]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
26 years agoupdate dc utility as well
Andreas Klemm [Wed, 29 Apr 1998 22:15:05 +0000 (22:15 +0000)]
update dc utility as well
after importing the new bc-1.0.4 sources,
which contain the latest DNU dc version

26 years agoremove old bc-1.0.3 stuff
Andreas Klemm [Wed, 29 Apr 1998 22:05:39 +0000 (22:05 +0000)]
remove old bc-1.0.3 stuff
modify Makefile, to get bc-1.0.4 from src/contrib/bc
adjust config.h
PR: closes 4183

26 years agoMake 'y' command 8bit clean
Andrey A. Chernov [Wed, 29 Apr 1998 21:58:36 +0000 (21:58 +0000)]
Make 'y' command 8bit clean
PR: 6458

26 years agoAdded short instruction for FreeBSD how to import
Andreas Klemm [Wed, 29 Apr 1998 21:57:35 +0000 (21:57 +0000)]
Added short instruction for FreeBSD how to import

26 years agoThis commit was generated by cvs2svn to compensate for changes in r35516,
Andreas Klemm [Wed, 29 Apr 1998 21:53:01 +0000 (21:53 +0000)]
This commit was generated by cvs2svn to compensate for changes in r35516,
which included commits to RCS files with non-trunk default branches.

26 years agoImport GNU bc 1.04
Andreas Klemm [Wed, 29 Apr 1998 21:53:01 +0000 (21:53 +0000)]
Import GNU bc 1.04
PR: 4183

26 years agoAdd note about BLOCKSIZE variable to -s flag.
Warner Losh [Wed, 29 Apr 1998 18:39:51 +0000 (18:39 +0000)]
Add note about BLOCKSIZE variable to -s flag.

26 years agoUse ${.TARGET} rather than $@. i386 tested, pc98 untested because config
Warner Losh [Wed, 29 Apr 1998 18:21:35 +0000 (18:21 +0000)]
Use ${.TARGET} rather than $@.  i386 tested, pc98 untested because config
on my box doesn't grok machine type pc98.
PR: 3272
Submitted by: jhs

26 years agoFixes incorrect company and product names in uha(4) manpage and
Andreas Klemm [Wed, 29 Apr 1998 17:09:41 +0000 (17:09 +0000)]
Fixes incorrect company and product names in uha(4) manpage and
LINT config file. Should be merged to -STABLE as well.
PR: closes 6447
Submitted by: Steven Plite splite@purdue.edu

26 years agoUse DFLTBSIZE instead of MAXBSIZE for pm_fatblksize.
Dmitrij Tejblum [Wed, 29 Apr 1998 12:55:51 +0000 (12:55 +0000)]
Use DFLTBSIZE instead of MAXBSIZE for pm_fatblksize.

In msdosfs_sync: spelling fix, formatting changes; fix MNT_LAZY (sync
modified denodes, don't sync device)

Mostly submitted by (and with hints from): bde

Increase limit for maximum disk size: as far as I can see previous limit was
gratuitously too low.

26 years agoAdd spinlock.
John Birrell [Wed, 29 Apr 1998 11:03:34 +0000 (11:03 +0000)]
Add spinlock.

26 years agoChange signal model to match POSIX (i.e. one set of signal handlers
John Birrell [Wed, 29 Apr 1998 09:59:34 +0000 (09:59 +0000)]
Change signal model to match POSIX (i.e. one set of signal handlers
for the process, not a separate set for each thread). By default, the
process now only has signal handlers installed for SIGVTALRM, SIGINFO
and SIGCHLD. The thread kernel signal handler is installed for other
signals on demand. This means that SIG_IGN and SIG_DFL processing is now
left to the kernel, not the thread kernel.

Change the signal dispatch to no longer use a signal thread, and
call the signal handler using the stack of the thread that has the
signal pending.

Change the atomic lock method to use test-and-set asm code with
a yield if blocked. This introduces separate locks for each type
of object instead of blocking signals to prevent a context
switch. It was this blocking of signals that caused the performance
degradation the people have noted.

This is a *big* change!

26 years agoSync with sys/i386/i386/trap.c revision 1.127.
KATO Takenori [Wed, 29 Apr 1998 09:45:38 +0000 (09:45 +0000)]
Sync with sys/i386/i386/trap.c revision 1.127.

26 years agoDon't need wrappers for longjmp/setjmp anymore.
John Birrell [Wed, 29 Apr 1998 09:40:51 +0000 (09:40 +0000)]
Don't need wrappers for longjmp/setjmp anymore.

26 years agoRemove empty files that were renamed some time ago.
John Birrell [Wed, 29 Apr 1998 09:39:34 +0000 (09:39 +0000)]
Remove empty files that were renamed some time ago.

26 years agoAtomic lock source.
John Birrell [Wed, 29 Apr 1998 09:36:03 +0000 (09:36 +0000)]
Atomic lock source.

26 years agoStop renaming these in libc_r because wrappered versions don't make sense.
John Birrell [Wed, 29 Apr 1998 09:14:35 +0000 (09:14 +0000)]
Stop renaming these in libc_r because wrappered versions don't make sense.
PR: i386/4826, bin/5953

26 years agoReplace the threaded locking with spinlock calls for both threaded
John Birrell [Wed, 29 Apr 1998 09:10:58 +0000 (09:10 +0000)]
Replace the threaded locking with spinlock calls for both threaded
and non-threaded programs. This makes malloc thread safe for linking
with libpthread and kernel threads.

Reviewed by: phk

26 years agoReference an external variable in threaded programs so that the
John Birrell [Wed, 29 Apr 1998 09:08:43 +0000 (09:08 +0000)]
Reference an external variable in threaded programs so that the
autoinitialiser gets linked in and therefore called before main().

26 years agoUse signal() in both the threaded and non-threaded cases.
John Birrell [Wed, 29 Apr 1998 09:06:13 +0000 (09:06 +0000)]
Use signal() in both the threaded and non-threaded cases.

26 years agoStubs are required in libc so that it can be used with libpthread
John Birrell [Wed, 29 Apr 1998 09:02:16 +0000 (09:02 +0000)]
Stubs are required in libc so that it can be used with libpthread
(and kernel threads), but weak symbols and non-weak symbols of the
same name built into libc_r result in unpredictable linking.

26 years agoAdd a needed prototype, and fix a panic problem with the new
John Dyson [Wed, 29 Apr 1998 06:59:08 +0000 (06:59 +0000)]
Add a needed prototype, and fix a panic problem with the new
memory code.

26 years agoAdd "flags 0x10" to the sio0 entry so that SMP-GENERIC acts in thesame
Masafumi Max NAKANE [Wed, 29 Apr 1998 06:58:43 +0000 (06:58 +0000)]
Add "flags 0x10" to the sio0 entry so that SMP-GENERIC acts in thesame
manner as the GENERIC kernel to the -h option.

26 years agoTighten up management of memory and swap space during map allocation,
John Dyson [Wed, 29 Apr 1998 04:28:22 +0000 (04:28 +0000)]
Tighten up management of memory and swap space during map allocation,
deallocation cycles.  This should provide a measurable improvement
on swap and memory allocation on loaded systems.  It is unlikely a
complete solution.  Also, provide more map info with procfs.
Chuck Cranor spurred on this improvement.

26 years agoTranslate T_PROTFLT to SIGSEGV instead of SIGBUS when running under
Eivind Eklund [Tue, 28 Apr 1998 18:15:08 +0000 (18:15 +0000)]
Translate T_PROTFLT to SIGSEGV instead of SIGBUS when running under
Linux emulation.  This make Allegro Common Lisp 4.3 work under
FreeBSD!

Submitted by: Fred Gilham <gilham@csl.sri.com>
Commented on by: bde, dg, msmith, tg
Hoping he got everything right:  eivind

26 years agoFix merge errors
Andrey A. Chernov [Tue, 28 Apr 1998 07:37:54 +0000 (07:37 +0000)]
Fix merge errors

26 years agoUpgrade to 2.2
Andrey A. Chernov [Tue, 28 Apr 1998 07:16:24 +0000 (07:16 +0000)]
Upgrade to 2.2

26 years agoOops, revert part of a diff that wasn't supposed to have been committed.
David Greenman [Tue, 28 Apr 1998 07:02:33 +0000 (07:02 +0000)]
Oops, revert part of a diff that wasn't supposed to have been committed.

26 years agoCache the results of the ps_strings sysctl so that it doesn't have to be
David Greenman [Tue, 28 Apr 1998 06:59:14 +0000 (06:59 +0000)]
Cache the results of the ps_strings sysctl so that it doesn't have to be
redone for every call of setproctitle().

26 years agoMerge local changes
Andrey A. Chernov [Tue, 28 Apr 1998 06:49:42 +0000 (06:49 +0000)]
Merge local changes

26 years agoThis commit was generated by cvs2svn to compensate for changes in r35486,
Andrey A. Chernov [Tue, 28 Apr 1998 06:22:20 +0000 (06:22 +0000)]
This commit was generated by cvs2svn to compensate for changes in r35486,
which included commits to RCS files with non-trunk default branches.

26 years agoV2.2 import
Andrey A. Chernov [Tue, 28 Apr 1998 06:22:20 +0000 (06:22 +0000)]
V2.2 import

26 years agoFix a pseudo-swap leak problem. This mitigates "leaks" due to
John Dyson [Tue, 28 Apr 1998 05:54:47 +0000 (05:54 +0000)]
Fix a pseudo-swap leak problem.  This mitigates "leaks" due to
freeing partial objects, not freeing entire objects didn't
free any of it.  Simple fix to the map code.
Reviewed by: dg

26 years agoFixed syntax error for `make checkdpadd'. libfl.a should never be
Bruce Evans [Tue, 28 Apr 1998 05:31:41 +0000 (05:31 +0000)]
Fixed syntax error for `make checkdpadd'.  libfl.a should never be
used.  ${LIBFL} is set to a weird value in an attempt to inhibit
its use, but only breaks properly in some contexts.

Fixed the usual style bugs for DPADD and LDADD (disorder, and += for the
initial assignment).

26 years agoUse historically correct error message in some cases, optionally.
Warner Losh [Tue, 28 Apr 1998 05:08:10 +0000 (05:08 +0000)]
Use historically correct error message in some cases, optionally.

26 years agoSet TCP_NODELAY on the control channel to improve performance a bit.
David Greenman [Tue, 28 Apr 1998 03:37:23 +0000 (03:37 +0000)]
Set TCP_NODELAY on the control channel to improve performance a bit.

26 years agoSubmitted by: Doug Ambrisko <ambrisko@whistle.com>
Julian Elischer [Tue, 28 Apr 1998 00:10:53 +0000 (00:10 +0000)]
Submitted by: Doug Ambrisko <ambrisko@whistle.com>

make standard links for the first detected sound card.

26 years agoDon't build dependencies unless NOCLEAN is set. This speeds up
Bruce Evans [Mon, 27 Apr 1998 16:21:33 +0000 (16:21 +0000)]
Don't build dependencies unless NOCLEAN is set.  This speeds up
`make world' by about 14% here (down to 4490 seconds real on a
K6/233).  Temporarily skip this optimization when building with
-j, since there are still many broken makefiles.

Fixed NOCLEANDIR option.  Cleaning of `.depend' was broken.

Put -nostdinc in CFLAGS, not in CC, and don't override the default
CC.  This fixes enforcing use of ${WORLDTMP}/usr/include.

Don't install library man pages in ${WORLDTMP}.

26 years agoDon't include ../Makefile.inc here. bsd.man.mk is always included
Bruce Evans [Mon, 27 Apr 1998 15:08:07 +0000 (15:08 +0000)]
Don't include ../Makefile.inc here.  bsd.man.mk is always included
in another .mk file which has already included ../Makefile.inc, so
including ../Makefike.inc again mainly messed up `+=' definitions.

26 years agoCast return values to the appropriate fp_*_t. Note that the man page
Dag-Erling Smørgrav [Mon, 27 Apr 1998 14:17:30 +0000 (14:17 +0000)]
Cast return values to the appropriate fp_*_t. Note that the man page
incorrectly refers to them as e.g. fp_except rather than fp_except_t.

PR: misc/6310
Submitted by: Niall Smart

26 years agoFixed spelling of "fictitious" in diagnostics and comments
Dag-Erling Smørgrav [Mon, 27 Apr 1998 11:36:06 +0000 (11:36 +0000)]
Fixed spelling of "fictitious" in diagnostics and comments
PR: kern/6429

26 years agoFixed a bug where if MAXUSRARGS amount of args were passed in, the argv[]
David Greenman [Mon, 27 Apr 1998 10:51:26 +0000 (10:51 +0000)]
Fixed a bug where if MAXUSRARGS amount of args were passed in, the argv[]
array would end up without the NULL pointer termination, causing the glob
code to glob whatever garbage happend to follow on the stack.

26 years ago*** empty log message ***
Peter Wemm [Mon, 27 Apr 1998 04:26:56 +0000 (04:26 +0000)]
*** empty log message ***

26 years agoSupport more than 256 tun devices:
Brian Somers [Mon, 27 Apr 1998 01:44:22 +0000 (01:44 +0000)]
Support more than 256 tun devices:

$ ls -l /dev/tun25[4-7]
crw-------  1 fax  dialer   52, 254 Apr 27 02:27 /dev/tun254
crw-------  1 fax  dialer   52, 255 Apr 27 02:27 /dev/tun255
crw-------  1 fax  dialer   52, 0x00010000 Apr 27 02:31 /dev/tun256
crw-------  1 fax  dialer   52, 0x00010001 Apr 27 02:31 /dev/tun257

26 years agoCheck that user really typed "who am {I,i}", and that no stray
Dag-Erling Smørgrav [Sun, 26 Apr 1998 19:10:51 +0000 (19:10 +0000)]
Check that user really typed "who am {I,i}", and that no stray
arguments are given.
Note that usage() and the manpage disagree...

PR: bin/6294
Suggested by: Ruslan Ermilov and Bruce Evans
Submitted by: Ruslan Ermilov (partly)

26 years agosay a few words about the -b option.
Poul-Henning Kamp [Sun, 26 Apr 1998 18:15:38 +0000 (18:15 +0000)]
say a few words about the -b option.

Reviewed by: Bill Trost <trost@cloud.rain.com>

26 years agoAdd warning about root-fs blocksize expectations.
Poul-Henning Kamp [Sun, 26 Apr 1998 17:44:23 +0000 (17:44 +0000)]
Add warning about root-fs blocksize expectations.
PR: 4485
Reviewed by: phk
Submitted by: Kees Jan Koster <kjk1@ukc.ac.uk>

26 years agoAdd pa=none to give a hint that they exist.
Poul-Henning Kamp [Sun, 26 Apr 1998 17:28:26 +0000 (17:28 +0000)]
Add pa=none to give a hint that they exist.
PR: 2368 (somewhat)

26 years agoFixed `make -j9' by putting generated headers in ${SRCS}. Removed bogus
Bruce Evans [Sun, 26 Apr 1998 16:31:14 +0000 (16:31 +0000)]
Fixed `make -j9' by putting generated headers in ${SRCS}.  Removed bogus
dependencies of generated .c files on generated headers (these made
`make -JN' work provided `depend' was made first).  Sorted sources lists.

26 years agoBacked out most of rev.1.19 (explicit dependencies of object files
Bruce Evans [Sun, 26 Apr 1998 16:12:23 +0000 (16:12 +0000)]
Backed out most of rev.1.19 (explicit dependencies of object files
on generated headers).  This is now handled generally in bsd.prog.mk.

26 years agoWhen all transformation rules to or from a suffix disappeared, make tries
Poul-Henning Kamp [Sun, 26 Apr 1998 09:44:48 +0000 (09:44 +0000)]
When all transformation rules to or from a suffix disappeared, make tries
to free the suffix. I think, it is a very strange idea. (Or, maybe, it is a
POSIX requirement?) And it is done incorrectly. Apparently, it even don't
update the list of known suffixes (but it is an other bug).

PR: 4254, 4692, 4783
Reviewed by: phk
Submitted by: Dmitrij Tejblum <dima@tejblum.dnttm.rssi.ru>

26 years agoUse more reasonable defaults for SUP variables (and print them
Jordan K. Hubbard [Sun, 26 Apr 1998 08:40:25 +0000 (08:40 +0000)]
Use more reasonable defaults for SUP variables (and print them
correctly) now that sup is long since dead and buried.

26 years agoAdd the whatis database to catpages too
Jordan K. Hubbard [Sun, 26 Apr 1998 08:34:26 +0000 (08:34 +0000)]
Add the whatis database to catpages too

26 years agotelnetd does not recognize the if (display file before login) gettytab flag.
Poul-Henning Kamp [Sun, 26 Apr 1998 06:51:36 +0000 (06:51 +0000)]
telnetd does not recognize the if (display file before login) gettytab flag.

PR: 6365
Reviewed by: phk
Submitted by: Jason Garman <init@risen.org>

26 years agoJean-Simon Pendry's paper on amd refers to the use of "ypcat -k"
Poul-Henning Kamp [Sun, 26 Apr 1998 06:32:13 +0000 (06:32 +0000)]
Jean-Simon Pendry's paper on  amd refers to the use of "ypcat -k"
against the "master map" to get the list of mount point/amd map
correspondences, and using that list as command-line arguments to start
amd.

When I tried to do this with the existing /etc/rc* scripts, I found that
I couldn't do this by modifying only /etc/rc.conf:  that file gets
sourced very early by /etc/rc, well before any networking functionality
is present, let alone NIS.  Further, I wasn't able to figure out a way
to use various levels & types of quoting to defer evaluation of the
string to a point subsequent to NIS initialization.

As a result, I resorted to hacking /etc/rc.network -- but I did it in a
way that ought to be reasonably general, and avoid breakage for anyone
else.

PR: 6387
Reviewed by: phk
Submitted by: David Wolfskill <dhw@whistle.com>

26 years agofix ncal.1 grammar
Poul-Henning Kamp [Sun, 26 Apr 1998 06:21:29 +0000 (06:21 +0000)]
fix ncal.1 grammar

PR: 6394
Reviewed by: phk
Submitted by: Josh Gilliam <josh@quick.net>

26 years ago`sigprocmask()' man page references `sigmask()' in synopsis.
Poul-Henning Kamp [Sun, 26 Apr 1998 06:19:24 +0000 (06:19 +0000)]
`sigprocmask()' man page references `sigmask()' in synopsis.
Note odd `sigmask()' line in synopsis.  `sigsetops(3)' is better suited
for `sigprocmask' and is already referenced from the manual page.
(`sigmask()' is useful for the older (& deprecated) `sigsetmask()' API).
PR: 6395
Reviewed by: phk
Submitted by: Joseph Koshy <koshy@india.hp.com>

26 years agoAdd the PAT cpuid feature.
John Dyson [Sun, 26 Apr 1998 03:18:38 +0000 (03:18 +0000)]
Add the PAT cpuid feature.

26 years agoDon't clobber errno when reaping child processes.
Tor Egge [Sun, 26 Apr 1998 01:00:00 +0000 (01:00 +0000)]
Don't clobber errno when reaping child processes.

26 years agoKeywords "iosize" and "memsize" are unused since 1.5. Removed.
Tatsumi Hosokawa [Sat, 25 Apr 1998 18:10:10 +0000 (18:10 +0000)]
Keywords "iosize" and "memsize" are unused since 1.5.  Removed.

26 years ago#define'd KWD_XXX for keywords.
Tatsumi Hosokawa [Sat, 25 Apr 1998 17:52:15 +0000 (17:52 +0000)]
#define'd KWD_XXX for keywords.

26 years agomake includes has to execute "make beforeinstall" in src/lib/libcalendar
Andreas Klemm [Sat, 25 Apr 1998 14:32:22 +0000 (14:32 +0000)]
make includes has to execute "make beforeinstall" in src/lib/libcalendar
as well, otherwise "make depend" during "make buildworld" fails in
usr.bin/ncal

26 years agoQuieten gcc-2.8.1
Brian Somers [Sat, 25 Apr 1998 09:16:23 +0000 (09:16 +0000)]
Quieten gcc-2.8.1

26 years agoSync with sys/i386/isa/wd.c revision up to 1.163.
KATO Takenori [Sat, 25 Apr 1998 05:19:33 +0000 (05:19 +0000)]
Sync with sys/i386/isa/wd.c revision up to 1.163.

26 years agoCorrect copyright.
John Dyson [Sat, 25 Apr 1998 04:50:03 +0000 (04:50 +0000)]
Correct copyright.

26 years agoSync with sys/i386/isa/wd.c revision 1.161.
KATO Takenori [Sat, 25 Apr 1998 04:27:54 +0000 (04:27 +0000)]
Sync with sys/i386/isa/wd.c revision 1.161.

26 years agoSync with sys/i386/conf/GENERIC revision 1.108.
KATO Takenori [Sat, 25 Apr 1998 04:27:15 +0000 (04:27 +0000)]
Sync with sys/i386/conf/GENERIC revision 1.108.

26 years agoStrengthen the rules governing the 127.0.0.0/8 subnet. The previous rules
Alexander Langer [Sat, 25 Apr 1998 00:40:55 +0000 (00:40 +0000)]
Strengthen the rules governing the 127.0.0.0/8 subnet.  The previous rules
allowed external hosts to send packets to the 127.0.0.0/8 subnet on the
firewall host.

Renumber the lo0 rules to guarantee they appear first.

PR: 6406
Submitted by: Archie Cobbs <archie@whistle.com>

26 years agoFix diagnostic typo.
Brian Somers [Sat, 25 Apr 1998 00:20:23 +0000 (00:20 +0000)]
Fix diagnostic typo.

26 years agoFix nasty error in len_octal, result not initialized
Andrey A. Chernov [Sat, 25 Apr 1998 00:12:32 +0000 (00:12 +0000)]
Fix nasty error in len_octal, result not initialized

26 years agoFix ctype error #1 - chars must be unsigned
Andrey A. Chernov [Sat, 25 Apr 1998 00:10:24 +0000 (00:10 +0000)]
Fix ctype error #1 - chars must be unsigned

26 years agoMake asctime_r static if !_THREAD_SAFE to prevent namespace pollution and
Andrey A. Chernov [Sat, 25 Apr 1998 00:00:57 +0000 (00:00 +0000)]
Make asctime_r static if !_THREAD_SAFE to prevent namespace pollution and
prototype mismatch

26 years agoRemove the bogus -? option after reading the getopt() manual :)
Dag-Erling Smørgrav [Fri, 24 Apr 1998 20:15:43 +0000 (20:15 +0000)]
Remove the bogus -? option after reading the getopt() manual :)

26 years agoOops- osreldate.h is not around like I thought it would be... hmmm.. quick fix to...
Matt Jacob [Fri, 24 Apr 1998 18:23:20 +0000 (18:23 +0000)]
Oops- osreldate.h is not around like I thought it would be... hmmm.. quick fix to get around this

26 years agoAugment the last change after some deliberation with rgrimes & peter.
Poul-Henning Kamp [Fri, 24 Apr 1998 17:32:23 +0000 (17:32 +0000)]
Augment the last change after some deliberation with rgrimes & peter.
It is important that we keep the ability to send packets to a remote
server and that the packets come from our well-known port, also in
that case.

Reviewed by: peter, rgrimes.

26 years agoDon't build shared libraries twice for `make world'. This reduces
Bruce Evans [Fri, 24 Apr 1998 17:13:33 +0000 (17:13 +0000)]
Don't build shared libraries twice for `make world'.  This reduces
`make world' times by about 10%.  Use a new `bootstrap-libraries'
target to build just enough (static) libraries to bootstrap the
tools, and delay building of shared libraries until all the tools
have been built.

Fixed `/usr/bin/make world'.  The version of `make' that we buildis
nand use is named "make", not ${MAKE}.

26 years agoo Removed two unused variables (f_dirname and f_newline) in ls.c
Dag-Erling Smørgrav [Fri, 24 Apr 1998 12:43:26 +0000 (12:43 +0000)]
o Removed two unused variables (f_dirname and f_newline) in ls.c

o Added the -H and -P options for treatment of symbolic links.

o Removed the #ifdef BSD4_4_LITE, since it does not seem to do anything
  useful

o Fixed up prn_octal() so its output looks more like that of AT&T Unices
  when -b is given.

The next two lines apply only to the first two changes above:

PR: bin/6140
Submitted by: Max Euston

26 years agoUpdated references to the behaviour of ls(1).
Dag-Erling Smørgrav [Fri, 24 Apr 1998 12:38:18 +0000 (12:38 +0000)]
Updated references to the behaviour of ls(1).

PR: bin/6140
Submitted by: Max Euston

26 years agoDiscussed with: bde
David E. O'Brien [Fri, 24 Apr 1998 11:50:30 +0000 (11:50 +0000)]
Discussed with: bde

26 years agoCreate virgin disklabels with 8 (MAXPARTITIONS) partitions rather than
David E. O'Brien [Fri, 24 Apr 1998 11:49:57 +0000 (11:49 +0000)]
Create virgin disklabels with 8 (MAXPARTITIONS) partitions rather than
three (RAW_PART + 1);
This makes ``disklabel -Brw sdN auto'' do the Right Thing.

26 years agoAt the request of Garrett, changed sysctl:
David Greenman [Fri, 24 Apr 1998 10:08:57 +0000 (10:08 +0000)]
At the request of Garrett, changed sysctl:

net.inet.tcp.delack_enabled -> net.inet.tcp.delayed_ack

26 years agoEnsure that TCP_REXMTVAL doesn't return a value less than t_rttmin. This
David Greenman [Fri, 24 Apr 1998 09:25:39 +0000 (09:25 +0000)]
Ensure that TCP_REXMTVAL doesn't return a value less than t_rttmin. This
is believed to have been broken with the Brakmo/Peterson srtt
calculation changes. The result of this bug is that TCP connections
could time out extremely quickly (in 12 seconds).
Also backed out jdp's partial fix for this problem in rev 1.17 of
tcp_timer.c as it is obsoleted by this commit.
Bug was pointed out by Kevin Lehey <kml@roller.nas.nasa.gov>.

PR: 6068

26 years agoslice_device.c: permissions changes for SLICE devices
Julian Elischer [Fri, 24 Apr 1998 07:54:00 +0000 (07:54 +0000)]
slice_device.c: permissions changes for SLICE devices
vn.c: change time of SYSINIT scheduling.
wd.c don't revert to fully closed state. ( may require more)
all in SLICE mode only.

26 years agoo Renamed '-b' (show unprintables in octal) to '-B'
Dag-Erling Smørgrav [Fri, 24 Apr 1998 07:49:51 +0000 (07:49 +0000)]
o Renamed '-b' (show unprintables in octal) to '-B'

o Added a new '-b' which behaves as in AT&T Unices (show unprintables in
  octal, using C escape codes when possible)

o Added '?' to the getopt() string, since the code in the switch considers
  it as a valid option.

26 years agoOutput mbuf clusters as current/peak/max. Requires new kern.ipc.nmbclusters
David Greenman [Fri, 24 Apr 1998 04:30:27 +0000 (04:30 +0000)]
Output mbuf clusters as current/peak/max. Requires new kern.ipc.nmbclusters
sysctl variable (previously committed).

26 years agoAdded kern.ipc.nmbclusters
David Greenman [Fri, 24 Apr 1998 04:15:52 +0000 (04:15 +0000)]
Added kern.ipc.nmbclusters

26 years agoWhen calling the open function, specify either FREAD and/or FWRITE
Julian Elischer [Thu, 23 Apr 1998 22:09:55 +0000 (22:09 +0000)]
When calling the open function, specify either FREAD and/or FWRITE
as leaving them both 0 has the same effect as not openning the device at all.

26 years agoOops forgot to include opt_bootp.h
Julian Elischer [Thu, 23 Apr 1998 20:59:42 +0000 (20:59 +0000)]
Oops forgot to include opt_bootp.h
Bootp works a lot better with devfs if this is present.

26 years agoThere is no "rwho" service, it's "who"
Dag-Erling Smørgrav [Thu, 23 Apr 1998 19:12:41 +0000 (19:12 +0000)]
There is no "rwho" service, it's "who"
PR: bin/6396
Submitted by: Ruslan Ermilov <ru@ucb.crimea.ua>

26 years agoDon't generate code with syntax errors (hard newlines in string constants
Bruce Evans [Thu, 23 Apr 1998 16:37:06 +0000 (16:37 +0000)]
Don't generate code with syntax errors (hard newlines in string constants
in config.c).

26 years agoFixed missing dependencies. The critical missing one was of bi-lexer.o
Bruce Evans [Thu, 23 Apr 1998 16:30:51 +0000 (16:30 +0000)]
Fixed missing dependencies.  The critical missing one was of bi-lexer.o
on bi-parser.h.  Not having it sometimes (only with `make -j') triggered
a bug suite that led to ordinary cpp output being put in .depend files.
Various bugs (the main one only with `make -j') prevented timely detection
of failure to build and install gnu/usr.bin/cc.  Eventually the missing
${WORLDTMP}/usr/libexec/cpp caused ${WORLDTMP}/usr/bin/cpp to be invoked
by cc, and this version of cpp is not suitable for such invocation.
Ordinary cpp output was put in .depend files when cpp terminated options
processing before seeing the -M flag.

Cleaned up.

26 years agoConvert timespecclear() back to an expression (this time without
Eivind Eklund [Thu, 23 Apr 1998 14:12:06 +0000 (14:12 +0000)]
Convert timespecclear() back to an expression (this time without
sideeffects).

26 years agoBack out last change
Paul Traina [Thu, 23 Apr 1998 04:44:08 +0000 (04:44 +0000)]
Back out last change

26 years agoWrap timerclear in do { ... } while(0)
Eivind Eklund [Thu, 23 Apr 1998 00:11:32 +0000 (00:11 +0000)]
Wrap timerclear in do { ... } while(0)

26 years agoAdd devfs support.
Eivind Eklund [Thu, 23 Apr 1998 00:06:43 +0000 (00:06 +0000)]
Add devfs support.

Submitted by: somebody whose name escapes me :-(

26 years agoMask the interrupt before setting the corresponding bit in ipending if
Tor Egge [Wed, 22 Apr 1998 22:49:29 +0000 (22:49 +0000)]
Mask the interrupt before setting the corresponding bit in ipending if
the interrupt is already active.
Don't use lock prefix for operations on ipending.
Always use lock prefix for operations on iactive.

26 years agoImprove description.
Robert Nordier [Wed, 22 Apr 1998 19:59:55 +0000 (19:59 +0000)]
Improve description.
Submitted by: Arne Henrik Juul <arnej@imf.unit.no>

26 years agoBad144 support for the slice system (!)
Julian Elischer [Wed, 22 Apr 1998 19:27:54 +0000 (19:27 +0000)]
Bad144 support for the slice system (!)
Submitted by: luoqi@watermarkgroup.com (Luoqi Chen)

I'm amazed by this. Slice has only been checked in for 2 days..

26 years agoAdd support for the Qlogic ISP SCSI && FC/AL Adapters
Matt Jacob [Wed, 22 Apr 1998 18:18:16 +0000 (18:18 +0000)]
Add support for the Qlogic ISP SCSI && FC/AL Adapters

26 years agoAdd support for the Qlogic ISP SCSI && FC/AL Adapters
Matt Jacob [Wed, 22 Apr 1998 18:12:29 +0000 (18:12 +0000)]
Add support for the Qlogic ISP SCSI && FC/AL Adapters