISA bridge.

25 years agoWrap the pcm driver for old isa probes.
Peter Wemm [Sat, 17 Apr 1999 19:38:38 +0000 (19:38 +0000)]
Wrap the pcm driver for old isa probes.
Correct #ifdef typo for sbxvi (Thanks Brian!)

25 years agoDisconnect lkm subdir so we don't build a lkm module, in case something
Peter Wemm [Sat, 17 Apr 1999 16:31:13 +0000 (16:31 +0000)]
Disconnect lkm subdir so we don't build a lkm module, in case something
I do breaks world or the nightly snapshots.  (you can still do it by hand
though for the moment)

25 years agoGet out the blow torch and hack away all the unused stuff. Note that
Peter Wemm [Sat, 17 Apr 1999 14:41:41 +0000 (14:41 +0000)]
Get out the blow torch and hack away all the unused stuff.  Note that
I zapped the MACHINE_MIPS stuff, it isn't likely to be useful apart from
recognition of the machine name.  It would be reasonable to expect new
ports would look something like the alpha/i386 from a config perspective.

25 years agoConvert the dummynet lkm code to be kld aware (this isn't actually used
Peter Wemm [Sat, 17 Apr 1999 11:09:08 +0000 (11:09 +0000)]
Convert the dummynet lkm code to be kld aware (this isn't actually used
anywhere that I can see).

25 years agoAllocate space for struct isa_device's, not for pointers thereto.
Bruce Evans [Sat, 17 Apr 1999 09:56:35 +0000 (09:56 +0000)]
Allocate space for struct isa_device's, not for pointers thereto.
This fixes memory corruption that caused calls to address 0 here.

25 years agoGC some unused code.
Peter Wemm [Sat, 17 Apr 1999 09:12:35 +0000 (09:12 +0000)]
GC some unused code.

25 years agoOops, forgot this part of lkm code that's been replaced with kld.
Peter Wemm [Sat, 17 Apr 1999 08:56:38 +0000 (08:56 +0000)]
Oops, forgot this part of lkm code that's been replaced with kld.

25 years agoWell folks, this is it - The second stage of the removal for build support
Peter Wemm [Sat, 17 Apr 1999 08:36:07 +0000 (08:36 +0000)]
Well folks, this is it - The second stage of the removal for build support
for LKM's..

25 years agoImage activators use EXEC_SET(), not TEXT_SET(). (The first is a macro
Peter Wemm [Sat, 17 Apr 1999 07:55:15 +0000 (07:55 +0000)]
Image activators use EXEC_SET(), not TEXT_SET().  (The first is a macro
wrapper for DECLARE_MODULE(), the second is a linker set declaration)

25 years agoReverse last checkin, which was a regression. Grrr.
Greg Lehey [Sat, 17 Apr 1999 07:48:34 +0000 (07:48 +0000)]
Reverse last checkin, which was a regression.  Grrr.

25 years agofree_vinum: clear up a warning message.
Greg Lehey [Sat, 17 Apr 1999 04:15:50 +0000 (04:15 +0000)]
free_vinum: clear up a warning message.

25 years agoReplace with the correct version which sources in the other .gdbinits.
Greg Lehey [Sat, 17 Apr 1999 04:14:40 +0000 (04:14 +0000)]
Replace with the correct version which sources in the other .gdbinits.

25 years agoWhen looking at all objects, look at all objects.
Greg Lehey [Sat, 17 Apr 1999 04:13:04 +0000 (04:13 +0000)]
When looking at all objects, look at all objects.

main:
  If the wrong version is started, execve the correct version.

25 years ago When looking at all objects, look at all objects.
Greg Lehey [Sat, 17 Apr 1999 04:12:33 +0000 (04:12 +0000)]
  When looking at all objects, look at all objects.

vinum_lsi:
  Print rough size of plex in parentheses.
  Print drive info with -v option (previously required -V).

25 years agovinum_init: count dead children correctly when waiting for completion.
Greg Lehey [Sat, 17 Apr 1999 04:11:48 +0000 (04:11 +0000)]
vinum_init: count dead children correctly when waiting for completion.

vinum_resetstats: count objects correctly.

25 years agoX11 include directory may not actually contain any header file, check for
Luoqi Chen [Sat, 17 Apr 1999 01:57:06 +0000 (01:57 +0000)]
X11 include directory may not actually contain any header file, check for
a specific one (X11/X.h).

25 years agoMerge up to rev 1.234 (nice revision number!) of i386/isa/sio.c.
Peter Wemm [Sat, 17 Apr 1999 01:02:27 +0000 (01:02 +0000)]
Merge up to rev 1.234 (nice revision number!) of i386/isa/sio.c.

25 years agoMerge up to rev 1.226 of i386/isa/sio.c
Peter Wemm [Sat, 17 Apr 1999 00:37:01 +0000 (00:37 +0000)]
Merge up to rev 1.226 of i386/isa/sio.c

25 years agooops, there was no tek390.c from the last commit, that's something on
Peter Wemm [Sat, 17 Apr 1999 00:06:24 +0000 (00:06 +0000)]
oops, there was no tek390.c from the last commit, that's something on
one of Doug's machines apparently.

25 years agoWe use the generic isa version now.
Peter Wemm [Sat, 17 Apr 1999 00:02:56 +0000 (00:02 +0000)]
We use the generic isa version now.

25 years agoWe now use the generic isa version..
Peter Wemm [Sat, 17 Apr 1999 00:01:59 +0000 (00:01 +0000)]
We now use the generic isa version..

25 years agovga_isa.c is now generic, not i386 specific (src/sys/isa/vga_isa.c)
Peter Wemm [Sat, 17 Apr 1999 00:00:31 +0000 (00:00 +0000)]
vga_isa.c is now generic, not i386 specific (src/sys/isa/vga_isa.c)

25 years agoCopy across some apparently missing cleanups from i386/isa/vga_isa.c
Peter Wemm [Fri, 16 Apr 1999 23:54:24 +0000 (23:54 +0000)]
Copy across some apparently missing cleanups from i386/isa/vga_isa.c

25 years agoAs a temporary anti-foot-shooting measure, don't let the user attach
Peter Wemm [Fri, 16 Apr 1999 23:39:15 +0000 (23:39 +0000)]
As a temporary anti-foot-shooting measure, don't let the user attach
the atkbd device to isa, as was in the old (and 3.x) GENERIC config.

25 years agoAdd a sample loadable pci driver Makefile..
Peter Wemm [Fri, 16 Apr 1999 21:52:27 +0000 (21:52 +0000)]
Add a sample loadable pci driver Makefile..

25 years agoCorresponding minimal changes for kernel configuration after new-bus
Peter Wemm [Fri, 16 Apr 1999 21:28:10 +0000 (21:28 +0000)]
Corresponding minimal changes for kernel configuration after new-bus
commit.

25 years agoBring the 'new-bus' to the i386. This extensively changes the way the
Peter Wemm [Fri, 16 Apr 1999 21:22:55 +0000 (21:22 +0000)]
Bring the 'new-bus' to the i386.  This extensively changes the way the
i386 platform boots, it is no longer ISA-centric, and is fully dynamic.
Most old drivers compile and run without modification via 'compatability
shims' to enable a smoother transition.  eisa, isapnp and pccard* are
not yet using the new resource manager.  Once fully converted, all drivers
will be loadable, including PCI and ISA.

(Some other changes appear to have snuck in, including a port of Soren's
 ATA driver to the Alpha.  Soren, back this out if you need to.)

This is a checkpoint of work-in-progress, but is quite functional.

The bulk of the work was done over the last few years by Doug Rabson and
Garrett Wollman.

Approved by: core

25 years agoAdd SYSVSEM so that newer versions of Xaccel don't require a kernel
Jordan K. Hubbard [Fri, 16 Apr 1999 18:27:18 +0000 (18:27 +0000)]
Add SYSVSEM so that newer versions of Xaccel don't require a kernel
compile just to work.  We have the room now, so what the heck.

Reqested by:    Thomas Roell <roell@xig.com>

25 years agoRemove the entries for umodem and ucom. These drivers only probe
Nick Hibma [Fri, 16 Apr 1999 16:17:05 +0000 (16:17 +0000)]
Remove the entries for umodem and ucom. These drivers only probe
and attach, nothing else. This is confusing to people.

25 years agozero out switchtime.tv_sec. This fixes a fairly serious problem where
Andrew Gallatin [Fri, 16 Apr 1999 13:57:38 +0000 (13:57 +0000)]
zero out switchtime.tv_sec.  This fixes a fairly serious problem where
processes were being charged for time when the system was idle.

25 years ago- Add another adhoc frame_size detection for stack trace.
Hidetoshi Shimokawa [Fri, 16 Apr 1999 12:04:39 +0000 (12:04 +0000)]
- Add another adhoc frame_size detection for stack trace.
egcs manupilates stack pointer with 'subq/addq' rather than 'lha'.
- Surpress warnings

25 years agoRevise list of interfaces on which bridging is supported. mdoc-ify a few
Guy Helmer [Fri, 16 Apr 1999 02:09:38 +0000 (02:09 +0000)]
Revise list of interfaces on which bridging is supported.  mdoc-ify a few
things while I'm here.

25 years agoAdd bridging support (tested in 3.1-RELEASE by Steven Vetzal
Guy Helmer [Fri, 16 Apr 1999 01:56:06 +0000 (01:56 +0000)]
Add bridging support (tested in 3.1-RELEASE by Steven Vetzal
<svetzal@icom.ca>).

25 years agoBetter handling for ARP/source routing on Token Ring
Eivind Eklund [Thu, 15 Apr 1999 17:58:24 +0000 (17:58 +0000)]
Better handling for ARP/source routing on Token Ring

Submitted by: Larry Lile <lile@stdio.com>

25 years agoRemoved dead code and cleaned up. setconf() now just asks for the
Bruce Evans [Thu, 15 Apr 1999 15:35:32 +0000 (15:35 +0000)]
Removed dead code and cleaned up.  setconf() now just asks for the
root device name.  The parser for the name is still too simple (it
forces slice = none, partition = 'a').

25 years agoMade booting with -a work for all configurations. Previously it
Bruce Evans [Thu, 15 Apr 1999 14:52:24 +0000 (14:52 +0000)]
Made booting with -a work for all configurations.  Previously it
only worked for configurations with "swap on generic".

usr.sbin/config/config.y:
- ignore all "swap [on] device ...' specifications except for
  warning about them.  They haven't done anything related to swap
  for almost 4 years, and were previously silently ignored,
  except for "swap on generic" which stopped swap${KERNEL}.c
  from being generated.  Code to support swapping is now deader
  than before.

usr.sbin/config/mkswapconf.c:
- don't generate a dummy setconf() function in swap${KERNEL}.c.

sys/i386/conf/files.i386:
- swapgeneric.c is now standard.  It should be merged into autoconf.c
  so that it doesn't conflict with swap${KERNEL}.c for kernels named
  "generic".

sys/i386/i386/autoconf.c:
- don't call setroot() for mfs roots.  Since setroot() doesn't do anything
  harmful, this was just a waste of time, except possibly for booting with
  -a it may have helped prevent an undesireable call to setconf() by
  finding a bogus rootdev.
- honor -a for ffs roots.  -a now overrides all other ways of specifying
  the root device.  Previously, -r had precedence over -a, and the -a
  handling was usually a no-op.
- don't honor -a for non-ffs roots, since it would currently just get in
  the way of a clean panic.

sys/i386/i386/swapgeneric.c:
- don't declare things that are now always declared in swap${KERNEL}.c.
  Don't decide things that are now decided in autoconf.c.  Code to
  support the "generic" case is now dead instead of useless.

25 years agoNuke lsock(1).
Dag-Erling Smørgrav [Thu, 15 Apr 1999 13:43:22 +0000 (13:43 +0000)]
Nuke lsock(1).

25 years agolsock(1) -> sockstat(1)
Dag-Erling Smørgrav [Thu, 15 Apr 1999 13:42:48 +0000 (13:42 +0000)]
lsock(1) -> sockstat(1)

25 years agoRename lsock(1) to sockstat(1).
Dag-Erling Smørgrav [Thu, 15 Apr 1999 13:40:43 +0000 (13:40 +0000)]
Rename lsock(1) to sockstat(1).

Suggested by: Sheldon Hearn <sheldonh@iafrica.com>
Approved by: Peter Wemm <peter@netplex.com.au>

25 years agoFoo. jdp changed .so to .So during the time I did the EGCS work and I
David E. O'Brien [Thu, 15 Apr 1999 08:08:56 +0000 (08:08 +0000)]
Foo.  jdp changed .so to .So during the time I did the EGCS work and I
missed bring in his changes into my work.

25 years agoCreate /var/db/port.mkversion rather than /var/db/pkg/.mkversion to be in
David E. O'Brien [Thu, 15 Apr 1999 07:19:56 +0000 (07:19 +0000)]
Create /var/db/port.mkversion rather than /var/db/pkg/.mkversion to be in
sync with bsd.port.mk rev 1.309.

Submitted by: make world

25 years agoClarify the current state of cc -aout:
Warner Losh [Thu, 15 Apr 1999 04:36:45 +0000 (04:36 +0000)]
Clarify the current state of cc -aout:
Compiling -g sometimes doesn't work (from Bruce)
libgcc_r won't build (Jos Backus)
Also mention that make -j n world is working again

25 years ago- Close PR #11136: add PCI ID for another new cyclone device: the
Bill Paul [Thu, 15 Apr 1999 03:18:33 +0000 (03:18 +0000)]
- Close PR #11136: add PCI ID for another new cyclone device: the
  3c900B-TPC (twisted pair and coax). Treated similarly to the
  3c900B-COMBO, except no AUI port.

- Fix media selection so that it's possible to select the AUI and BNC
  ports on the 3c905B-COMBO. This board is now fully supported.

- Change TX queueing strategy to hopefully be more efficient by avoiding
  register accesses in xl_start(). Should provide small performance
  improvement and a little better reliability.

25 years agoAdded missing newline to "no ports found!" message.
Alexander Langer [Thu, 15 Apr 1999 00:13:20 +0000 (00:13 +0000)]
Added missing newline to "no ports found!" message.

25 years agoRemove references to libg++ and libf2c that aren't needed post-EGCS.
David E. O'Brien [Wed, 14 Apr 1999 23:59:09 +0000 (23:59 +0000)]
Remove references to libg++ and libf2c that aren't needed post-EGCS.

25 years agoGrrr. Make all modes work properly with the 82c168 chip and built-in
Bill Paul [Wed, 14 Apr 1999 19:40:08 +0000 (19:40 +0000)]
Grrr. Make all modes work properly with the 82c168 chip and built-in
transceiver. Note in the manual page that autoselection doesn't
work on the 82c168 because the built-in NWAY support is horribly
broken. Manual mode selection works fine, but autoneg is broken for
everything except maybe 10Mbps half-duplex. There's no simple way
to fix this at the moment, so I have to settle for documenting the
bug for now. Fortunately, there aren't anywhere near as many 82c168
boards around as there are 82c169s.

25 years agogetnewbuf(): check return value from tsleep(). Interruptible NFS may pass
Dmitrij Tejblum [Wed, 14 Apr 1999 18:51:52 +0000 (18:51 +0000)]
getnewbuf(): check return value from tsleep(). Interruptible NFS may pass
PCATCH to slpflag.

25 years agoFix typos.
Nik Clayton [Wed, 14 Apr 1999 18:49:15 +0000 (18:49 +0000)]
Fix typos.

PR: docs/8682
Submitted by: hnokubi@yyy.or.jp
Reviewed by: nik

25 years agoChange description of file(1) following symlinks to the truth (namely,
Nik Clayton [Wed, 14 Apr 1999 18:45:07 +0000 (18:45 +0000)]
Change description of file(1) following symlinks to the truth (namely,
by default, file(1) does not follow symlinks, the -L flag must be
specified.

PR: docs/8602
Submitted by: Kazuo Horikawa <k-horik@yk.rim.or.jp>
Reviewed by: nik

25 years agotemp fix for internal queue overflow problem
Matt Jacob [Wed, 14 Apr 1999 17:37:36 +0000 (17:37 +0000)]
temp fix for internal queue overflow problem

25 years agoAdd example for 'makeoptions DEBUG' and some notes. I have not activated
Peter Wemm [Wed, 14 Apr 1999 16:54:00 +0000 (16:54 +0000)]
Add example for 'makeoptions DEBUG' and some notes.  I have not activated
it here since a -g LINT kernel is 100% useless as it won't run and hence
doesn't need debug capabilities (and would just waste disk space :-).

25 years agoActivate lsock(1).
Dag-Erling Smørgrav [Wed, 14 Apr 1999 16:17:52 +0000 (16:17 +0000)]
Activate lsock(1).

25 years agoImport lsock(1), a Perl script for listing open Internet sockets.
Dag-Erling Smørgrav [Wed, 14 Apr 1999 16:17:34 +0000 (16:17 +0000)]
Import lsock(1), a Perl script for listing open Internet sockets.

25 years agoInstall kld.4.
Bruce Evans [Wed, 14 Apr 1999 16:04:55 +0000 (16:04 +0000)]
Install kld.4.

Sorted lists.

25 years agoSearch bdevsw[] instead of the half-baked builtin table of devices
Bruce Evans [Wed, 14 Apr 1999 15:20:03 +0000 (15:20 +0000)]
Search bdevsw[] instead of the half-baked builtin table of devices
in the RB_ASKNAME case.  I had thought that I made this change in
rev.1.18, but rev.1.18 only affects obscure subcases of the
RB_DFLTROOT case (subcases where the compiled in default root is
not found in bdevsw[] -- then the root device is set to the first
device in the half-baked table that is also in bdevsw[]).

Removed yet more vestiges of config-time swap configuration.

25 years agoGenerate intrnames[] dynamically. This should be new-bus friendly.
Bruce Evans [Wed, 14 Apr 1999 14:26:36 +0000 (14:26 +0000)]
Generate intrnames[] dynamically.  This should be new-bus friendly.

Old version reviewed by: se

25 years agoDocument the "-D" option (UTSL).
David E. O'Brien [Wed, 14 Apr 1999 05:37:54 +0000 (05:37 +0000)]
Document the "-D" option (UTSL).

25 years agoClose PR #11122: check key length before calling strncmp()
Bill Paul [Wed, 14 Apr 1999 04:05:59 +0000 (04:05 +0000)]
Close PR #11122: check key length before calling strncmp()
in yp_next_record().

25 years agoSupport linking against libgcc_r.
David E. O'Brien [Wed, 14 Apr 1999 00:25:45 +0000 (00:25 +0000)]
Support linking against libgcc_r.

Submitted by: Luoqi Chen <luoqi@chen.ml.org>

25 years agoCreate a libgcc with POSIX threads support.
David E. O'Brien [Wed, 14 Apr 1999 00:24:03 +0000 (00:24 +0000)]
Create a libgcc with POSIX threads support.

Submitted by: Luoqi Chen <luoqi@chen.ml.org> (reorganized by me)

25 years agoAdd a commented-out example on using the makeoptions command to get a
Peter Wemm [Tue, 13 Apr 1999 20:59:36 +0000 (20:59 +0000)]
Add a commented-out example on using the makeoptions command to get a
kernel.debug.

25 years agoBuild the functionality of the wdc_p hack into the ide_pci.c code.
Peter Wemm [Tue, 13 Apr 1999 20:22:34 +0000 (20:22 +0000)]
Build the functionality of the wdc_p hack into the ide_pci.c code.
All it did was match a specific device ID and turn on a quirk for
the wdc driver.

Incidently, at line 1462 there is a return that prevents the generic
ide_pci code from trying to look at the device.  I'd be interested
to know if we can take out the return and let the generic code "see" it.
I've left the return in because that's the way it worked before.

(Be sure to rerun config after cvsup or you'll get undefined files!)

25 years agoShoot the LKM support in the old wd/wdc/atapi driver set in the head and
Peter Wemm [Tue, 13 Apr 1999 19:38:12 +0000 (19:38 +0000)]
Shoot the LKM support in the old wd/wdc/atapi driver set in the head and
perform a cleanup/unifdef sweep over it to tidy things up.  The atapi
code is permanently attached to the wd driver and is always probed.

I will add an extra option bit in the flags to disable an atapi probe on
either the master or slave if needed, if people want this.

Remember, this driver is destined to die some time.  It's possible that
it will loose all atapi support down the track and only be used for
dumb non-ATA disks and all ata/atapi devices will be handled by the new
ata system.

ATAPI, ATAPI_STATIC and CMD640 are no longer options, all are implicit.

Previously discussed with:  sos

25 years agoMove the declaration of faultin() from the vm headers to proc.h, since
Peter Wemm [Tue, 13 Apr 1999 19:17:15 +0000 (19:17 +0000)]
Move the declaration of faultin() from the vm headers to proc.h, since
it is now referenced from a macro there (PHOLD()).

25 years agoDrop the atapi module from the build, it's just wasting time.
Peter Wemm [Tue, 13 Apr 1999 18:53:54 +0000 (18:53 +0000)]
Drop the atapi module from the build, it's just wasting time.

25 years agoHopefully replicate the install.debug changes from the i386 version.
Peter Wemm [Tue, 13 Apr 1999 18:35:28 +0000 (18:35 +0000)]
Hopefully replicate the install.debug changes from the i386 version.

25 years agoClean up the kernel/kernel.debug/install.debug etc stuff.
Peter Wemm [Tue, 13 Apr 1999 18:25:08 +0000 (18:25 +0000)]
Clean up the kernel/kernel.debug/install.debug etc stuff.
Sense ${DEBUG} to decide on building kernel.debug or not.
Use a common install and install.debug target to minimize duplication.
(I deleted the ELF transition kernel warning, it was getting dated..)

25 years agoClean up the -g/DEBUG handling. This logic can go in the Makefile
Peter Wemm [Tue, 13 Apr 1999 18:22:57 +0000 (18:22 +0000)]
Clean up the -g/DEBUG handling.  This logic can go in the Makefile
so that config -g can work the same as:  makeoptions DEBUG="-g"

25 years agoAs the INFO stuff un doc/ works, we may as well build/install it.
Mark Murray [Tue, 13 Apr 1999 17:56:28 +0000 (17:56 +0000)]
As the INFO stuff un doc/ works, we may as well build/install it.

OK'ed by: obrien

25 years agoSome more fixes:
Bill Paul [Tue, 13 Apr 1999 17:29:55 +0000 (17:29 +0000)]
Some more fixes:

- It turns out that the 'promiscuous mode' bug what I discovered with the
  PNIC is not restricted to promiscuous mode. I've been doing some remote
  debugging for someone with a P75 system, and at 100Mbps, the receiver
  screws up even when the NIC is in normal mode. Thus, enable the workaround
  for this bug all the time. Note that the workaround is still not enabled
  for the PNIC II, since I haven't tested one yet.

- Set the 'arbitration' bit in the bus configuration register and set the
  maximum burst size to 16 longwords. This seems to fix problems with
  transmit corruption on the P75 system mentioned above. (It probably hurts
  performance a bit too, but I've given up trying to make the PNIC perform
  well.)

25 years agoUse u_int32_t for sin_addr.s_addr rather than u_long to avoid
Hidetoshi Shimokawa [Tue, 13 Apr 1999 16:26:21 +0000 (16:26 +0000)]
Use u_int32_t for sin_addr.s_addr rather than u_long to avoid
unaligned access on alpha.

25 years agoInitialize alpha_unaligned_* as I intended.
Hidetoshi Shimokawa [Tue, 13 Apr 1999 15:42:34 +0000 (15:42 +0000)]
Initialize alpha_unaligned_* as I intended.
I misunderstood SYSCTL_INT in previous revision.

25 years agoRemoved DIAGNOSTIC opion redefinition.
Semen Ustimenko [Tue, 13 Apr 1999 14:54:20 +0000 (14:54 +0000)]
Removed DIAGNOSTIC opion redefinition.

Submitted by: Eivind Eklund <eivind@FreeBSD.org>

25 years agoReference kldfind(2), kldfirstmod(2), kldload(2), kldnext(2), kldstat(2),
Guy Helmer [Tue, 13 Apr 1999 14:51:51 +0000 (14:51 +0000)]
Reference kldfind(2), kldfirstmod(2), kldload(2), kldnext(2), kldstat(2),
and kldunload(2).

Submitted by: Chris Costello <chris@holly.dyndns.org>

25 years agoAdd kld(4) man page. Copyright notice in PR submission was "???", which
Guy Helmer [Tue, 13 Apr 1999 13:23:11 +0000 (13:23 +0000)]
Add kld(4) man page.  Copyright notice in PR submission was "???", which
I changed to "Christopher G. Demetriou" since the page appears to be a
revision of lkm(4).

PR: docs/8611
Submitted by: Rajesh Vaidheeswarran <rv@fore.com>

25 years agoRemove my temporary detection for PR 7059, Tor Egge fixed this bug.
Martin Cracauer [Tue, 13 Apr 1999 12:43:55 +0000 (12:43 +0000)]
Remove my temporary detection for PR 7059, Tor Egge fixed this bug.

PR: 7059

25 years agoMake sending the postcard optional (as suggested by phk).
Hellmuth Michaelis [Tue, 13 Apr 1999 12:24:33 +0000 (12:24 +0000)]
Make sending the postcard optional (as suggested by phk).

25 years agoFixed bogus SRCS definition which broke `make depend'. Plain `make'
Bruce Evans [Tue, 13 Apr 1999 09:11:20 +0000 (09:11 +0000)]
Fixed bogus SRCS definition which broke `make depend'.  Plain `make'
worked accidentally because the rule to generate ${OBJS} from ${SRCS}
is sloppy.

Fixed style bugs.

25 years agoNote that cc -aout has been broken for a while and a reminder to
Warner Losh [Tue, 13 Apr 1999 06:10:53 +0000 (06:10 +0000)]
Note that cc -aout has been broken for a while and a reminder to
recompile now that it has been fixed.

Submitted by: Peter Wemm

25 years agoFix a misspelling.
Jordan K. Hubbard [Tue, 13 Apr 1999 05:14:44 +0000 (05:14 +0000)]
Fix a misspelling.

25 years agoDuring variable expansion, the internal representation of the expression
Tor Egge [Tue, 13 Apr 1999 04:13:09 +0000 (04:13 +0000)]
During variable expansion, the internal representation of the expression
might be relocated.  Handle this case.
PR: 7059

25 years agoBackout early start of APs since it caused some machines to hang.
Tor Egge [Tue, 13 Apr 1999 03:24:47 +0000 (03:24 +0000)]
Backout early start of APs since it caused some machines to hang.

25 years agoCorrectly ref editline(3).
Guy Helmer [Tue, 13 Apr 1999 02:42:53 +0000 (02:42 +0000)]
Correctly ref editline(3).

PR: docs/10973

25 years agoTypo fix.
Guy Helmer [Tue, 13 Apr 1999 02:32:12 +0000 (02:32 +0000)]
Typo fix.

PR: docs/11022

25 years agoChange LKM/modload to KLD/kldload.
Guy Helmer [Tue, 13 Apr 1999 02:20:07 +0000 (02:20 +0000)]
Change LKM/modload to KLD/kldload.

Submitted by: Nathan Ahlstrom <nrahlstr@winternet.com>

25 years agoUpdate description of how to analyse kernel dumps.
Greg Lehey [Tue, 13 Apr 1999 01:18:51 +0000 (01:18 +0000)]
Update description of how to analyse kernel dumps.

25 years agoMerge in some updates for the RealTek driver, mainly:
Bill Paul [Mon, 12 Apr 1999 21:37:00 +0000 (21:37 +0000)]
Merge in some updates for the RealTek driver, mainly:

- Rewrite the transmit section to be a little less bogus.
- Set ifq_maxlen correctly. RL_TX_LIST_CNT - 1 is wrong, because for the
  RealTek, RL_TX_LIST_CNT is 4. Set it to IFQ_MAXLEN instead.