]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
29 years agoSigh. My error message printing was still wrong. Wipe the sleep from
jkh [Sun, 6 Nov 1994 02:24:45 +0000 (02:24 +0000)]
Sigh.  My error message printing was still wrong.  Wipe the sleep from
my eyes and do this properly (debugging is especially difficult when you have
to burn a floppy every time you want to test your changes! :-( ).

29 years agoDon't use a shell, silly, we don't have one!
jkh [Sun, 6 Nov 1994 01:34:14 +0000 (01:34 +0000)]
Don't use a shell, silly, we don't have one!

29 years agoNuke the losing version of microtime. The assembler version now works
bde [Sun, 6 Nov 1994 01:33:03 +0000 (01:33 +0000)]
Nuke the losing version of microtime.  The assembler version now works
for all reasonable HZ's.  HZ > 1000 doesn't work because of sloppy
conversions in hzto() (division by (tick / 1000) == 0).  This was
fixed in 1.1.5.

Eliminate some extern declarations by including the appropriate header
files that now contain appropriate declarations.

29 years agoGet this braindead, mongoloid shell look in /stand for pwd if it can't
jkh [Sun, 6 Nov 1994 01:29:26 +0000 (01:29 +0000)]
Get this braindead, mongoloid shell look in /stand for pwd if it can't
find it in /bin.  This is something of a kludge, I know, but consider
my limited alternatives:  I can't make this an execvp() without making
people scream that I introduced a failure point or slowed down pwd,
and I can't make it an optional macro since crunch doesn't let you pass
arbitrary command-line args to the build of one of its crunch-ees.
This is the simplest, if not the nicest looking, solution I could come up
with.

29 years agoDon't strip kernel after all - perhaps a bad idea. People may be stuck
jkh [Sun, 6 Nov 1994 01:19:48 +0000 (01:19 +0000)]
Don't strip kernel after all - perhaps a bad idea.  People may be stuck
with it for awhile.

29 years agoTweak tweak..
jkh [Sun, 6 Nov 1994 01:16:30 +0000 (01:16 +0000)]
Tweak tweak..

29 years agoMake /usr/bin/crontab install setuid root. (doesn't work otherwise)
smace [Sun, 6 Nov 1994 01:13:59 +0000 (01:13 +0000)]
Make /usr/bin/crontab install setuid root. (doesn't work otherwise)
The distributed makefile in the package installs it setuid root..
Reviewed by: jkh

29 years agoPublic function declarations moved to <machine/npx.h>.
bde [Sun, 6 Nov 1994 00:58:06 +0000 (00:58 +0000)]
Public function declarations moved to <machine/npx.h>.

29 years agoAbort writes if a signal is received (don't ignore the value returned by
bde [Sun, 6 Nov 1994 00:46:21 +0000 (00:46 +0000)]
Abort writes if a signal is received (don't ignore the value returned by
tsleep()).  Try `dd if=/dev/zero of=/dev/pcaudio bs=640k count=1'.  The
write takes a few hundred seconds to drain, and if it is killed by a
signal, it still takes a few hundred seconds to drain and all of those
seconds are spent busy-waiting.

Clean up includes and declarations.  Remove bogus casts of args to
timeout functions.

29 years agoThe style in my patch011 is inconsistent and out of date. It should
bde [Sun, 6 Nov 1994 00:30:16 +0000 (00:30 +0000)]
The style in my patch011 is inconsistent and out of date.  It should
have been changed _before_ committing the patch.

29 years agoWork around microtime() enabling interrupts.
bde [Sun, 6 Nov 1994 00:23:45 +0000 (00:23 +0000)]
Work around microtime() enabling interrupts.

29 years agoTwiddle newfs parameters to select for space optimization. We're
jkh [Sun, 6 Nov 1994 00:21:39 +0000 (00:21 +0000)]
Twiddle newfs parameters to select for space optimization.  We're
back under 1.2MB again.

29 years agoStrip kernel before installation. Anything to bum a few more bytes..
jkh [Sun, 6 Nov 1994 00:10:00 +0000 (00:10 +0000)]
Strip kernel before installation.  Anything to bum a few more bytes..

29 years agoAdd -lreadline for ncftp.
jkh [Sun, 6 Nov 1994 00:00:17 +0000 (00:00 +0000)]
Add -lreadline for ncftp.

29 years agoI misunderstood dialog_prgbox() - fix it.
jkh [Sat, 5 Nov 1994 23:56:40 +0000 (23:56 +0000)]
I misunderstood dialog_prgbox() - fix it.

29 years agoMaintain a new variable `timer0_overflow_threshold' so that microtime()
bde [Sat, 5 Nov 1994 23:55:07 +0000 (23:55 +0000)]
Maintain a new variable `timer0_overflow_threshold' so that microtime()
doesn't have to calculate it every call.

Rename `timer0_prescale' to `timer0_prescaler_count' and maintain it
correctly.  Previously we lost a few 8253 cycles for every "prescaled"
clock interrupt, and the lossage grows rapidly at 16 KHz.  Now we
only lose a few cycles for every standard clock interrupt.

Rename `*_divisor' to `*_max_count'.

Do the calculation of TIMER_DIV(rate) only once instead of 3 times each
time the rate is changed.

Don't allow preposterously large interrupt rates.  Bug fixes elsewhere
should allow the system to survive rates that saturate the system, however.

Clean up declarations.

Include <machine/clock.h> to check our own declarations.

29 years agoFix a bug introduced between 1.1 and 1.1.5. Loading the time was moved
bde [Sat, 5 Nov 1994 23:53:46 +0000 (23:53 +0000)]
Fix a bug introduced between 1.1 and 1.1.5.  Loading the time was moved
outside the critical region.

Make it work with 2.0.  It wasn't designed to be called at splclock().

Make it work with prescaling.  The overflow threshold was bogus.

Make it work for any HZ.  Side effect of fixing prescaling.

Speed it up.  Allocate registers better.  Reduce multiplication and
division to multiplication and a shift.  Speed is now 5-6 usec on a
486DX/33, was about 3 usec more.

Optimize for the non-pentium case.  The pentium code got moved around
a bit and hasn't been tested.

Change #include's to 2.0 style.

29 years agoMake ShowFile() now show gzip'd text files. This should save a fair
jkh [Sat, 5 Nov 1994 23:12:10 +0000 (23:12 +0000)]
Make ShowFile() now show gzip'd text files.  This should save a fair
bit of space on the install floppy.  Modify stage0 and the target names
accordingly.

29 years agoEnable back dialog_msgbox in TellEm, real problem was delwin
ache [Sat, 5 Nov 1994 23:09:26 +0000 (23:09 +0000)]
Enable back dialog_msgbox in TellEm, real problem was delwin
into edit_line

29 years agoInstall docs gzip'd.
jkh [Sat, 5 Nov 1994 23:05:53 +0000 (23:05 +0000)]
Install docs gzip'd.

29 years agoDeclare all functions exported by the npx driver.
bde [Sat, 5 Nov 1994 22:59:09 +0000 (22:59 +0000)]
Declare all functions exported by the npx driver.

Uniformize idempotency ifdefs.

29 years agoDeclare the full uglyness of the interfaces to the clock driver (except
bde [Sat, 5 Nov 1994 22:51:17 +0000 (22:51 +0000)]
Declare the full uglyness of the interfaces to the clock driver (except
things declared in machine-independent files).

29 years agoDisable the direct call from hardclock() to softclock(). Support
bde [Sat, 5 Nov 1994 22:44:34 +0000 (22:44 +0000)]
Disable the direct call from hardclock() to softclock().  Support
for it is incomplete and buggy.  There is no problem unless Xintr0()
is reentered or should be reentered, but high clock interrupt
frequencies for pcaudio cause Xintr0() to be reentered (or clock
ticks to be lost when Xintr0() should have been reentered but
wasn't), and we lose little by delaying the call to softclock().

Move declarations related to the clock driver to clock.h.

Move declarations related to the npx driver to npx.h.

Clean up the remaining declarations.

29 years agoChange -O2 to -O
ache [Sat, 5 Nov 1994 21:06:16 +0000 (21:06 +0000)]
Change -O2 to -O
With each gcc version -O2 can cause absolutely unpredicatable things
Second stage boot still fits in allowed size

29 years agoFix group permissions check for group list size > 1
ache [Sat, 5 Nov 1994 20:48:06 +0000 (20:48 +0000)]
Fix group permissions check for group list size > 1
Pointed by Bruce

29 years agoFix root -w case according to manpage (-x & -r cases already accords)
ache [Sat, 5 Nov 1994 20:24:49 +0000 (20:24 +0000)]
Fix root -w case according to manpage (-x & -r cases already accords)
Pointed by Bruce.

29 years agoDate: Wed, 26 Oct 1994 15:44:49 -0600
nate [Sat, 5 Nov 1994 18:49:34 +0000 (18:49 +0000)]
Date: Wed, 26 Oct 1994 15:44:49 -0600
From: Chris Torek <torek@bsdi.com>
Here is a semi-official patch (apply to /usr/src/lib/libc/stdio/fseek.c,
rebuild libc, install).  The current code fails when the seek:

  - is optimized, and
  - is to just past the end of the block currently in the buffer, and
  - is followed by another seek with no intervening read operation, and
  - the destination of subsequent seek is within the block left in the
    buffer (seeking to the beginning of a block does not force a read,
    so the buffer still contains the previous block)

so it is indeed rather obscure.

I may have a different `final' fix, as this one `loses' the buffer
contents on a seek that goes just past the end of the current block.

[Footnote: seeks are optimized only on read-only opens of regular
files that are buffered by the file's optimal I/O size.  This is
what you get with fopen(path, "r") and no call to setvbuf().]

Obtained from: [ BSDI mailing list ]

29 years agoCosmetique, missing tab/spaces after patch
ache [Sat, 5 Nov 1994 17:31:23 +0000 (17:31 +0000)]
Cosmetique, missing tab/spaces after patch

29 years agoMore complex implementation of previos superuser fix
ache [Sat, 5 Nov 1994 17:28:03 +0000 (17:28 +0000)]
More complex implementation of previos superuser fix
because rw and x bit treats differently infact

29 years agotest produce wrong results for superuser, i.e. tells
ache [Sat, 5 Nov 1994 17:07:14 +0000 (17:07 +0000)]
test produce wrong results for superuser, i.e. tells
that file is unreadable when it is readable infact.

29 years agoAdd targets to crunch floppy to match Makefile.
jkh [Sat, 5 Nov 1994 09:06:10 +0000 (09:06 +0000)]
Add targets to crunch floppy to match Makefile.

29 years agoAdd ncftp. I needed it for "hands-free" ftp operation from the install
jkh [Sat, 5 Nov 1994 08:49:05 +0000 (08:49 +0000)]
Add ncftp.  I needed it for "hands-free" ftp operation from the install
script.

29 years agoWhoops! Knew I forgot something.. Bump the menu size.
jkh [Sat, 5 Nov 1994 08:20:54 +0000 (08:20 +0000)]
Whoops!  Knew I forgot something..  Bump the menu size.

29 years agoAdd a chance to read the release notes, too.
jkh [Sat, 5 Nov 1994 08:20:08 +0000 (08:20 +0000)]
Add a chance to read the release notes, too.

29 years agoAdd the release notes, so I can have the user read these too.
jkh [Sat, 5 Nov 1994 08:16:01 +0000 (08:16 +0000)]
Add the release notes, so I can have the user read these too.

29 years agoPut this here instead.
jkh [Sat, 5 Nov 1994 08:11:22 +0000 (08:11 +0000)]
Put this here instead.

29 years agoAdd new mount_* commands.
jkh [Sat, 5 Nov 1994 07:55:30 +0000 (07:55 +0000)]
Add new mount_* commands.

29 years agoAdd slattach.
jkh [Sat, 5 Nov 1994 07:52:14 +0000 (07:52 +0000)]
Add slattach.

29 years agoArgh! Missing quotes.
jkh [Sat, 5 Nov 1994 07:41:08 +0000 (07:41 +0000)]
Argh!  Missing quotes.

29 years agoWe need CD9660 and MSDOS filesystems built-in if the floppy is to have
jkh [Sat, 5 Nov 1994 07:38:50 +0000 (07:38 +0000)]
We need CD9660 and MSDOS filesystems built-in if the floppy is to have
a hope of getting at these types of filesystems without dragging all
the LKM stuff in.

29 years agoI'm not sure if we're just supposed to fall off the end like this,
jkh [Sat, 5 Nov 1994 07:34:02 +0000 (07:34 +0000)]
I'm not sure if we're just supposed to fall off the end like this,
but at least return something.

29 years agoMake stage5 go either to a shell, or on to the bindist installation.
jkh [Sat, 5 Nov 1994 07:23:58 +0000 (07:23 +0000)]
Make stage5 go either to a shell, or on to the bindist installation.

29 years agoMake this work the way it was really supposed to - less intrusively.
jkh [Sat, 5 Nov 1994 07:03:55 +0000 (07:03 +0000)]
Make this work the way it was really supposed to - less intrusively.

29 years agoDe-Danish this. Add a lot more chatty commentary in my own inimitable style.
jkh [Sat, 5 Nov 1994 06:54:49 +0000 (06:54 +0000)]
De-Danish this.  Add a lot more chatty commentary in my own inimitable style.

29 years agoUse /DISKSPACE.FAQ as help-file in the disk-editor.
phk [Sat, 5 Nov 1994 06:03:06 +0000 (06:03 +0000)]
Use /DISKSPACE.FAQ as help-file in the disk-editor.

29 years agoput the DISKSPACE.FAQ on the boot.flp
phk [Sat, 5 Nov 1994 06:02:18 +0000 (06:02 +0000)]
put the DISKSPACE.FAQ on the boot.flp

29 years agoThis file is the documentation for the fdisk/disklabel part of sysinstall,
phk [Sat, 5 Nov 1994 05:54:21 +0000 (05:54 +0000)]
This file is the documentation for the fdisk/disklabel part of sysinstall,
and will be available from the boot.flp with just one key-stroke.  Somebody
will want to purge any danishisms from it before is is let loose...

29 years agoChange the logic for how error messages are printed so we can actually
jkh [Sat, 5 Nov 1994 05:46:12 +0000 (05:46 +0000)]
Change the logic for how error messages are printed so we can actually
see them.

29 years agoCosmetics: edit_line() does standend() when done;
phk [Sat, 5 Nov 1994 03:59:34 +0000 (03:59 +0000)]
Cosmetics:  edit_line() does standend() when done;

29 years agoAdded a (H)elp function to the fdisk/disklabel code, much needed, now I only
phk [Sat, 5 Nov 1994 03:34:22 +0000 (03:34 +0000)]
Added a (H)elp function to the fdisk/disklabel code, much needed, now I only
have to write the text-file it will dump on you.
Stopped using cleartoeol in AskEm(), just as a test to see if the screen
looks more sane.
Added a attempted auto-recognition of /, swap and /usr for the first disk
where it looks sensible.  Logic of this might need to be improved.
Made a "ShowFile()" which will not bomb/ignore you if the file isn't there.

29 years agoFixed a ghastly mistake. Edit_line() would "delwin(window)". Bad move
phk [Sat, 5 Nov 1994 03:13:33 +0000 (03:13 +0000)]
Fixed a ghastly mistake.  Edit_line() would "delwin(window)".  Bad move
buddy, you'd better not call with stdwin then!   fixed now.
Seems like dialog is a LOT more rational now.

29 years agoExplicitly tell cpp where to look for include files.
phk [Fri, 4 Nov 1994 21:40:39 +0000 (21:40 +0000)]
Explicitly tell cpp where to look for include files.

29 years agoFix a potential bug in stage2.
phk [Fri, 4 Nov 1994 21:38:37 +0000 (21:38 +0000)]
Fix a potential bug in stage2.

disable the dialog_msgbox in TellEm(), until we find out why it
hangs on second or third invocation.

29 years agoFrom njw@cs.city.ac.uk (Nick Williams):
dg [Fri, 4 Nov 1994 17:51:37 +0000 (17:51 +0000)]
From njw@cs.city.ac.uk (Nick Williams):

/sbin/umount does not return the correct exit status due to incorrect
logic in its internals.

Further, because of the nature of the code, you *cannot* use it to
umount a directory from a union mountpoint. Well, you can sometimes,
it depends on if the directory is at the top of the union stack or not :)

Submitted by: njw@cs.city.ac.uk (Nick Williams)

29 years agoAdd const to termcap prototypes to help libg++ 2.6.1 compiling,
ache [Fri, 4 Nov 1994 15:14:03 +0000 (15:14 +0000)]
Add const to termcap prototypes to help libg++ 2.6.1 compiling,
this change must not affect other curses pgms

29 years agoFrom tim@cs.city.ac.uk (Tim Wilkinson):
dg [Fri, 4 Nov 1994 14:41:46 +0000 (14:41 +0000)]
From tim@cs.city.ac.uk (Tim Wilkinson):

Find enclosed a short bugfix to get the union filesystem up and running
in FreeBSD-current.  We don't think we've got all the problems yet but
these fixes sort out the major ones (which mostly concert bad locking
of vnodes), no doubt we'll post others as necessary.  Known problems
include the inability of the umount command (not the system call) to unmount
unions in certain circumstances (this is due the way "realpath" works),
and the failure of direntries to always get all available files in
unioned subdirectories.  We are, as they say, working on it.

Submitted by: tim@cs.city.ac.uk (Tim Wilkinson)

29 years agoBetter tuning for FreeBSD 2.0
ache [Fri, 4 Nov 1994 06:59:17 +0000 (06:59 +0000)]
Better tuning for FreeBSD 2.0

29 years agoMerge changes with new version
ache [Fri, 4 Nov 1994 06:11:39 +0000 (06:11 +0000)]
Merge changes with new version

29 years agoThis commit was generated by cvs2svn to compensate for changes in r4135,
ache [Fri, 4 Nov 1994 05:36:00 +0000 (05:36 +0000)]
This commit was generated by cvs2svn to compensate for changes in r4135,
which included commits to RCS files with non-trunk default branches.

29 years agoInitial import 1.8.6
ache [Fri, 4 Nov 1994 05:36:00 +0000 (05:36 +0000)]
Initial import 1.8.6

29 years agoPrepare for 1.8.6 import
ache [Fri, 4 Nov 1994 05:34:16 +0000 (05:34 +0000)]
Prepare for 1.8.6 import

29 years agoLast commit was bogus. Changed b_bsize to b_bcount.
phk [Fri, 4 Nov 1994 05:21:17 +0000 (05:21 +0000)]
Last commit was bogus.  Changed b_bsize to b_bcount.

29 years agoSimplify __FreeBSD__ block since __386BSD_ eliminated
ache [Fri, 4 Nov 1994 04:53:13 +0000 (04:53 +0000)]
Simplify __FreeBSD__ block since __386BSD_ eliminated

29 years ago__386BSD__ -> __FreeBSD__
jkh [Fri, 4 Nov 1994 02:14:13 +0000 (02:14 +0000)]
__386BSD__ -> __FreeBSD__

I know that many of these entries are bogus and need to be revisited,
but let's get the tree working again for now and then do a pass through
looking at all the __FreeBSD__ entries, shall we?

29 years agoBacked out mbuf performance improvement. mbufs are allocated with various
dg [Fri, 4 Nov 1994 00:28:38 +0000 (00:28 +0000)]
Backed out mbuf performance improvement. mbufs are allocated with various
different types, and with the 'local cache', what is freed isn't necessarily
what was originally malloced. This screws malloc's statistics and type
allocation limits, resulting eventually in a deadlock when one of the
limits is bogusly reached. Recent performance tests on a Pentium machine
indicate no improvement with this optimization anyway (this is something
to be looked at further).

29 years agoCorrect my hosing...:(
gclarkii [Thu, 3 Nov 1994 23:50:43 +0000 (23:50 +0000)]
Correct my hosing...:(

29 years agoNuked #ifdef __386BSD__ since this is not appropriate and no longer defined
dg [Thu, 3 Nov 1994 21:46:55 +0000 (21:46 +0000)]
Nuked #ifdef __386BSD__ since this is not appropriate and no longer defined
in our 2.6.1 gcc port.

29 years agoFix off-by-one error reported to NetBSD by Karl Fox in
wollman [Thu, 3 Nov 1994 21:04:21 +0000 (21:04 +0000)]
Fix off-by-one error reported to NetBSD by Karl Fox in
<9411031449.AA11102@gefilte.MorningStar.Com>.

29 years agoBump the version-number of the "shared-gcc-internals" library. Will only
phk [Thu, 3 Nov 1994 20:37:05 +0000 (20:37 +0000)]
Bump the version-number of the "shared-gcc-internals" library.  Will only
affect you if you decided to use that feature.

29 years agoChange libs to install in /usr/lib/perl
gclarkii [Thu, 3 Nov 1994 20:07:46 +0000 (20:07 +0000)]
Change libs to install in /usr/lib/perl

29 years agoChange libs to /usr/lib/perl
gclarkii [Thu, 3 Nov 1994 20:03:53 +0000 (20:03 +0000)]
Change libs to /usr/lib/perl
Reviewed by:
Submitted by:
Obtained from:

29 years agoInstall libs into /usr/lib/perl. The discussion we had did not come to
gclarkii [Thu, 3 Nov 1994 20:02:15 +0000 (20:02 +0000)]
Install libs into /usr/lib/perl.  The discussion we had did not come to
a point so...  If any thing besides here I would vote for /usr/include.

29 years agoFor submakes, don't do a `${MAKE} ${MAKEFILE} all' or
jkh [Thu, 3 Nov 1994 19:53:46 +0000 (19:53 +0000)]
For submakes, don't do a `${MAKE} ${MAKEFILE} all' or
`${MAKE} ${MAKEFILE} install', indirect the target through ALL_TARGET and
INSTALL_TARGET now, so they can be easily overridden.

29 years agoAdd reinstall target for Poul-H. This lets you install again from
jkh [Thu, 3 Nov 1994 19:14:08 +0000 (19:14 +0000)]
Add reinstall target for Poul-H.  This lets you install again from
the same build, ignoring the INSTALL_COOKIE.

29 years ago|Both, the wd and the sd driver do not reject invalid request of odd
joerg [Thu, 3 Nov 1994 18:20:15 +0000 (18:20 +0000)]
|Both, the wd and the sd driver do not reject invalid request of odd
|sizes. They simply pass them to the disks, which usually causes
|fatal errors then.

29 years agoAssign character device 20 to be the user reserved device.
pst [Thu, 3 Nov 1994 17:49:01 +0000 (17:49 +0000)]
Assign character device 20 to be the user reserved device.

29 years agoEliminate USERCONFIG. This option is now standard.
jkh [Thu, 3 Nov 1994 15:51:38 +0000 (15:51 +0000)]
Eliminate USERCONFIG.  This option is now standard.

29 years agoDon't include map.h - it's been obsoleted.
dg [Thu, 3 Nov 1994 15:34:24 +0000 (15:34 +0000)]
Don't include map.h - it's been obsoleted.

29 years agoUnconditionalize USERCONFIG. Uh, thanks, David.
jkh [Thu, 3 Nov 1994 14:57:54 +0000 (14:57 +0000)]
Unconditionalize USERCONFIG.  Uh, thanks, David.

29 years agoRemove obsolete __386BSD__ predefine.
jkh [Thu, 3 Nov 1994 14:29:57 +0000 (14:29 +0000)]
Remove obsolete __386BSD__ predefine.

29 years agoFix for 'pstat -t' works on vtys
ache [Thu, 3 Nov 1994 11:41:34 +0000 (11:41 +0000)]
Fix for 'pstat -t' works on vtys
Submitted by: jhay@mikom.csir.co.za

29 years ago----------------------------------
phk [Thu, 3 Nov 1994 06:52:42 +0000 (06:52 +0000)]
----------------------------------
GCC-2.6.1 COMES TO FREEBSD-current
----------------------------------
Everybody needs to 'make world'.

Oakland, Nov 2nd 1994.  In a surprise move this sunny afternoon, the release-
engineer for the slightly delayed FreeBSD-2.0, Poul-Henning Kamp (28),
decided to pull in the new version 2.6.1 of the GNU C-compiler.
The new version of the compiler was release today at noon, and hardly 9
hours later it was committed into the FreeBSD-current source-repository.
"It's is simply because we have had too much trouble with the version 2.6.0
of the compiler" Poul-Henning told the FreeBSD-Gazette, "we took a gamble
when we decided to use that as our compiler for the 2.0 release, but it
seems to pay of in the end now" he concludes.
The move has not been discussed on the "core" list at all, and will come as
a surprise for most Poul-Hennings peers.  "I have only discussed it with
Jordan [J. K. Hubbard, the FreeBSD's resident humourist], and we agreed that
we needed to do it, so ... I did it!".  After a breath he added with a grin:
"My email will probably get an all time 'disk-full' now!".
This will bring quite a flag-day to the FreeBSD developers, the patch-file
is almost 1.4 Megabyte, and they will have to run "make world" to get
entirely -current again.  "Too bad, but we just had to do this."  Was
the only comment from Poul-Henning to these problems.
When asked how this move would impact the 2.0 release-date, Poul-Hennings
face grew dark, he mumbled some very Danish words while he moved his fingers
in strange geometrical patterns.  Immediately something ecclipsed the Sun, a
minor tremor shook the buildings, and the temperature fell significantly.
We decided not to pursure the question.

-----------
JOB-SECTION
-----------
Are you a dedicated GCC-hacker ?
We BADLY need somebody to look at the 'freebsd' OS in gcc, sanitize it and
carry the patches back to the GNU people.  In particular, we need to get
out of the "i386-only" spot we are in now.  I have the stuff to take a
gnu-dist into bmake-form, and will do that part.

Please apply to phk@freebsd.org

No Novice Need Apply.

29 years agoWhoops - one two many zeros in the initializer. This should have everybody
jkh [Thu, 3 Nov 1994 05:37:53 +0000 (05:37 +0000)]
Whoops - one two many zeros in the initializer.  This should have everybody
enabled by default now.

29 years agoWhoops - make sure TRUE and FALSE are defined now.
jkh [Thu, 3 Nov 1994 05:02:41 +0000 (05:02 +0000)]
Whoops - make sure TRUE and FALSE are defined now.

29 years agoChanges to initialize the id_enabled flag to 1.
jkh [Thu, 3 Nov 1994 04:30:01 +0000 (04:30 +0000)]
Changes to initialize the id_enabled flag to 1.
Also initialize some fields that were never initialized before, and
simply defaulted to 0.  I've never looked at this code before, now
I know why.  Config needs to die.  Horribly.

29 years agoAdd extra id_enabled flag for userconfig to manipulate. If id_enabled
jkh [Thu, 3 Nov 1994 04:15:04 +0000 (04:15 +0000)]
Add extra id_enabled flag for userconfig to manipulate.  If id_enabled
is FALSE, the device will not be probed.  id_enabled is TRUE by default.

29 years agoMake the enable & disable commands finally work.
jkh [Thu, 3 Nov 1994 04:14:23 +0000 (04:14 +0000)]
Make the enable & disable commands finally work.

29 years ago-i case goes to background pause and acts like -a case on receiving
ache [Thu, 3 Nov 1994 03:16:16 +0000 (03:16 +0000)]
-i case goes to background pause and acts like -a case on receiving
SIGTERM. This helps to keep CMOS clock updated before reboot.
Idea from J.Wunsch.

29 years agoCosmetic changes in comment at start (it's no longer a GENERICAH config
gpalmer [Thu, 3 Nov 1994 01:46:59 +0000 (01:46 +0000)]
Cosmetic changes in comment at start (it's no longer a GENERICAH config
file!)

29 years agoCompletely replace JTW's idea with my (incompletely implemented) original
wollman [Thu, 3 Nov 1994 01:05:34 +0000 (01:05 +0000)]
Completely replace JTW's idea with my (incompletely implemented) original
idea.  This is les likely to crash your machine.  As before, this code is only
enabled under `options IN_RMX'.

29 years agoCollapse two fields so that we have space for another 32 flags.
wollman [Thu, 3 Nov 1994 01:04:32 +0000 (01:04 +0000)]
Collapse two fields so that we have space for another 32 flags.
NB: You will have to recompile programs which use the `rt_use' member in
order to get the correct values.  This should not cause incorrect operation,
but the statistics may look a little confusing.

29 years agoRestore jkh's changes, now it works for const strings
ache [Thu, 3 Nov 1994 00:30:27 +0000 (00:30 +0000)]
Restore jkh's changes, now it works for const strings

29 years agoNow autosizing code works with const strings too.
ache [Thu, 3 Nov 1994 00:28:05 +0000 (00:28 +0000)]
Now autosizing code works with const strings too.
Prototype added.

29 years agoDon't set hostname if we already have it. (diskless).
phk [Wed, 2 Nov 1994 23:50:54 +0000 (23:50 +0000)]
Don't set hostname if we already have it.  (diskless).

29 years agoSubmitted by:
se [Wed, 2 Nov 1994 23:47:14 +0000 (23:47 +0000)]
Submitted by:
Added hooks for "lsdev" ...
PCI devices should need no individual code for lsdev.

29 years agoSubmitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de>
se [Wed, 2 Nov 1994 23:44:11 +0000 (23:44 +0000)]
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de>
Modifications required for PCI support.

29 years agoSubmitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de>
se [Wed, 2 Nov 1994 23:27:36 +0000 (23:27 +0000)]
Submitted by: Wolfgang Stanglmeier <wolf@dentaro.GUN.de>
Added PCI support code.

29 years agoDon't use autosizing code after all - it doesn't like const strings.
jkh [Wed, 2 Nov 1994 22:39:57 +0000 (22:39 +0000)]
Don't use autosizing code after all - it doesn't like const strings.

29 years agoa missing star.
phk [Wed, 2 Nov 1994 22:06:24 +0000 (22:06 +0000)]
a missing star.

29 years agoFix from Gary Jennejohn - use 'cp' not 'buf' in read call. Oops.
dg [Wed, 2 Nov 1994 16:38:51 +0000 (16:38 +0000)]
Fix from Gary Jennejohn - use 'cp' not 'buf' in read call. Oops.