]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
24 years agoDo not specify the size to move. Allow the assembler to figure it out.
David E. O'Brien [Wed, 10 May 2000 01:06:39 +0000 (01:06 +0000)]
Do not specify the size to move.  Allow the assembler to figure it out.

24 years agoSupport non-i386 architectures.
David E. O'Brien [Wed, 10 May 2000 01:05:11 +0000 (01:05 +0000)]
Support non-i386 architectures.

24 years agocorrect possible security issue(s) in name resolution, due to use of
Jun-ichiro itojun Hagino [Wed, 10 May 2000 00:47:20 +0000 (00:47 +0000)]
correct possible security issue(s) in name resolution, due to use of
pre-4.9.7 BIND resolver code.
ftp://ftp.kame.net/pub/mail-list/snap-users/2348 for details.

Reviewed by: ume

24 years agoFix bug in dealing with "hlen == 1 and opt > 1"
Darren Reed [Tue, 9 May 2000 23:35:24 +0000 (23:35 +0000)]
Fix bug in dealing with "hlen == 1 and opt > 1"

24 years agoTidy up the offset calculations.
Paul Saab [Tue, 9 May 2000 22:20:14 +0000 (22:20 +0000)]
Tidy up the offset calculations.

Submitted by: bde

24 years agoWhen setting up an iomem resources, remember to set the MEM_ASSIGNED bit
Bill Paul [Tue, 9 May 2000 22:01:16 +0000 (22:01 +0000)]
When setting up an iomem resources, remember to set the MEM_ASSIGNED bit
so that pccardd actually does something to set up the memory range.

24 years agoRegenerated (fixed the calculation of sy_nargs in sysent tables).
Bruce Evans [Tue, 9 May 2000 21:52:02 +0000 (21:52 +0000)]
Regenerated (fixed the calculation of sy_nargs in sysent tables).

24 years agoDon't forget to back up svr4_syscallnames.c. Don't depend on side effects
Bruce Evans [Tue, 9 May 2000 21:40:01 +0000 (21:40 +0000)]
Don't forget to back up svr4_syscallnames.c.  Don't depend on side effects
to generate it.

24 years agoActually, overriding DISTFILES at all is just a bad idea, so remove the
John Baldwin [Tue, 9 May 2000 21:21:42 +0000 (21:21 +0000)]
Actually, overriding DISTFILES at all is just a bad idea, so remove the
variable all together.

PR: misc/14104
Submitted by: Don Croyle <croyle@gelemna.ft-wayne.in.us>

24 years agoFixed the calculation of sy_nargs in sysent tables. We attempted to do
Bruce Evans [Tue, 9 May 2000 21:18:30 +0000 (21:18 +0000)]
Fixed the calculation of sy_nargs in sysent tables.  We attempted to do
this in awk using the hack of counting args of type off_t twice and args
of all other types once.  This is too simple to work.  It gave benignly
wrong results on alphas (off_t shouldn't be counted twice) and for
svr4_sys_mmap64() on i386's (off64_t should be counted twice).  It gave
fatally wrong results for i386's with 64-bit longs (longs should be
counted twice).  The correct value for sy_nargs is easier to determine
from the size of the args struct anyway, except for complications to
make the generated code almost readable.

Improved formatting of sysent tables by lining up the comments where
possible.

24 years agoRegen.
Nick Hibma [Tue, 9 May 2000 20:45:14 +0000 (20:45 +0000)]
Regen.

24 years agoAdd Alcor Keyboard with built-in hub and its keyboard.
Nick Hibma [Tue, 9 May 2000 20:43:37 +0000 (20:43 +0000)]
Add Alcor Keyboard with built-in hub and its keyboard.

24 years agoSmall grammar gnit.
Tim Vanderhoek [Tue, 9 May 2000 20:35:13 +0000 (20:35 +0000)]
Small grammar gnit.

24 years agoAdd reference to new name DIGITAL Server 330x for AlphaServer 800
Wilko Bulte [Tue, 9 May 2000 20:30:04 +0000 (20:30 +0000)]
Add reference to new name DIGITAL Server 330x for AlphaServer 800

24 years agoThis appears to have been repo-copied to lpd a while ago and never
Warner Losh [Tue, 9 May 2000 19:41:18 +0000 (19:41 +0000)]
This appears to have been repo-copied to lpd a while ago and never
deleted here.  Diff shows there is only one change to this file, which
matches the repo histories.

24 years agoChange the "bdev-whiner" to whine when open is attempted and extend
Poul-Henning Kamp [Tue, 9 May 2000 18:53:57 +0000 (18:53 +0000)]
Change the "bdev-whiner" to whine when open is attempted and extend
the deadline a month.

24 years agoSpelling fix: despatching -> dispatching. This makes the
Mike Pritchard [Tue, 9 May 2000 18:52:26 +0000 (18:52 +0000)]
Spelling fix:  despatching -> dispatching.  This makes the
spelling consistent throughout the man page.

PR: docs/18467
Submitted by: Alexander N. Kabaev <ak03@gte.com>

24 years agoCorrect a patch name and use the correct macro for displaying it.
Mike Pritchard [Tue, 9 May 2000 18:44:34 +0000 (18:44 +0000)]
Correct a patch name and use the correct macro for displaying it.

PR: docs/18468
Submitted by: Anatoly Vorobey <mellon@pobox.com>

24 years agoRemove metion of the 'magic options.' Apparently PXE doesn't require the
Doug White [Tue, 9 May 2000 18:28:07 +0000 (18:28 +0000)]
Remove metion of the 'magic options.'  Apparently PXE doesn't require the
special options on 0.99c (it mumbles something about 'PXE server not found'
but works anyway), and it won't work at all with PXE 2.0.

24 years agoMove UKBD_DFLT_KEYMAP into the options file since USB is machine
John Baldwin [Tue, 9 May 2000 18:23:38 +0000 (18:23 +0000)]
Move UKBD_DFLT_KEYMAP into the options file since USB is machine
independent.  This allows the ukbd0 driver to compile on the Alpha.

Approved by: o`brien (in principle)

24 years agoFixed the return type for exit() and the args struct tag several NOPROTO
Bruce Evans [Tue, 9 May 2000 18:23:16 +0000 (18:23 +0000)]
Fixed the return type for exit() and the args struct tag several NOPROTO
syscalls including exit().  These entries were unused, so the bugs had no
effect, but the the args struct tag will be used to calculate sy_nargs
correctly.  exit() was wrong in all emulators.

24 years agoCorrect "it's" to "its".
Matthew Hunt [Tue, 9 May 2000 18:08:52 +0000 (18:08 +0000)]
Correct "it's" to "its".

PR: 18469
Submitted by: Anatoly Vorobey <mellon@pobox.com>

24 years agoFixed the return type and args struct tag for exit(). They were wrong in
Bruce Evans [Tue, 9 May 2000 18:08:51 +0000 (18:08 +0000)]
Fixed the return type and args struct tag for exit().  They were wrong in
all emulators.  These entries were unused, so the bug had no effect, but
the the args struct tag will be used to calculate sy_nargs correctly.

24 years agoFixed the return type and args struct tag for exit(). They were wrong in
Bruce Evans [Tue, 9 May 2000 17:54:30 +0000 (17:54 +0000)]
Fixed the return type and args struct tag for exit().  They were wrong in
all emulators.  These entries were unused, so the bug had no effect, but
the the args struct tag will be used to calculate sy_nargs correctly.

24 years ago Some ioctl routines assume that the ioctl buffer is aligned, but a
Matthew Dillon [Tue, 9 May 2000 17:43:21 +0000 (17:43 +0000)]
Some ioctl routines assume that the ioctl buffer is aligned, but a
    char[] declaration makes no such guarentee.  A union is used to force
    alignment of the char buffer.

24 years agoChange floating point exception type to match the i386 one.
David E. O'Brien [Tue, 9 May 2000 17:43:20 +0000 (17:43 +0000)]
Change floating point exception type to match the i386 one.

Submitted by: Mark Abene <phiber@radicalmedia.com>

24 years agoAdd missing include machine/in_cksum.h.
Paul Saab [Tue, 9 May 2000 16:56:51 +0000 (16:56 +0000)]
Add missing include machine/in_cksum.h.

Submitted by: n_hibma

24 years agoThe printenv(1) command is now a builtin(1) in csh(1).
Sheldon Hearn [Tue, 9 May 2000 15:02:03 +0000 (15:02 +0000)]
The printenv(1) command is now a builtin(1) in csh(1).

24 years agoUpdate the builtin(1) manpage to reflect changes since the csh(1)
Sheldon Hearn [Tue, 9 May 2000 14:59:06 +0000 (14:59 +0000)]
Update the builtin(1) manpage to reflect changes since the csh(1)
to tcsh(1) upgrade.  The following commands were added as builtins:

bindkey
builtins
complete
echotc
filetest
hup
log
ls-F
printenv
sched
settc
setty
telltc
uncomplete
where

The printf builtin was removed.

24 years agoFix miscellaneous mdoc macro argument limit infringements.
Sheldon Hearn [Tue, 9 May 2000 14:02:06 +0000 (14:02 +0000)]
Fix miscellaneous mdoc macro argument limit infringements.

PR: 18465
Reported by: Kazu TAKAMUNE <takamune@avrl.mei.co.jp>

24 years agoAdd '-DPC98' to CFLAGS if MACHINE == pc98.
Yoshihiro Takahashi [Tue, 9 May 2000 13:46:14 +0000 (13:46 +0000)]
Add '-DPC98' to CFLAGS if MACHINE == pc98.

24 years agoWork around the mdoc argument limit in the Nd name description.
Sheldon Hearn [Tue, 9 May 2000 13:01:04 +0000 (13:01 +0000)]
Work around the mdoc argument limit in the Nd name description.

PR: 18465
Submitted by: Kazu TAKAMUNE <takamune@avrl.mei.co.jp>

24 years agoCreate a separate target for the 'modules'. Without this the modules
Nick Hibma [Tue, 9 May 2000 10:39:38 +0000 (10:39 +0000)]
Create a separate target for the 'modules'. Without this the modules
are only rebuilt if the kernel has changed as well.

While there make 'all' and 'modules' a PHONY target.

24 years agoRegenerated (fixed the type of mmap()'s padding arg).
Bruce Evans [Tue, 9 May 2000 08:35:51 +0000 (08:35 +0000)]
Regenerated (fixed the type of mmap()'s padding arg).

24 years agoFixed the declaration of mmap(). The crufty padding arg had the wrong
Bruce Evans [Tue, 9 May 2000 08:31:06 +0000 (08:31 +0000)]
Fixed the declaration of mmap().  The crufty padding arg had the wrong
type.  This gave an inconsistent amount of crufty padding on i386's with
64-bit longs (8 bytes instead of 4).  On alphas it gives a consistent
amount of crufty padding (8 bytes) in addition to the 4 bytes of normal
padding caused by passing int args as register_t's.

Fixed the args struct tag for the NOPROTO syscalls (netbsd_lchown() and
netbsd_msync()).  The tag is currently unused for NOPROTO syscalls, so
the bug has no effect, but it will be used even in the NOPROTO case to
calculate sy_nargs correctly.

24 years agoFix nits in previous commit: hard sentence break and grammar.
Sheldon Hearn [Tue, 9 May 2000 08:09:31 +0000 (08:09 +0000)]
Fix nits in previous commit:  hard sentence break and grammar.

24 years agoSome prettifying for the quirk comments.
Matt Jacob [Tue, 9 May 2000 04:54:10 +0000 (04:54 +0000)]
Some prettifying for the quirk comments.

Add a SA_QUIRK_NO_MODESEL type and use it for the OnStream real SCSI
device (not the broken one). This one is still broken in that it can't
be set to the same fixed block size it reports [ unflattering comments
about this company elided ].

If we're unable to set buffered mode on, complain, but drive on. It's
not a fatal error to not be in buffered mode.

24 years agoExit if we read two EOFs from the keyboard input stream. Why two and not
Tim Vanderhoek [Tue, 9 May 2000 04:13:49 +0000 (04:13 +0000)]
Exit if we read two EOFs from the keyboard input stream.  Why two and not
one?  Just 'cause.

PR: misc/14932

24 years agoDon't segv if viewing "#" and there is no prev-file (long-standing).
Tim Vanderhoek [Tue, 9 May 2000 03:53:13 +0000 (03:53 +0000)]
Don't segv if viewing "#" and there is no prev-file (long-standing).
Staticize a variable that was meant to be static (broken in r.1.11).

PR: bin/11370 Henry Whincup <henry@techiebod.com> (for the segv)

24 years ago.Xr vgl 3 ,
Tim Vanderhoek [Tue, 9 May 2000 02:52:39 +0000 (02:52 +0000)]
.Xr vgl 3 ,

PR: kern/14566

24 years agoChange references/comments about 'secondary' to reflect that while we'd
Matt Jacob [Tue, 9 May 2000 02:20:44 +0000 (02:20 +0000)]
Change references/comments about 'secondary' to reflect that while we'd
like to see the true SRM bus number be passed to us, instead, we get FreeBSD's
PCI bus instance number (Brzzt! Wrong Answer!).

Also, once we've seen the MCPCIA that has the EISA bus on it, call
dec_kn300_cons_init just before configuring devices on this bus.

24 years agoUnhide dec_kn300_cons_init and don't make it something we call early.
Matt Jacob [Tue, 9 May 2000 02:19:04 +0000 (02:19 +0000)]
Unhide dec_kn300_cons_init and don't make it something we call early.
We'll call it later when, in the natural order of things, we configure
the MCPCIA that has the EISA bus that a serial console and/or VGA and/or
keyboard.

24 years agoFix some whitespace brokeness.. (usio and ctx)
Peter Wemm [Tue, 9 May 2000 01:15:55 +0000 (01:15 +0000)]
Fix some whitespace brokeness.. (usio and ctx)

24 years agoFix some breakage about how we build WWNs. Do some other fabric related
Matt Jacob [Tue, 9 May 2000 01:14:43 +0000 (01:14 +0000)]
Fix some breakage about how we build WWNs. Do some other fabric related
changes: consider a new PDB entry different if Class 3 service parameter
roles change (!!!). Do some checking as we're getting a port database
that traps whether things change while we're doing so. Handle N-port
and F-ports correctly. Fix the fabric login loop to retain a login/binding
if things haven't changed (I mean, why logout a device only to log it back
in). No longer accept, after fabric logins, garbage if we can't get a PDB
entry that matches the device we've just logged into- if it doesn't, log
it out as it is very unlikely to still be what we thought it was. Get rid
of some of the debounce loops because we could get stuck there.

24 years agoroll platform minor
Matt Jacob [Tue, 9 May 2000 01:09:46 +0000 (01:09 +0000)]
roll platform minor

24 years agoRoll core minor version. Change our 'fabdev' tag to 'loggedin'.
Matt Jacob [Tue, 9 May 2000 01:09:23 +0000 (01:09 +0000)]
Roll core minor version. Change our 'fabdev' tag to 'loggedin'.

24 years agoAdd in a watchdog routine to catch cases where we've dropped the command.
Matt Jacob [Tue, 9 May 2000 01:08:21 +0000 (01:08 +0000)]
Add in a watchdog routine to catch cases where we've dropped the command.
Apparently the f/w has finished the command, but somehow an interrupt is
being lost. So, we just plain wedge when booting alphas.

This is a general routine we've needed for a while.

24 years agoRead the correct size into kdumplo.
Paul Saab [Tue, 9 May 2000 01:08:00 +0000 (01:08 +0000)]
Read the correct size into kdumplo.

Botched/Noticed by: peter

24 years agoThe storage for WWN from NVRAM is actually the PORT WWN, not the NODE WWN.
Matt Jacob [Tue, 9 May 2000 01:06:47 +0000 (01:06 +0000)]
The storage for WWN from NVRAM is actually the PORT WWN, not the NODE WWN.

24 years agoConrrect a macro with parenthesis.
Matt Jacob [Tue, 9 May 2000 01:06:18 +0000 (01:06 +0000)]
Conrrect a macro with parenthesis.

24 years agoMake issetugid return correctly. It was returning -1 with
Peter Wemm [Tue, 9 May 2000 00:58:34 +0000 (00:58 +0000)]
Make issetugid return correctly.  It was returning -1 with
errno == 1 if it was set?id!

Submitted by:  Valentin Nechayev <netch@segfault.kiev.ua>

24 years agoFinish off replacing the rest of the duplicate mknod's with hard links.
Peter Wemm [Tue, 9 May 2000 00:07:46 +0000 (00:07 +0000)]
Finish off replacing the rest of the duplicate mknod's with hard links.

24 years agoInclude machine/in_cksum.h to unbreak options MROUTING.
Kenneth D. Merry [Mon, 8 May 2000 23:56:30 +0000 (23:56 +0000)]
Include machine/in_cksum.h to unbreak options MROUTING.

24 years agoAllow savecore to work on > 2GB partitions.
Paul Saab [Mon, 8 May 2000 22:57:35 +0000 (22:57 +0000)]
Allow savecore to work on > 2GB partitions.

Submitted by: peter

24 years agoUse printf (shell builtin) instead of cut - cut isn't in /bin.
Peter Wemm [Mon, 8 May 2000 22:56:17 +0000 (22:56 +0000)]
Use printf (shell builtin) instead of cut - cut isn't in /bin.
This should fix the ttyA* and cuaA* up to the maximum possible 128 nodes.

24 years agoAdd a DOCDISTFILES variable to point to a directory containing the
John Baldwin [Mon, 8 May 2000 22:21:17 +0000 (22:21 +0000)]
Add a DOCDISTFILES variable to point to a directory containing the
distfiles to build the docproj port during release.  Prior to this, the
distfiles were copied froom ${DISTFILES}.  However, if like me, you needed
to store the distfiles in some directory other than /usr/ports/distfiles,
for example /usr/docdistfiles, then when you overrode ${DISTFILES} it broke
the release build when it tried to build docproj port in the chroot'd
release area.  To preserve backwards compatibility, DOCDISTFILES defaults
to the value of DISTFILES.

Ok'd by: jkh

24 years agoAdd [en]sa devices for othoginality.
David E. O'Brien [Mon, 8 May 2000 21:32:24 +0000 (21:32 +0000)]
Add [en]sa devices for othoginality.

24 years agoDescribe firewall_logging option
Andrey A. Chernov [Mon, 8 May 2000 21:10:33 +0000 (21:10 +0000)]
Describe firewall_logging option

24 years agoAdd an explicit rule number to natd so you do not end up with two
Paul Saab [Mon, 8 May 2000 20:28:20 +0000 (20:28 +0000)]
Add an explicit rule number to natd so you do not end up with two
rule 100's.

Submitted by: Jan Koum <jkb@yahoo-inc.com>

24 years agoRemove the MLINKS entry for cwd, since it's no longer listed in the
Sheldon Hearn [Mon, 8 May 2000 16:23:56 +0000 (16:23 +0000)]
Remove the MLINKS entry for cwd, since it's no longer listed in the
builtin.1 page.

24 years agoRemove the bogusly inserted ``cwd'' which isn't a builtin command
Sheldon Hearn [Mon, 8 May 2000 15:59:23 +0000 (15:59 +0000)]
Remove the bogusly inserted ``cwd'' which isn't a builtin command
to sh, csh nor tcsh.

Fix the header line of the table to use cross-references and
emphasis.

24 years agoUse the new and shiny module versioning. This enables you to load the
Nick Hibma [Mon, 8 May 2000 15:10:05 +0000 (15:10 +0000)]
Use the new and shiny module versioning. This enables you to load the
ums driver (mouse) and get the usb driver at the same time.

This also shows that
MODULE_VERSION(usb, 1);
MODULE_DEPEND(usb, usb, 1, 1, 1);
in the same module in one source file works.

24 years agoAdded PC-98 supports.
Yoshihiro Takahashi [Mon, 8 May 2000 14:55:21 +0000 (14:55 +0000)]
Added PC-98 supports.

Obtained from: FreeBSD(98) 4.0R-Rev01

24 years agoFix DELAY function to use the RPCC cycle counter register.
Matt Jacob [Sun, 7 May 2000 22:44:27 +0000 (22:44 +0000)]
Fix DELAY function to use the RPCC cycle counter register.

24 years agoFor tape devices, 'r' means REWIND, not RAW.
Mike Smith [Sun, 7 May 2000 21:47:07 +0000 (21:47 +0000)]
For tape devices, 'r' means REWIND, not RAW.

24 years agoWhile we're at it, add a length field too in case we want to fully implement
Dag-Erling Smørgrav [Sun, 7 May 2000 20:52:40 +0000 (20:52 +0000)]
While we're at it, add a length field too in case we want to fully implement
partial fetches later.

24 years agoImplement restart
Dag-Erling Smørgrav [Sun, 7 May 2000 20:51:31 +0000 (20:51 +0000)]
Implement restart

24 years agoForgot to mention: The previous patch undoes rev1.21 - rev1.22 change by
Nick Hibma [Sun, 7 May 2000 20:48:16 +0000 (20:48 +0000)]
Forgot to mention: The previous patch undoes rev1.21 - rev1.22 change by
jlemon.

24 years agoIf PERIPH_1284 is not defined, don't even bother calling the BUS_*_INTR
Nick Hibma [Sun, 7 May 2000 20:46:49 +0000 (20:46 +0000)]
If PERIPH_1284 is not defined, don't even bother calling the BUS_*_INTR
functions.

If it is defined, check whether bus_alloc_resource has succeeded. If it
hasn't, it is in polled mode. Mike Nowlin reports that this change makes
the geek port (whatever _that_ is :-) work again on his machine.

Submitted by: Mike Nowlin <mike@argos.org>

24 years agoAdd the fields needed to store the cookies.
Nick Hibma [Sun, 7 May 2000 20:17:30 +0000 (20:17 +0000)]
Add the fields needed to store the cookies.

24 years agoDo the attach (and detach) properly. Store the cookies.
Nick Hibma [Sun, 7 May 2000 20:16:25 +0000 (20:16 +0000)]
Do the attach (and detach) properly. Store the cookies.

This is a first step to make usb unloadable.

24 years agoUse sizeof more consistently.
Dag-Erling Smørgrav [Sun, 7 May 2000 20:02:52 +0000 (20:02 +0000)]
Use sizeof more consistently.

24 years agoUse sizeof more consistently.
Dag-Erling Smørgrav [Sun, 7 May 2000 20:01:55 +0000 (20:01 +0000)]
Use sizeof more consistently.
Outline sanity checking of server port spec.

24 years agoImplement restart.
Dag-Erling Smørgrav [Sun, 7 May 2000 20:00:12 +0000 (20:00 +0000)]
Implement restart.
Use sizeof more consistently.
Outline sanity checking of server port spec.

24 years agoImplement restart
Dag-Erling Smørgrav [Sun, 7 May 2000 19:56:07 +0000 (19:56 +0000)]
Implement restart

24 years agoAdd the Prolific PL2301 and PL2302 Link cables.
Nick Hibma [Sun, 7 May 2000 19:49:48 +0000 (19:49 +0000)]
Add the Prolific PL2301 and PL2302 Link cables.

Submitted by: Dirk van Gulik <dirkx@webweaving.org>

24 years agoAdd offset field to struct url
Dag-Erling Smørgrav [Sun, 7 May 2000 19:46:33 +0000 (19:46 +0000)]
Add offset field to struct url

24 years agoWe match on vendor/product so return the fact that udbp is vendor/product
Nick Hibma [Sun, 7 May 2000 19:46:20 +0000 (19:46 +0000)]
We match on vendor/product so return the fact that udbp is vendor/product
specific.

24 years agoRegen.
Nick Hibma [Sun, 7 May 2000 19:44:44 +0000 (19:44 +0000)]
Regen.

24 years agoSync with NetBSD
Nick Hibma [Sun, 7 May 2000 19:44:16 +0000 (19:44 +0000)]
Sync with NetBSD

24 years agoRemove debugging cruft.
Nick Hibma [Sun, 7 May 2000 19:13:19 +0000 (19:13 +0000)]
Remove debugging cruft.

24 years agoInitialise the debug value.
Nick Hibma [Sun, 7 May 2000 18:40:44 +0000 (18:40 +0000)]
Initialise the debug value.

24 years agoIncrease the debugging level on a printf.
Nick Hibma [Sun, 7 May 2000 18:38:41 +0000 (18:38 +0000)]
Increase the debugging level on a printf.

24 years agoAdd a define for the Zip 250
Nick Hibma [Sun, 7 May 2000 18:37:13 +0000 (18:37 +0000)]
Add a define for the Zip 250

24 years agoDon't claim devices that we fail to probe for. Reverse the sense of a test
Warner Losh [Sun, 7 May 2000 18:20:40 +0000 (18:20 +0000)]
Don't claim devices that we fail to probe for.  Reverse the sense of a test
so that we return failure when we've failed and succeed when we succeed.
Otherwise, users always get a sn device :-(.

PR: 18431
Submitted by: <sanpei@sanpei.org> MIHIRA Yoshiro-san

24 years ago*sigh* I must have been on something that night. Make xpt_periph an
Nick Hibma [Sun, 7 May 2000 18:04:50 +0000 (18:04 +0000)]
*sigh* I must have been on something that night. Make xpt_periph an
extern with the original in cam_xpt.c instead of replicating
xpt_periph in all the sources using it (and hence not initialising it)

24 years agooops! final piece for Rawhides
Matt Jacob [Sun, 7 May 2000 17:44:50 +0000 (17:44 +0000)]
oops! final piece for Rawhides

24 years agos/ffs_unmonut/ffs_unmount/ in a gratuitous ufs_extattr printf.
Robert Watson [Sun, 7 May 2000 17:21:08 +0000 (17:21 +0000)]
s/ffs_unmonut/ffs_unmount/ in a gratuitous ufs_extattr printf.

Reported by: knu

24 years agoFix checksum calculations. This should fix the network problems
Paul Saab [Sun, 7 May 2000 16:41:15 +0000 (16:41 +0000)]
Fix checksum calculations.  This should fix the network problems
in current where all packets were returning with bad checksums.
(observed with netstat -s).

Reviewed by: alfred

24 years agoMove _detach to be underneath _attach.
Duncan Barclay [Sun, 7 May 2000 16:03:36 +0000 (16:03 +0000)]
Move _detach to be underneath _attach.

Move the code that checks the card into _probe and out of _attach and
shuffle a couple of things around in attach.

Set ifp in the variable block - catches unused variables.

24 years agoMove variables to if_rayvar.h
Duncan Barclay [Sun, 7 May 2000 15:15:06 +0000 (15:15 +0000)]
Move variables to if_rayvar.h

24 years agoSplit out structures to make it a bit easier to find my way around and
Duncan Barclay [Sun, 7 May 2000 15:12:18 +0000 (15:12 +0000)]
Split out structures to make it a bit easier to find my way around and
be more like sys/dev coding.

24 years agoSync with an early version of the -current code. Mainly a tidy up
Duncan Barclay [Sun, 7 May 2000 15:00:06 +0000 (15:00 +0000)]
Sync with an early version of the -current code. Mainly a tidy up
of some function names, dropping the ISA bits that were never used and
removing __P

We've renamed the _pccard_ stuff as nothing else probes or attaches.

In the next commit we will move some routines around and split out some of the
structure definitions to another file.

I've dropped the SYSCTL as I never got it working. APM is gone too as it
is broken.

Note, that the attr_read/write will be in both this and NEWCARD/NEWBUS so
I've moved the prototypes to the right place.
Add a few casts.

24 years agoFixed bus_space_{read,write}_region_* functions.
Yoshihiro Takahashi [Sun, 7 May 2000 14:46:15 +0000 (14:46 +0000)]
Fixed bus_space_{read,write}_region_* functions.

Pointed out: Motomichi Matsuzaki <mzaki@e-mail.ne.jp>

24 years agoInclude a RFC 1533 "Maximum DHCP Message Size" option in our request.
Poul-Henning Kamp [Sun, 7 May 2000 14:29:19 +0000 (14:29 +0000)]
Include a RFC 1533 "Maximum DHCP Message Size" option in our request.

ISC DHCP will limit the reply length to 64 bytes for bootp replies
unless we explicitly tell it we can do more.  We tell it that we can do
1200 bytes.

24 years agoMore static -> Static to keep OpenBSD happy.
Nick Hibma [Sun, 7 May 2000 14:23:32 +0000 (14:23 +0000)]
More static -> Static to keep OpenBSD happy.

24 years agoRemove stale file - moved to new location
Andrey A. Chernov [Sun, 7 May 2000 13:34:40 +0000 (13:34 +0000)]
Remove stale file - moved to new location

24 years agoReserve 145 for DRI graphics system.
Doug Rabson [Sun, 7 May 2000 13:31:41 +0000 (13:31 +0000)]
Reserve 145 for DRI graphics system.

24 years agoCorrect a bad bug in m_prepend()
Brian Somers [Sun, 7 May 2000 10:08:24 +0000 (10:08 +0000)]
Correct a bad bug in m_prepend()

Submitted by: luoqi

24 years agoWork-around a couple of C1010 quirks:
Gerard Roudier [Sun, 7 May 2000 09:54:33 +0000 (09:54 +0000)]
Work-around a couple of C1010 quirks:
- Reload SCNTL3 after selection from host   (C1010-33).
- Reload SCNTL4 prior to any DATA OUT phase (C1010-66).
- Use max SCSI offset 31 for ST but 62 for DT.