]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
27 years agoAdd back netns
Jordan K. Hubbard [Thu, 17 Oct 1996 18:44:56 +0000 (18:44 +0000)]
Add back netns

27 years agoBring back netns so that Netcon can take over support for it, as agreed.
Jordan K. Hubbard [Thu, 17 Oct 1996 18:42:33 +0000 (18:42 +0000)]
Bring back netns so that Netcon can take over support for it, as agreed.
This was actually easier than I thought - CVS is smart enough to handle
"resurrections" now. :-)

27 years agoWhen freeing buffers in the db routines, also zeroize them
Guido van Rooij [Thu, 17 Oct 1996 18:27:58 +0000 (18:27 +0000)]
When freeing buffers in the db routines, also zeroize them
This should solve the bug where a coredumping ftpd reveals
encrypted passwords.
Obtained from: OpenBSD

27 years agoI'm not sure why, but Netcon's TFS filesystem code doesn't want to
Jordan K. Hubbard [Thu, 17 Oct 1996 17:56:07 +0000 (17:56 +0000)]
I'm not sure why, but Netcon's TFS filesystem code doesn't want to
add free vnodes back to the freelist.  They must do their own vnode
management.  Anyway, this change is *only* activated with their filesystem
and doesn't affect anyone else.  Whoops, forgot the submitted-by lines
in my previous commits too.. :-(
Submitted-By: Tony Ardolino <tony@netcon.com>
27 years agoadd km for syscons so folks with meta maps can take advantage of them
Paul Traina [Thu, 17 Oct 1996 17:50:15 +0000 (17:50 +0000)]
add km for syscons so folks with meta maps can take advantage of them

27 years agoFix xterm termcap entry to support km and new xterms
Paul Traina [Thu, 17 Oct 1996 17:47:28 +0000 (17:47 +0000)]
Fix xterm termcap entry to support km and new xterms

27 years agoOops, fix my previous commit, now tell user his s/key parameters
Andrey A. Chernov [Thu, 17 Oct 1996 17:46:00 +0000 (17:46 +0000)]
Oops, fix my previous commit, now tell user his s/key parameters

27 years agoAdded missing extern declaration of timer_freq.
Bruce Evans [Thu, 17 Oct 1996 17:31:25 +0000 (17:31 +0000)]
Added missing extern declaration of timer_freq.
Sorted declarations of scalars.

27 years agoNetcon's changes for their extended NS support. This only effects
Jordan K. Hubbard [Thu, 17 Oct 1996 17:16:58 +0000 (17:16 +0000)]
Netcon's changes for their extended NS support.  This only effects
people compiling with NS, so the effects on everyone else are nil.

27 years agoSome very small changes to support Netcon's TFS filesystem.
Jordan K. Hubbard [Thu, 17 Oct 1996 17:12:04 +0000 (17:12 +0000)]
Some very small changes to support Netcon's TFS filesystem.
These patches were formerly applied by the Netcon installer
before rebuilding your kernel.

27 years agoDon't ever ask for password if it is impossible to confirm it
Andrey A. Chernov [Thu, 17 Oct 1996 17:06:04 +0000 (17:06 +0000)]
Don't ever ask for password if it is impossible to confirm it
It happens if 1) regular passwords not allowed, 2) skey database
not activated for given user.
Under some rare circumstanes skey_challenge can return empty
diagnostic or even previous buffer, fix it.

27 years agoChange "nope" to ":"
Andrey A. Chernov [Thu, 17 Oct 1996 16:47:22 +0000 (16:47 +0000)]
Change "nope" to ":"
Previous variant not work well, if you have a user with name nope

27 years agoAdded Igor Vinokurov <igor@zynaps.ru>, the maintainer of trafshow port
Masafumi Max NAKANE [Thu, 17 Oct 1996 16:42:12 +0000 (16:42 +0000)]
Added Igor Vinokurov <igor@zynaps.ru>, the maintainer of trafshow port
to the additional contributor list.

27 years agoAdd -D_SKEY_INTERNAL to pick up non-API functions
Andrey A. Chernov [Thu, 17 Oct 1996 15:30:21 +0000 (15:30 +0000)]
Add -D_SKEY_INTERNAL to pick up non-API functions

27 years agoAdd -D_SKEY_INTERNAL to pick up non-API skey functions
Andrey A. Chernov [Thu, 17 Oct 1996 15:25:12 +0000 (15:25 +0000)]
Add -D_SKEY_INTERNAL to pick up non-API skey functions

27 years agoPrevent namespace pollution cause by this library by splitting to
Andrey A. Chernov [Thu, 17 Oct 1996 15:22:46 +0000 (15:22 +0000)]
Prevent namespace pollution cause by this library by splitting to
API and non-API functions and giving _sk_ prefix for all non-API
functions. Old names will be available by _SKEY_INTERNAL define
it allows not change anything in keyinit & key, only -D_SKEY_INTERNAL required.
Staticise some things.

27 years agoIf 1) regular Unix passwords not permitted and 2) skey database
Andrey A. Chernov [Thu, 17 Oct 1996 13:43:58 +0000 (13:43 +0000)]
If 1) regular Unix passwords not permitted and 2) skey database
for this user not activated, there no correct password exist,
so don't bother ask for it and return "" immediately

27 years agoThis patch attempts to make the pccard probe in if_ed.c work with
Nate Williams [Thu, 17 Oct 1996 13:42:13 +0000 (13:42 +0000)]
This patch attempts to make the pccard probe in if_ed.c work with
non-memory-mapped cards.

Submitted by: Michael Smith <msmith@atrad.adelaide.edu.au>
Reviewed by: phk,dg

27 years agoMake it work in localized environment, replace strftime's %b which can be
Andrey A. Chernov [Thu, 17 Oct 1996 12:45:39 +0000 (12:45 +0000)]
Make it work in localized environment, replace strftime's %b which can be
in national form with months names table

27 years agoFix diagnostic: skey->key, we have key command
Andrey A. Chernov [Thu, 17 Oct 1996 12:06:05 +0000 (12:06 +0000)]
Fix diagnostic: skey->key, we have key command

27 years agoImport NFSv3 support from NetBSD
Doug Rabson [Thu, 17 Oct 1996 10:53:34 +0000 (10:53 +0000)]
Import NFSv3 support from NetBSD

Obtained from: NetBSD

27 years agoFix a problem that could cause msync (or many other things) to deadlock.
John Dyson [Thu, 17 Oct 1996 03:04:43 +0000 (03:04 +0000)]
Fix a problem that could cause msync (or many other things) to deadlock.
The heuristic for managment of memory backing the buffer cache was
nice, but didn't work due to some architectural problems.  Simplify
and improve the algorithm.

27 years agoMake processes waken up eligible for immediate swap-in.
John Dyson [Thu, 17 Oct 1996 02:58:20 +0000 (02:58 +0000)]
Make processes waken up eligible for immediate swap-in.

27 years agoClean up the rundown of the object backing a vnode. This should fix
John Dyson [Thu, 17 Oct 1996 02:49:35 +0000 (02:49 +0000)]
Clean up the rundown of the object backing a vnode.  This should fix
NFS problems associated with forcible dismounts.

27 years agoIssue a warning if the user specifies an invalid interface in a rule.
Alexander Langer [Thu, 17 Oct 1996 01:05:03 +0000 (01:05 +0000)]
Issue a warning if the user specifies an invalid interface in a rule.
The rule is still added to the chain since the interface may get
created later on after loading an LKM.

27 years agoCorrently null-terminate the path being passed to the opendir() calls,
Peter Wemm [Thu, 17 Oct 1996 00:40:04 +0000 (00:40 +0000)]
Corrently null-terminate the path being passed to the opendir() calls,
malloc() does is not defined to return a zeroed buffer leading to
"strange" problems.

Submitted by: Karl Denninger <karl@mcs.com>, PR#1826

27 years agoactually impliment permissions for teh chown/chmod/etc functions
Julian Elischer [Wed, 16 Oct 1996 23:14:00 +0000 (23:14 +0000)]
actually impliment permissions for teh chown/chmod/etc functions
that resemble what the man pages say.
also fix braino (type last week)

27 years agoadd a man page for suser() because its usage is not as obvious
Julian Elischer [Wed, 16 Oct 1996 21:14:32 +0000 (21:14 +0000)]
add a man page for suser() because its usage is not as obvious
as you would first think.

27 years agoUse a (signed) int32_t counter instead of an `unsigned int' counter
Bruce Evans [Wed, 16 Oct 1996 21:02:49 +0000 (21:02 +0000)]
Use a (signed) int32_t counter instead of an `unsigned int' counter
for the GPROF4 case.  This allows a simpler method to be used for
non-statistical profiling (it allows overhead adjustments to be
subtracted from one counter without harm if that counter goes
negative; otherwise the adjustment would have to be distributed).

32 bit counters were already too small for GPROF4 with a 200MHz
clock.  int64_t counters should be used.

27 years agoOops forgot to remove a debug printf.
Søren Schmidt [Wed, 16 Oct 1996 20:59:08 +0000 (20:59 +0000)]
Oops forgot to remove a debug printf.

27 years agoBring back ns_* routines; we need them for ifconfig and route.
Jordan K. Hubbard [Wed, 16 Oct 1996 19:38:51 +0000 (19:38 +0000)]
Bring back ns_* routines; we need them for ifconfig and route.

27 years agoBring back NS support for this utility - we need it for the commercial Novell
Jordan K. Hubbard [Wed, 16 Oct 1996 19:37:58 +0000 (19:37 +0000)]
Bring back NS support for this utility - we need it for the commercial Novell
server software.

27 years agoFixed miscounting for non-statistical (GUPROF) profiling:
Bruce Evans [Wed, 16 Oct 1996 18:37:58 +0000 (18:37 +0000)]
Fixed miscounting for non-statistical (GUPROF) profiling:
- use CROSSJUMP() and CROSSJUMP_LABEL() for conditional jumps from idle()
  into cpu_switch() and vice versa.
- moved badsw code to after cpu_switch().

Cosmetic changes:
- moved sw0 string to be immediately after its caller (badsw).
- removed unused #include.

27 years agoUtility to "brand" ELF binaries.
Søren Schmidt [Wed, 16 Oct 1996 18:16:22 +0000 (18:16 +0000)]
Utility to "brand" ELF binaries.
Man page to come...
For now use: <brandelf -t Linux linuxbin> to brand, and just
<brandelf> to verify branding on a ELF file. FreeBSD native is
set with <brandelf -t FreeBSD freebsdbin>.

27 years agoAdded macros CROSSJUMP(), CROSSJUMP_LABEL() and GPROF_RET. These will
Bruce Evans [Wed, 16 Oct 1996 18:13:25 +0000 (18:13 +0000)]
Added macros CROSSJUMP(), CROSSJUMP_LABEL() and GPROF_RET.  These will
be used to fix some benign(?) bugs in GUPROF profiling.

Fixed stale comments and long lines.

27 years agoAllow the utimes call to work on devfs nodes.
Julian Elischer [Wed, 16 Oct 1996 18:02:53 +0000 (18:02 +0000)]
Allow the utimes call to work on devfs nodes.

27 years agoPrepare kernel to take advantage of "branded" ELF binaries.
Søren Schmidt [Wed, 16 Oct 1996 17:51:08 +0000 (17:51 +0000)]
Prepare kernel to take advantage of "branded" ELF binaries.

27 years agoUse newly defined entity for the ``Last modified by'' line.
Masafumi Max NAKANE [Wed, 16 Oct 1996 12:28:37 +0000 (12:28 +0000)]
Use newly defined entity for the ``Last modified by'' line.

27 years agoAdded entity for Bill Lloyd, the author of the latest isdn.sgml.
Masafumi Max NAKANE [Wed, 16 Oct 1996 12:06:31 +0000 (12:06 +0000)]
Added entity for Bill Lloyd, the author of the latest isdn.sgml.
And put it in the list of additional contributor.

27 years agoAdded following people to the additional contributor list:
Masafumi Max NAKANE [Wed, 16 Oct 1996 05:59:27 +0000 (05:59 +0000)]
Added following people to the additional contributor list:

Dmitry Kohmanyuk <dk@farm.org>
Lars Koeller <Lars_Koeller@odie.physik2.uni-rostock.de>
Michael Searle <searle@longacre.demon.co.uk>
MOROHOSHI Akihiko <moro@race.u-tokyo.ac.jp>
Ronald Kuehn <kuehn@rz.tu-clausthal.de>
Ville Eerola <ve@sci.fi>
Yukihiro Nakai <nakai@mlab.t.u-tokyo.ac.jp>

Updated following people's addresses:
Brian Clapper, Hideaki Ohmon, Jian-Da Li, Jim Lowe

Removed &a.andreas; as he's listed in the developpers section.

27 years agoSimplify and improve parser.
Andrey A. Chernov [Wed, 16 Oct 1996 03:12:22 +0000 (03:12 +0000)]
Simplify and improve parser.
Clarify manpage.

27 years agoBenign | --> || correction.
Steve Price [Wed, 16 Oct 1996 02:30:39 +0000 (02:30 +0000)]
Benign | --> || correction.

27 years agoAdd support for embedded operation withou console
Julian Elischer [Wed, 16 Oct 1996 00:19:40 +0000 (00:19 +0000)]
Add support for embedded operation withou console
The boot.c patch is applied only to teh JULIAN_HACK branch
the muted console is controlable by a sysctl variable kern.consmute

27 years agoSave half of LC_COLLATE space
Andrey A. Chernov [Tue, 15 Oct 1996 22:05:18 +0000 (22:05 +0000)]
Save half of LC_COLLATE space
ASCIIfy data files, use RFC1345 symbolic names instead of hex codes
Improve parser and manpage

27 years agoSave half of space in LC_COLLATE and remove unneded code.
Andrey A. Chernov [Tue, 15 Oct 1996 21:53:23 +0000 (21:53 +0000)]
Save half of space in LC_COLLATE and remove unneded code.
This change is not compatible with previous variant, however proper
error code returned in both cases.
Colldef changes will follows.

27 years agoCorrected the setting of winsize on open().
Søren Schmidt [Tue, 15 Oct 1996 20:27:07 +0000 (20:27 +0000)]
Corrected the setting of winsize on open().
Problem pointed out by David Nugent - davidn@blaze.net.au.

27 years agoPreserve file flags in accept(2).
Garrett Wollman [Tue, 15 Oct 1996 19:28:44 +0000 (19:28 +0000)]
Preserve file flags in accept(2).

Submitted by: fredriks@mcs.com in PR#1775 (this implmentaion is different)

27 years agoUpdated #includes to 4.4lite style.
Bruce Evans [Tue, 15 Oct 1996 19:22:46 +0000 (19:22 +0000)]
Updated #includes to 4.4lite style.

27 years agoWow!, i finally found a usable vgrind definition for Perl.
Joerg Wunsch [Tue, 15 Oct 1996 18:58:45 +0000 (18:58 +0000)]
Wow!, i finally found a usable vgrind definition for Perl.

27 years agoRemoved nested include of <sys/proc.h> from <vm/vm_object.h> and fixed
Bruce Evans [Tue, 15 Oct 1996 18:24:34 +0000 (18:24 +0000)]
Removed nested include of <sys/proc.h> from <vm/vm_object.h> and fixed
the one place that depended on it.  wakeup() is now prototyped in
<sys/systm.h> so that it is normally visible.

Added nested include of <sys/queue.h> in <vm/vm_object.h>.  The queue
macros are a more fundamental prerequisite for <vm/vm_object.h> than
the wakeup prototype and previously happened to be included by
namespace pollution from <sys/proc.h> or elsewhere.

27 years agoMondo-merge of new 2.2 installation technology.
Jordan K. Hubbard [Tue, 15 Oct 1996 17:04:56 +0000 (17:04 +0000)]
Mondo-merge of new 2.2 installation technology.

27 years agoForward-declared `struct route' for the KERNEL case so that <net/route.h>
Bruce Evans [Tue, 15 Oct 1996 16:54:47 +0000 (16:54 +0000)]
Forward-declared `struct route' for the KERNEL case so that <net/route.h>
isn't a prerequisite.

Fixed style of ifdefs.

27 years agoSynchronized with biosboot/Makefile:
Bruce Evans [Tue, 15 Oct 1996 16:26:05 +0000 (16:26 +0000)]
Synchronized with biosboot/Makefile:
- use gcc-2.7.2 alignment features
- get CWARNFLAGS form <bsd.kern.mk>
- fixed whitespace differences.

27 years agoMove much of the machine dependent code from vm_glue.c into
John Dyson [Tue, 15 Oct 1996 03:16:45 +0000 (03:16 +0000)]
Move much of the machine dependent code from vm_glue.c into
pmap.c.  Along with the improved organization, small proc fork
performance is now about 5%-10% faster.

27 years agoSet the effective user id to that of the real user id when
Alexander Langer [Tue, 15 Oct 1996 01:53:48 +0000 (01:53 +0000)]
Set the effective user id to that of the real user id when
saving/restoring a game (necessary when adventure is invoked via dm).

27 years agoFix my yesterday's brain-o: don't account for the kilobytes twice.
Joerg Wunsch [Mon, 14 Oct 1996 22:56:01 +0000 (22:56 +0000)]
Fix my yesterday's brain-o: don't account for the kilobytes twice.

27 years agoDon't casually blow away FTP site settings now that it's easy for user
Jordan K. Hubbard [Mon, 14 Oct 1996 21:50:38 +0000 (21:50 +0000)]
Don't casually blow away FTP site settings now that it's easy for user
to override them.

27 years agoFixed setting of the keyboard LEDs. For some keyboards and/or keyboard
Bruce Evans [Mon, 14 Oct 1996 21:44:05 +0000 (21:44 +0000)]
Fixed setting of the keyboard LEDs.  For some keyboards and/or keyboard
controllers, it is an error to issue a command before the keyboard
has had time to reply to the previous command.  Setting the LEDs
involves issueing 2 commands, so it never worked on these keyboards.

Fixed resetting of keyboard.  It is possible for unprocessed
scancodes to be present when the reset routine is called.  This
usually occurs after switching from one console driver to another
in userconfig.  pcvt and syscons attempt to flush any stale scancodes,
but sometimes fail to do so because keyboard and/or keyboard
controller takes a long time to send the scancodes after reset
(scancodes are apparently not flushed by reset!).  syscons handles
this later by not checking for errors at strategic places, but pcvt
was confused.

Use an impossible initial and failure mode setting for the LEDs
so that the LEDs always get set if they are possibly out of sync.

Added missing spltty() in update_led().

27 years agoRemove the now-unusable ports configuration code. Swapping CDs is bogus.
Jordan K. Hubbard [Mon, 14 Oct 1996 21:32:39 +0000 (21:32 +0000)]
Remove the now-unusable ports configuration code.  Swapping CDs is bogus.

27 years agoChange the boot-time menu.
Jordan K. Hubbard [Mon, 14 Oct 1996 21:19:43 +0000 (21:19 +0000)]
Change the boot-time menu.

27 years agoFix PR#1557 - chained package dependencies using URLs didn't work.
Jordan K. Hubbard [Mon, 14 Oct 1996 19:41:44 +0000 (19:41 +0000)]
Fix PR#1557 - chained package dependencies using URLs didn't work.

27 years agocatch a mismatch after final review...
Paul Traina [Mon, 14 Oct 1996 17:27:53 +0000 (17:27 +0000)]
catch a mismatch after final review...

27 years agoExpose the control mechanism for serial console boot so that the default
Paul Traina [Mon, 14 Oct 1996 17:25:53 +0000 (17:25 +0000)]
Expose the control mechanism for serial console boot so that the default
shipped with freebsd can be changed without modifying the Makefiles directly.

Creates: BOOT_FORCE_COMCONSOLE
BOOT_PROBE_KEYBOARD
BOOT_PROBE_KEYBOARD_LOCK
BOOT_COMCONSOLE (port value for console)

27 years agoFixed a bug that got introduced when I changed the CY16_RESET and
David Greenman [Mon, 14 Oct 1996 16:43:11 +0000 (16:43 +0000)]
Fixed a bug that got introduced when I changed the CY16_RESET and
CY_CLEAR_INTR definitions. This might have affected probing of ISA
versions of the Cyclom 16/32-Y.

27 years agoMake this compile again for the CyDebug case.
David Greenman [Mon, 14 Oct 1996 16:10:00 +0000 (16:10 +0000)]
Make this compile again for the CyDebug case.

27 years agoFixed a minor typo.
Masafumi Max NAKANE [Mon, 14 Oct 1996 14:48:02 +0000 (14:48 +0000)]
Fixed a minor typo.

27 years agoAdjust the boot-time userconfig options to match reality.
Jordan K. Hubbard [Mon, 14 Oct 1996 13:29:40 +0000 (13:29 +0000)]
Adjust the boot-time userconfig options to match reality.

27 years agopci_map_mem() did a too restrictive check on the mapping type:
Stefan Eßer [Mon, 14 Oct 1996 13:04:34 +0000 (13:04 +0000)]
pci_map_mem() did a too restrictive check on the mapping type:
PCI_MAP_MEMORY_TYPE_32BIT_1M should be accepted as well as
PCI_MAP_MEMORY_TYPE_32BIT (and now is).
(Problem reported by David Greenman.)

27 years agooops, how did this get into the reinstall target? :-]
Peter Wemm [Mon, 14 Oct 1996 12:58:47 +0000 (12:58 +0000)]
oops, how did this get into the reinstall target? :-]
Must have been a cut/paste foulup...  The whole idea of 'make reinstall'
is that it doesn't build anything from the source server mount.

27 years agoOops, PROBE_KEYBOARD_LOCK wasn't supposed to be enabled by default.
Bruce Evans [Mon, 14 Oct 1996 12:37:47 +0000 (12:37 +0000)]
Oops, PROBE_KEYBOARD_LOCK wasn't supposed to be enabled by default.

27 years agoMove the initialization of np->ns_sync and ns_async out of ncr_getclock(),
Stefan Eßer [Mon, 14 Oct 1996 10:09:52 +0000 (10:09 +0000)]
Move the initialization of np->ns_sync and ns_async out of ncr_getclock(),
which does no longer get called for non-Ultra cards.
Fix suggested by Gerard Roudier, slightly modified by me.

27 years agoA couple of minor nits.
Poul-Henning Kamp [Mon, 14 Oct 1996 08:34:02 +0000 (08:34 +0000)]
A couple of minor nits.

Submitted by: Craig Leres <leres@ee.lbl.gov>
Submitted by: Theo Deraadt <deraadt@theos.com>

27 years agoYikes! Originally, i intended to apply the patch from PR # 1322,
Joerg Wunsch [Sun, 13 Oct 1996 20:55:45 +0000 (20:55 +0000)]
Yikes!  Originally, i intended to apply the patch from PR # 1322,
but Bill has beaten me on this. ;-)

However, he missed the part to compute the kernelsize in kilobytes,
so the minfree consideration was now overcautious.  (I've also
changed the return type of dump_size() to void since int was useless.)

Being here, the fact that `vmcore' was written world-readable was just
a plain security hole: everybody who was able to crash a kernel could
later read any confidential information out of it at his will.  Create
it with umask 077 instead.

27 years agoThe dos() function needs a new second argument, containing the size
Bruce Evans [Sun, 13 Oct 1996 18:18:50 +0000 (18:18 +0000)]
The dos() function needs a new second argument, containing the size
of the partition. Only if the size is 0 should the
special handling of 0 as first argument be triggered.

[This bug caused offset 0 to give C/H/S = 0/0/0 instead of 0/0/1.]

The init_sector0 function needs to decrease the first argument
to the second call to dos() by one to be consistent with the
calls to dos() in change_part().

[This bug caused fdisk -i to create bogus partition tables with
the ending C/H/S value 1 too high.  This usually gives S = 1
instead of S = maximum, so the geometry guessing in the slice
code and perhaps in SCSI BIOSes was defeated.]

Submitted by: Tor Egge <tegge@itea.ntnu.no>

27 years agoMake the savecore command work like the man page says:
Bill Fenner [Sun, 13 Oct 1996 18:12:20 +0000 (18:12 +0000)]
Make the savecore command work like the man page says:
- make minfree work by getting the dump size before checking to see
  if the dump will fit on the filesystem
- also fail to dump if no minfree is specified but there are not enough
  free blocks.

Fix a typo in the man page.

Fixes PR bin/1322

Submitted by: "Philippe C." <charnier@lirmm.fr>

27 years agoBuild/install c++filt from g++, we already have all the sources and
Peter Wemm [Sun, 13 Oct 1996 17:37:03 +0000 (17:37 +0000)]
Build/install c++filt from g++, we already have all the sources and
it's and useful. (Ever tried to read 'nm' and 'ldd -v' output on a c++
object or library? :-) This filter decodes the mangled symbol names.)

Requested by: Chuck Robey <chuckr@glue.umd.edu>

27 years agoUncomment the "raw" lp entry in /etc/printcap. We start lpd by
Joerg Wunsch [Sun, 13 Oct 1996 16:52:40 +0000 (16:52 +0000)]
Uncomment the "raw" lp entry in /etc/printcap.  We start lpd by
default, so there's no use in running it without any printer
definition in printcap.  Also added a bunch of hints about the printer
setup, to guide the admin about the printer setup (handbook,
"apsfilter"), and a commented-out sample setup for a remote printer.

In the same line, add /var/spool/lpd/output to BSD.var.dist since it
is referred to by the "lp" entry in printcap.

27 years agoFixed the security leek I introduced the other day, now
Søren Schmidt [Sun, 13 Oct 1996 15:05:14 +0000 (15:05 +0000)]
Fixed the security leek I introduced the other day, now
shell command is only executabel from command files, not
from the commandline.

27 years agoRemove some old compatibility names.
Poul-Henning Kamp [Sun, 13 Oct 1996 14:36:37 +0000 (14:36 +0000)]
Remove some old compatibility names.

27 years agoStop using old compatibility names from buf.h
Poul-Henning Kamp [Sun, 13 Oct 1996 14:36:06 +0000 (14:36 +0000)]
Stop using old compatibility names from buf.h

27 years agoRemove a stale comment.
Poul-Henning Kamp [Sun, 13 Oct 1996 07:16:50 +0000 (07:16 +0000)]
Remove a stale comment.

27 years agoPmap_resident_count was mistakenly removed from pmap.h, thereby
John Dyson [Sun, 13 Oct 1996 03:14:57 +0000 (03:14 +0000)]
Pmap_resident_count was mistakenly removed from pmap.h, thereby
disabling the RSS listing in ps and ^T.  This commit re-inserts
the macro defn.

27 years agoAdjust release notes for upcoming re-SNAP.
Jordan K. Hubbard [Sun, 13 Oct 1996 01:53:24 +0000 (01:53 +0000)]
Adjust release notes for upcoming re-SNAP.

27 years ago8-Bit character support.
Wolfram Schneider [Sun, 13 Oct 1996 01:44:43 +0000 (01:44 +0000)]
8-Bit character support.

Old locate(1) programs still works with the new database format, print
some garbage for 8 bit characters, but don't core (maybe except char 30).

7-Bit Puritan should not notice any difference. Same speed,
Same database size if the database contain only ASCII characters.

Reviewed by: ache

27 years agoMinor optimization for final rundown of a pmap.
John Dyson [Sun, 13 Oct 1996 01:38:37 +0000 (01:38 +0000)]
Minor optimization for final rundown of a pmap.

27 years agoChange DEVFS device naming convention for cuac*, ttyc*.
David Greenman [Sun, 13 Oct 1996 01:30:32 +0000 (01:30 +0000)]
Change DEVFS device naming convention for cuac*, ttyc*.

27 years agoChanged ttyc? and cuac? specifications to support multiple controllers.
David Greenman [Sun, 13 Oct 1996 01:12:01 +0000 (01:12 +0000)]
Changed ttyc? and cuac? specifications to support multiple controllers.
A "MAKEDEV ttyc0" now creates 32 units for controller 0, ttyc1 for
controller 1, etc..

27 years agoChanges to add support for the PCI version of the Cyclades Cyclom-Y
David Greenman [Sun, 13 Oct 1996 01:09:24 +0000 (01:09 +0000)]
Changes to add support for the PCI version of the Cyclades Cyclom-Y
serial adapter, and support for multiple Cyclom controllers.

27 years agoBe less draconian about distribution extraction errors - still write
Jordan K. Hubbard [Sat, 12 Oct 1996 23:48:33 +0000 (23:48 +0000)]
Be less draconian about distribution extraction errors - still write
/etc file changes out.

27 years agoFixed lblktosize(). It overflowed at 2G. This bug only affected
Bruce Evans [Sat, 12 Oct 1996 22:12:51 +0000 (22:12 +0000)]
Fixed lblktosize().  It overflowed at 2G.  This bug only affected
ufs_read() and ufs_write().

Found by: looking at warnings for comparing the result of lblktosize()
(which is usually daddr_t = long) with file sizes (which are u_quad_t
for ufs).  File sizes should probably be off_t's to avoid warnings
when the are compared with file offsets, so the fixed lblktosize()
casts to off_t instead of u_quad_t.

Added definition of smalllblksize().  It is the same as the old
lblksize() and is more efficient for small block numbers on 32-bit
machines.

Use smalllblktosize() instead of its expansion in blksize() and
dblksize().  This keeps the line length short and makes it more
obvious that the shift can't overflow.

27 years agoUndo a mistaken change from splhigh to splvm. Subr_rlist has been
John Dyson [Sat, 12 Oct 1996 21:37:59 +0000 (21:37 +0000)]
Undo a mistaken change from splhigh to splvm.  Subr_rlist has been
capable of being used for things other than swap space allocation,
and splvm would have been appropriate for only swap space allocation
and other VM things.  My commit broke that (and was actually a mistake.)

27 years agoPerformance optimizations. One of which was meant to go in before the
John Dyson [Sat, 12 Oct 1996 21:35:25 +0000 (21:35 +0000)]
Performance optimizations.  One of which was meant to go in before the
previous snap.  Specifically, kern_exit and kern_exec now makes a
call into the pmap module to do a very fast removal of pages from the
address space.  Additionally, the pmap module now updates the PG_MAPPED
and PG_WRITABLE flags.  This is an optional optimization, but helpful
on the X86.

27 years agoCleaned up:
Bruce Evans [Sat, 12 Oct 1996 20:36:15 +0000 (20:36 +0000)]
Cleaned up:
- fixed a sloppy common-style declaration.
- removed an unused macro.
- moved once-used macros to the one file where they are used.
- removed unused forward struct declarations.
- removed __pure.
- declared inline functions as inline in their prototype as well
  as in theire definition (gcc unfortunately allows the prototype
  to be inconsistent).
- staticized.

27 years agoexport variable TMPDIR (which maybe set, but not exported in /etc/locate.rc)
Wolfram Schneider [Sat, 12 Oct 1996 20:29:25 +0000 (20:29 +0000)]
export variable TMPDIR (which maybe set, but not exported in /etc/locate.rc)

27 years agoGet the pysical address for id_maddr when opening an -incore kernel.
Eric L. Hernes [Sat, 12 Oct 1996 20:23:26 +0000 (20:23 +0000)]
Get the pysical address for id_maddr when opening an -incore kernel.
This fixes the kernel panic when propagating userconfig changes to
arbitrary kernels.

Remove obsoleted `#include <tcl.h>' added a few <stdio.h> where
necessary.

Fix getting scsi bus information from an -incore kernel.

Turned on SAVE_USERCONFIG by default.

27 years agoRemoved __pure's and __pure2's. __pure is a no-op for recent versions
Bruce Evans [Sat, 12 Oct 1996 20:09:48 +0000 (20:09 +0000)]
Removed __pure's and __pure2's.  __pure is a no-op for recent versions
of gcc by definition, and __pure2 is a no-op in effect (presumably the
compiler can see when an inline function has no side effects).

27 years agoRemoved nested include if <sys/socket.h> from <net/if.h> and
Bruce Evans [Sat, 12 Oct 1996 19:49:43 +0000 (19:49 +0000)]
Removed nested include if <sys/socket.h> from <net/if.h> and
<net/if_arp.h> and fixed the things that depended on it.  The nested
include just allowed unportable programs to compile and made my
simple #include checking program report that networking code doesn't
need to include <sys/socket.h>.

27 years agoLog the interface name which received the packet.
Alexander Langer [Sat, 12 Oct 1996 19:38:50 +0000 (19:38 +0000)]
Log the interface name which received the packet.

Suggested by: Hal Snyder <hsndyer@thoughtport.com>

27 years agoSome cosmetic tweaks, attempt to fix package wait problem by checking
Jordan K. Hubbard [Sat, 12 Oct 1996 19:30:26 +0000 (19:30 +0000)]
Some cosmetic tweaks, attempt to fix package wait problem by checking
for error return as well (not sure if this is it, but it's one possible
explanation).

27 years agoClose PR#1781 - time returns with EXIT_SUCCESS on some failures
Jordan K. Hubbard [Sat, 12 Oct 1996 18:41:26 +0000 (18:41 +0000)]
Close PR#1781 - time returns with EXIT_SUCCESS on some failures
Submitted-By: Tor Egge <Tor.Egge@idt.ntnu.no>