]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
26 years agoDon't forget to clean up after an error reading the directory entry
Bruce Evans [Sun, 17 May 1998 18:09:28 +0000 (18:09 +0000)]
Don't forget to clean up after an error reading the directory entry
in deget().

26 years agoRemoved vestiges of pre-Lite2 locking.
Bruce Evans [Sun, 17 May 1998 18:00:45 +0000 (18:00 +0000)]
Removed vestiges of pre-Lite2 locking.

26 years agoClear kernel stack pages before usage.
Tor Egge [Sun, 17 May 1998 17:43:13 +0000 (17:43 +0000)]
Clear kernel stack pages before usage.
Correct panic message in pmap_zero_page (s/CMAP /CMAP2 /).

26 years agoBe more restrictive about when to disregard MP config table values
Tor Egge [Sun, 17 May 1998 17:32:10 +0000 (17:32 +0000)]
Be more restrictive about when to disregard MP config table values
(previous commit broke support for Dec Personal Workstation).

26 years agos/nanoruntime/nanouptime/g
Poul-Henning Kamp [Sun, 17 May 1998 11:53:46 +0000 (11:53 +0000)]
s/nanoruntime/nanouptime/g
s/microruntime/microuptime/g

Reviewed by: bde

26 years agoCleanup after Garret, include unpch.h to get at various macros..
Søren Schmidt [Sun, 17 May 1998 09:37:39 +0000 (09:37 +0000)]
Cleanup after Garret, include unpch.h to get at various macros..

26 years agoTurn on TCP_NODELAY on the remote socket, to turn off sender silly window
Bill Fenner [Sun, 17 May 1998 06:39:38 +0000 (06:39 +0000)]
Turn on TCP_NODELAY on the remote socket, to turn off sender silly window
 syndrome avoidance.  The combination of SWS avoidance and ack-every-other
 causes low throughput if the block size divided by the MSS is odd (which
 is true with the default block size and MSS).
Turning on TCP_NODELAY disables the Nagle algorithm and sender SWS avoidance.
 The rdump request/response protocol can not invoke Nagle and cannot cause
 SWS, so this has no negative effects.

26 years agoAn important fix for proper inheritance of backing objects for
John Dyson [Sat, 16 May 1998 23:03:20 +0000 (23:03 +0000)]
An important fix for proper inheritance of backing objects for
object splits.  Another excellent detective job by Tor.
Submitted by: Tor Egge <Tor.Egge@idi.ntnu.no>

26 years agoNow that FreeBSD/Alpha can be installed on a disk and run on its own,
John Birrell [Sat, 16 May 1998 22:20:44 +0000 (22:20 +0000)]
Now that FreeBSD/Alpha can be installed on a disk and run on its own,
a bootstrap build under NetBSD is an option.

This makefile will require further changes to ignore aout tools.

26 years agoInclude correct header to get definition of struct sockaddr_un.
Mark Murray [Sat, 16 May 1998 21:35:37 +0000 (21:35 +0000)]
Include correct header to get definition of struct sockaddr_un.

26 years agoMoved from directory above. It is i386 specific.
John Birrell [Sat, 16 May 1998 21:29:17 +0000 (21:29 +0000)]
Moved from directory above. It is i386 specific.

26 years agoAdd an alpha specific ttys
John Birrell [Sat, 16 May 1998 21:28:45 +0000 (21:28 +0000)]
Add an alpha specific ttys

26 years agoMake ttys machine specific.
John Birrell [Sat, 16 May 1998 21:28:17 +0000 (21:28 +0000)]
Make ttys machine specific.

26 years agoReturn back initial tzset() must be before first chroot
Andrey A. Chernov [Sat, 16 May 1998 21:23:33 +0000 (21:23 +0000)]
Return back initial tzset() must be before first chroot

26 years agoFix rev. 1.7: get offsetof from the correct header file.
Garrett Wollman [Sat, 16 May 1998 18:04:00 +0000 (18:04 +0000)]
Fix rev. 1.7: get offsetof from the correct header file.

26 years agoDon't use "ffs" in an ext2fs sleep message string.
Bruce Evans [Sat, 16 May 1998 17:47:44 +0000 (17:47 +0000)]
Don't use "ffs" in an ext2fs sleep message string.

Don't forget to clear the inode hash lock before returning from ext2_vget()
after getnewvnode() fails.  Obtained from: rev.1.24 of ffs_vfsops.c (the
original patch for the getnewvnode() race).  Forgotten in: rev.1.4 here.

Removed a duplicate comment.  Duplicated in: rev.1.4 here.

Fixed the MALLOC() vs getnewvnode() race in ext2_vget().  Obtained from:
rev.1.39 of ffs_vfsops.c.

26 years agoAbbreviate "ext2fs_fsync" as "e2fsyn" instead of as "extfsn" in a sleep
Bruce Evans [Sat, 16 May 1998 16:52:20 +0000 (16:52 +0000)]
Abbreviate "ext2fs_fsync" as "e2fsyn" instead of as "extfsn" in a sleep
message string.

26 years agoGet timespecs directly instead of via timevals.
Bruce Evans [Sat, 16 May 1998 16:20:50 +0000 (16:20 +0000)]
Get timespecs directly instead of via timevals.

26 years agoDon't abuse `+' to combine flags.
Bruce Evans [Sat, 16 May 1998 16:03:10 +0000 (16:03 +0000)]
Don't abuse `+' to combine flags.

26 years agoBacked out rev.1.76. It just added style bugs.
Bruce Evans [Sat, 16 May 1998 15:21:29 +0000 (15:21 +0000)]
Backed out rev.1.76.  It just added style bugs.

26 years agoGet timespecs directly instead of via timevals.
Bruce Evans [Sat, 16 May 1998 15:11:24 +0000 (15:11 +0000)]
Get timespecs directly instead of via timevals.

26 years agoSome of newer PC-98 may cause "Windows Protection Fault" when booting
KATO Takenori [Sat, 16 May 1998 14:44:11 +0000 (14:44 +0000)]
Some of newer PC-98 may cause "Windows Protection Fault" when booting
Windows 95 after rebooting FreeBSD without power off.  In PC-98
system, reboot mode is set via I/O port 0x37 in cpu_reset(), and
accessing of this port is the reason of the problem.  To avnoid the
fault, current status of reboot mode should be checked before
accessing the I/O port.

26 years agoDisable local APIC in UP kernel. Intel specification update describes
KATO Takenori [Sat, 16 May 1998 14:38:10 +0000 (14:38 +0000)]
Disable local APIC in UP kernel.  Intel specification update describes
that local APIC should be disabled in UP system.  However, some of old
BIOS does not disable local APIC, and virtual wire mode through local
APIC may cause int 15.

26 years agoSpotted a misplaced comma in my previous patch.
Dag-Erling Smørgrav [Sat, 16 May 1998 14:10:12 +0000 (14:10 +0000)]
Spotted a misplaced comma in my previous patch.

26 years agoTypo nits in SUIDDIR comment (removed some parentheses, moved some
Dag-Erling Smørgrav [Sat, 16 May 1998 14:08:31 +0000 (14:08 +0000)]
Typo nits in SUIDDIR comment (removed some parentheses, moved some
commas, replaced "partition" with "filesystem", reformatted the
paragraph)

26 years agoFix broken sun_len usage: it is not sun_path length but bigger
Andrey A. Chernov [Sat, 16 May 1998 08:31:49 +0000 (08:31 +0000)]
Fix broken sun_len usage: it is not sun_path length but bigger

26 years ago1.31 Randall Hopper <rhh@ct.picker.com>
Amancio Hasty [Sat, 16 May 1998 07:07:10 +0000 (07:07 +0000)]
1.31                       Randall Hopper <rhh@ct.picker.com>
                           submitted ioctl to clear the video buffer
                           prior to starting video capture
   Amancio : clean up yuv12 so that it does not
                           affect rgb capture. Basically, fxtv after
                           capturing in yuv12 mode , switching to rgb
                           would cause the video capture to be too bright.
1.32                       disable inverse gamma function for rgb and yuv
                           capture. fixed meteor brightness ioctl it now
                           converts the brightness value from unsigned to
                           signed.
1.33                       added sysctl: hw.bt848.tuner, hw.bt848.reverse_mute,
                           hw.bt848.card
   card takes a value from 0 to bt848_max_card
                           tuner takes a value from 0 to bt848_max_tuner
                           reverse_mute : 0 no effect, 1 reverse tuner
                           mute function some tuners are wired reversed :(

26 years agoWhen a timeval is stored at the beginning of icmp_data, the fields
John Birrell [Sat, 16 May 1998 00:19:27 +0000 (00:19 +0000)]
When a timeval is stored at the beginning of icmp_data, the fields
are unaligned for access by the alpha, so copy the value to a variable
that is aligned.

When checking the returned data, be careful to avoid confusing the
size of the icmp header with the size of a timeval. On i386 these
are both 8, but on alpha, a timeval is 16 bytes. This means that
a packet sent from an alpha contains 48 bytes of data, not 56 like
on i386.

26 years agoPatch RPC library to avoid possible denial of service attacks as described
Bill Paul [Fri, 15 May 1998 22:57:31 +0000 (22:57 +0000)]
Patch RPC library to avoid possible denial of service attacks as described
recently in BUGTRAQ. The set_input_fragment() routine in the XDR record
marking code blindly trusts that the first two bytes it sees will in fact
be an actual record header and that the specified size will be sane. In
fact, if you just telnet to a listening port of an RPC service and send a
few carriage returns, set_input_fragment() will obtain a ridiculously large
record size and sit there for a long time trying to read from the network.

A sanity test is required: if the record size is larger than the receive
buffer, punt.

26 years agoPatch RPC library to avoid possible denial of service attacks as described
Bill Paul [Fri, 15 May 1998 22:53:47 +0000 (22:53 +0000)]
Patch RPC library to avoid possible denial of service attacks as described
recently in BUGTRAQ. If a stream oriented transport fails to properly decode
an RPC message header structure where there should be one, it should mark
the stream as dead so that the connection will be dropped.

26 years ago Low level use of 'vidattr()' can cause a NULL pointer to be
Andrey A. Chernov [Fri, 15 May 1998 21:35:53 +0000 (21:35 +0000)]
Low level use of 'vidattr()' can cause a NULL pointer to be
dereferenced.  This is because 'SP' is only initialized via 'newterm()'
(which is not required if you are going to interact with the 'terminfo'
database without using 'ncurses').

PR: 6648
Submitted by: Max Euston <meuston@jmrodgers.com>

26 years agoChange an arbitarly hardcoded `10' to MAX_NO_DISKS, which is currently "20".
David E. O'Brien [Fri, 15 May 1998 21:17:01 +0000 (21:17 +0000)]
Change an arbitarly hardcoded `10' to MAX_NO_DISKS, which is currently "20".
Otherwise can't see sd10 (obtained using wired down sd unit numbers),
reguardless of the number of actual disks you have.

26 years agombuf, inet, and unix modules no longer read kvm.
Garrett Wollman [Fri, 15 May 1998 20:19:21 +0000 (20:19 +0000)]
mbuf, inet, and unix modules no longer read kvm.

26 years agoConvert socket structures to be type-stable and add a version number.
Garrett Wollman [Fri, 15 May 1998 20:11:40 +0000 (20:11 +0000)]
Convert socket structures to be type-stable and add a version number.

Define a parameter which indicates the maximum number of sockets in a
system, and use this to size the zone allocators used for sockets and
for certain PCBs.

Convert PF_LOCAL PCB structures to be type-stable and add a version number.

Define an external format for infomation about socket structures and use
it in several places.

Define a mechanism to get all PF_LOCAL and PF_INET PCB lists through
sysctl(3) without blocking network interrupts for an unreasonable
length of time.  This probably still has some bugs and/or race
conditions, but it seems to work well enough on my machines.

It is now possible for `netstat' to get almost all of its information
via the sysctl(3) interface rather than reading kmem (changes to follow).

26 years agoFix an obvious parameter-order bogon. (Don't know what happened to
Garrett Wollman [Fri, 15 May 1998 20:02:47 +0000 (20:02 +0000)]
Fix an obvious parameter-order bogon.  (Don't know what happened to
the warning message before.)

26 years agoSmall typo in T/TCP patch ("speicfy" -> "specify").
Pierre Beyssac [Fri, 15 May 1998 19:16:35 +0000 (19:16 +0000)]
Small typo in T/TCP patch ("speicfy" -> "specify").

26 years agoDon't use `&&' in any shell commands here. Using it to give conditional
Bruce Evans [Fri, 15 May 1998 17:12:00 +0000 (17:12 +0000)]
Don't use `&&' in any shell commands here.  Using it to give conditional
execution is usually unnecessary in BSD Makefiles because BSD make
invokes shells with -e.  Using it to give conditional execution is
often wrong in BSD makefiles because BSD make joins shell commands
when invoked in certain ways (in particular, as `make -jN').  Example
makefile:
---
clean:
cd /
false && true
rm -rf * # a dangerous command
---
This should terminate after the `false && true' command fails, but
it doesn't when the commands are joined (`false && true' is a non-
simple command, so -e doesn't cause termination).  The b-maked version:
---
clean:
cd /
false; true
rm -rf * # a dangerous command
---
terminates after the `false' command fails (`false' is a simple
command, so -e causes termination).  However, for versions of
make like gnu make that don't invoke shells with -e, this change
completely breaks the makefile.

This is one of the fixes for the bug suite that caused `make world'
to sometimes put raw cpp output in .depend files.  Building of cc
sometimes failed, but the failure did not terminate the build
immediately, and various wrong versions of the cc components were
used until one was wrong enough to cause a fatal error.

26 years agoReturn back vfork and use execve with TZ="" environment in vfork case
Andrey A. Chernov [Fri, 15 May 1998 16:51:06 +0000 (16:51 +0000)]
Return back vfork and use execve with TZ="" environment in vfork case

26 years agoUse fork instead of vfork since setenv clobber parent environment
Andrey A. Chernov [Fri, 15 May 1998 16:30:09 +0000 (16:30 +0000)]
Use fork instead of vfork since setenv clobber parent environment
Fork already used for INTERNAL_LS in anycase

26 years agoMove TZ="" assignment just before exec to not touch other time stuff
Andrey A. Chernov [Fri, 15 May 1998 16:08:52 +0000 (16:08 +0000)]
Move TZ="" assignment just before exec to not touch other time stuff

26 years agoDo TZ= as first thing, since FTP protocol is unable to tell zone offset in
Andrey A. Chernov [Fri, 15 May 1998 15:06:58 +0000 (15:06 +0000)]
Do TZ= as first thing, since FTP protocol is unable to tell zone offset in
any case.

It makes no difference for anon account (since chroot already makes it GMT),
but if you do mirror with special non-anon login, in old variant
your mirror will be wholy retransmitted twice in the year due to
time zone changes (/etc/localtime plays bad role here)

26 years agoPR: 6641
Daniel O'Callaghan [Fri, 15 May 1998 12:38:07 +0000 (12:38 +0000)]
PR: 6641
Submitted by: Andre Albsmeier <andre.albsmeier@mchp.siemens.de>
Make -q work for zeroing a specific rule.

26 years agoMore programs are now broken on alpha due to the libbind changes
John Birrell [Fri, 15 May 1998 12:33:56 +0000 (12:33 +0000)]
More programs are now broken on alpha due to the libbind changes
that prevent the programs from being linked static (duplicated
symbols).

Other programs depend on kernel internals. These will have to wait
for a custom alpha kernel. For now, let's just make the build safe.

26 years agoChange variable from long to time_t where they are passed by reference
John Birrell [Fri, 15 May 1998 12:28:43 +0000 (12:28 +0000)]
Change variable from long to time_t where they are passed by reference
to time().

26 years agoRemove a big hack after adding a small one to libc/gen/getcwd.c to
John Birrell [Fri, 15 May 1998 12:01:06 +0000 (12:01 +0000)]
Remove a big hack after adding a small one to libc/gen/getcwd.c to
handle the lack of __getcwd syscall in NetBSD.

26 years agoNetBSD doesn't have a __getcwd syscall, so set have__getcwd to `no'
John Birrell [Fri, 15 May 1998 11:59:00 +0000 (11:59 +0000)]
NetBSD doesn't have a __getcwd syscall, so set have__getcwd to `no'
when building libc with NetBSD syscalls.

26 years agoThe FreeBSD lkm design is aout specific.
John Birrell [Fri, 15 May 1998 11:57:05 +0000 (11:57 +0000)]
The FreeBSD lkm design is aout specific.

26 years agosyscons is i386 specific due to machine/console.h
John Birrell [Fri, 15 May 1998 11:55:47 +0000 (11:55 +0000)]
syscons is i386 specific due to machine/console.h

26 years agoAdd cross references for find(1), locate(1), whereis(1) and which(1).
Joseph Koshy [Fri, 15 May 1998 11:22:42 +0000 (11:22 +0000)]
Add cross references for find(1), locate(1), whereis(1) and which(1).

Submitted by: Josh Gillam <josh@quick.net>
PR: docs/6642

26 years agoOops, forgot references in previous commit.
Bruce Evans [Fri, 15 May 1998 09:34:48 +0000 (09:34 +0000)]
Oops, forgot references in previous commit.

Submitted by: Brian Cully <shmit@erols.com>
PR: 6178

26 years agoSupport Objective C almost as well as C++.
Bruce Evans [Fri, 15 May 1998 09:30:13 +0000 (09:30 +0000)]
Support Objective C almost as well as C++.

Notes:
- We no longer use -fgnu-runtime in bsd.lib.mk, since it is the default
  and bsd.lib.mk is the wrong place to override it.
- Gnu C doesn't have a special compiler driver for Objective C like it
  does for C++.  The defaults are suitable for Gnu C.  Use `OBJCLIBS='
  in /etc/make.conf for POC.

26 years agoChange the name of a variable from _start to s_start. On alpha there
John Birrell [Fri, 15 May 1998 09:26:28 +0000 (09:26 +0000)]
Change the name of a variable from _start to s_start. On alpha there
must be a library function called _start.

26 years agoRemoved misleading claim that login classes are no longer implemented.
Joseph Koshy [Fri, 15 May 1998 09:19:03 +0000 (09:19 +0000)]
Removed misleading claim that login classes are no longer implemented.
Added Xref to pw(8).

26 years agoDisable the auto-Write Combining setup for the pmap code. This
John Dyson [Fri, 15 May 1998 07:25:25 +0000 (07:25 +0000)]
Disable the auto-Write Combining setup for the pmap code.  This
worked on a couple of machines of mine, but appears to cause problems
on others.

26 years agoChange the order of the include paths so that the architecture specific
John Birrell [Fri, 15 May 1998 06:42:37 +0000 (06:42 +0000)]
Change the order of the include paths so that the architecture specific
directory is searched before the generic one. You can guess what was
happening.. all the world's assumed to be 32-bit.

26 years agoCorrect use of .Nm. Add rcsid.
Philippe Charnier [Fri, 15 May 1998 06:30:58 +0000 (06:30 +0000)]
Correct use of .Nm. Add rcsid.

26 years agocorrect use of .Nm. Add rcsid.
Philippe Charnier [Fri, 15 May 1998 06:23:45 +0000 (06:23 +0000)]
correct use of .Nm. Add rcsid.

26 years agoCorrect use of .Nm. Typo.
Philippe Charnier [Fri, 15 May 1998 06:22:30 +0000 (06:22 +0000)]
Correct use of .Nm. Typo.

26 years agoCorrect use of .Nm. Add rcsid.
Philippe Charnier [Fri, 15 May 1998 06:19:05 +0000 (06:19 +0000)]
Correct use of .Nm. Add rcsid.

26 years agoNetBSD kernels don't support TCP_NOPUSH, so on alpha don't try setting
John Birrell [Fri, 15 May 1998 03:23:28 +0000 (03:23 +0000)]
NetBSD kernels don't support TCP_NOPUSH, so on alpha don't try setting
this socket option. This is temporary code while the alpha still uses
NetBSD socket code in the kernel.

26 years agoint -> long changes that reduce the diffs with the NetBSD version to
John Birrell [Thu, 14 May 1998 21:45:18 +0000 (21:45 +0000)]
int -> long changes that reduce the diffs with the NetBSD version to
work in a 64-bit environment.

26 years agoOn request of Garrett, ad a way to specify that a service should be
Guido van Rooij [Thu, 14 May 1998 20:26:16 +0000 (20:26 +0000)]
On request of Garrett, ad a way to specify that a service should be
reachable via T/TCP
Reviewed by: Garrett Wollman

26 years agoFix list corruption and memory leak that could occur when
Justin T. Gibbs [Thu, 14 May 1998 19:47:38 +0000 (19:47 +0000)]
Fix list corruption and memory leak that could occur when
releasing EISA irqs.

26 years agoRegenerated syscalls list after signanosleep removing
Andrey A. Chernov [Thu, 14 May 1998 14:49:58 +0000 (14:49 +0000)]
Regenerated syscalls list after signanosleep removing

26 years agoRemove reference to signanosleep
Andrey A. Chernov [Thu, 14 May 1998 14:39:58 +0000 (14:39 +0000)]
Remove reference to signanosleep

26 years agoSync with sys/i386conf/majors.i386 revision 1.42.
KATO Takenori [Thu, 14 May 1998 12:42:11 +0000 (12:42 +0000)]
Sync with sys/i386conf/majors.i386 revision 1.42.

26 years agoNuke signanosleep()
Peter Wemm [Thu, 14 May 1998 11:36:16 +0000 (11:36 +0000)]
Nuke signanosleep()

26 years agonuke signanosleep().
Peter Wemm [Thu, 14 May 1998 11:33:04 +0000 (11:33 +0000)]
nuke signanosleep().

26 years agoNuke signanosleep(). (I've left nanosleep1() seperate to nanosleep()
Peter Wemm [Thu, 14 May 1998 11:31:08 +0000 (11:31 +0000)]
Nuke signanosleep().  (I've left nanosleep1() seperate to nanosleep()
as I don't want to mess with the multiple returns)

26 years agoregen after signanosleep nuke
Peter Wemm [Thu, 14 May 1998 11:29:06 +0000 (11:29 +0000)]
regen after signanosleep nuke

26 years agodeep-six signanosleep(). It sounded like a good idea at the time.
Peter Wemm [Thu, 14 May 1998 11:28:11 +0000 (11:28 +0000)]
deep-six signanosleep().  It sounded like a good idea at the time.

26 years agoCommit an old change that has been sitting around for a long while.
Peter Wemm [Thu, 14 May 1998 10:38:52 +0000 (10:38 +0000)]
Commit an old change that has been sitting around for a long while.
signanosleep() did not deal with signal masks properly.  This change was
based on a discussion with bde some time ago (at least 6 months or more).

signanosleep() should probably go away since it was never really used for
more than a few weeks and doesn't appear in released code.  It should
probably be killed before somebody uses it and it becomes a gratuitous
nonstandard feature.

26 years agoAllow `w' to treat a corrupted "utmp" as a non fatal error.
Joseph Koshy [Thu, 14 May 1998 10:13:24 +0000 (10:13 +0000)]
Allow `w' to treat a corrupted "utmp" as a non fatal error.

PR: bin/2832

26 years agoRemove a bogus prototype for time() and let time.h do that.
John Birrell [Thu, 14 May 1998 10:07:29 +0000 (10:07 +0000)]
Remove a bogus prototype for time() and let time.h do that.

Change pointer casts from int to long. The code that looks to index -1
of argv is still broken on alpha.

26 years agoAdd an example showing how to build a UFS floppy
Paul Traina [Thu, 14 May 1998 06:42:47 +0000 (06:42 +0000)]
Add an example showing how to build a UFS floppy

26 years agoMake a bit of FreeBSD-specific code a bit more specific => i386 only.
John Birrell [Wed, 13 May 1998 22:55:44 +0000 (22:55 +0000)]
Make a bit of FreeBSD-specific code a bit more specific => i386 only.

26 years agolibtcl now builds (with lots of pointer to int cast warnings) on alpha.
John Birrell [Wed, 13 May 1998 22:54:03 +0000 (22:54 +0000)]
libtcl now builds (with lots of pointer to int cast warnings) on alpha.

26 years agolibtcl expects to see ieeefp.h definitions through this header.
John Birrell [Wed, 13 May 1998 22:40:42 +0000 (22:40 +0000)]
libtcl expects to see ieeefp.h definitions through this header.

26 years agoFix bogus "cleanup" in bufq_remove. The "switch point" for tqdisksort was
Justin T. Gibbs [Wed, 13 May 1998 16:03:33 +0000 (16:03 +0000)]
Fix bogus "cleanup" in bufq_remove.  The "switch point" for tqdisksort was
getting mangled.

Submitted by: Bruce Evans <bde@zeta.org.au>

26 years agoName cronyx devices correctly.
Jordan K. Hubbard [Wed, 13 May 1998 12:38:26 +0000 (12:38 +0000)]
Name cronyx devices correctly.

26 years agoSync with sys/i386/isa/sio.c revision 1.201.
KATO Takenori [Wed, 13 May 1998 10:42:36 +0000 (10:42 +0000)]
Sync with sys/i386/isa/sio.c revision 1.201.

26 years agoSync with sys/i386/conf/majors.i386 revision 1.41.
KATO Takenori [Wed, 13 May 1998 10:40:32 +0000 (10:40 +0000)]
Sync with sys/i386/conf/majors.i386 revision 1.41.

26 years agoFreeBSD -> %s/%m
Poul-Henning Kamp [Wed, 13 May 1998 08:10:07 +0000 (08:10 +0000)]
FreeBSD -> %s/%m

PR: 6600
Reviewed by: phk
Submitted by: Josh Gilliam <josh@quick.net>

26 years agoThe SYNOPSYS of mount_umap(8) says
Poul-Henning Kamp [Wed, 13 May 1998 08:02:50 +0000 (08:02 +0000)]
The SYNOPSYS of mount_umap(8) says

     mount_umap [-o options] target mount-point uid-mapfile gid-mapfile

This should read

     mount_umap [-o options] -u uid-mapfile -g gid-mapfile target mount-point

PR: 6586
Reviewed by: phk
Submitted by: Amakawa Shuhei amakawa@hoh.t.u-tokyo.ac.jp

26 years agoThe description of the format of the id mapfile is wrong. You have
Poul-Henning Kamp [Wed, 13 May 1998 08:01:00 +0000 (08:01 +0000)]
The description of the format of the id mapfile is wrong.  You have
to write an original-id and a local-id in the other way around.

PR: 6593
Reviewed by: phk
Submitted by: Amakawa Shuhei <amakawa@hoh.t.u-tokyo.ac.jp>

26 years ago.Sh ENVIRONMENT VARIABLES -> .Sh ENVIRONMENT
Poul-Henning Kamp [Wed, 13 May 1998 07:57:49 +0000 (07:57 +0000)]
.Sh ENVIRONMENT VARIABLES -> .Sh ENVIRONMENT

PR: 6599
Reviewed by: phk
Submitted by: Josh Gilliam <josh@quick.net>

26 years agoAdd missing arg to vget().. Serves me right for committing a 2.2 patch to
Peter Wemm [Wed, 13 May 1998 07:49:08 +0000 (07:49 +0000)]
Add missing arg to vget().. Serves me right for committing a 2.2 patch to
-current without testing it there.. :-(

Submitted by: Michael Hancock <michaelh@cet.co.jp>

26 years agoCorrect use of .Nm. Add rcsid.
Philippe Charnier [Wed, 13 May 1998 07:47:35 +0000 (07:47 +0000)]
Correct use of .Nm. Add rcsid.

26 years agoCorrect use of .Nm.
Philippe Charnier [Wed, 13 May 1998 07:43:56 +0000 (07:43 +0000)]
Correct use of .Nm.

26 years agoConvert to mdoc.
Philippe Charnier [Wed, 13 May 1998 07:41:13 +0000 (07:41 +0000)]
Convert to mdoc.

26 years agoCorrect use of .Nm. Add rcsid.
Philippe Charnier [Wed, 13 May 1998 07:39:11 +0000 (07:39 +0000)]
Correct use of .Nm. Add rcsid.

26 years agoCorrect use of .Nm.
Philippe Charnier [Wed, 13 May 1998 07:35:44 +0000 (07:35 +0000)]
Correct use of .Nm.

26 years agoRestore Lite-2 sccsid. Restore include of sys/types.h.
Philippe Charnier [Wed, 13 May 1998 07:33:54 +0000 (07:33 +0000)]
Restore Lite-2 sccsid. Restore include of sys/types.h.

26 years agoRestore Lite-2 sccsid. Restore include of sys/param.h.
Philippe Charnier [Wed, 13 May 1998 07:31:42 +0000 (07:31 +0000)]
Restore Lite-2 sccsid. Restore include of sys/param.h.

26 years agoRestore include of sys/types.h and sys/param.h.
Philippe Charnier [Wed, 13 May 1998 07:29:15 +0000 (07:29 +0000)]
Restore include of sys/types.h and sys/param.h.

26 years agoThe PnP code in 2.2.6 detects the Motorola ModemSurfer 56K,
Poul-Henning Kamp [Wed, 13 May 1998 07:26:55 +0000 (07:26 +0000)]
The PnP code in 2.2.6 detects the Motorola ModemSurfer 56K,
but doesn't do much of anything with it.  I added it to siopnp_ids[]
and it was found and recognized as a serial port.

PR: 6605
Reviewed by: phk
Submitted by: Dave Marquardt <marquard@zilker.net>

26 years agoRestore Lite-2 sccsid.
Philippe Charnier [Wed, 13 May 1998 07:25:17 +0000 (07:25 +0000)]
Restore Lite-2 sccsid.

26 years agoRestore original Lite-2 sccsid. Restore include of sys/types.h.
Philippe Charnier [Wed, 13 May 1998 07:22:11 +0000 (07:22 +0000)]
Restore original Lite-2 sccsid. Restore include of sys/types.h.

26 years agoDo not remove include of <sys/param.h> and <sys/types.h>. They should
Philippe Charnier [Wed, 13 May 1998 07:19:45 +0000 (07:19 +0000)]
Do not remove include of <sys/param.h> and <sys/types.h>. They should
be here before including almost any POSIX header.
Requested by: Bruce

26 years agoRestore Lite2 sccsids by removing `const'. Improve documentation a little.
Philippe Charnier [Wed, 13 May 1998 07:16:37 +0000 (07:16 +0000)]
Restore Lite2 sccsids by removing `const'. Improve documentation a little.
Suggested by: Bruce