]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
26 years agoSynchronize with sys/i386/i386/machdep.c revision 1.268.
kato [Sun, 12 Oct 1997 11:58:09 +0000 (11:58 +0000)]
Synchronize with sys/i386/i386/machdep.c revision 1.268.

26 years agoSynchronize with sys/i386/i386/trap.c revision 1.113.
kato [Sun, 12 Oct 1997 11:57:23 +0000 (11:57 +0000)]
Synchronize with sys/i386/i386/trap.c revision 1.113.

26 years agoSynchronize with sys/i386/boot/netboot/Makefile revision 1.14.
kato [Sun, 12 Oct 1997 11:56:06 +0000 (11:56 +0000)]
Synchronize with sys/i386/boot/netboot/Makefile revision 1.14.

26 years agoTypo fix.
max [Sun, 12 Oct 1997 11:51:25 +0000 (11:51 +0000)]
Typo fix.

26 years agoSubmitted by: Joachim Kuebart, thanks.
helbig [Sun, 12 Oct 1997 09:52:49 +0000 (09:52 +0000)]
Submitted by: Joachim Kuebart, thanks.
Add -u option to force unbuffered output

26 years agoRemove the call to scsi_start_unit() from sd_open(). It was causing
joerg [Sun, 12 Oct 1997 08:54:47 +0000 (08:54 +0000)]
Remove the call to scsi_start_unit() from sd_open().  It was causing
much grieve to owners of IBM drives when used in conjunction with
tagged command queuing, and didn't serve any purpose at all (since
experiments have proven that it simply didn't work).

Instead, call scsi_start_unit() once in sd_attach(), so in case the
drive has been configured to `remote start', it will spin up there.
(If it has spun down later, it must have been because of administrator
action (scsi(8)) anyway.)

While i was at it, bump the timeout for scsi_start_unit() to 30
seconds.  10 seconds were way too few for most drives.

26 years agoAllow "foo0: disabled, not probed" message to stay, but make it conditional
jkh [Sun, 12 Oct 1997 08:31:41 +0000 (08:31 +0000)]
Allow "foo0: disabled, not probed" message to stay, but make it conditional
on bootverbose so that those who _really wanna know_ still can.
Compromise suggested by: joerg

26 years agoUpdate de.4 to list new media options. Hopefully this is reasonably
peter [Sun, 12 Oct 1997 02:20:54 +0000 (02:20 +0000)]
Update de.4 to list new media options.  Hopefully this is reasonably
correct.

26 years agoCosmetic: Mention that ppp is suspending before
brian [Sun, 12 Oct 1997 01:49:37 +0000 (01:49 +0000)]
Cosmetic: Mention that ppp is suspending before
          executing a "shell" command.

26 years agoreference rc.conf(5) instead of a bogus rc.conf(8)
jkh [Sun, 12 Oct 1997 01:00:19 +0000 (01:00 +0000)]
reference rc.conf(5) instead of a bogus rc.conf(8)
Submitted by: "Vanill I. Shu" <vanilla@MinJe.com.TW>

26 years agoStore temporary files in /var/tmp instead /tmp. This should avoid
wosch [Sat, 11 Oct 1997 22:05:44 +0000 (22:05 +0000)]
Store temporary files in /var/tmp instead /tmp. This should avoid
a possible disk overflow for enormous large mails.
Submitted by: grog

26 years agoTeach dd(1) about an option to write sparse files. Can be useful for
joerg [Sat, 11 Oct 1997 20:09:05 +0000 (20:09 +0000)]
Teach dd(1) about an option to write sparse files.  Can be useful for
things like diskless clients' swap files etc.

Submitted by: pascal@zuo.dec.com (Pascal Pederiva) (ages ago, with many
  stylistic changes by me)

26 years agoMention the improved sppp(4).
joerg [Sat, 11 Oct 1997 18:41:12 +0000 (18:41 +0000)]
Mention the improved sppp(4).

26 years agosppp(4) is now also a consumer of the kernel MD5 code.
joerg [Sat, 11 Oct 1997 18:35:26 +0000 (18:35 +0000)]
sppp(4) is now also a consumer of the kernel MD5 code.

26 years agoDistribute and statizice a lot of the malloc M_* types.
phk [Sat, 11 Oct 1997 18:31:40 +0000 (18:31 +0000)]
Distribute and statizice a lot of the malloc M_* types.

Substantial input from: bde

26 years agoDo not use an intermediate file. The intermediate filename will go
wosch [Sat, 11 Oct 1997 17:30:18 +0000 (17:30 +0000)]
Do not use an intermediate file. The intermediate filename will go
into the info output which may breaks other programs (e.g. info2html).

26 years agoFreeing with unknown type is a panic kind of thing.
phk [Sat, 11 Oct 1997 13:13:09 +0000 (13:13 +0000)]
Freeing with unknown type is a panic kind of thing.

26 years agoRemove all traces of M_VFSCONF, which were for all practical
phk [Sat, 11 Oct 1997 13:11:32 +0000 (13:11 +0000)]
Remove all traces of M_VFSCONF, which were for all practical
purposes unused.

26 years agoEnable spppcontrol.
joerg [Sat, 11 Oct 1997 11:34:00 +0000 (11:34 +0000)]
Enable spppcontrol.

26 years agoThis commit was generated by cvs2svn to compensate for changes in r30302,
joerg [Sat, 11 Oct 1997 11:30:30 +0000 (11:30 +0000)]
This commit was generated by cvs2svn to compensate for changes in r30302,
which included commits to RCS files with non-trunk default branches.

26 years agoThe spppcontrol(8) utility that is required to set or display things
joerg [Sat, 11 Oct 1997 11:30:30 +0000 (11:30 +0000)]
The spppcontrol(8) utility that is required to set or display things
like PAP and CHAP secrets with sppp(4).  This is the first utility
using the new SIOC[SG]IFGENERIC ioctls (and the reason for inventing
them in the first place).

26 years agoUpdate to match reality. PAP/CHAP are now implemented, a number of bugs
joerg [Sat, 11 Oct 1997 11:27:25 +0000 (11:27 +0000)]
Update to match reality.  PAP/CHAP are now implemented, a number of bugs
fixed as well.

26 years agoJumbo patch to implement PAP and CHAP for sppp(4). Partially based on
joerg [Sat, 11 Oct 1997 11:25:28 +0000 (11:25 +0000)]
Jumbo patch to implement PAP and CHAP for sppp(4).  Partially based on
Serge's (Cronyx's) code in the vendor branch.  (FR support not yet
merged.)

26 years agoRemove a debug printf entirely.
phk [Sat, 11 Oct 1997 10:49:43 +0000 (10:49 +0000)]
Remove a debug printf entirely.

26 years agoDisable an extremely annoying printf.
peter [Sat, 11 Oct 1997 10:41:44 +0000 (10:41 +0000)]
Disable an extremely annoying printf.

26 years agoAttempt to fix the previous fix to the contigmalloc1 prototype.
peter [Sat, 11 Oct 1997 10:39:19 +0000 (10:39 +0000)]
Attempt to fix the previous fix to the contigmalloc1 prototype.
struct malloc_type isn't defined in all cases (eg: from ddb), and the line
wrapping was very badly mangled.

26 years agoUpdate malloc.9 to reflect new reality.
phk [Sat, 11 Oct 1997 08:11:31 +0000 (08:11 +0000)]
Update malloc.9 to reflect new reality.

26 years agoBe a little closer to current reality.
phk [Sat, 11 Oct 1997 07:39:45 +0000 (07:39 +0000)]
Be a little closer to current reality.

26 years agoRemove a #ifndef __FreeBSD__ chunk.
phk [Sat, 11 Oct 1997 07:35:25 +0000 (07:35 +0000)]
Remove a #ifndef __FreeBSD__ chunk.

26 years agoDike out a weird warning.
phk [Sat, 11 Oct 1997 07:34:27 +0000 (07:34 +0000)]
Dike out a weird warning.

26 years agoFix a misleading comment. Rename a variable to make more sense.
phk [Sat, 11 Oct 1997 07:32:43 +0000 (07:32 +0000)]
Fix a misleading comment.  Rename a variable to make more sense.
Pointed out by by:  bde

26 years agoMake this file p-make clean. (Use "ld -O foo" instead of "ld; mv
asami [Sat, 11 Oct 1997 02:37:42 +0000 (02:37 +0000)]
Make this file p-make clean.  (Use "ld -O foo" instead of "ld; mv
a.out foo".)

Reviewed by: bde (actually more like "Suggested by")

26 years agomake a string longer so we can handle deeper source trees.
julian [Sat, 11 Oct 1997 01:07:42 +0000 (01:07 +0000)]
make a string longer so we can handle deeper source trees.
if global is in 2.2 can we do this there as well?

Submitted by: doug ambrisko (ambrisko@whistle.com)

26 years agoMake the target for the number of AIO daemons work.
dyson [Sat, 11 Oct 1997 01:07:03 +0000 (01:07 +0000)]
Make the target for the number of AIO daemons work.

26 years agoNIS+ compatibility kludge. A long time ago, I set up innetgr() so
wpaul [Sat, 11 Oct 1997 00:03:25 +0000 (00:03 +0000)]
NIS+ compatibility kludge. A long time ago, I set up innetgr() so
that if searching through the special netgroup.byhost or netgroup.byuser
maps didn't work, we would roll over to the 'slow' method of grovelling
though the netgroup map and working out the dependencies on the fly.
But I left this option hidden inside an #ifdef CHARITABLE since I
didn't think I'd ever need it.

Well, the Sun rpc.nisd NIS+ server in YP compat mode doesn't support
the .byhost and .byuser reverse maps, so the  failover is necessary
in order to be compatible. *sigh*

This closes PR #3891, and should be merged into RELENG_2_2.

26 years agostruct kmemstats was renamed.
phk [Fri, 10 Oct 1997 18:19:48 +0000 (18:19 +0000)]
struct kmemstats was renamed.

26 years agoFix contigmalloc() and contigmalloc1() arguments.
phk [Fri, 10 Oct 1997 18:18:47 +0000 (18:18 +0000)]
Fix contigmalloc() and contigmalloc1() arguments.

26 years agoMake ufs_reclaim free the underlying inode.
phk [Fri, 10 Oct 1997 18:18:13 +0000 (18:18 +0000)]
Make ufs_reclaim free the underlying inode.

26 years agoUse generic ufs_reclaim().
phk [Fri, 10 Oct 1997 18:17:42 +0000 (18:17 +0000)]
Use generic ufs_reclaim().

26 years agoAdd type arg to ffs_mountfs and avoid examining v_tag to find out
phk [Fri, 10 Oct 1997 18:17:00 +0000 (18:17 +0000)]
Add type arg to ffs_mountfs and avoid examining v_tag to find out
if MFS is getting a free ride.

Use generic ufs_reclaim().

26 years agoRemove a bunch of unused malloc types.
phk [Fri, 10 Oct 1997 18:15:47 +0000 (18:15 +0000)]
Remove a bunch of unused malloc types.
A couple of potential bogons flagged.
Various prototypes changed.

26 years agoRename "struct kmemstats" to "struct malloc_type" it makes more sense now.
phk [Fri, 10 Oct 1997 18:14:23 +0000 (18:14 +0000)]
Rename "struct kmemstats" to "struct malloc_type" it makes more sense now.

Fix type argument to hashinit() and phashinit()

26 years agoMega commit to cleanup the "remaining nits" after my malloc change.
phk [Fri, 10 Oct 1997 18:13:06 +0000 (18:13 +0000)]
Mega commit to cleanup the "remaining nits" after my malloc change.

Introduce a M_EXT2NODE for ext2fs vnodes.
Use generic ufs_reclaim instead of hijacking ffs_reclaim.

26 years agoMake vmstat understand the dynamic malloc stuff. There are a number
phk [Fri, 10 Oct 1997 14:08:07 +0000 (14:08 +0000)]
Make vmstat understand the dynamic malloc stuff.  There are a number
of ugly hardcoded constants in there.

26 years agoMake malloc more extensible. The malloc type is now a pointer to
phk [Fri, 10 Oct 1997 14:06:34 +0000 (14:06 +0000)]
Make malloc more extensible.  The malloc type is now a pointer to
the struct kmemstats that describes the type.

This allows subsystems to declare their malloc types locally
and <sys/malloc.h> doesn't need tweaked everytime somebody
gets an idea.  You can even have a type local to a lkm...

I don't know if we really need the longdesc, comments welcome.

TODO: There is a single nit in ext2fs, that will be fixed later,
and I intend to remove all unused malloc types and distribute
the rest closer to their use.

26 years agoDo not use -B in initial build of "make". It doesn't exist in 2.1.x.
asami [Fri, 10 Oct 1997 13:02:36 +0000 (13:02 +0000)]
Do not use -B in initial build of "make".  It doesn't exist in 2.1.x.
Instead, list the targets one per command line.

Reviewed by: bde
Submitted by: bde (part)

26 years ago#include <machine/tss.h> explicitly
peter [Fri, 10 Oct 1997 12:53:27 +0000 (12:53 +0000)]
#include <machine/tss.h> explicitly

26 years agoCompensate for pcb.h tweaks.
peter [Fri, 10 Oct 1997 12:42:54 +0000 (12:42 +0000)]
Compensate for pcb.h tweaks.

(Bruce pointed out the nesting)

26 years agoDon't #include unneeded includes here. pcb_ext.h picks up lots of other
peter [Fri, 10 Oct 1997 12:40:09 +0000 (12:40 +0000)]
Don't #include unneeded includes here.  pcb_ext.h picks up lots of other
stuff with it.

26 years agoGPROC0_SEL isn't used in any *.s files it seems..
peter [Fri, 10 Oct 1997 12:38:27 +0000 (12:38 +0000)]
GPROC0_SEL isn't used in any *.s files it seems..

26 years agoThis has always bugged me. At single user, the implied example it gives
peter [Fri, 10 Oct 1997 12:14:48 +0000 (12:14 +0000)]
This has always bugged me.  At single user, the implied example it gives
is not valid - it says that "sh" is the default, but you can't
actually type "sh" at this prompt - it has to be /bin/sh or some other
full pathname.

26 years agoGive PPP_FILTER a whirl.
peter [Fri, 10 Oct 1997 11:59:02 +0000 (11:59 +0000)]
Give PPP_FILTER a whirl.

26 years agoTry out PPP_FILTER
peter [Fri, 10 Oct 1997 11:57:43 +0000 (11:57 +0000)]
Try out PPP_FILTER

26 years agoAlso install pcap-int.h so things can call pcap_compile on arbitary
peter [Fri, 10 Oct 1997 11:55:42 +0000 (11:55 +0000)]
Also install pcap-int.h so things can call pcap_compile on arbitary
expressions (eg: in pppd)

26 years agoAdd IDs for Intel 82371MX/82437MX (mobile PCI chipset).
asami [Fri, 10 Oct 1997 11:52:17 +0000 (11:52 +0000)]
Add IDs for Intel 82371MX/82437MX (mobile PCI chipset).

Reviewed by: se

26 years agoChange default keymap (I left all the international ones alone) so
jkh [Fri, 10 Oct 1997 10:32:53 +0000 (10:32 +0000)]
Change default keymap (I left all the international ones alone) so
that LALT is META by default.  This will make the emacs users happy.
Approved by: sos

26 years agoConvert the VM86 option from a global option to an option only depended
peter [Fri, 10 Oct 1997 09:44:12 +0000 (09:44 +0000)]
Convert the VM86 option from a global option to an option only depended
on by the files that use it.  Changing the VM86 option now only causes
a recompile of a dozen files or so rather than the entire kernel.

26 years agoRevive the connect-max-attempts option. When running in persist mode,
peter [Fri, 10 Oct 1997 09:28:38 +0000 (09:28 +0000)]
Revive the connect-max-attempts option.  When running in persist mode,
this limits the number of retries before aborting.  This can save
expensive phone bills. :-)  See rev 1.6 of main.c for original.

26 years agoAllow a deleted deveice to delte it's nodes in other mounted devfs
julian [Fri, 10 Oct 1997 07:54:05 +0000 (07:54 +0000)]
Allow a deleted deveice to delte it's nodes in other mounted devfs
filesystems even if not in SPLIT_DEVS mode.

26 years agoUse err(3). Silent -Wall.
charnier [Fri, 10 Oct 1997 06:31:07 +0000 (06:31 +0000)]
Use err(3). Silent -Wall.

26 years agoRevive rev 1.10, which documents the secrets files overrides for the
peter [Fri, 10 Oct 1997 06:30:08 +0000 (06:30 +0000)]
Revive rev 1.10, which documents the secrets files overrides for the
local:remote IP addresses on the command line or the options.ttynn files.

26 years agoStaticize usage(). Cosmetics.
charnier [Fri, 10 Oct 1997 06:27:07 +0000 (06:27 +0000)]
Staticize usage(). Cosmetics.

26 years agoUse err(3) instead of local redefinition. Add rcsid string.
charnier [Fri, 10 Oct 1997 06:23:42 +0000 (06:23 +0000)]
Use err(3) instead of local redefinition. Add rcsid string.

26 years agoRevive some things that were lost during the ppp-2.3.1 update.
peter [Fri, 10 Oct 1997 06:02:57 +0000 (06:02 +0000)]
Revive some things that were lost during the ppp-2.3.1 update.
- (see auth.c rev 1.13) allow the pap/chap secrets file to specify an
override for the otherwise hard coded IP addresses.  This allows specific
users to dial in on a rotary which would otherwise get a dynamic address
forced to authenticate and get their own fixed addresses.
- (see options.c rev 1.9) recognize the old dns1 and dns2 options.  This
is a hack (TM). :-)

26 years agoAdd the AIO/LIO to libc. They aren't fully done yet, but have been in the
dyson [Fri, 10 Oct 1997 05:48:16 +0000 (05:48 +0000)]
Add the AIO/LIO to libc.  They aren't fully done yet, but have been in the
kernel for a few months.

26 years agoIf kerberos is installed and enabled in make.conf, activate cvs's kserver
peter [Fri, 10 Oct 1997 04:03:17 +0000 (04:03 +0000)]
If kerberos is installed and enabled in make.conf, activate cvs's kserver
hooks.  This is a lot safer than 'cvs pserver'.

26 years agoFix bug in __yp_ping(): keep proper count of assigned request structures.
wpaul [Thu, 9 Oct 1997 20:29:39 +0000 (20:29 +0000)]
Fix bug in __yp_ping(): keep proper count of assigned request structures.
The test to see if a request struct had been assigned after looping
though the server list was bogus.

Submitted by: "Erik E. Rantapaa" <rantapaa@math.umn.edu>

26 years agoBugfix for rev 1.41:
wosch [Thu, 9 Oct 1997 18:14:18 +0000 (18:14 +0000)]
Bugfix for rev 1.41:
makeinfo does not understand the tex command \input

26 years agoMake sure lfp is non-NULL before closing it.
jkh [Thu, 9 Oct 1997 14:11:19 +0000 (14:11 +0000)]
Make sure lfp is non-NULL before closing it.
PR: 4652

26 years agoUn-clobber mktemp->mkstemp patch.
eivind [Thu, 9 Oct 1997 11:05:16 +0000 (11:05 +0000)]
Un-clobber mktemp->mkstemp patch.

26 years agoAdd <sys/errno.h>
eivind [Thu, 9 Oct 1997 10:50:54 +0000 (10:50 +0000)]
Add <sys/errno.h>

26 years agoNote Comtrol rocketport driver.
jkh [Thu, 9 Oct 1997 09:11:40 +0000 (09:11 +0000)]
Note Comtrol rocketport driver.

26 years agoUse err(3). Add usage().
charnier [Thu, 9 Oct 1997 07:22:08 +0000 (07:22 +0000)]
Use err(3). Add usage().
Default source of tables (-M) is /dev/mem, not /dev/kmem.

26 years agoUse err(3). Add usage().
charnier [Thu, 9 Oct 1997 07:17:36 +0000 (07:17 +0000)]
Use err(3). Add usage().
Correct the .Nd string: in real life, portmap is translating rpc prog numbers
to darpa ports, not the opposite.

26 years agoMajor cleanup and debugging of the new AIO/LIO code. The LIO code is
dyson [Thu, 9 Oct 1997 04:14:41 +0000 (04:14 +0000)]
Major cleanup and debugging of the new AIO/LIO code.  The LIO code is
now corrected.  New tunables/instrumentation added.  The code is now
likely "good enough to use."  I will add the userland support soon.
The "high performance" mode for raw devices is still missing, and will
be added next.  POSIX system calls that now appear to work:
aio_cancel, aio_error, aio_read, aio_return, aio_suspend, aio_write,
lio_listio.  Missing, but to be added soon: aio_fsync.

26 years agoFixed scrolling. Newline at the "--More--" prompt scrolled 2 lines.
bde [Thu, 9 Oct 1997 00:20:00 +0000 (00:20 +0000)]
Fixed scrolling.  Newline at the "--More--" prompt scrolled 2 lines.

26 years agoInterpret multiple slashes as a single slash in ftp: URL's. This
fenner [Wed, 8 Oct 1997 18:43:53 +0000 (18:43 +0000)]
Interpret multiple slashes as a single slash in ftp: URL's.  This
is contrary to RFC1738, which specifies that empty elements between
slashes really mean to send a CWD command with an empty argument.
Most FTP servers consider this an error, so strict RFC compliance
is advisable in this case.

26 years ago- Enable PS/2 mouse support by default. Given that almost all new hardware
nate [Wed, 8 Oct 1997 17:05:03 +0000 (17:05 +0000)]
- Enable PS/2 mouse support by default.  Given that almost all new hardware
  has a PS/2 port, this is a good thing.  Note, older 386/486 boxes may
  lockup the keyboard controller with this enabled, but most of these kinds
  of machines don't run -current, so the benefits outweigh the downsides.

Discussed with: Kazutaka YOKOTA <yokota@zodiac.mech.utsunomiya-u.ac.jp>

26 years agoDocument extra tk*/tcl* package categories.
jkh [Wed, 8 Oct 1997 15:31:13 +0000 (15:31 +0000)]
Document extra tk*/tcl* package categories.

26 years agomake signal handlers errno safe
eivind [Wed, 8 Oct 1997 14:47:56 +0000 (14:47 +0000)]
make signal handlers errno safe

Obtained from: OpenBSD (file rev 1.11) - Theo de Raadt <deraadt@openbsd.org>

26 years agomktemp -> mkstemp
eivind [Wed, 8 Oct 1997 13:46:39 +0000 (13:46 +0000)]
mktemp -> mkstemp

Obtained from: OpenBSD (file rev 1.4 by Theo de Raadt <deraadt@openbsd.org>)

26 years agomake signal handlers errno safe
eivind [Wed, 8 Oct 1997 12:59:14 +0000 (12:59 +0000)]
make signal handlers errno safe

Obtained from: OpenBSD (file rev 1.5) Theo de Raadt <deraadt@openbsd.org>

26 years agoSlightly improved fix compared to my insiration
eivind [Wed, 8 Oct 1997 12:10:33 +0000 (12:10 +0000)]
Slightly improved fix compared to my insiration

don't munge argv strings -- copy them first. avoids 'w' showing:
    deraadt  p8 zeus.theos.com    8:26AM     3 dd if of bs

Obtained from: OpenBSD (file rev 1.3) by Theo de Raadt <deraadt@openbsd.org>

26 years agoBack out my last commit - the trailing / should stay and fetch(1) should
jkh [Wed, 8 Oct 1997 12:07:09 +0000 (12:07 +0000)]
Back out my last commit - the trailing / should stay and fetch(1) should
stop puking on null directory components.

26 years agoRemove trailing slash from MASTER_SITE_BACKUP example since it
jkh [Wed, 8 Oct 1997 10:56:27 +0000 (10:56 +0000)]
Remove trailing slash from MASTER_SITE_BACKUP example since it
can lead to two trailing /'s, which is in turn an invalid URL (just
bit me :-).

26 years agoMention SIOC[SG]IFGENERIC.
joerg [Wed, 8 Oct 1997 09:14:15 +0000 (09:14 +0000)]
Mention SIOC[SG]IFGENERIC.

Prodded by: jkh

26 years agoAdd pop
ache [Wed, 8 Oct 1997 08:45:35 +0000 (08:45 +0000)]
Add pop

26 years agoUse err(3). Cosmetics in usage string and man page.
charnier [Wed, 8 Oct 1997 07:48:21 +0000 (07:48 +0000)]
Use err(3). Cosmetics in usage string and man page.

26 years agoStaticise a variable.
markm [Wed, 8 Oct 1997 07:02:48 +0000 (07:02 +0000)]
Staticise a variable.
PR: 4722
Submitted by: Karl Denninger

26 years agoDespam 1.11, 1.12 and 1.13 caused by no sync between author/committer/CURRENT
markm [Wed, 8 Oct 1997 06:54:43 +0000 (06:54 +0000)]
Despam 1.11, 1.12 and 1.13 caused by no sync between author/committer/CURRENT

26 years agoInstall ${PKGDIR}/MESSAGE as /var/db/pkg/${PKGNAME}/+DISPLAY if it exists.
asami [Wed, 8 Oct 1997 05:04:48 +0000 (05:04 +0000)]
Install ${PKGDIR}/MESSAGE as /var/db/pkg/${PKGNAME}/+DISPLAY if it exists.

Reviewed by: kiri

26 years agoPR: bin/771 and bin/1037 are resolved by this change
uhclem [Wed, 8 Oct 1997 03:14:34 +0000 (03:14 +0000)]
PR: bin/771 and bin/1037 are resolved by this change
This change changes the default handling of linemode so that older and/or
stupider telnet clients can still get wakeup characters like <ESC> and
<CTRL>D to work correctly multiple times on the same line, as in csh
"set filec" operations.   It also causes CR and LF characters to be read by
apps in certain terminal modes consistently, as opposed to returning
CR sometimes and LF sometimes, which broke existing apps.  The change
was shown to fix the problem demonstrated in the FreeBSD telnet client,
along with the telnet client in Solaris, SCO, Windows '95 & NT, DEC OSF,
NCSA, and others.

A similar change was incorporated in the non-crypto version of telnetd.

This resolves bin/771 and bin/1037.

26 years agoPR: bin/771 and bin/1037 are resolved by this change
uhclem [Wed, 8 Oct 1997 03:10:32 +0000 (03:10 +0000)]
PR: bin/771 and bin/1037 are resolved by this change
This change changes the default handling of linemode so that older and/or
stupider telnet clients can still get wakeup characters like <ESC> and
<CTRL>D to work correctly multiple times on the same line, as in csh
"set filec" operations.   It also causes CR and LF characters to be read by
apps in certain terminal modes consistently, as opposed to returning
CR sometimes and LF sometimes, which broke existing apps.  The change
was shown to fix the problem demonstrated in the FreeBSD telnet client,
along with the telnet client in Solaris, SCO, Windows '95 & NT, DEC OSF,
NCSA, and others.

A similar change will be incorporated in the crypto version of telnetd.

This resolves bin/771 and bin/1037.

26 years agoDon't allow the window to be increased beyond what is possible to
fenner [Tue, 7 Oct 1997 21:10:06 +0000 (21:10 +0000)]
Don't allow the window to be increased beyond what is possible to
represent in the TCP header.  The old code did effectively:
win = min(win, MAX_ALLOWED);
win = max(win, what_i_think_i_advertised_last_time);
so if what_i_think_i_advertised_last_time is bigger than can be
represented in the header (e.g. large buffers and no window scaling)
then we stuff a too-big number into a short.  This fix reverses the
order of the comparisons.

PR: kern/4712

26 years agosave errno in sigchld handler
eivind [Tue, 7 Oct 1997 16:43:16 +0000 (16:43 +0000)]
save errno in sigchld handler

Obtained from: Theo deRaadt <theo@openbsd.org> (OpenBSD rev 1.6)

26 years agovsprintf -> vnsprintf paranoia.
eivind [Tue, 7 Oct 1997 16:33:22 +0000 (16:33 +0000)]
vsprintf -> vnsprintf paranoia.

Obtained from: Todd Miller <millert@openbsd.org> (OpenBSD rev 1.3)

26 years agoDon't honor $HOME if >= MAXPATHLEN
eivind [Tue, 7 Oct 1997 16:27:48 +0000 (16:27 +0000)]
Don't honor $HOME if >= MAXPATHLEN

Obtained from: Todd Miller <millert@openbsd.org> (OpenBSD csh.c rev 1.3 and 1.4)

26 years agoRemoved uneeded ${.MAKEFLAGS} again.
bde [Tue, 7 Oct 1997 11:32:05 +0000 (11:32 +0000)]
Removed uneeded ${.MAKEFLAGS} again.
Originally committed in: rev.1.141
Spammed by: rev.1.149 (mouldy parallel make diffs)

26 years ago`numdirtybuffers' was not maintained properly. This caused excessive
bde [Tue, 7 Oct 1997 11:10:18 +0000 (11:10 +0000)]
`numdirtybuffers' was not maintained properly.  This caused excessive
flushing of buffers in an attempt to reduce numdirtybuffers, and
perhaps other problems.

26 years agoPreserve old SC_STATIC value after units exchange
ache [Tue, 7 Oct 1997 09:13:06 +0000 (09:13 +0000)]
Preserve old SC_STATIC value after units exchange