unix-history/.git
30 years agoSpecify library names explicitly. Bug reported by hd@world.std.com.
Garrett Wollman [Fri, 4 Mar 1994 20:27:02 +0000 (20:27 +0000)]
Specify library names explicitly.  Bug reported by hd@world.std.com.

30 years agoAdded missing Kerberos password changing functionality.
Garrett Wollman [Fri, 4 Mar 1994 20:25:16 +0000 (20:25 +0000)]
Added missing Kerberos password changing functionality.

30 years agoMemcmp compares SIGNED characters! (change movsbl to movzbl)
Andrey Chernov [Fri, 4 Mar 1994 19:47:11 +0000 (19:47 +0000)]
Memcmp compares SIGNED characters! (change movsbl to movzbl)

30 years agoI got a lot of
Andrey Chernov [Fri, 4 Mar 1994 17:51:39 +0000 (17:51 +0000)]
I got a lot of
"getty repeating too quickly on port %s, sleeping"
from init.bsdi, it means that getty start and exit in five seconds.
This is common situation for poor quality Russian phone lines:
modem got CONNECT message and after retries got NO CARRIER.
So I introduce spacing count, it means that this warning and
sleep occurse only after GETTY_NSPACE times of sequental attempts.

30 years ago1) My previous fix does nothing, now Rod's fix rewrited to implement
Andrey Chernov [Fri, 4 Mar 1994 16:43:07 +0000 (16:43 +0000)]
1) My previous fix does nothing, now Rod's fix rewrited to implement
my previous fix too (using wdp_heads controller value) and check
0 heads case too.

Other fixes from Bruce:
2) Fix dk_timeout from 2 to about 4 seconds.
3) wdcontrol not retried on internal error.
4) wdwait return check changed ( "!=0" to "<0" into wdsetctlr,
"<0" to "!=0" into wdgetctlr).

30 years agoFix unsigned comparation (movsbl --> movzbl)
Andrey Chernov [Fri, 4 Mar 1994 15:50:29 +0000 (15:50 +0000)]
Fix unsigned comparation (movsbl --> movzbl)

30 years agoFix typo in man page
Steven Wallace [Fri, 4 Mar 1994 08:07:08 +0000 (08:07 +0000)]
Fix typo in man page

30 years agoDenmark has shuffled telephone numbers, no clear picture for the next year.
Poul-Henning Kamp [Fri, 4 Mar 1994 00:11:50 +0000 (00:11 +0000)]
Denmark has shuffled telephone numbers, no clear picture for the next year.

30 years agoAdded kernel NTP PLL code from Dave Mills.
Garrett Wollman [Thu, 3 Mar 1994 20:23:34 +0000 (20:23 +0000)]
Added kernel NTP PLL code from Dave Mills.

30 years agoMade a failure when committing the last version.
Guido van Rooij [Wed, 2 Mar 1994 20:35:09 +0000 (20:35 +0000)]
Made a failure when committing the last version.

30 years agoTtys structures are now allocated dynamically via ttymalloc/ttyfree.
Guido van Rooij [Wed, 2 Mar 1994 20:29:03 +0000 (20:29 +0000)]
Ttys structures are now allocated dynamically via ttymalloc/ttyfree.
This inetrface should be used from now on.
pseudo device pty xx still keeps its meaning: a maximum of
xx ptys is allowed.
A ringbuffer is now 2040 bytes long, per Garrett Wollman's request.
The changes are inspired by the way NetBSD did it (thanks for that!),
though I made it slihghtly different, including the interface so
at least 75% of the allocated space is deallocated when the tty is
closed.
Note further that it is easy to modify the ringbuffer length runtime.
This will have to wait untill some later date...

-Guido

30 years agoTtys structures are now allocated dynamically via ttymalloc/ttyfree.
Guido van Rooij [Wed, 2 Mar 1994 20:18:09 +0000 (20:18 +0000)]
Ttys structures are now allocated dynamically via ttymalloc/ttyfree.
This inetrface should be used from now on.
pseudo device pty xx still keeps its meaning: a maximum of
xx ptys is allowed.
A ringbuffer is now 2040 bytes long, per Garrett Wollman's request.
The changes are inspired by the way NetBSD did it (thanks for that!),
though I made it slihghtly different, including the interface so
at least 75% of the allocated space is deallocated when the tty is
closed.
Note further that it is easy to modify the ringbuffer length runtime.
This will have to wait untill some later date...

-Guido

30 years agoWARNING! Records layout in *pwd.db now changed (for reasons see fix
Andrey Chernov [Wed, 2 Mar 1994 19:46:20 +0000 (19:46 +0000)]
WARNING! Records layout in *pwd.db now changed (for reasons see fix
list below), so you need some sort of upgrade after SUPing -current
sources.

Here strategy step by step:
1) Re-compile and re-install pwd_mkdb.
2) Re-compile libc (but not install it at this step).
2) Run passwd or any pwd_mkdb related program to produce new *pwd.db
3) Re-install libc.
(all shared pgms with getpw*() functions automatically becomes correct).
4) Re-compile and re-install all non-shared pgms, which use
getpw*() functions (more simple step is to rebuild /sbin & /bin)

Fix list:

1) Patch from bag@demos.su, original description included:

>Problems:

>Previous version of passwd dbase work ok for small number of users (<500).
>On our site we have > 2500 logins for uucp hosts and in this case size
>of /etc/master.passwd > 300Kb and size of pwd.db+spwd.db > 2.5 Mb.

>Original FreeBSD pwd_mkdb work > 15 min to rebuild passwd db and
>require > 5 M on root fs. Passwd changing (system administrator
>make/delete > 20 accouns per day) is a big problem and require
>too many resources.

>BSDI patch for fasting pwd_mkdb due to db caching require 4Mb cache
>(2Mb per each db - pwd.db/spwd.db). When empty real 4Mb mem not available
>swaping eat all our speed.

>Building pwd.db/spwd.db sequentialy require only 2Mb cache, but additional
>reading of /etc/master.passwd. If 2Mb mem available pwd_mkdb works only 20s
>in our case, but 5Mb space requied on disk for rebuilding and in the case
>you have 8Mb RAM only < 1Mb available and swaping very intensive.

>New features:

>With modified passwd db structure on our site we have sizes of
>pwd.db ~320Kb, of spwd.db ~80Kb (sum ~400Kb) and rebuilding time 5s.

>pwd.db/spwd.db rebuilding simultaneously and use only 640Kb cache.

>We found that work time for lib's functions with old and new passwd
>structure practicaly equals.

>Without #define PW_COMPACT in pwd_mkdb.c and getpwent.c sources
>compatible with old pwd.db/spwd.db structure with BSDI patch.

>      Alex G. Bulushev, bag@demos.su

30 years agoWARNING! Records layout in *pwd.db now changed (for reasons see fix
Andrey Chernov [Wed, 2 Mar 1994 19:40:01 +0000 (19:40 +0000)]
WARNING! Records layout in *pwd.db now changed (for reasons see fix
list below), so you need some sort of upgrade after SUPing -current
sources.

Here strategy step by step:
1) Re-compile and re-install pwd_mkdb.
2) Re-compile libc (but not install it at this step).
2) Run passwd or any pwd_mkdb related program to produce new *pwd.db
3) Re-install libc.
(all shared pgms with getpw*() functions automatically becomes correct).
4) Re-compile and re-install all non-shared pgms, which use
getpw*() functions (more simple step is to rebuild /sbin & /bin)

Fix list:

1) Fix return code and type for scan() function, remove unnedeed exit
from there.
2) Fix fsync argument.
3) Patch from sms@wlv.iipo.gtegsc.com (BSDI patch) to improve hash
parameters for build speedup and build *pwd.bd in sequental order to
prevent disk trashing.
4) Patch from bag@demos.su, original description included:

>Problems:

>Previous version of passwd dbase work ok for small number of users (<500).
>On our site we have > 2500 logins for uucp hosts and in this case size
>of /etc/master.passwd > 300Kb and size of pwd.db+spwd.db > 2.5 Mb.

>Original FreeBSD pwd_mkdb work > 15 min to rebuild passwd db and
>require > 5 M on root fs. Passwd changing (system administrator
>make/delete > 20 accouns per day) is a big problem and require
>too many resources.

>BSDI patch for fasting pwd_mkdb due to db caching require 4Mb cache
>(2Mb per each db - pwd.db/spwd.db). When empty real 4Mb mem not available
>swaping eat all our speed.

>Building pwd.db/spwd.db sequentialy require only 2Mb cache, but additional
>reading of /etc/master.passwd. If 2Mb mem available pwd_mkdb works only 20s
>in our case, but 5Mb space requied on disk for rebuilding and in the case
>you have 8Mb RAM only < 1Mb available and swaping very intensive.

>New features:

>With modified passwd db structure on our site we have sizes of
>pwd.db ~320Kb, of spwd.db ~80Kb (sum ~400Kb) and rebuilding time 5s.

>pwd.db/spwd.db rebuilding simultaneously and use only 640Kb cache.

>We found that work time for lib's functions with old and new passwd
>structure practicaly equals.

>Without #define PW_COMPACT in pwd_mkdb.c and getpwent.c sources
>compatible with old pwd.db/spwd.db structure with BSDI patch.

>      Alex G. Bulushev, bag@demos.su

30 years ago360 DD entry speed fixed to 250KBPS
Andrey Chernov [Wed, 2 Mar 1994 18:34:41 +0000 (18:34 +0000)]
360 DD entry speed fixed to 250KBPS

30 years agokrcmd fix.
Garrett Wollman [Wed, 2 Mar 1994 18:01:05 +0000 (18:01 +0000)]
krcmd fix.

30 years agoProvide kerberized rcmd source file.
Garrett Wollman [Wed, 2 Mar 1994 17:59:47 +0000 (17:59 +0000)]
Provide kerberized rcmd source file.

30 years agoFixed comments.
Garrett Wollman [Wed, 2 Mar 1994 17:58:29 +0000 (17:58 +0000)]
Fixed comments.

30 years agoAdded missing file for Kerberized r* utilities.
Garrett Wollman [Wed, 2 Mar 1994 17:17:46 +0000 (17:17 +0000)]
Added missing file for Kerberized r* utilities.
cVS: ----------------------------------------------------------------------

30 years agofixed typoo in kerberos (kdc -> kdb). Added macro which points to CC-source.
Poul-Henning Kamp [Wed, 2 Mar 1994 12:32:44 +0000 (12:32 +0000)]
fixed typoo in kerberos (kdc -> kdb).  Added macro which points to CC-source.

30 years agofloppy tape support shouldn't interfere with floppy disk driver on
Andrew Moore [Wed, 2 Mar 1994 08:10:42 +0000 (08:10 +0000)]
floppy tape support shouldn't interfere with floppy disk driver on
systems with no floppy tape drive (patch from Jim Babb).

30 years agoMinor comment change.
David Greenman [Wed, 2 Mar 1994 05:50:01 +0000 (05:50 +0000)]
Minor comment change.

30 years agoFix from John Dyson: the last param in a call to swap_pager_diskaddr
David Greenman [Wed, 2 Mar 1994 02:53:27 +0000 (02:53 +0000)]
Fix from John Dyson: the last param in a call to swap_pager_diskaddr
was incorrect.

30 years agoDon't sync a ro-mounted fs.
Andrew Moore [Wed, 2 Mar 1994 00:03:41 +0000 (00:03 +0000)]
Don't sync a ro-mounted fs.

30 years agoDon't offer to mount a to-be-overwritten DOS partition.
Andrew Moore [Tue, 1 Mar 1994 23:51:48 +0000 (23:51 +0000)]
Don't offer to mount a to-be-overwritten DOS partition.

30 years agoRemove last relic of the dcfclk driver.
Poul-Henning Kamp [Tue, 1 Mar 1994 23:21:44 +0000 (23:21 +0000)]
Remove last relic of the dcfclk driver.

30 years agoAdded the missing bzero after the malloc in aha1542.c that seems to
Andreas Schulz [Tue, 1 Mar 1994 16:06:37 +0000 (16:06 +0000)]
Added the missing bzero after the malloc in aha1542.c that seems to
produce the mbox not free messages on many systems. Bug reported from
hd@world.std.com.

30 years agoFix from Christoph Robitschko: the '790 interrupt table was wrong.
David Greenman [Tue, 1 Mar 1994 12:23:33 +0000 (12:23 +0000)]
Fix from Christoph Robitschko: the '790 interrupt table was wrong.

30 years agofloppy tape support still not quite there - it prevents
Andrew Moore [Tue, 1 Mar 1994 01:27:03 +0000 (01:27 +0000)]
floppy tape support still not quite there - it prevents
mounting the floppy disk  drives on some systems that don't have
a floppy tape drive.

30 years agoI found (and fix) ugly bugs in init.bsdi (this bugs not present
Andrey Chernov [Mon, 28 Feb 1994 21:53:52 +0000 (21:53 +0000)]
I found (and fix) ugly bugs in init.bsdi (this bugs not present
in old init)

1) Init don't setup TERM environment variable for default terminal
type from /etc/ttys before calling getty/window.

2) When "kill -1 1" issued, init don't restart getty when
/etc/ttys parameters was changed (it only kill "off" end empty entries).

3) Small memory leak if "window" /etc/ttys parameter specified and
"kill -1 1" issued.

30 years agoAdd floppy tape support.
Andrew Moore [Mon, 28 Feb 1994 04:19:39 +0000 (04:19 +0000)]
Add floppy tape support.

30 years agoChanges backed out until discussion stopped.
Andrey Chernov [Sun, 27 Feb 1994 22:29:41 +0000 (22:29 +0000)]
Changes backed out until discussion stopped.

30 years agoChanges backed out, until discussion stopped
Andrey Chernov [Sun, 27 Feb 1994 22:28:35 +0000 (22:28 +0000)]
Changes backed out, until discussion stopped

30 years agodcfclk driver obsoleted, sio/TIOCTIMESTAMP took over.
Poul-Henning Kamp [Sun, 27 Feb 1994 21:53:54 +0000 (21:53 +0000)]
dcfclk driver obsoleted, sio/TIOCTIMESTAMP took over.

30 years agodcfclk driver obsoleted by sio/TIOCTIMESTAMP.
Poul-Henning Kamp [Sun, 27 Feb 1994 21:51:05 +0000 (21:51 +0000)]
dcfclk driver obsoleted by sio/TIOCTIMESTAMP.

30 years agoReceived a bleesed patch from GNU for the ((packed)) problem. Got rid of
Poul-Henning Kamp [Sun, 27 Feb 1994 21:42:35 +0000 (21:42 +0000)]
Received a bleesed patch from GNU for the ((packed)) problem.  Got rid of
a couple of '.5' suffixes which had slipped through.

30 years agoAdded /etc/kerberosIV directory. NB: the sample files should NOT be
Garrett Wollman [Sun, 27 Feb 1994 21:29:40 +0000 (21:29 +0000)]
Added /etc/kerberosIV directory.  NB: the sample files should NOT be
installed by default, because then everybody would suddenly start
trying to authenticate themselves in the CS.BERKELEY.EDU realm, which
is really not a very good idea.  Maybe the README could get installed.

30 years agoSample configuration files for Kerberos. (Not restricted.)
Garrett Wollman [Sun, 27 Feb 1994 21:17:55 +0000 (21:17 +0000)]
Sample configuration files for Kerberos.  (Not restricted.)

30 years agoIf Kerberos is present,use it.
Garrett Wollman [Sun, 27 Feb 1994 21:09:57 +0000 (21:09 +0000)]
If Kerberos is present,use it.

30 years agoIf Kerberos is present, use it.
Garrett Wollman [Sun, 27 Feb 1994 20:59:49 +0000 (20:59 +0000)]
If Kerberos is present, use it.

30 years agoIf Kerberos support is present, use it.
Garrett Wollman [Sun, 27 Feb 1994 20:52:11 +0000 (20:52 +0000)]
If Kerberos support is present, use it.

30 years agoIf Kerberos is present, use it.
Garrett Wollman [Sun, 27 Feb 1994 20:37:04 +0000 (20:37 +0000)]
If Kerberos is present, use it.

30 years agoIf Kerberos is installed, use it.
Garrett Wollman [Sun, 27 Feb 1994 20:27:44 +0000 (20:27 +0000)]
If Kerberos is installed, use it.

30 years agoIf compiling with DESTDIR set (i.e., into another tree), get includes and
Garrett Wollman [Sun, 27 Feb 1994 20:24:27 +0000 (20:24 +0000)]
If compiling with DESTDIR set (i.e., into another tree), get includes and
libraries from that tree in preference to the usual locations.

30 years agoSince we now support Kerberos v4, make those ports the default
Garrett Wollman [Sun, 27 Feb 1994 19:45:37 +0000 (19:45 +0000)]
Since we now support Kerberos v4, make those ports the default
rather than the official ones for v5.

30 years agoIf Kerberos is installed and NOCRYPT is not set, use the Kerberized version.
Garrett Wollman [Sun, 27 Feb 1994 19:39:16 +0000 (19:39 +0000)]
If Kerberos is installed and NOCRYPT is not set, use the Kerberized version.

30 years agoTo: current-users@sun-lamp.cs.berkeley.edu
Nate Willams [Sun, 27 Feb 1994 19:28:44 +0000 (19:28 +0000)]
To: current-users@sun-lamp.cs.berkeley.edu
Subject: bug in bsd.dep.mk
Date: Sun, 27 Feb 1994 10:59:28 CST
From: "Michael L. VanLoon -- Iowa State University" <michaelv@iastate.edu>

The bsd.README file in /usr/share/mk says that make clean isn't
supposed to remove .depend files, but make cleandir is.  However, it
hasn't worked this way for a long time.

[ Fix deleted ]

30 years agoUpdated to reflect FreeBSD 1.1.
Garrett Wollman [Sun, 27 Feb 1994 18:57:47 +0000 (18:57 +0000)]
Updated to reflect FreeBSD 1.1.

30 years agorestore the possibility to "undef" chars. Who has made the reverse patch
Andreas Schulz [Sun, 27 Feb 1994 18:54:30 +0000 (18:54 +0000)]
restore the possibility to "undef" chars. Who has made the reverse patch
the last time :-). Look into the CVS logs.

30 years agoAdd goo to install target to manufacture directory in case it does not
Garrett Wollman [Sun, 27 Feb 1994 18:04:29 +0000 (18:04 +0000)]
Add goo to install target to manufacture directory in case it does not
already exist.

30 years agoAdded tickadj to start of xntpd, this improves performance to nominal.
Poul-Henning Kamp [Sun, 27 Feb 1994 15:32:09 +0000 (15:32 +0000)]
Added tickadj to start of xntpd, this improves performance to nominal.

30 years agosubr_prf.c is fixed in -current so it have been removed from the patch.
Poul-Henning Kamp [Sun, 27 Feb 1994 15:24:07 +0000 (15:24 +0000)]
subr_prf.c is fixed in -current so it have been removed from the patch.

30 years agoA litte trouble with declarations of boot() and panic() resolved. gcc-2.5.8
Poul-Henning Kamp [Sun, 27 Feb 1994 15:07:20 +0000 (15:07 +0000)]
A litte trouble with declarations of boot() and panic() resolved. gcc-2.5.8
uncovered this.

30 years agoThis commit was generated by cvs2svn to track changes on a CVS vendor
Poul-Henning Kamp [Sun, 27 Feb 1994 14:23:52 +0000 (14:23 +0000)]
This commit was generated by cvs2svn to track changes on a CVS vendor
branch.

30 years agoImported the means for migrating to gcc-2.5.8 when we feel the urge and
Poul-Henning Kamp [Sun, 27 Feb 1994 14:23:52 +0000 (14:23 +0000)]
Imported the means for migrating to gcc-2.5.8 when we feel the urge and
need for it.  Read the README for information.  This should *NOT* be
included in the Makefile above.

30 years agoPrevious version of passwd dbase works well for small number of users (<500).
Andrey Chernov [Sat, 26 Feb 1994 23:34:19 +0000 (23:34 +0000)]
Previous version of passwd dbase works well for small number of users (<500).
On our site we have > 2500 logins for uucp hosts and in this case size
of pwd.db + spwd.db > 2.5 M.
pwd_mkdb ("passwd user" issued f.e.)
works ~20 min to rebuild passwd db and require > 5 M on root fs.

Passwd changing (system administrator make > 20 new accouns per day) becomes
a big problem and requires too many resources.

With modified passwd db structure on our site we have sizes of pwd.db +
+ spwd.db == ~400k and rebuilding time is ~2.5 min now!

We found that work time for libc functions with old and new passwd structure
almost equals.

      Alex Bulushev, bag@demos.su

30 years agoI have made the stuffm, which was pulled into libc for the sake of gcc1,
Poul-Henning Kamp [Sat, 26 Feb 1994 23:24:04 +0000 (23:24 +0000)]
I have made the stuffm, which was pulled into libc for the sake of gcc1,
conditional.  You will have to define GCC1_IN_LIBC to get it in the future.
If this is needed it should be removed to another library, with this specific
contents&purpose, if noone needs it we should nuke it entirely.  Make World
misses nothing on my system.

30 years agoAutomatically build Kerberos stuff, if it is present and NOCRYPT is not
Garrett Wollman [Sat, 26 Feb 1994 19:32:43 +0000 (19:32 +0000)]
Automatically build Kerberos stuff, if it is present and NOCRYPT is not
defined.  Someone sufficiently Rod-like will have to frob etc/Makefile to get
the right tarballs to appear in the right places and make sure that the
Kerberos code doesn't get shipped to the wrong places.

30 years agoAdd LBL's kernel random-number generator. Nothing currently uses this,
Garrett Wollman [Sat, 26 Feb 1994 19:03:32 +0000 (19:03 +0000)]
Add LBL's kernel random-number generator.  Nothing currently uses this,
so it isn't even compiled in, but I expect that some things will want to
use it in the mid-term future.

30 years agoAdded support for "Conrad DCF-77 mobil" radio clock. This is the cheapest
Poul-Henning Kamp [Sat, 26 Feb 1994 12:32:19 +0000 (12:32 +0000)]
Added support for "Conrad DCF-77 mobil" radio clock.  This is the cheapest
(DM130) way to get precise ticks if you are within range of the xmitter in
Mainflingen.  I'm some 1000Km away and have +/- 2ms clock using this.

30 years agoPatch from John Dyson to fix a bug where a panic will occur if a process
David Greenman [Sat, 26 Feb 1994 07:17:53 +0000 (07:17 +0000)]
Patch from John Dyson to fix a bug where a panic will occur if a process
is being traced and some of its page tables get paged out. The problem
was reported by Adam David.

30 years agoBump CLK_TCK to more precise value (128)
Andrey Chernov [Sat, 26 Feb 1994 00:56:02 +0000 (00:56 +0000)]
Bump CLK_TCK to more precise value (128)
If you want more precise, use directly getrusage(),
because clock() emulated via it.

30 years agoRTS shouldn't be touched unless CS_RTS_IFLOW is set.
Poul-Henning Kamp [Sat, 26 Feb 1994 00:04:03 +0000 (00:04 +0000)]
RTS shouldn't be touched unless CS_RTS_IFLOW is set.

30 years agovalues.h added for applications (from g++ include)
Andrey Chernov [Fri, 25 Feb 1994 23:45:20 +0000 (23:45 +0000)]
values.h added for applications (from g++ include)

30 years agoextract() is more friendly
Andrew Moore [Fri, 25 Feb 1994 23:34:14 +0000 (23:34 +0000)]
extract() is more friendly
added load_dos (just calls load_fd)

30 years agoBetter fix for >16 heads problem, right value from
Andrey Chernov [Fri, 25 Feb 1994 23:17:40 +0000 (23:17 +0000)]
Better fix for >16 heads problem, right value from
controller restored.

30 years ago"date | wc -c" didn't work.
Poul-Henning Kamp [Fri, 25 Feb 1994 22:24:42 +0000 (22:24 +0000)]
"date | wc -c" didn't work.

30 years agoOne instance of ${CC} hadn't got ${CFLAGS}. No impact for normal environments.
Poul-Henning Kamp [Fri, 25 Feb 1994 21:50:29 +0000 (21:50 +0000)]
One instance of ${CC} hadn't got ${CFLAGS}.  No impact for normal environments.

30 years agoadded the -C/-check again.
Poul-Henning Kamp [Fri, 25 Feb 1994 21:46:07 +0000 (21:46 +0000)]
added the -C/-check again.

30 years ago>From: Andreas Schulz <ats@g386bsd.first.gmd.de>
Rod Grimes [Fri, 25 Feb 1994 14:11:16 +0000 (14:11 +0000)]
>From: Andreas Schulz <ats@g386bsd.first.gmd.de>
Subject: failure in /usr/src/etc/group

The /usr/src/etc/group file is missing a colon in the line
"dialer:*:117" at the end.

30 years agoKerberos version 4 from the Net/2 tape.
Garrett Wollman [Fri, 25 Feb 1994 01:14:51 +0000 (01:14 +0000)]
Kerberos version 4 from the Net/2 tape.
WARNING: Contains export-restricted software.

30 years agoThis commit was generated by cvs2svn to track changes on a CVS vendor
Garrett Wollman [Fri, 25 Feb 1994 01:14:51 +0000 (01:14 +0000)]
This commit was generated by cvs2svn to track changes on a CVS vendor
branch.

30 years agoTIOCTIMESTAMP capability. The port in question must be configured with
Poul-Henning Kamp [Thu, 24 Feb 1994 16:39:48 +0000 (16:39 +0000)]
TIOCTIMESTAMP capability.  The port in question must be configured with
"vector siointrts".  Thus only a compare is added for the other ports.

30 years agoRoot must be able to do TIOCSPGRP on non-ctty-tty's or xntpd cannot see it's
Poul-Henning Kamp [Thu, 24 Feb 1994 16:13:09 +0000 (16:13 +0000)]
Root must be able to do TIOCSPGRP on non-ctty-tty's or xntpd cannot see it's
clock(s).

30 years agoI have added
Poul-Henning Kamp [Thu, 24 Feb 1994 16:09:09 +0000 (16:09 +0000)]
I have added
TIOCTIMESTAMP _IOR('t', 89, struct timeval)
This is for precision timekeeping using xntpd and similar stuff.  The semantics
of this call are: return a timestamp, taken as soon as possible after the
last interrupt on the device.  Patch to sio.c will be forthcomming.

30 years agoUse double precision floating point in calculation of total, used, avail
Steven Wallace [Thu, 24 Feb 1994 10:11:06 +0000 (10:11 +0000)]
Use double precision floating point in calculation of total, used, avail
so that FS > 2 GB report accurate results.

30 years agoAdd Jeff Hsu and Cury Mayer to list of contrib folks..
Jordan K. Hubbard [Thu, 24 Feb 1994 06:49:44 +0000 (06:49 +0000)]
Add Jeff Hsu and Cury Mayer to list of contrib folks..

30 years agoCorrect definitions of flags used by sigreturn to validate sigcontext.
Jeffrey Hsu [Thu, 24 Feb 1994 00:21:12 +0000 (00:21 +0000)]
Correct definitions of flags used by sigreturn to validate sigcontext.

30 years agovalidate sigcontext before restoring it
Jeffrey Hsu [Thu, 24 Feb 1994 00:18:04 +0000 (00:18 +0000)]
validate sigcontext before restoring it

30 years agoIt is not nice to extract libc.so.1.0 using tar when tar is using it!!
Rod Grimes [Wed, 23 Feb 1994 16:47:04 +0000 (16:47 +0000)]
It is not nice to extract libc.so.1.0 using tar when tar is using it!!
UGLY HACK TO GET 1.1 BETA OUT :-(

30 years agoAdd diclaimer of copyright to _osname() manual page.
Garrett Wollman [Wed, 23 Feb 1994 15:55:45 +0000 (15:55 +0000)]
Add diclaimer of copyright to _osname() manual page.

30 years agoMake it say Preparing for dos fs instead of Initializing.. the later sounds
Rod Grimes [Wed, 23 Feb 1994 15:43:20 +0000 (15:43 +0000)]
Make it say Preparing for dos fs instead of Initializing.. the later sounds
scary to some people!

30 years agoUpdate to generate the ISO9660 names for the floppy images, use fd1200
Rod Grimes [Wed, 23 Feb 1994 15:36:05 +0000 (15:36 +0000)]
Update to generate the ISO9660 names for the floppy images, use fd1200
inplace of floppy5.

30 years ago>From: batie@agora.rain.com (Alan Batie)
Rod Grimes [Wed, 23 Feb 1994 12:06:09 +0000 (12:06 +0000)]
>From: batie@agora.rain.com (Alan Batie)
Subject: bug in elvis man page

The elvis man page (and "vi -r") says to use "elvrec" to recover a file
that was being edited when you got dumped off the system or it crashed.
The actual command is "elvisrecover", at least as installed on FreeBSD.

Also, there needs to be a man page link to vi.1 so you can say "man vi".
[editors note, link not done, we do have a real vi now...]

30 years agoFix off by one error
Rod Grimes [Wed, 23 Feb 1994 11:14:26 +0000 (11:14 +0000)]
Fix off by one error

30 years agoPut back the bt0 that some one deleted....
Rod Grimes [Wed, 23 Feb 1994 10:59:44 +0000 (10:59 +0000)]
Put back the bt0 that some one deleted....

30 years agoMain branch is not at version:
Rod Grimes [Wed, 23 Feb 1994 10:02:35 +0000 (10:02 +0000)]
Main branch is not at version:
kernvers="FreeBSD 1.1.0(Current)"

About to tag BETA_0_1_1_0 :-)

30 years agoAdd $Id$ to everything, fix one compiler warning about multiple #defines
Rod Grimes [Wed, 23 Feb 1994 09:56:56 +0000 (09:56 +0000)]
Add $Id$ to everything, fix one compiler warning about multiple #defines
by nuking the local copy and using the one in machine/param.h!  Note I
know about the copyright on osname.3 and have informed Garratt to please
correct it.

>From: conklin@ngai.kaleida.com (J.T. Conklin)

In pwcashe.c, the "name" entry in the group cashe was allocated as follows:
        char name[UT_NAMESIZE];

instead of:
char name[UT_NAMESIZE + 1];

This only causes problems when a name hashes to a value following a
hash table entry containing a 8 character group name.  When the value
is stored, it overwrites the trailing zero of the previous name (since
c_gid[n].name[UT_NAMESIZE] overlaps the same location in memory as
c_gid[n+1].uid).

This causes some wierd output in ls(1), given the right circumstances.

30 years ago>From: Andreas Schulz <ats@g386bsd.first.gmd.de>
Rod Grimes [Tue, 22 Feb 1994 23:34:30 +0000 (23:34 +0000)]
>From: Andreas Schulz <ats@g386bsd.first.gmd.de>
Subject: missing man page link
Date: Wed, 23 Feb 1994 00:35:46 +0100 (MET)

A manual page link is missing for "cc.1". We have only "gcc.1".
The Makefile in /usr/src/gnu/usr.bin/cc/cc needs a
"MLINKS += gcc.1 cc.1".

30 years agoBump boot block revision to 1.9 by a dummy edit on disk.c and:
Rod Grimes [Tue, 22 Feb 1994 22:59:40 +0000 (22:59 +0000)]
Bump boot block revision to 1.9 by a dummy edit on disk.c and:

>From: bde@kralizec.zeta.org.au (Bruce Evans)
Subject: cache botch in bootstrap
Date: Sun, 20 Feb 94 18:14:35 +1100
The cache in the bootstrap loader didn't allow for the device changing.

This caused surprisingly little trouble - the cache is only for a single
track (or part of a track), and the first access to a new device is
always for an early sector, so there is no problem unless the last access
to the old device was for an early sector.  I saw the problem on a system
with BSD on wd1 and no label on wd0.  Everything worked if the the device
name wd(1,a) was specified before the default of wd(0,a) was tried, but
when the default was tried, it failed on the first sector because there
was no label, and then the first sector on wd0 was always used instead
of the first sector on wd1, so there was apparently no label on wd1
either.

Bruce

30 years ago/sys/i386/boot
Rod Grimes [Tue, 22 Feb 1994 22:34:52 +0000 (22:34 +0000)]
/sys/i386/boot
You can not have the root file system mounted on /dev/fd1a, it
panics on you.  This occurs when booting from floppy and answering
the question for what drive the root file system is on with B.
This will be fixed by the 1.1 release date.
cVS: ----------------------------------------------------------------------

30 years agoMore clean up of the drive Id string, trancate trailing spaces.
Rod Grimes [Tue, 22 Feb 1994 22:13:37 +0000 (22:13 +0000)]
More clean up of the drive Id string, trancate trailing spaces.

30 years agoReplaced version w/out copyright with something I whipped up based on
Nate Willams [Tue, 22 Feb 1994 21:54:03 +0000 (21:54 +0000)]
Replaced version w/out copyright with something I whipped up based on
hostname(1) and id(1).  This file has a BSD copyright since it's based on
those sources.

30 years agoThe model string can be 40 bytes long with no terminating null, so
Rod Grimes [Tue, 22 Feb 1994 18:51:27 +0000 (18:51 +0000)]
The model string can be 40 bytes long with no terminating null, so
convert all nulls to spaces and then print it with a %40s.

30 years agoAdd Christoph P. Kukulies mirror site at gil.physik.rwth-aachen.de:/pub/FreeBSD
Rod Grimes [Tue, 22 Feb 1994 18:23:27 +0000 (18:23 +0000)]
Add Christoph P. Kukulies mirror site at gil.physik.rwth-aachen.de:/pub/FreeBSD
Add Remy CARD mirror site at ftp.ibp.fr:/pub/FreeBSD.

Clean up the format so we can fit these long names and email address into
80 columns.

30 years agoIMPORTANT:
Rod Grimes [Tue, 22 Feb 1994 17:52:23 +0000 (17:52 +0000)]
IMPORTANT:
/sys/i386/boot
DO NOT ATTEMPT TO REBUILD THE BOOT BLOCKS.  There is currently
a problem in that the boot block code has grown and if you build
it with the new version of as and install them you will NOT BE
ABLE TO BOOT YOUR SYSTEM.  This problem will be fixed before
the 1.1 release is ship, and should only exist with the 1.1 BETA
release.

The boot block binaries shipped with the 1.1 BETA release were
built using an older version of as and can be installed safely.

30 years agoAdd the NE2100, STB 4 port, Boca 6 port. Delete the mirror stuff and
Rod Grimes [Tue, 22 Feb 1994 17:23:48 +0000 (17:23 +0000)]
Add the NE2100, STB 4 port, Boca 6 port.  Delete the mirror stuff and
refer them to the external file.  Some other details cleaned up.

Everyone please read this thing over and lets get it up to date!!!

30 years agoAdd NE2100 Per Paul Richards, and whoopsss.... I did this in /usr/src
Rod Grimes [Tue, 22 Feb 1994 17:11:27 +0000 (17:11 +0000)]
Add NE2100 Per Paul Richards, and whoopsss.... I did this in /usr/src
on the WRONG system (actually, just the wrong window :-)).

30 years agoCleanup the lptprobe and lpt_port_test routine to be KNF.
Rod Grimes [Tue, 22 Feb 1994 09:05:13 +0000 (09:05 +0000)]
Cleanup the lptprobe and lpt_port_test routine to be KNF.

30 years agoPull in some of the lpt_port_test fixes from lpt.c.
Rod Grimes [Tue, 22 Feb 1994 09:04:08 +0000 (09:04 +0000)]
Pull in some of the lpt_port_test fixes from lpt.c.