]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
26 years agoFix the signal behaviour for internal states which set the thread
jb [Fri, 5 Jun 1998 23:31:55 +0000 (23:31 +0000)]
Fix the signal behaviour for internal states which set the thread
state to running despite the SA_RESTART flag which is really just for
syscalls.

26 years agoMake the Man page reflect teh new reality.
julian [Fri, 5 Jun 1998 23:20:12 +0000 (23:20 +0000)]
Make the Man page reflect teh new reality.

26 years agoReverse the default sense of the IPFW/DIVERT reinjection code
julian [Fri, 5 Jun 1998 22:40:01 +0000 (22:40 +0000)]
Reverse the default sense of the IPFW/DIVERT reinjection code
so that the new behaviour is now default.
Solves the "infinite loop in diversion" problem when more than one diversion
is active.
Man page changes follow.

The new code is in -stable as the NON default option.

26 years agoChanged the log() of "Out of mbuf clusters - increase maxusers" to a
dg [Fri, 5 Jun 1998 21:48:45 +0000 (21:48 +0000)]
Changed the log() of "Out of mbuf clusters - increase maxusers" to a
printf() of "Out of mbuf clusters - adjust NMBCLUSTERS or increase
maxusers" so that the message is more informative and so that it will
appear in the kernel message buffer.

26 years agoMoved limit frobbing (and the resulting limcopy()) that occurs for
dg [Fri, 5 Jun 1998 21:44:20 +0000 (21:44 +0000)]
Moved limit frobbing (and the resulting limcopy()) that occurs for
accounting to the accounting function so that this isn't needlessly
done for some process exits.
Reviewed by: bde,phk

26 years agoIf we are out of mb_map space and we failed to m_reclaim() anything and
dg [Fri, 5 Jun 1998 21:41:48 +0000 (21:41 +0000)]
If we are out of mb_map space and we failed to m_reclaim() anything and
the alloc is not M_DONTWAIT, then panic with "Out of mbuf clusters".
Callers that specify M_WAIT can't deal with getting a NULL buffer, so this
is a more graceful failure than randomly page faulting in the socket code
or elsewhere.

26 years agoRecognize ".cpp" as a C++ suffix.
dt [Fri, 5 Jun 1998 18:38:55 +0000 (18:38 +0000)]
Recognize ".cpp" as a C++ suffix.

PR: 3476

26 years agoAdd mtree to bootstrap targets now that peter has added a new
jkh [Fri, 5 Jun 1998 16:50:45 +0000 (16:50 +0000)]
Add mtree to bootstrap targets now that peter has added a new
keyword to it which will otherwise call the next target to fall over
on transitioning systems.

26 years agoHands off my /usr/src and /usr/obj! If they are symlinks (since they are
peter [Fri, 5 Jun 1998 14:47:10 +0000 (14:47 +0000)]
Hands off my /usr/src and /usr/obj!  If they are symlinks (since they are
hard coded into too many things), it's not nice to go and change /home/src
etc.  This means they will be created if missing (so it shouldn't break
the releases), but won't touch them once they are changed.

26 years agoCreate a new flag, "nochange", which will specify that a directory is
peter [Fri, 5 Jun 1998 14:43:42 +0000 (14:43 +0000)]
Create a new flag, "nochange", which will specify that a directory is
to be created if it's missing, otherwise completely ignore it's modes and
owners.  Primary intended targets:  /usr/src and /usr/obj.

Adjust the 'not created: File exists' message to mention that it's a
directory that's the problem, otherwise it doesn't make sense.

I had created chown-style  -L and -P flag to control logical/physical mode
(ie: whether symlinks were followed), but the nochange flag is enough to
get the blasted thing out of my hair so I took them back out.

26 years agoMention that syslog.conf(5) does not log LOG_FTP messages by default.
jkoshy [Fri, 5 Jun 1998 10:31:24 +0000 (10:31 +0000)]
Mention that syslog.conf(5) does not log LOG_FTP messages by default.

PR: 5287

26 years agoFix handling more than 1 char with non-C locale
ache [Fri, 5 Jun 1998 09:49:51 +0000 (09:49 +0000)]
Fix handling more than 1 char with non-C locale
Misc. cleanup
PR: 6825
Submitted by: Sergey Gershtein <sg@mplik.ru>

26 years agoMention that tftpd(8) also uses syslog(3).
jkoshy [Fri, 5 Jun 1998 09:20:19 +0000 (09:20 +0000)]
Mention that tftpd(8) also uses syslog(3).

26 years agoReviewed by: hasty@star-gate.com
ahasty [Fri, 5 Jun 1998 08:59:11 +0000 (08:59 +0000)]
Reviewed by: hasty@star-gate.com
Submitted by:  Roger Hardiman <roger@cs.strath.ac.uk>

I left out a line of code  from Roger's last patch :(

26 years agoSync with sys/i386/isa/sio.c revision up to 1.205.
kato [Fri, 5 Jun 1998 08:31:01 +0000 (08:31 +0000)]
Sync with sys/i386/isa/sio.c revision up to 1.205.

26 years agoSync with sys/i386/conf/GENERIC revision 1.110.
kato [Fri, 5 Jun 1998 08:29:55 +0000 (08:29 +0000)]
Sync with sys/i386/conf/GENERIC revision 1.110.

26 years agoAdd missing uu_lock_txfr() prototype
ache [Fri, 5 Jun 1998 08:21:33 +0000 (08:21 +0000)]
Add missing uu_lock_txfr() prototype

26 years agospurious ntohs calls were stopping some packets from being recognised
julian [Fri, 5 Jun 1998 06:55:37 +0000 (06:55 +0000)]
spurious ntohs calls were stopping some packets from being recognised
for what they are..
Taken from the netatalk mailing list. from a NetBSD user.

26 years agoFix a spelling error.
steve [Thu, 4 Jun 1998 21:09:08 +0000 (21:09 +0000)]
Fix a spelling error.

PR: 6857
Submitted by: Josh Gilliam <josh@quick.net>

26 years agoFix a bunch of spelling errors.
steve [Thu, 4 Jun 1998 21:06:07 +0000 (21:06 +0000)]
Fix a bunch of spelling errors.

PR: 6856
Submitted by: Josh Gilliam <josh@quick.net>

26 years agoUse size_t instead of u_int for sizes.
dfr [Thu, 4 Jun 1998 17:21:39 +0000 (17:21 +0000)]
Use size_t instead of u_int for sizes.

26 years agoIf the filesystem blocksize is less than the VM page size, use the generic
dfr [Thu, 4 Jun 1998 17:04:44 +0000 (17:04 +0000)]
If the filesystem blocksize is less than the VM page size, use the generic
getpages code.  This happens for filesystems with 4k pages on the alpha since
the normal alpha pagesize is 8k.

26 years agoDon't cast a pointer to an int in DQHASH.
dfr [Thu, 4 Jun 1998 17:03:16 +0000 (17:03 +0000)]
Don't cast a pointer to an int in DQHASH.

26 years agoDarnit, that last commit was only supposed to change files in the
wpaul [Thu, 4 Jun 1998 15:37:05 +0000 (15:37 +0000)]
Darnit, that last commit was only supposed to change files in the
-current branch. I mistakenly checked out the 2.2.x rpc.ypxfrd into my
-current working directory. No harm done, but I got really confused
when I went to check out rpc.ypxfrd again and found the changes I
wanted to make were already there. (I'm going to fix the other 2.2.x
versions of the other programs in a minute.)

Anyway: protect errno in the signal handler, in the -current version
of rpc.ypxfrd this time.

26 years agoMake the whereobj target print it's output without the overhead of
peter [Thu, 4 Jun 1998 15:31:55 +0000 (15:31 +0000)]
Make the whereobj target print it's output without the overhead of
starting another make..
Besides, make -V is broken, since it will print unexpanded variables.

26 years agoHard coded /usr/lib reference
peter [Thu, 4 Jun 1998 15:15:38 +0000 (15:15 +0000)]
Hard coded /usr/lib reference

26 years agoProtect errno in signal handlers, like in portmap.
wpaul [Thu, 4 Jun 1998 15:11:32 +0000 (15:11 +0000)]
Protect errno in signal handlers, like in portmap.

26 years agoBuild aout bootstrap tools only on i386.
jb [Thu, 4 Jun 1998 12:02:52 +0000 (12:02 +0000)]
Build aout bootstrap tools only on i386.

`make world' now works on alpha!

26 years agoRemoved generated files. named used the wrong version of ns_parser.h.
bde [Thu, 4 Jun 1998 07:55:55 +0000 (07:55 +0000)]
Removed generated files.  named used the wrong version of ns_parser.h.

26 years agoSimplified using new yacc rules. This will fix bugs when we stop picking
bde [Thu, 4 Jun 1998 07:25:52 +0000 (07:25 +0000)]
Simplified using new yacc rules.  This will fix bugs when we stop picking
up a possibly stale version of ns_parser.h from contrib/bind.

26 years agoFix errors that crept into the previous commit.
jkoshy [Thu, 4 Jun 1998 07:02:59 +0000 (07:02 +0000)]
Fix errors that crept into the previous commit.

26 years agoUse .Pa for filenames. Use .Bl/.El in FILES section. Remove unused #includes.
charnier [Thu, 4 Jun 1998 06:51:14 +0000 (06:51 +0000)]
Use .Pa for filenames. Use .Bl/.El in FILES section. Remove unused #includes.

26 years agoDon't use beforedepend or give dependencies on generated headers explicitly.
bde [Thu, 4 Jun 1998 06:50:58 +0000 (06:50 +0000)]
Don't use beforedepend or give dependencies on generated headers explicitly.
Just put generated headers in SRCS.

26 years agoAdd section number to .Xr. Use of .Nm. Typo. Add rcsid. Remove unused
charnier [Thu, 4 Jun 1998 06:49:13 +0000 (06:49 +0000)]
Add section number to .Xr. Use of .Nm. Typo. Add rcsid. Remove unused
#includes. Use err(3).

26 years ago1. `ps' output now shows 3 characters in the `TT' field, not 2, after
jkoshy [Thu, 4 Jun 1998 06:46:13 +0000 (06:46 +0000)]
1. `ps' output now shows 3 characters in the `TT' field, not 2, after
    rev 1.6 of "ps.c".
2. Reword description of `-f' option.

PR: 5340
Submitted by: Jorge Goncalves <j@bug.fe.up.pt>

26 years agoCorrect .Nm. Spelling. Add rcsid, remove unused #includes. Add usage(). Do not
charnier [Thu, 4 Jun 1998 06:44:18 +0000 (06:44 +0000)]
Correct .Nm. Spelling. Add rcsid, remove unused #includes. Add usage(). Do not
dot-terminate errx() strings.

26 years agoAdd support for baudrates higher than 37800.
dfr [Thu, 4 Jun 1998 06:43:01 +0000 (06:43 +0000)]
Add support for baudrates higher than 37800.

26 years agoCorrect use of .Nm, use .Bl/.El in FILES section. Add rcsid. Remove unused
charnier [Thu, 4 Jun 1998 06:41:26 +0000 (06:41 +0000)]
Correct use of .Nm, use .Bl/.El in FILES section. Add rcsid. Remove unused
#includes and make it a little more -Wall-friendly.

26 years agoRemoved superfluous beforedepends. Their files were already in SRCS.
bde [Thu, 4 Jun 1998 06:38:17 +0000 (06:38 +0000)]
Removed superfluous beforedepends.  Their files were already in SRCS.

26 years agoCorrect use of .Nm. Add section number to .Xr. Add rcsid, usage(). Use err(3).
charnier [Thu, 4 Jun 1998 06:38:03 +0000 (06:38 +0000)]
Correct use of .Nm. Add section number to .Xr. Add rcsid, usage(). Use err(3).

26 years agoCorrect use of .Nm. Add missing .El. Add rcsid. Remove unused #include.
charnier [Thu, 4 Jun 1998 06:35:12 +0000 (06:35 +0000)]
Correct use of .Nm. Add missing .El. Add rcsid. Remove unused #include.

26 years agoUse .Ar instead of .ar. Correct incorrect use of .Sm (detected by Paul (pst)).
charnier [Thu, 4 Jun 1998 06:30:50 +0000 (06:30 +0000)]
Use .Ar instead of .ar. Correct incorrect use of .Sm (detected by Paul (pst)).

26 years agoDon't use beforedepend; just put generated headers in SRCS. ldgram.h
bde [Thu, 4 Jun 1998 06:26:23 +0000 (06:26 +0000)]
Don't use beforedepend; just put generated headers in SRCS.  ldgram.h
was already put in SRCS by the general yacc rules.  Putting ldemul-list.h
in SRCS fixes races in `make -jN' when .depend hasn't been made.

Don't forget to clean ldemul-list.h.

26 years agoTypos.
charnier [Thu, 4 Jun 1998 06:25:22 +0000 (06:25 +0000)]
Typos.

26 years agoDon't forget to pretend that `make -n -jN' makes the targets that it
bde [Thu, 4 Jun 1998 05:48:57 +0000 (05:48 +0000)]
Don't forget to pretend that `make -n -jN' makes the targets that it
says it pretends to make.

This bug was apparently harmless except for normal cases involving
.ORDER statements when it made debugging of -jN using -n very
confusing.  E.g., for:

.ORDER: beforedepend .depend
depend: beforedepend .depend

where beforedepend depends on something so that it is not initially
up to date, `make [-n] -j2 depend' causes `make' to wait for itself
to make beforedepend.  This works fine without -n.  The job to make
beforedepend has normally been started, and beforedepend is marked
as made when the job completes.  However, with -n, the pseudo-job
for making beforedepend has normally completed, and in any case
there was no chance of beforedepend being marked as made.  `make'
actually exited almost immediately with status 0 instead of waiting
forever.

26 years agoRemoved special rules for lex- and yacc- generated files. The general
bde [Thu, 4 Jun 1998 01:09:28 +0000 (01:09 +0000)]
Removed special rules for lex- and yacc- generated files.  The general
rules work identically, except that they actually work if the yacc-
generated header somehow becomes out of date.

26 years agoFix my last commit which got caught in the cvs/cvsup cycle.
jb [Wed, 3 Jun 1998 22:07:26 +0000 (22:07 +0000)]
Fix my last commit which got caught in the cvs/cvsup cycle.
I'm told that I botched the tabs too. I'm not sure they are fixed now
because I find that rather tedious.

26 years agoUse gethostname(3) to obtain the local host name, as opposed to uname(3).
joerg [Wed, 3 Jun 1998 20:01:28 +0000 (20:01 +0000)]
Use gethostname(3) to obtain the local host name, as opposed to uname(3).
Otherwise the length of the name is limited to 32 characters only.

26 years agoMissing DPADD's.
peter [Wed, 3 Jun 1998 18:57:59 +0000 (18:57 +0000)]
Missing DPADD's.

26 years agoAdd missing DPADD's.
peter [Wed, 3 Jun 1998 18:01:04 +0000 (18:01 +0000)]
Add missing DPADD's.

26 years agoAdd the DPT driver here. It's kinda ironic that it got enabled in -stable
jkh [Wed, 3 Jun 1998 13:33:34 +0000 (13:33 +0000)]
Add the DPT driver here.  It's kinda ironic that it got enabled in -stable
first. :)
PR: 6848

26 years agoForce success of the probe (after doing it as before except in one
bde [Wed, 3 Jun 1998 12:30:10 +0000 (12:30 +0000)]
Force success of the probe (after doing it as before except in one
miscconfigured case) if the port is the console.  This fixes several
bugs:
- if all sioprobe()s failed, then the console driver followed null
  pointers in cdevsw[].
- if the sioprobe() for the console failed but another sioprobe()
  succeeded, then init hung early when the console couldn't be
  opened.
- it was silly for the console to not be there after printing boot
  messages on it.
Bugs introduced by this are hopefully no worse than old ones caused
by forcing the success of the `cn' level probe.

26 years agosignal() returns SIG_ERR on error, not int.
jb [Wed, 3 Jun 1998 11:33:44 +0000 (11:33 +0000)]
signal() returns SIG_ERR on error, not int.

time() requires a time_t pointer, not a long.

26 years agoFixed a printf() arg botch in the previous commit.
bde [Wed, 3 Jun 1998 09:43:38 +0000 (09:43 +0000)]
Fixed a printf() arg botch in the previous commit.

Only complain about an irq mismatch in the probe if the configured
irq doesn't become active, and then print the bitmap of irqs that
became active (including clock irqs) instead of just the first
(not including clock irqs).

Bugs reported by: msmith

26 years agoNo longer need the architecture specific include from libc now that
jb [Wed, 3 Jun 1998 08:55:20 +0000 (08:55 +0000)]
No longer need the architecture specific include from libc now that
NetBSD syscalls are defined in sys.mk if they are being used.

26 years agoUh, the NetBSD syscall define spam for alpha has to go after /etc/make.conf
jb [Wed, 3 Jun 1998 08:51:39 +0000 (08:51 +0000)]
Uh, the NetBSD syscall define spam for alpha has to go after /etc/make.conf
so that it works all the time.

26 years agoSync with sys/i386/i386/machdep.c revision 1.298.
kato [Wed, 3 Jun 1998 08:48:00 +0000 (08:48 +0000)]
Sync with sys/i386/i386/machdep.c revision 1.298.

26 years agoIf using NetBSD syscalls the rtprio syscall doesn't exist, so just
jb [Wed, 3 Jun 1998 08:12:00 +0000 (08:12 +0000)]
If using NetBSD syscalls the rtprio syscall doesn't exist, so just
don't try to use it to set special priorities.

26 years agoIn the case of alpha (only), more than just libc needs to know that
jb [Wed, 3 Jun 1998 08:07:31 +0000 (08:07 +0000)]
In the case of alpha (only), more than just libc needs to know that
NetBSD syscalls are being used, so spam all compiles with the define
that indicates this.

26 years agoIfdefed the netisr support.
bde [Wed, 3 Jun 1998 07:56:59 +0000 (07:56 +0000)]
Ifdefed the netisr support.

PR: 6760
Reviewed by: joerg

26 years agoCorrect wrong claim about `telldir()' cookie lifetimes.
jkoshy [Wed, 3 Jun 1998 06:54:24 +0000 (06:54 +0000)]
Correct wrong claim about `telldir()' cookie lifetimes.

PR: 4043
Submitted by: Joe Orthoefer <j_orthoefer@tia.net>

26 years agoSpelling corrections.
jkoshy [Wed, 3 Jun 1998 04:21:41 +0000 (04:21 +0000)]
Spelling corrections.

PR: 6829
Submitted by: Josh Gilliam <josh@quick.net>

26 years agoCorrect description of -crtscts.
jkoshy [Wed, 3 Jun 1998 04:07:59 +0000 (04:07 +0000)]
Correct description of -crtscts.

PR: 6836
Submitted by: Andrew <andrew@ugh.net.au>

26 years agoCorrect typo.
jkoshy [Wed, 3 Jun 1998 03:59:00 +0000 (03:59 +0000)]
Correct typo.

PR: 6844
Submitted by: Adrian Filipi-Martin <adrian@ubergeeks.com>

26 years agoIf vm86 services are available, use these to perform the APM BIOS
msmith [Wed, 3 Jun 1998 01:59:42 +0000 (01:59 +0000)]
If vm86 services are available, use these to perform the APM BIOS
probe and intialisation.  This will ultimately remove the grubby (but
functional) hack that copies a real-mode function into low memory
early in locore.s.

26 years agoPR: 6831
danny [Tue, 2 Jun 1998 23:02:09 +0000 (23:02 +0000)]
PR: 6831
Submitted by: zerium@webindex.no
MF22: remove '.' after weekday abbreviation.

26 years agoIncrement __FreeBSD_version because the semctl() prototype changed
jdp [Tue, 2 Jun 1998 20:58:22 +0000 (20:58 +0000)]
Increment __FreeBSD_version because the semctl() prototype changed
in <sys/sem.h>.

26 years agoDon't build perl in build-tools: if NOPERL is defined.
jhay [Tue, 2 Jun 1998 18:28:55 +0000 (18:28 +0000)]
Don't build perl in build-tools: if NOPERL is defined.

26 years agoPR: bin/6828
thepish [Tue, 2 Jun 1998 13:11:04 +0000 (13:11 +0000)]
PR: bin/6828
Submitted by: njs3@doc.ic.ac.uk
Obtained from: NetBSD

Implement embedded variable expansion

26 years ago".. x11amp appears to be calling shmctl(id, IPC_RMID, 0) and the emulation
jkh [Tue, 2 Jun 1998 12:38:31 +0000 (12:38 +0000)]
".. x11amp appears to be calling shmctl(id, IPC_RMID, 0) and the emulation
layer does not like the null shmid_ds buffer pointer.  The emulation layer
returned an error without ever calling FreeBSD's shmctl, so the segments
were not being deleted when the reference count went to zero."

Submitted by: Kevin Street <street@iname.com>

26 years agoTreat an EOPNOTSUPP from fchflags() as a non-fatal case. Only warn about
peter [Tue, 2 Jun 1998 12:00:08 +0000 (12:00 +0000)]
Treat an EOPNOTSUPP from fchflags() as a non-fatal case.  Only warn about
it if flags were explicitly specified on the command line.  Do not warn
if we were merely trying to preserve flags or remove UF_NODUMP.  NFS does
not support flags.

I'm not sure that this is ideal, but it should do for now.  Installing
a plain file onto a NFS server must work, we used to silently ignore the
attempt.  Doing a binary install looses the flags anyway since cpio
doens't preserve them with the cdrom/network images.
XXX make world should not use flags or chown/chgrp in the obj/tmp area.

This is based on a suggestion from Ken Merry <ken@plutotech.com>.

26 years agoFix brain-fart. "usr.bin/reinstall" should be "sbin/ldconfig".
peter [Tue, 2 Jun 1998 11:22:08 +0000 (11:22 +0000)]
Fix brain-fart. "usr.bin/reinstall" should be "sbin/ldconfig".
No comment. *-<:-)

Submitted by: "Pierre Y. Dampure" <pierre.dampure@k2c.co.uk>

26 years agoCut through the discussions and settle the issue: Everything in /var/run
phk [Tue, 2 Jun 1998 11:02:16 +0000 (11:02 +0000)]
Cut through the discussions and settle the issue:  Everything in /var/run
is blown away at boot.  If sudo or other programs needs directories, they
had better make them in /usr/local/etc/rc.d/mumble.sh

PR: 4138 5619 6793

26 years agoCleanup and remove some dead code from the initialization.
dyson [Tue, 2 Jun 1998 05:50:08 +0000 (05:50 +0000)]
Cleanup and remove some dead code from the initialization.

26 years agoCorrect sleep priority.
dyson [Tue, 2 Jun 1998 05:39:13 +0000 (05:39 +0000)]
Correct sleep priority.

26 years agoAdd a reference to the original softupdates paper
julian [Tue, 2 Jun 1998 01:30:51 +0000 (01:30 +0000)]
Add a reference to the original softupdates paper

26 years agoAdd a reference to the Ganger/Patt paper
julian [Tue, 2 Jun 1998 01:27:27 +0000 (01:27 +0000)]
Add a reference to the Ganger/Patt paper

26 years agoo Return error when the controller can't accept commands.
eivind [Tue, 2 Jun 1998 00:32:38 +0000 (00:32 +0000)]
o Return error when the controller can't accept commands.
o Make driver less chatty on boot (only announce version under
  bootverbose)

Submitted by: Simon Shapiro <shimon@simon-shapiro.org>

26 years agoSet PAGE_SIZE for _SC_PAGESIZE sysconf().
dufault [Mon, 1 Jun 1998 21:54:43 +0000 (21:54 +0000)]
Set PAGE_SIZE for _SC_PAGESIZE sysconf().

26 years agoAdd a trivial mechanism for returning a useful default value if one is
msmith [Mon, 1 Jun 1998 20:58:03 +0000 (20:58 +0000)]
Add a trivial mechanism for returning a useful default value if one is
available and the kernel MIB setting is zero.

Return the result from getpagesize() if the p1003_1b.pagesize MIB
value is zero.

Suggested by: Joerg Schilling <schilling@fokus.gmd.de>

26 years agoAllow devfs to support the 'uk' device
julian [Mon, 1 Jun 1998 19:44:54 +0000 (19:44 +0000)]
Allow devfs to support the 'uk' device
Submitted by: Micha Class <michael_class@hp.com>

26 years agoadd libcipher, and update comments about libdes, libkdb, libkrb etc
peter [Mon, 1 Jun 1998 19:38:20 +0000 (19:38 +0000)]
add libcipher, and update comments about libdes, libkdb, libkrb etc
which do exist but are in the secure dist rather than the base.

26 years agoLook for kzhead.o and kztail.o in /usr/lib/aout.
jhay [Mon, 1 Jun 1998 18:34:19 +0000 (18:34 +0000)]
Look for kzhead.o and kztail.o in /usr/lib/aout.

26 years agoldconfig is used at the end of installworld.
peter [Mon, 1 Jun 1998 18:29:11 +0000 (18:29 +0000)]
ldconfig is used at the end of installworld.

26 years agoSUBDIR += rtld-aout
peter [Mon, 1 Jun 1998 14:33:06 +0000 (14:33 +0000)]
SUBDIR += rtld-aout

26 years agoSUBDIR += ldd
peter [Mon, 1 Jun 1998 14:30:58 +0000 (14:30 +0000)]
SUBDIR += ldd

26 years agoSUBDIR += ldconfig
peter [Mon, 1 Jun 1998 14:12:48 +0000 (14:12 +0000)]
SUBDIR += ldconfig

26 years agoThese files were moved elsewhere (sbin/ldconfig, libexec/rtld-aout,
peter [Mon, 1 Jun 1998 14:11:28 +0000 (14:11 +0000)]
These files were moved elsewhere (sbin/ldconfig, libexec/rtld-aout,
usr.bin/ldd) as they stand alone and are under bsd-style license.

26 years agoLook for files in their new locations.
peter [Mon, 1 Jun 1998 14:08:01 +0000 (14:08 +0000)]
Look for files in their new locations.

26 years agoBuild ldconfig in /sbin environment directly, rather than under the gnu ld
peter [Mon, 1 Jun 1998 13:58:19 +0000 (13:58 +0000)]
Build ldconfig in /sbin environment directly, rather than under the gnu ld
directory.  The rcs files were repository copied.

26 years agoldd standalone build, after repository copy.
peter [Mon, 1 Jun 1998 13:55:12 +0000 (13:55 +0000)]
ldd standalone build, after repository copy.

26 years agoSeperate the Paul Kranenburg a.out rtld stuff into a stand-alone area away
peter [Mon, 1 Jun 1998 13:00:32 +0000 (13:00 +0000)]
Seperate the Paul Kranenburg a.out rtld stuff into a stand-alone area away
from the gpl ld code.  This is part 2 of something that I began in 1996.
A repository copy has happened behind cvs's back.

26 years agoSync with sys/i386/isa/sio.c revision 1.203.
kato [Mon, 1 Jun 1998 12:40:24 +0000 (12:40 +0000)]
Sync with sys/i386/isa/sio.c revision 1.203.

26 years agoMake sure we go a nfs_fsinfo() in get/putpages before calling
peter [Mon, 1 Jun 1998 11:32:53 +0000 (11:32 +0000)]
Make sure we go a nfs_fsinfo() in get/putpages before calling
readrpc/writerpc, since they assume it's already been done.  This could
break if the first read/write access to a nfs filesystem was an exec() or
mmap() instead of a read(), write() syscall.  (or statfs()).
nfs_getpages() could return an errno (EOPNOTSUPP) instead of a VM_PAGER_*
return code.  Some layout tweaks for the get/putpages code.

26 years agoFix post-test pre-commit cleanup typo.
peter [Mon, 1 Jun 1998 11:07:16 +0000 (11:07 +0000)]
Fix post-test pre-commit cleanup typo.

26 years agoreadlink() returns EINVAL rather than EPERM if called on a non-symlink.
peter [Mon, 1 Jun 1998 10:59:23 +0000 (10:59 +0000)]
readlink() returns EINVAL rather than EPERM if called on a non-symlink.

26 years agoPreset the maximum file size before we get to nfs_fsinfo(), based on
peter [Mon, 1 Jun 1998 10:01:31 +0000 (10:01 +0000)]
Preset the maximum file size before we get to nfs_fsinfo(), based on
an (over?) conservative assumption about what the client can store in it's
buffer cache using a signed 32-bit 512-byte block number index.  Otherwise
it's possible for some file access when maxfilesize = 0 (eg: /usr is nfs
mounted and doing an execve())
Pointed out by:  bde

XXX It might make sense to do a preemptive nfs_fsinfo() call at mount time.

26 years agoTrim a domain part for wtmp as same as showed by "netstat -r".
amurai [Mon, 1 Jun 1998 08:47:04 +0000 (08:47 +0000)]
Trim a domain part for wtmp as same as showed by "netstat -r".
Here is a some example for avoiding a confusion.

 It asssumes a logged host domain is "spec.co.jp". All
example is longer than UT_HOSTNAMELEN value.

   1) turbo.tama.spec.co.jp: 192.19.0.2  -> trubo.tama
   2) turbo.tama.foo.co.jp : 192.19.0.2  -> 192.19.0.2
   3) specgw.spec.co.jp    : 202.32.13.1 -> specgw

Submitted by: Atsushi Murai <amurai@spec.co.jp>

26 years agoRemove bogus #defines and redefintions for KERNEL, mount, nfs_init, the
peter [Mon, 1 Jun 1998 07:36:20 +0000 (07:36 +0000)]
Remove bogus #defines and redefintions for KERNEL, mount, nfs_init, the
beastie builds just fine without them and gets less warnings.

26 years agoHide more kernel stuff from userland. This stops nethostaddr etc being
peter [Mon, 1 Jun 1998 07:23:26 +0000 (07:23 +0000)]
Hide more kernel stuff from userland.  This stops nethostaddr etc being
wanted by mount_nfs.c.