]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
27 years agoFixed some bugs in BREAK handling. If BRKINT is set, then always flush
Bruce Evans [Fri, 29 Nov 1996 15:23:42 +0000 (15:23 +0000)]
Fixed some bugs in BREAK handling.  If BRKINT is set, then always flush
the queues and generate a SIGINT.  Previously, this wasn't done if ISIG
was clear or the VINTR character was disabled, and it was done by
converting the BREAK to a VINTR character and sometimes bogusly echoing
this character.

Found by: NIST-PCTS

27 years agoFixed handling of non-POSIX control characters. They must not do
Bruce Evans [Fri, 29 Nov 1996 15:06:17 +0000 (15:06 +0000)]
Fixed handling of non-POSIX control characters.  They must not do
anything special unless IEXTEN is set.

Found by: NIST-PCTS

27 years agoChanged _POSIX_VERSION from 198808L to 199009L. We're close to conforming
Bruce Evans [Fri, 29 Nov 1996 14:48:17 +0000 (14:48 +0000)]
Changed _POSIX_VERSION from 198808L to 199009L.  We're close to conforming
to the 1990 version according to NIST-PCTS.  Many man pages still refer to
the 1988 version.

27 years agoFixed EFAULT handling in i586_copyin() and i586_copyout(). Use a
Bruce Evans [Fri, 29 Nov 1996 14:32:30 +0000 (14:32 +0000)]
Fixed EFAULT handling in i586_copyin() and i586_copyout().  Use a
consistent stack frame in fastmove() so that only one new fault handler
is necessary.

Should be in 2.2.  Harmless until the i586 versions are reenabled.

27 years agoDon't print bootinfo.bi_n_bios_used in cpu_startup() since it is always
Bruce Evans [Fri, 29 Nov 1996 13:19:26 +0000 (13:19 +0000)]
Don't print bootinfo.bi_n_bios_used in cpu_startup() since it is always
zero because no drivers have had a chance to change it.

27 years agoDon't clobber the SIGCONT bit in the signal mask in sigreturn(). Use
Bruce Evans [Fri, 29 Nov 1996 13:12:14 +0000 (13:12 +0000)]
Don't clobber the SIGCONT bit in the signal mask in sigreturn().  Use
the `sigcantmask' macro to get the correct set of unmaskable signals.

Found by: NIST-PCTS.

27 years ago${COPY} -> -c. ${COPY} should never be used for installing sources since
Bruce Evans [Fri, 29 Nov 1996 12:54:10 +0000 (12:54 +0000)]
${COPY} -> -c.  ${COPY} should never be used for installing sources since
it might be empty.

27 years agoCorrect name in probe type. Closes PR# 2122 (with a change, since the PR
Jordan K. Hubbard [Fri, 29 Nov 1996 12:26:30 +0000 (12:26 +0000)]
Correct name in probe type.  Closes PR# 2122 (with a change, since the PR
did not actually use the correct name either :-).
Submitted-By: Toyonori Fujiura <toyo@exiv.pearnet.org>
27 years agoAdd Sander Vesik for his tkcron port.
Thomas Gellekum [Fri, 29 Nov 1996 08:06:48 +0000 (08:06 +0000)]
Add Sander Vesik for his tkcron port.

27 years agoMore involved description for fe0 driver using cards. Closes PR#2099.
Jordan K. Hubbard [Fri, 29 Nov 1996 07:19:57 +0000 (07:19 +0000)]
More involved description for fe0 driver using cards.  Closes PR#2099.
Submitted-By: seki@sysrap.cs.fujitsu.co.jp
27 years agoClarified the comment about removing other CPU defs. Specifically,
John Dyson [Fri, 29 Nov 1996 07:08:48 +0000 (07:08 +0000)]
Clarified the comment about removing other CPU defs.  Specifically,
I added the suggestion to remove the I386_CPU def if possible.

27 years agoSupport the appropriate use of bswap instruction on non-I386 builds.
John Dyson [Fri, 29 Nov 1996 07:04:03 +0000 (07:04 +0000)]
Support the appropriate use of bswap instruction on non-I386 builds.
Per Wayne Scott of Intel, the old sequence took 20cycles!!! on a P6.
Another nice side-benefit is that the kernel is about 3K smaller!!!
Submitted by: Wayne Scott <wscott@ichips.intel.com>

27 years agoAdd a new feature which Garrett pointed out.
Jordan K. Hubbard [Fri, 29 Nov 1996 04:55:47 +0000 (04:55 +0000)]
Add a new feature which Garrett pointed out.

27 years agoNote unpleasant interaction between rawrite.exe and various flavors
Jordan K. Hubbard [Fri, 29 Nov 1996 03:33:27 +0000 (03:33 +0000)]
Note unpleasant interaction between rawrite.exe and various flavors
of Windows.

27 years agoMake the kernel smaller with at worst a neutral effect on perf by
John Dyson [Thu, 28 Nov 1996 23:15:07 +0000 (23:15 +0000)]
Make the kernel smaller with at worst a neutral effect on perf by
de-inlining some VM calls.  (Actually, I measured a small improvement.)

27 years agoSome new sections.
John Fieber [Thu, 28 Nov 1996 18:09:30 +0000 (18:09 +0000)]
Some new sections.
Submitted by: William Lloyd <wlloyd@tolstoy.mpd.ca>

27 years agoI broke psm.c and none of the 5 reviewers noticed :-)
Poul-Henning Kamp [Thu, 28 Nov 1996 17:18:56 +0000 (17:18 +0000)]
I broke psm.c and none of the 5 reviewers noticed :-)

27 years agoUpdate to more current reality.
Poul-Henning Kamp [Thu, 28 Nov 1996 09:57:31 +0000 (09:57 +0000)]
Update to more current reality.

Reviewed by: phk
Submitted by: Robin Carey <robin@mailgate.dtc.rankxerox.co.uk>

27 years agoAdd a timeout here, just like in if_ed.c
Poul-Henning Kamp [Thu, 28 Nov 1996 09:54:38 +0000 (09:54 +0000)]
Add a timeout here, just like in if_ed.c

Reviewed by: phk
Submitted by: Luigi Rizzo <luidi@labinfo.iet.unipi.it>

27 years agoRogue entry for DEC DLT2700
Poul-Henning Kamp [Thu, 28 Nov 1996 09:51:03 +0000 (09:51 +0000)]
Rogue entry for DEC DLT2700

Reviewed by: julian & phk
Submitted by: Andre Albsmeier <Andre.Albsmeier@mchp.siemens.de>

27 years agoFew fixit improvements:
Joerg Wunsch [Thu, 28 Nov 1996 08:56:43 +0000 (08:56 +0000)]
Few fixit improvements:

. hint about the usability of the arrow keys,
. export TAPE to an obscure dev node so mt(1) will work correctly.

27 years agoPotentially fix a problem, whereby MSDOSFS can request buffers
John Dyson [Thu, 28 Nov 1996 04:26:04 +0000 (04:26 +0000)]
Potentially fix a problem, whereby MSDOSFS can request buffers
larger than the vfs layer can provide.  We now automatically support
32K clusters if MSDOSFS is installed, and panic if a filesystem tries
to allocate a buffer larger than MAXBSIZE.

This commit is a result of some "prodding" by BDE.

27 years agoPotentially fix a problem, whereby MSDOSFS can request buffers
John Dyson [Thu, 28 Nov 1996 04:07:44 +0000 (04:07 +0000)]
Potentially fix a problem, whereby MSDOSFS can request buffers
larger than the vfs layer can provide.  We now automatically support
32K clusters if MSDOSFS is installed, and panic if a filesystem tries
to allocate a buffer larger than MAXBSIZE.

This commit is a result of some "prodding" by BDE.

27 years agoAdd missing ${COPY} to install options, zone.tab
Andrey A. Chernov [Wed, 27 Nov 1996 23:32:10 +0000 (23:32 +0000)]
Add missing ${COPY} to install options, zone.tab

27 years agoMake intro command a NO-OP if no VISUAL_USERCONFIG
Poul-Henning Kamp [Wed, 27 Nov 1996 22:53:10 +0000 (22:53 +0000)]
Make intro command a NO-OP if no VISUAL_USERCONFIG

27 years agoTry to impreve DD mode.
Poul-Henning Kamp [Wed, 27 Nov 1996 22:52:34 +0000 (22:52 +0000)]
Try to impreve DD mode.
Try to make 4Mb floppies work again.

27 years agoWaste less space.
Poul-Henning Kamp [Wed, 27 Nov 1996 22:52:25 +0000 (22:52 +0000)]
Waste less space.

27 years agoRemove warning at AHC_SCBPAGING_ENABLE, not needed now
Andrey A. Chernov [Wed, 27 Nov 1996 22:52:09 +0000 (22:52 +0000)]
Remove warning at AHC_SCBPAGING_ENABLE, not needed now

27 years agoImprove the Dangerously Dedidcated mode a bit. Not much, but a bit better.
Poul-Henning Kamp [Wed, 27 Nov 1996 22:44:43 +0000 (22:44 +0000)]
Improve the Dangerously Dedidcated mode a bit.  Not much, but a bit better.

27 years agoAdd rcsid[]
Andrey A. Chernov [Wed, 27 Nov 1996 22:30:44 +0000 (22:30 +0000)]
Add rcsid[]

Since locale reading code not resistent against stack overflowing or
similar intruder attacks, don't allow PATH_LOCALE env variable action
for s-bit programs (non-standard locale path setting)

27 years agoDon't show on the screen just securely entered password
Andrey A. Chernov [Wed, 27 Nov 1996 22:04:55 +0000 (22:04 +0000)]
Don't show on the screen just securely entered password
(in /etc/adduser.message text)
Sending password by E-mail on local machine is joke in any case

27 years agoAdd -s flag to always supply mapping if known, ignoring the
Bill Fenner [Wed, 27 Nov 1996 20:45:10 +0000 (20:45 +0000)]
Add -s flag to always supply mapping if known, ignoring the
presence or absence of files in /tftpboot.

27 years agoMake the man page reflect reality. Add BUGS section about DNS.
Bill Fenner [Wed, 27 Nov 1996 20:42:09 +0000 (20:42 +0000)]
Make the man page reflect reality.  Add BUGS section about DNS.

Logging cleanups (including logging the requestor's MAC address instead
 of the server's).

27 years agoSimplified. Some fts related bugs removed. Made less verbose. The default
Stephen McKay [Wed, 27 Nov 1996 13:06:51 +0000 (13:06 +0000)]
Simplified.  Some fts related bugs removed.  Made less verbose.  The default
number of mail messages sent per run was lowered from 2 to 1.  Why?  Well,
some numbers just give you the warm fuzzies, like zero and one.  Zero isn't
much use here, so I picked my all time favourite, one.

27 years agoRemoved unnecessary locking. Simplified. Tidied.
Stephen McKay [Wed, 27 Nov 1996 12:58:44 +0000 (12:58 +0000)]
Removed unnecessary locking.  Simplified.  Tidied.

27 years agoCorrect some typos relating to PC98 server selection.
Jordan K. Hubbard [Wed, 27 Nov 1996 12:44:43 +0000 (12:44 +0000)]
Correct some typos relating to PC98 server selection.
Submitted-By: HOSOKAWA Tatsumi <hosokawa@mt.cs.keio.ac.jp>
27 years agoReflect the change made to the English version (Rev.1.17 -> Rev.1.18.)
Masafumi Max NAKANE [Wed, 27 Nov 1996 04:11:00 +0000 (04:11 +0000)]
Reflect the change made to the English version (Rev.1.17 -> Rev.1.18.)

27 years agoTypo.
Masafumi Max NAKANE [Wed, 27 Nov 1996 03:54:50 +0000 (03:54 +0000)]
Typo.

Some reformatting in a part of the sgml.

27 years agoo Update for upcoming BETA release of 2.2
Jordan K. Hubbard [Wed, 27 Nov 1996 01:44:57 +0000 (01:44 +0000)]
o Update for upcoming BETA release of 2.2
o Fix Adaptec controllers description error, spotted by Richard Wiwatowski
  <rjwiwat@adelaide.on.net>

27 years agoRecommend my new Lieblingcdromlaufwerke, the Plasmon PX-12CS.
Jordan K. Hubbard [Wed, 27 Nov 1996 01:06:40 +0000 (01:06 +0000)]
Recommend my new Lieblingcdromlaufwerke, the Plasmon PX-12CS.

27 years agoOpen devices read-only to test whether they are available. This
Joerg Wunsch [Wed, 27 Nov 1996 01:01:52 +0000 (01:01 +0000)]
Open devices read-only to test whether they are available.  This
allows to use write-protected tapes...

27 years agoChange number of ports to the correct 690 (was 710, but the ports tree
Satoshi Asami [Tue, 26 Nov 1996 12:03:20 +0000 (12:03 +0000)]
Change number of ports to the correct 690 (was 710, but the ports tree
has never had 710 ports!).

27 years agoAdd share/misc. Definite 2.2 candidate, if we want it to be consistent
Satoshi Asami [Tue, 26 Nov 1996 10:01:28 +0000 (10:01 +0000)]
Add share/misc.  Definite 2.2 candidate, if we want it to be consistent
with the ports.

Found by: obrien and his mutt port

27 years agoUse LANG=C, LC_CTYPE=C, LC_TIME=C
Andrey A. Chernov [Tue, 26 Nov 1996 08:58:57 +0000 (08:58 +0000)]
Use LANG=C, LC_CTYPE=C, LC_TIME=C

27 years agoUse LC_TIME=C
Andrey A. Chernov [Tue, 26 Nov 1996 08:53:48 +0000 (08:53 +0000)]
Use LC_TIME=C

27 years agoUse LC_TIME=C
Andrey A. Chernov [Tue, 26 Nov 1996 08:45:28 +0000 (08:45 +0000)]
Use LC_TIME=C

27 years agoAdd Id
Andrey A. Chernov [Tue, 26 Nov 1996 08:00:17 +0000 (08:00 +0000)]
Add Id
Optimize/improve recently added locale restoration on failure

27 years agoMore fixes to Russian catalog
Andrey A. Chernov [Tue, 26 Nov 1996 07:00:11 +0000 (07:00 +0000)]
More fixes to Russian catalog

Should go to 2.2
(wrong number of % formatting found in old version)

27 years agoReplace LC_TIME= with LC_TIME=C
Andrey A. Chernov [Tue, 26 Nov 1996 06:55:35 +0000 (06:55 +0000)]
Replace LC_TIME= with LC_TIME=C

27 years agoUse LC_TIME=C
Andrey A. Chernov [Tue, 26 Nov 1996 03:41:39 +0000 (03:41 +0000)]
Use LC_TIME=C

27 years agoUse LC_TIME=C instead of LC_TIME=
Andrey A. Chernov [Tue, 26 Nov 1996 03:31:26 +0000 (03:31 +0000)]
Use LC_TIME=C instead of LC_TIME=

27 years agoUse LC_TIME=C instead of LC_TIME= (POSIX)
Andrey A. Chernov [Tue, 26 Nov 1996 03:26:31 +0000 (03:26 +0000)]
Use LC_TIME=C instead of LC_TIME= (POSIX)

27 years agoMove PATH_LOCALE handling to setlocale.c, simplify locale path building,
Andrey A. Chernov [Tue, 26 Nov 1996 02:52:52 +0000 (02:52 +0000)]
Move PATH_LOCALE handling to setlocale.c, simplify locale path building,
don't treat empty encoding as C encoding

27 years agoPATH_LOCALE: use this non-standard env variable first time only, i.e.
Andrey A. Chernov [Tue, 26 Nov 1996 02:49:53 +0000 (02:49 +0000)]
PATH_LOCALE: use this non-standard env variable first time only, i.e.
strdup() it to prevent unsetenv() or setenv() effects. Check its length to
not allow user to overflow internal locale buffer. Move PATH_LOCALE
handling code into one place.

POSIX: make better stub for LC_MONETARY & LC_NUMERIC, now it check
locale directory existance instead of refusing all non-C non-POSIX
locales. POSIX treats empty locale env variable as unset variable
while our old code treats it as "C" locale, fix it. Implement previous locale
restoring, if locale setting fails. Old code assumes success if some
of LC_ALL subset is successed even other fails, POSIX treats it as
failure with previous locale restoring, fix it.

Remove unneccessary length checking in currentlocale()

27 years agoDon't close f->f_file on F_TTY types on error returns from ttymsg() since
Peter Wemm [Tue, 26 Nov 1996 02:35:08 +0000 (02:35 +0000)]
Don't close f->f_file on F_TTY types on error returns from ttymsg() since
it wasn't open in the first place..

27 years agoMake the handling of the /dev/ prefix for tty names more consistant.
Peter Wemm [Tue, 26 Nov 1996 02:24:42 +0000 (02:24 +0000)]
Make the handling of the /dev/ prefix for tty names more consistant.
ttymsg() insists on them not being there.

Also, since ttymsg() opens the tty "on demand", don't keep an fd open
ourselves.  This would interfere with HUPCL etc.

This should close PR#2103 from <xaa@stack.nl>

27 years agoLBL chose to use IPPROTO_ENCAP to represent IP proto 4 encapsulation
Bill Fenner [Tue, 26 Nov 1996 01:43:20 +0000 (01:43 +0000)]
LBL chose to use IPPROTO_ENCAP to represent IP proto 4 encapsulation
as performed by the multicast kernel modifications.  However, BSD
already had IPPROTO_ENCAP defined as 98 (RFC 1241 encapsulation).

This changes the use of IPPROTO_ENCAP to IPPROTO_IPIP, which is
the BSD name for IP proto 4.

27 years agoUpdate the French message catalog to use ISO 8859-1 instead of IBM437
Joerg Wunsch [Mon, 25 Nov 1996 22:45:27 +0000 (22:45 +0000)]
Update the French message catalog to use ISO 8859-1 instead of IBM437
characters.

2.2 candidate.

Submitted by: roberto@keltia.freenix.fr (Ollivier Robert)

27 years agobin/1974: core dump from amd when phkmalloc is set to AJ.
Poul-Henning Kamp [Mon, 25 Nov 1996 19:36:26 +0000 (19:36 +0000)]
bin/1974: core dump from amd when phkmalloc is set to AJ.

27 years agoMake a kernel with DDB but without sio possible again. This is only
Poul-Henning Kamp [Mon, 25 Nov 1996 18:39:25 +0000 (18:39 +0000)]
Make a kernel with DDB but without sio possible again.  This is only
a stopgap measure, a more complete solution is on somebodys whiteboard
(and we all know that THAT means :-).

Reviewed by: pst

27 years agoFix the IGMP printer. A lost pair of parenthesis caused all IGMP
Bill Fenner [Mon, 25 Nov 1996 06:20:54 +0000 (06:20 +0000)]
Fix the IGMP printer.  A lost pair of parenthesis caused all IGMP
reports to be printed as IGMPv1, and a misapplied patch caused
IGMP Leave messages to be partially interpreted as DVMRP routing
messages.

27 years agoCorrect the syntax of the example to reflect reality, and note that the
Mike Smith [Mon, 25 Nov 1996 01:52:28 +0000 (01:52 +0000)]
Correct the syntax of the example to reflect reality, and note that the
'host:' element is mandatory.
A better fix would involve reworking bootparamd not to require a ':'
in the response string.

27 years agoMerged change from 1.15 to 1.17 in the English version in -current.
Masafumi Max NAKANE [Sun, 24 Nov 1996 22:19:50 +0000 (22:19 +0000)]
Merged change from 1.15 to 1.17 in the English version in -current.

Cosmetic tweaks.

Typo.

27 years agoFixed two typos I made
David E. O'Brien [Sun, 24 Nov 1996 08:39:11 +0000 (08:39 +0000)]
Fixed two typos I made

Thanks:  Max

27 years agoSlow down the RIP and SAP broadcasts. At least some NW4.x machines can't
John Hay [Sun, 24 Nov 1996 08:35:23 +0000 (08:35 +0000)]
Slow down the RIP and SAP broadcasts. At least some NW4.x machines can't
handle them back-to-back. (We couldn't either without my ipx receive buffer
enlargement.)

27 years agoEnlarge the transmit and receive bufferspace of ipx. Make it tweakable
John Hay [Sun, 24 Nov 1996 08:25:48 +0000 (08:25 +0000)]
Enlarge the transmit and receive bufferspace of ipx. Make it tweakable
with sysctl.

27 years agoBack out my previous change here. aic7xxx_asm is only a stdio program
Peter Wemm [Sun, 24 Nov 1996 08:15:11 +0000 (08:15 +0000)]
Back out my previous change here.  aic7xxx_asm is only a stdio program
that runs in the host's build environment, not the kernel's environment.

27 years agoFix my previous (untested) commit. The probe routine now appears to work
Peter Wemm [Sun, 24 Nov 1996 08:06:01 +0000 (08:06 +0000)]
Fix my previous (untested) commit.  The probe routine now appears to work
again.

Something needs to be done about this in 2.2, since the code there doesn't
even compile anymore since the syscons/kbdio/psm merge into 2.2.

27 years agoAdded a translator's note on the term ``anonymous FTP''.
Masafumi Max NAKANE [Sat, 23 Nov 1996 23:49:40 +0000 (23:49 +0000)]
Added a translator's note on the term ``anonymous FTP''.

Submitted by: Yoshiaki Uchikawa <yoshiaki@kt.rim.or.jp>

27 years agoAdded forgotten directories zoneinfo/Antarctica and zoneinfo/Arctic.
Bruce Evans [Sat, 23 Nov 1996 20:04:02 +0000 (20:04 +0000)]
Added forgotten directories zoneinfo/Antarctica and zoneinfo/Arctic.

The zoneinfo makefile doesn't follow the rules.  It builds everything
at install time.  It dpends on zic to create the directories.  zic
doesn't know about the weird 555 permissions specified in BSD.usr.dist,
so it creates the directories with nonstandard permissions.

27 years agoAllocate a header mbuf for the start of the encapsulated packet.
Bill Fenner [Sat, 23 Nov 1996 19:07:02 +0000 (19:07 +0000)]
Allocate a header mbuf for the start of the encapsulated packet.
The rest of the code was treating it as a header mbuf, but it was
allocated as a normal mbuf.

This fixes the panic: ip_output no HDR when you have a multicast
tunnel configured.

27 years agoAdded `info' subdirectory to the section ``Subdirectories'' as the location
Masafumi Max NAKANE [Sat, 23 Nov 1996 19:02:25 +0000 (19:02 +0000)]
Added `info' subdirectory to the section ``Subdirectories'' as the location
to install info pages.
(It hasn't been clear in the handbook either share/info or info should be
used.)

27 years agoMinor typo.
Masafumi Max NAKANE [Sat, 23 Nov 1996 18:26:30 +0000 (18:26 +0000)]
Minor typo.

27 years agoName Pentium Pro Support.
Andreas Klemm [Sat, 23 Nov 1996 12:33:34 +0000 (12:33 +0000)]
Name Pentium Pro Support.
Changed an expression "one year ago", which is in fact only valid
for a short period of time.
Updated the section about ports. Make this more general, "hundreds of
ports" and give a concrete number of ports (over 710) with a timestamp
"at end of November 96".

I reformatted the "ports" paragraph using fmt, because I think it looks
nicer now in sgml source. So the diff shows more changed lines than was
actually changed. If this isn't suitable for you, then I'll do my best
in the future, to avoid this. My intention was, to make the source look
nicer as well.

27 years agoMake this compile again after the syscons/psm/kbdio changes that removed
Peter Wemm [Sat, 23 Nov 1996 07:38:25 +0000 (07:38 +0000)]
Make this compile again after the syscons/psm/kbdio changes that removed
the keyboard IO port definitions from <machine/console.h>.

Note: I do not have a machine available to test this yet,  I sent this
for review to several people who have not replied.

This is also totally busted in 2.2, which has half of the syscons/psm/kbdio
changes in it.  This really desperately needs to be tested and fixed
in 2.2 as well.

27 years agoAdd missing nls directories
Andrey A. Chernov [Sat, 23 Nov 1996 04:06:10 +0000 (04:06 +0000)]
Add missing nls directories

2.2 candidate

27 years agoUnderstand ru_RU too
Andrey A. Chernov [Sat, 23 Nov 1996 02:31:40 +0000 (02:31 +0000)]
Understand ru_RU too

27 years agoget_myaddress() wasn't following the interface array properly
Paul Traina [Fri, 22 Nov 1996 23:37:08 +0000 (23:37 +0000)]
get_myaddress() wasn't following the interface array properly

Cannidate for: 2.2

27 years agoFix minor typo/grammatical error
Gary Palmer [Fri, 22 Nov 1996 22:16:25 +0000 (22:16 +0000)]
Fix minor typo/grammatical error

Submitted by: John-Mark Gurney <jmg@nike.efn.org> in PR 2085

27 years agomake the Usage string match reality
Peter Wemm [Fri, 22 Nov 1996 13:58:03 +0000 (13:58 +0000)]
make the Usage string match reality

Submitted by:  faried nawaz <fn@uidaho.edu>, PR#2075

27 years agoBack out recent security patch for rexecd. After more careful analysis,
Paul Traina [Fri, 22 Nov 1996 08:59:07 +0000 (08:59 +0000)]
Back out recent security patch for rexecd.  After more careful analysis,
it is both uneeded and breaks certain lock-step timing in the rexec
protocol.

Yes, an attacker can "relay" connections using this trick,  but a properly
configured firewall that would make this sort of subterfuge necessary in the
first place (instead of direct packet spoofing) would also thwart useful
attacks based on this.

27 years agoAdd code to the SCSIINT handler for unexpected busfrees and handling immediate
Justin T. Gibbs [Fri, 22 Nov 1996 08:28:45 +0000 (08:28 +0000)]
Add code to the SCSIINT handler for unexpected busfrees and handling immediate
SCBs in response to a busfree.

When re-queueing an SCB that returned with QUEUE FULL status, reset its timeout.

Ensure that aborted SCBs have an error code set in there xs before it gets
passed back up with scsi_done.

Fix a few KNF nits.

27 years agoAdded reminder to edit the parent Makefile to add what you just imported.
David E. O'Brien [Fri, 22 Nov 1996 08:28:32 +0000 (08:28 +0000)]
Added reminder to edit the parent Makefile to add what you just imported.

27 years agoConditionally enable the busfree interrupt to catch unexpected busfrees.
Justin T. Gibbs [Fri, 22 Nov 1996 08:25:23 +0000 (08:25 +0000)]
Conditionally enable the busfree interrupt to catch unexpected busfrees.
Immediate SCBs, since they always send messages that tell the target to
transition to bus free now rely on the busfree interrupt instead of the
IMMEDDONE sequencer interrupt that was generated before.

Rearrange some code in the message out loop to give ATN a little more time
to drop before we ACK the last byte.

Use SPIORDY instead of REQINIT when snooping for a tag message on a reconnect.
This is done for the same reasons we use SPIORDY in the inb functions.

When going into BITBUCKET mode, turn off HDMAEN in the DFCNTRL register so
that we can "not care" what the value of HCNT is.  If HCNT is 0, BITBUCKET
mode won't transfer any data if HDMAEN is set.  Seeing as we don't want the
transfer to even think about touching the host, this seems more sane anyway.
Thanks to "Dan Willis" <dan@plutotech.com> for pointing out that this was
a problem.

27 years agoSet a more explicit #include path for building aix7xxx_asm, otherwise it
Peter Wemm [Fri, 22 Nov 1996 04:27:43 +0000 (04:27 +0000)]
Set a more explicit #include path for building aix7xxx_asm, otherwise it
uses /usr/include/sys/*, which may point to a different build tree.  I'm
not sure that this is necessary, but there was a question mark over what
/usr/include/sys points to when building the "user mode" binaries in the
kernel code, especially when building the smp tree.

I suspect that the "right" line here is to use ${INCLUDES}, but that
causes warnings about unused static inline functions in stdio.h and ctype.h

27 years agoFix ifconfig link flag handling
Guido van Rooij [Thu, 21 Nov 1996 21:04:50 +0000 (21:04 +0000)]
Fix ifconfig link flag handling
Submitted by: "Jon Morgan" <morgan@terminus.trailblazer.com>

27 years agoFix lots of non-bug questionable code:
Garrett Wollman [Thu, 21 Nov 1996 19:36:09 +0000 (19:36 +0000)]
Fix lots of non-bug questionable code:

- Don't link against libkvm and don't uselessly include <kvm.h>.
- Declare constant objects as const.
- Declare functions with the correct types.
- Call functions with the correct parameters.

Not fixed:

- The sysctl parsing remains ugly (but it may be the best we can do).
- atnetrange() should use strtoul() rather than sscanf() for better
  error checking.

27 years agoDon't dump core on zero-size files.
Poul-Henning Kamp [Thu, 21 Nov 1996 15:17:04 +0000 (15:17 +0000)]
Don't dump core on zero-size files.

27 years agoru_SU -> ru_RU
Andrey A. Chernov [Thu, 21 Nov 1996 10:17:45 +0000 (10:17 +0000)]
ru_SU -> ru_RU

27 years agoAdd ru_RU as alias to ru_SU
Andrey A. Chernov [Thu, 21 Nov 1996 10:13:25 +0000 (10:13 +0000)]
Add ru_RU as alias to ru_SU

27 years agochanges to make devfs more 'normal'
Julian Elischer [Thu, 21 Nov 1996 07:19:00 +0000 (07:19 +0000)]
changes to make devfs more 'normal'
also fixes a bug I've been chasing for a LONG TIME,
due to the fact that spec_bwrite is a NOP and I didn't realise it..

old symptom:
mount -t devfs devfs /mnt
mount /mnt/wd0e /mnt/mnt2
umount /mnt2 <process hangs>

there are some pretty large structural differences internal to devfs
but outwards it should look the same.
I have not yet tested extensively but will do so and fix 3 warnings tomorrow.

27 years agoBack out Russian zones/rulez awaiting for new Paul Eggert update
Andrey A. Chernov [Thu, 21 Nov 1996 07:17:08 +0000 (07:17 +0000)]
Back out Russian zones/rulez awaiting for new Paul Eggert update
Reviewed by: wollman

27 years agoUpdate comment and disable SCB paging for Rev E cards since paging, at least
Justin T. Gibbs [Thu, 21 Nov 1996 06:19:13 +0000 (06:19 +0000)]
Update comment and disable SCB paging for Rev E cards since paging, at least
in its current form, will not work with these cards.

27 years agoChange the way Automatic PIO is performed in the mesgin and mesgout phases.
Justin T. Gibbs [Thu, 21 Nov 1996 06:18:33 +0000 (06:18 +0000)]
Change the way Automatic PIO is performed in the mesgin and mesgout phases.
SPIORDY should go active on any REQ of the bus, so testing for REQINIT is
not necessary.  It also seems that testing for SPIORDY is more robust then
REQINIT since SPIORDY comes active after REQINIT and PHASEMIS seems to take
some time to come true after REQ is asserted if the phase has changed.  Of
course, none of this is documented.

This should give the code savings of my original changes, without breaking the
driver on fast peripherals.

27 years agoConditionalize setsockopt IP_PORTRANGE to make ftpd portable.
Paul Traina [Wed, 20 Nov 1996 22:13:51 +0000 (22:13 +0000)]
Conditionalize setsockopt IP_PORTRANGE to make ftpd portable.

27 years agoTruncate argument list to avoid buffer overflows.
Paul Traina [Wed, 20 Nov 1996 22:12:50 +0000 (22:12 +0000)]
Truncate argument list to avoid buffer overflows.

Cannidate for: 2.1 and 2.2

27 years agoOops, we still want to build/install the static libraries.
Garrett Wollman [Wed, 20 Nov 1996 20:55:11 +0000 (20:55 +0000)]
Oops, we still want to build/install the static libraries.

27 years agoCopy a slightly modified version of the `libgnumalloc hack' to allow
Garrett Wollman [Wed, 20 Nov 1996 20:54:28 +0000 (20:54 +0000)]
Copy a slightly modified version of the `libgnumalloc hack' to allow
us to finally phase out the 8K shared version of liby in favor of the
600-byte static one, as was done years ago for libl/libfl.

27 years agoFor some reason the merge of this file didn't get committed properly.
Garrett Wollman [Wed, 20 Nov 1996 15:23:40 +0000 (15:23 +0000)]
For some reason the merge of this file didn't get committed properly.
Who has the conical hat?

Make things compile again.

27 years agoTypo.
Masafumi Max NAKANE [Wed, 20 Nov 1996 14:49:59 +0000 (14:49 +0000)]
Typo.

Improvement in translation.

Submitted by: Mitsuru IWASAKI <iwasaki@pc.jaring.my>