]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
25 years agoAdd mbr: a standard master boot record intended to replace a couple
Robert Nordier [Mon, 24 May 1999 22:36:24 +0000 (22:36 +0000)]
Add mbr: a standard master boot record intended to replace a couple
that are presently embedded in the fdisk and sysinstall code.

25 years agoDon't try to allocate memory resource unless address is non-zero.
Doug Rabson [Mon, 24 May 1999 18:50:41 +0000 (18:50 +0000)]
Don't try to allocate memory resource unless address is non-zero.

25 years agoFix i4bteld entry, the section of the sh case for "i4teld*)" should be
Hellmuth Michaelis [Mon, 24 May 1999 18:01:03 +0000 (18:01 +0000)]
Fix i4bteld entry, the section of the sh case for "i4teld*)" should be
BEFORE the case for "i4tel*)". (match the longest prefix first!)
Recognized by: Harry Starr

25 years agoUpdate PicoBSD kernel config files. Changes are
Roger Hardiman [Mon, 24 May 1999 17:27:30 +0000 (17:27 +0000)]
Update PicoBSD kernel config files. Changes are
  removal of bio, tty, net
  removal of quotes
  switches from isa? to nexus? or atkbdc?
  additional comments

  These bring the kernel config files in sync with those in
  RELENG_3

25 years agoCorrect the ``ignoring sync/async'' warnings so that they show
Brian Somers [Mon, 24 May 1999 16:39:17 +0000 (16:39 +0000)]
Correct the ``ignoring sync/async'' warnings so that they show
up with the correct device type.
Reassign the correct tcpdevice or execdevice after transfering
a link in MP server mode.

25 years agoFix bug that can cause transmit corruption. There are actually two 'rings'
Bill Paul [Mon, 24 May 1999 14:56:55 +0000 (14:56 +0000)]
Fix bug that can cause transmit corruption. There are actually two 'rings'
in the transmit code: the TX descriptor ring, and a 'shadow' ring of mbuf
pointers, one for each TX descriptor. When transmitting a packet that
consists of several fragments in an mbuf chain, we link each fragment
to a descriptor in the TX ring, but we only save a pointer to the mbuf
chain. This pointer is saved in the shadow ring entry which corresponds
to the first fragment in the packet. Later, ti_txeof() can release the
whole chain with a single m_freem() call. (We need the second ring to
keep track of the virtual addresses of the mbuf chains.)

The problem with this is that the Tigon isn't actually through with the
mbuf chain until it reaches the last fragment (which has the TI_BDFLAG_END
bit set), however the current scheme releases the mbuf chain as soon as
the first fragment is consumed. This is wrong, since the mbufs can then
be yanked out from under the Tigon and modified before the other fragments
can be transmitted.

The fix is to make a one line change to ti_encap() so that it saves the
mbuf chain pointer in the shadow ring entry that corresponds to the last
fragment in TX ring instead of the first. This prevents the mbufs from
being released until the last fragment is transmitted.

Painstakingly diagnosed and fixed by: Robert Picco <picco@mail.wevinc.com>
Brought to my attention by: dg

25 years agoclose pr 10889:
Luigi Rizzo [Mon, 24 May 1999 10:01:22 +0000 (10:01 +0000)]
close pr 10889:
+ add a missing call to dn_rule_delete() when flushing firewall
  rules, thus preventing possible panics due to dangling pointers
  (this was already done for single rule deletes).
+ improve "usage" output in ipfw(8)
+ add a few checks to ipfw pipe parameters and make it a bit more
  tolerant of common mistakes (such as specifying kbit instead of Kbit)

PR: kern/10889
Submitted by: Ruslan Ermilov

25 years agoDon't warn about ``!'' lines being unindented commands in ppp.conf.
Brian Somers [Mon, 24 May 1999 09:04:35 +0000 (09:04 +0000)]
Don't warn about ``!'' lines being unindented commands in ppp.conf.

25 years agoAdded a missing comma to the static condition variable initialisation
John Birrell [Mon, 24 May 1999 07:22:55 +0000 (07:22 +0000)]
Added a missing comma to the static condition variable initialisation
definition.

Submitted by: David Leonard <David.Leonard@csee.uq.edu.au>, an OpenBSD guy.

25 years agoFix a [start,end] vs [start,count] botch that corrupted the resource
Peter Wemm [Mon, 24 May 1999 03:08:46 +0000 (03:08 +0000)]
Fix a [start,end] vs [start,count] botch that corrupted the resource
manager and prevented IOPort allocation beyond the first EISA slot from
working.  subr_rman.c should have trapped this on the way into the system
rather than tripping over the wreckage.

Head banged into wall repeatedly by:  "Matthew N. Dodd" <winter@jurai.net>

25 years agoRemove the static declaration from the line[] variable to allow
John Birrell [Mon, 24 May 1999 01:15:28 +0000 (01:15 +0000)]
Remove the static declaration from the line[] variable to allow
openpty() to be called from a threaded application.

25 years agoRemove the mfs stuff now that mfs_init() does the MFS_ROOT initialisation.
John Birrell [Mon, 24 May 1999 00:54:17 +0000 (00:54 +0000)]
Remove the mfs stuff now that mfs_init() does the MFS_ROOT initialisation.

25 years agoBack out my previous change (phk didn't like it) in favour of setting
John Birrell [Mon, 24 May 1999 00:37:26 +0000 (00:37 +0000)]
Back out my previous change (phk didn't like it) in favour of setting
rootdev in the mfs initialisation code iff MFS_ROOT (which Bruce doesn't
like). Damned if I do - damned if I don't.

25 years agoFix typo in dump reporting
Jean-Marc Zucconi [Mon, 24 May 1999 00:35:47 +0000 (00:35 +0000)]
Fix typo in dump reporting

PR: bin/10573
Submitted by: Christian Weisgerber <naddy@mips.rhein-neckar.de>

25 years agoRemove the test for bdevsw(dev) == NULL from bdevvp() because it fails
John Birrell [Mon, 24 May 1999 00:34:10 +0000 (00:34 +0000)]
Remove the test for bdevsw(dev) == NULL from bdevvp() because it fails
if there is no character device associated with the block device. In this
case that doesn't matter because bdevvp() doesn't use the character
device structure.

I can use the pointy bit of the axe too.

25 years ago- Make setroot() conditional on FFS etc, to avoid a compiler warning
John Birrell [Mon, 24 May 1999 00:30:49 +0000 (00:30 +0000)]
- Make setroot() conditional on FFS etc, to avoid a compiler warning
  on systems with no FFS.
- Remove all references to mfs from cpu_rootconf(). mfs_init is
  called prior to cpu_rootconf(), so it can set mountrootfsname to mfs
  and (more imporantly) set rootdev using the (bogus in Bruce's opinion)
  special major number of 255.

25 years ago- Back out Luoqi's cdevsw stuff. It panics on my system and is not required.
John Birrell [Mon, 24 May 1999 00:27:12 +0000 (00:27 +0000)]
- Back out Luoqi's cdevsw stuff. It panics on my system and is not required.
- Fix an error message.
- Do the MFS_ROOT setting of mountrootfsname in mfs_init() instead of
  cpu_rootconf().
- Set rootdev in mfs_init instead of later in mfs_mount() iff MFS_ROOT.

25 years agogetopt(3) returns -1 not EOF.
Warner Losh [Sun, 23 May 1999 23:24:26 +0000 (23:24 +0000)]
getopt(3) returns -1 not EOF.

25 years agoGetopt(3) returns -1 not EOF.
Warner Losh [Sun, 23 May 1999 23:18:38 +0000 (23:18 +0000)]
Getopt(3) returns -1 not EOF.

25 years agoGetopt(3) returns -1, not EOF.
Warner Losh [Sun, 23 May 1999 23:17:28 +0000 (23:17 +0000)]
Getopt(3) returns -1, not EOF.

25 years agoAdded rgb_vbi_prog() to capture VBI data and video at the
Roger Hardiman [Sun, 23 May 1999 21:40:51 +0000 (21:40 +0000)]
Added rgb_vbi_prog() to capture VBI data and video at the
same time. To capture VBI data, /dev/vbi must be opened
before starting video capture.
A partly on code from Hiroki Mori <mori@infocity.co.jp>

25 years agoUse RELENG_3 as an example, not RELENG_2_2.
Brian Somers [Sun, 23 May 1999 20:01:20 +0000 (20:01 +0000)]
Use RELENG_3 as an example, not RELENG_2_2.

25 years agoRely on the statistics in XPT_GDEV_STATS instead of the versions still
Justin T. Gibbs [Sun, 23 May 1999 18:57:29 +0000 (18:57 +0000)]
Rely on the statistics in XPT_GDEV_STATS instead of the versions still
retained in XPT_GDEV_TYPE for binary compatibility.  Mark the legacy
structure values for removal when we bump the major CAM revision.

25 years agoDon't reference our SCB until we have validated that the firmware has
Justin T. Gibbs [Sun, 23 May 1999 18:55:58 +0000 (18:55 +0000)]
Don't reference our SCB until we have validated that the firmware has
returned an SCB that is in range.

25 years agoPoll for interrupts in bt_cmd in case they are masked. A completing
Justin T. Gibbs [Sun, 23 May 1999 18:54:34 +0000 (18:54 +0000)]
Poll for interrupts in bt_cmd in case they are masked.  A completing
mailbox command will block the completion of an immediate command,
so we must service them for our command to succeed.

PR: 11799

25 years agoGet case right in "COMPAT3X".
David E. O'Brien [Sun, 23 May 1999 17:57:23 +0000 (17:57 +0000)]
Get case right in "COMPAT3X".

Submitted by: Kevin Street <street@iname.com>

25 years agoxargs was spliting the input in a very different way from everyone else,
Jean-Marc Zucconi [Sun, 23 May 1999 15:58:22 +0000 (15:58 +0000)]
xargs was spliting the input in a very different way from everyone else,
and was also a bit inconsistent: leading blanks, or any double blanks
generated empty arguments, but a trailing blank did not.

PR: bin/2630, bin/10914
Submitted by: Arne Henrik Juul <arnej@imf.unit.no>

25 years agobrucify
Brian Somers [Sun, 23 May 1999 13:52:05 +0000 (13:52 +0000)]
brucify
Mentioned by: sprice@hiwaay.net

25 years agoUse the correct width integer on both i386 and alpha.
Doug Rabson [Sun, 23 May 1999 13:43:04 +0000 (13:43 +0000)]
Use the correct width integer on both i386 and alpha.

Reviewed by: Steve Price <sprice@hiwaay.net>

25 years agoDon't reference non-existant ATAPI option..
Peter Wemm [Sun, 23 May 1999 11:11:31 +0000 (11:11 +0000)]
Don't reference non-existant ATAPI option..

PR: 11814
Submitted by: Sheldon Hearn <sheldonh@uunet.co.za>

25 years agoFix a problem with static initialisation of mutexes and condition
John Birrell [Sun, 23 May 1999 10:55:33 +0000 (10:55 +0000)]
Fix a problem with static initialisation of mutexes and condition
variables.

Submitted by: Dan Eischen <eischen@vigrid.com>

25 years agoMake MFS_ROOT work again. MFS_ROOT means that rootdev is not set.
John Birrell [Sun, 23 May 1999 10:51:33 +0000 (10:51 +0000)]
Make MFS_ROOT work again. MFS_ROOT means that rootdev is not set.

Broken by: phk
Problem ignored by: phk

25 years agoInstall g2c.h header incase others want to call libg2c from C.
David E. O'Brien [Sun, 23 May 1999 10:02:46 +0000 (10:02 +0000)]
Install g2c.h header incase others want to call libg2c from C.

Submitted by: Steve

25 years agoAdd a missing backslash from the last commit. (Broke installworld)
John Birrell [Sun, 23 May 1999 02:11:41 +0000 (02:11 +0000)]
Add a missing backslash from the last commit. (Broke installworld)

25 years agoConstruct the profile file name from the name of the executable. A program
Jean-Marc Zucconi [Sun, 23 May 1999 00:37:56 +0000 (00:37 +0000)]
Construct the profile file name from the name of the executable. A program
compiled with -pg and run will generate a file <executable-filename>.gmon,
not gmon.out.

PR: bin/8426

25 years agoInitialize exit code so that the program exits with code 0 on success.
Jean-Marc Zucconi [Sat, 22 May 1999 22:56:26 +0000 (22:56 +0000)]
Initialize exit code so that the program exits with code 0 on success.

PR: bin/10904
Submitted by: Ben Smithurst <ben@scientia.demon.co.uk>

25 years agoAdd the amd driver.
Justin T. Gibbs [Sat, 22 May 1999 22:10:07 +0000 (22:10 +0000)]
Add the amd driver.

25 years agoaic7xxx.c:
Justin T. Gibbs [Sat, 22 May 1999 22:04:11 +0000 (22:04 +0000)]
aic7xxx.c:
Honor the 'bus reset at startup' option now that the XPT properly
handles transfer negotiation in this scenario.

Honor the sync rate settings on Ultra2 controllers.  We would
always negotiate at the fastest speed.  Oops.

aic7xxx.h:
Whitespace.

aic7xxx.seq:
Fix a minor nit that would cause the controller to miss the update
of the negotiation required bitmask causing the negotiation to
be delayed by a command.

25 years agoAdd a default async handler funstion to cam_periph.c to remove duplicated
Justin T. Gibbs [Sat, 22 May 1999 22:00:24 +0000 (22:00 +0000)]
Add a default async handler funstion to cam_periph.c to remove duplicated
code in all initiator type peripheral drivers.

scsi_target.c:
Release ATIO structures that wind up in the 'unkown command queue'
for consumption by our userland counterpart, back to the controller
when the exception for that command is cleared.

25 years agoAdd the XPT_PATH_STATS and XPT_GDEV_STATS function codes. These ccb
Justin T. Gibbs [Sat, 22 May 1999 21:58:47 +0000 (21:58 +0000)]
Add the XPT_PATH_STATS and XPT_GDEV_STATS function codes.  These ccb
types allow the reporting of error counts and other statistics.  Currently
we provide information on the last BDR or bus reset as well as active
transaction inforamtion, but this will be expanded as more information is
added to aid in error recovery.

Use the 'last reset' information to better handle bus settle delays.
Peripheral drivers now control whether a bus settle delay occurs and
for how long.  This allows target mode peripheral drivers to avoid
having their device queue frozen by the XPT for what shoudl only be
initiator type behavior.

Don't perform a bus reset if the target device is incapable of performing
transfer negotiation (e.g. Fiber Channel).

If we don't perform a bus reset but the controller is capable of transfer
negotiations, force negotiations on the first transaction to go to the
device.  This ensures that we aren't tripped up by a left over negotiation
from the prom, BIOS, loader, etc.

Add a default async handler funstion to cam_periph.c to remove duplicated
code in all initiator type peripheral drivers.

Allow mapping of XPT_CONT_TARGET_IO ccbs from userland.  They are
itentical to XPT_SCSI_IO ccbs as far as data mapping is concerned.

25 years agoFirst cut at a driver for the amd53c974 PCI SCSI host adapter. This
Justin T. Gibbs [Sat, 22 May 1999 21:50:40 +0000 (21:50 +0000)]
First cut at a driver for the amd53c974 PCI SCSI host adapter.  This
driver lacks error recovery and still needs more testing, but it's
about time I got it under revision control.

Submitted by:  Tekram Inc.
Bus Space/DMA and cleanup: gibbs

25 years agoDon't call calcru() on a swapped-out process. calcru() access p_stats, which
Dmitrij Tejblum [Sat, 22 May 1999 20:10:31 +0000 (20:10 +0000)]
Don't call calcru() on a swapped-out process. calcru() access p_stats, which
is in U-area.

25 years agoUse BINOWN and BINGRP (some people do change these...)
Tim Vanderhoek [Sat, 22 May 1999 19:38:04 +0000 (19:38 +0000)]
Use BINOWN and BINGRP (some people do change these...)

25 years agoPort pnpinfo to alpha.
Doug Rabson [Sat, 22 May 1999 17:35:48 +0000 (17:35 +0000)]
Port pnpinfo to alpha.

25 years agoCope with non page-aligned arguments to ioperm.
Doug Rabson [Sat, 22 May 1999 17:35:04 +0000 (17:35 +0000)]
Cope with non page-aligned arguments to ioperm.

25 years agoI seem to be having some CVS problems... I don't know why this wasn't
David E. O'Brien [Sat, 22 May 1999 16:21:47 +0000 (16:21 +0000)]
I seem to be having some CVS problems... I don't know why this wasn't
commited with the Makefile fixes.  CVS/Entries clearly show it being added...

25 years ago- Fix up some comments in if_wi.c (no code changes)
Bill Paul [Sat, 22 May 1999 16:12:54 +0000 (16:12 +0000)]
- Fix up some comments in if_wi.c (no code changes)
- Mention that the 6Mbps turbo adapters are supported in HARDWARE.TXT
  and RELNOTES.TXT and the wi.4 man page
- Mention turbo adapters in the wicontrol.8 man page and provide a
  complete table of available transmit speed settings

25 years agoUse the resource apis to manipulate resources.
Doug Rabson [Sat, 22 May 1999 15:47:34 +0000 (15:47 +0000)]
Use the resource apis to manipulate resources.

25 years agoDon't use BUS_WRITE_IVAR to manipulate resources.
Doug Rabson [Sat, 22 May 1999 15:45:47 +0000 (15:45 +0000)]
Don't use BUS_WRITE_IVAR to manipulate resources.

25 years ago* Factor out the common code between the isa bus drivers for i386 and alpha.
Doug Rabson [Sat, 22 May 1999 15:18:28 +0000 (15:18 +0000)]
* Factor out the common code between the isa bus drivers for i386 and alpha.
* Re-work the resource allocation code to use helper functions in subr_bus.c.
* Add simple isa interface for manipulating the resource ranges which can be
  allocated and remove the code from isa_write_ivar() which was previously
  used for this purpose.

25 years agoAdd some helper functions to make it easier to write a driver for a bus
Doug Rabson [Sat, 22 May 1999 14:57:15 +0000 (14:57 +0000)]
Add some helper functions to make it easier to write a driver for a bus
which needs to manage resources for its children.

25 years agoFix a reference to the mysterious Mike Smisth.
Robert Nordier [Sat, 22 May 1999 13:18:28 +0000 (13:18 +0000)]
Fix a reference to the mysterious Mike Smisth.

25 years agoInclude a table of disk error codes and a note about the cylinder >
Robert Nordier [Sat, 22 May 1999 12:55:16 +0000 (12:55 +0000)]
Include a table of disk error codes and a note about the cylinder >
1023 issue.

25 years agoAdd seatbelt like in previous function..
Peter Wemm [Sat, 22 May 1999 09:52:21 +0000 (09:52 +0000)]
Add seatbelt like in previous function..

25 years agoRecover from removing the last (unshared) interrupt handler.
Peter Wemm [Sat, 22 May 1999 09:39:31 +0000 (09:39 +0000)]
Recover from removing the last (unshared) interrupt handler.

PR: 11806
Submitted by: Assar Westerlund <assar@sics.se>

25 years agoFixed style bugs in previous commit.
Bruce Evans [Sat, 22 May 1999 08:36:55 +0000 (08:36 +0000)]
Fixed style bugs in previous commit.

25 years agoFix a coredump when debugging corefiles from static programs.
Doug Rabson [Sat, 22 May 1999 08:25:02 +0000 (08:25 +0000)]
Fix a coredump when debugging corefiles from static programs.

25 years agoFixed disordering in previous commit.
Bruce Evans [Sat, 22 May 1999 07:54:38 +0000 (07:54 +0000)]
Fixed disordering in previous commit.

25 years agoIf we're going to do such a non-UNIX(tm)y thing as appending output
Tim Vanderhoek [Sat, 22 May 1999 06:57:22 +0000 (06:57 +0000)]
If we're going to do such a non-UNIX(tm)y thing as appending output
to a file instead of truncating, at least word the notice of output
redirection appropriately.

25 years agoMove arpcom structure be the first in softc structure. Needed
Semen Ustimenko [Sat, 22 May 1999 06:10:14 +0000 (06:10 +0000)]
Move arpcom structure be the first in softc structure. Needed
for ether_ioctl.

PR:   pending/11754

25 years agoCosmetic changes to make it compile without errors in gcc -Wall
Julian Elischer [Sat, 22 May 1999 04:43:04 +0000 (04:43 +0000)]
Cosmetic changes to make it compile without errors in gcc -Wall

25 years agoRestored rev.1.76 which was clobbered by rev.1.77.
Bruce Evans [Sat, 22 May 1999 04:34:59 +0000 (04:34 +0000)]
Restored rev.1.76 which was clobbered by rev.1.77.

25 years agoRemove `ix' driver. Justin Gibbs added support for the EtherExpress16 to
David E. O'Brien [Sat, 22 May 1999 02:33:08 +0000 (02:33 +0000)]
Remove `ix' driver.  Justin Gibbs added support for the  EtherExpress16 to
the `ie' driver in rev 1.41 of if_ie.c on 1997/04/14.  Thus retiring the
`ix' driver.

Forgotten by: gibbs

25 years agoAdd euro-currency and constant rates of EMU area.
Jean-Marc Zucconi [Sat, 22 May 1999 02:01:59 +0000 (02:01 +0000)]
Add euro-currency and constant rates of EMU area.

PR: bin/9263
Submitted by: Timo J. Rinne <tri@freebsd.org>

25 years agoRemove some #ifdef'ed code. The ADMtek doesn't have a 'enable reception of
Bill Paul [Fri, 21 May 1999 23:58:12 +0000 (23:58 +0000)]
Remove some #ifdef'ed code. The ADMtek doesn't have a 'enable reception of
broadcast frames' bit, so we don't need any code in al_init() to set/clear
it.

25 years agoFix a memory leak and a double free that could happen in certain error
Kenneth D. Merry [Fri, 21 May 1999 22:02:02 +0000 (22:02 +0000)]
Fix a memory leak and a double free that could happen in certain error
cases.  These were unlikely to happen in normal operation.

Noticed by: "Christopher R. Bowman" <crb@ChrisBowman.com>

25 years agoFix hostname lookup problems on 64-bit machines.
John Polstra [Fri, 21 May 1999 20:52:42 +0000 (20:52 +0000)]
Fix hostname lookup problems on 64-bit machines.

25 years agoAdd an a.out ld.so so these distributions are standalone.
David E. O'Brien [Fri, 21 May 1999 18:40:54 +0000 (18:40 +0000)]
Add an a.out ld.so so these distributions are standalone.

25 years agoNeed a few more dependacies.
David E. O'Brien [Fri, 21 May 1999 18:37:33 +0000 (18:37 +0000)]
Need a few more dependacies.

25 years agoAdd usr/libexec/ld.so
David E. O'Brien [Fri, 21 May 1999 18:35:57 +0000 (18:35 +0000)]
Add usr/libexec/ld.so

25 years agoRemove duplicated `ex0' entry.
David E. O'Brien [Fri, 21 May 1999 18:28:32 +0000 (18:28 +0000)]
Remove duplicated `ex0' entry.

25 years agoChflags was clearing all flags supplied on the command line after a
Jean-Marc Zucconi [Fri, 21 May 1999 17:04:13 +0000 (17:04 +0000)]
Chflags was clearing all flags supplied on the command line after a
clearing flag like dump or noschg, etc.

PR: bin/10071
Submitted by: Andreas Klussmann <andreas@infosys.heitec.net>

25 years agoFix a sign-extension bug in xntpdc's "kerninfo" output for on 64-bit
John Polstra [Fri, 21 May 1999 16:10:44 +0000 (16:10 +0000)]
Fix a sign-extension bug in xntpdc's "kerninfo" output for on 64-bit
machines.  Negative pll offset values were printed as large positive
numbers.

25 years agoRealy fix overflow on SO_*TIMEO
Andrey A. Chernov [Fri, 21 May 1999 15:54:40 +0000 (15:54 +0000)]
Realy fix overflow on SO_*TIMEO

Submitted by: bde

25 years agoFix a couple of small grammatical errors.
Bill Paul [Fri, 21 May 1999 14:59:48 +0000 (14:59 +0000)]
Fix a couple of small grammatical errors.

25 years agoImprove handling of "do ... while" following "else". This change
Robert Nordier [Fri, 21 May 1999 14:46:52 +0000 (14:46 +0000)]
Improve handling of "do ... while" following "else".  This change
eliminates some incorrect "Unmatched 'else'" errors.

25 years agoDon't do anything on alpha machines.
Jordan K. Hubbard [Fri, 21 May 1999 11:00:56 +0000 (11:00 +0000)]
Don't do anything on alpha machines.

25 years agoSilently return NULL from devclass_get_device if dc == NULL. The caller
Doug Rabson [Fri, 21 May 1999 08:23:58 +0000 (08:23 +0000)]
Silently return NULL from devclass_get_device if dc == NULL. The caller
should be handling NULL returns already.

Submitted by: Andrew Gallatin <gallatin@cs.duke.edu>

25 years agoftp3.fr.freebsd.org (UVSQ, Versailles, France) has been available for a long
Ollivier Robert [Fri, 21 May 1999 07:10:35 +0000 (07:10 +0000)]
ftp3.fr.freebsd.org (UVSQ, Versailles, France) has been available for a long
time.

Submitted by: Rémy Card <Remy.Card@csi.uvsq.fr>

25 years agoFix typo (#unset where #undef was meant).
John Polstra [Fri, 21 May 1999 04:56:32 +0000 (04:56 +0000)]
Fix typo (#unset where #undef was meant).

25 years agoAck! Cut & paste-o: xl -> al
Bill Paul [Fri, 21 May 1999 04:42:36 +0000 (04:42 +0000)]
Ack! Cut & paste-o: xl -> al

25 years agoThis commit adds driver support for PCI fast ethernet cards based on the
Bill Paul [Fri, 21 May 1999 04:37:48 +0000 (04:37 +0000)]
This commit adds driver support for PCI fast ethernet cards based on the
ADMtek AL981 "Comet" chipset. The AL981 is yet another DEC tulip clone,
except with simpler receive filter options. The AL981 has a built-in
transceiver, power management support, wake on LAN and flow control.
This chip performs extremely well; it's on par with the ASIX chipset
in terms of speed, which is pretty good (it can do 11.5MB/sec with TCP
easily).

I would have committed this driver sooner, except I ran into one problem
with the AL981 that required a workaround. When the chip is transmitting
at full speed, it will sometimes wedge if you queue a series of packets
that wrap from the end of the transmit descriptor list back to the
beginning. I can't explain why this happens, and none of the other tulip
clones behave this way. The workaround this is to just watch for the end
of the transmit ring and make sure that al_start() breaks out of its
packet queuing loop and waiting until the current batch of transmissions
completes before wrapping back to the start of the ring. Fortunately, this
does not significantly impact transmit performance.

This is one of those things that takes weeks of analysis just to come
up with two or three lines of code changes.

25 years agoFIx the output of long command names.
Jean-Marc Zucconi [Fri, 21 May 1999 01:09:45 +0000 (01:09 +0000)]
FIx the output of long command names.

PR: bin/10027
Submitted by: Arne Henrik Juul <arnej@math.ntnu.no>

25 years agoPrint SPANS addresses in the correct byte order.
Mike Spengler [Thu, 20 May 1999 23:52:15 +0000 (23:52 +0000)]
Print SPANS addresses in the correct byte order.

PR: 11691
Submitted by: Jim Pirzyk <Jim.Pirzyk@disney.com>

25 years agoSet CHAR_{MIN,MAX} according to -funsigned-char flag given or not
Andrey A. Chernov [Thu, 20 May 1999 23:33:35 +0000 (23:33 +0000)]
Set CHAR_{MIN,MAX} according to -funsigned-char flag given or not

PR: 11627
Submitted by: Petr Lampa <lampa@fee.vutbr.cz>

25 years agoIgnore the MAIL environment variable when the -u flag is set.
Jean-Marc Zucconi [Thu, 20 May 1999 22:23:04 +0000 (22:23 +0000)]
Ignore the MAIL environment variable when the -u flag is set.

PR: bin/8665

25 years agoMake incoming packets work as keepalives, too. This should fix problems
Eivind Eklund [Thu, 20 May 1999 20:20:24 +0000 (20:20 +0000)]
Make incoming packets work as keepalives, too.  This should fix problems
for some games.

Notified of problem by: tim@turbinegames.com

25 years agousbdi.h:
Nick Hibma [Thu, 20 May 1999 20:02:37 +0000 (20:02 +0000)]
usbdi.h:
        Implement priorities.
GENERIC, LINT, files:
        Remove remarks about ordering of device names.
GENERIC, LINT:
        Sort the devices alphabetically in LINT and GENERIC.

25 years agoAdd comment about split in driver
Nick Hibma [Thu, 20 May 1999 19:52:04 +0000 (19:52 +0000)]
Add comment about split in driver

25 years agoAdd support for multiple PCI "hoses" used on various alpha platforms.
Andrew Gallatin [Thu, 20 May 1999 15:33:33 +0000 (15:33 +0000)]
Add support for multiple PCI "hoses" used on various alpha platforms.
The specific intent of this commit is to pave the way for importing
Compaq XP1000 support.  These changes should not affect the i386 port.

Reviewed by: Doug Rabson <dfr@nlsystems.com>
(actually, he walked me through most of it & deserves more than reviewd-by
credit )

25 years agoupgrade isdn4bsd from version 0.71 to the just released version 0.81
Hellmuth Michaelis [Thu, 20 May 1999 10:14:57 +0000 (10:14 +0000)]
upgrade isdn4bsd from version 0.71 to the just released version 0.81

25 years agoRemove bio/cam/net/tty labels.
Kazutaka YOKOTA [Thu, 20 May 1999 09:56:18 +0000 (09:56 +0000)]
Remove bio/cam/net/tty labels.

OKed by: peter

25 years ago- Include isa/isareg.h rather than i386/isa/isa.h for i386.
Kazutaka YOKOTA [Thu, 20 May 1999 09:49:33 +0000 (09:49 +0000)]
- Include isa/isareg.h rather than i386/isa/isa.h for i386.
- Remove unused (thus, commented out) section of code.

25 years agoInclude sys/isa/isareg.h rather than i386/isa/isa.h for i386.
Kazutaka YOKOTA [Thu, 20 May 1999 09:48:13 +0000 (09:48 +0000)]
Include sys/isa/isareg.h rather than i386/isa/isa.h for i386.

25 years agoEigth update to the new ATA/ATAPI driver:
Søren Schmidt [Thu, 20 May 1999 09:12:06 +0000 (09:12 +0000)]
Eigth update to the new ATA/ATAPI driver:

Fixed problems:

    LS120/ZIP drives still currupted data.
Reworked once again, buffered I/O is just ignoring any sizehints
it is given :(
Now the atapifd driver splits up requests for devices that has
limitted transfer size.

    ISA only configs fails on boot with interrupt timeouts.
The new-bus integration introduced a bug where the softc ptr
was lost during the probe.

Some minor cleanups and rearrangements as well.

As usual USE AT YOUR OWN RISK!!, this is still pre alpha level code.
Especially the DMA support can hose your disk real bad if anything
goes wrong, again you have been warned :)
Notebook owners should be carefull that their machines dont suspend
as this might cause trouble...

But please tell me how it works for you!

Enjoy!

25 years agoConsistantly protect "#define FOO ..." with "#undef FOO".
David E. O'Brien [Thu, 20 May 1999 08:30:48 +0000 (08:30 +0000)]
Consistantly protect "#define FOO ..."  with "#undef FOO".

25 years agoSync with sys/boot/i386/btx/Makefile revision 1.5.
KATO Takenori [Thu, 20 May 1999 08:00:08 +0000 (08:00 +0000)]
Sync with sys/boot/i386/btx/Makefile revision 1.5.

25 years agoOops, set module->file..
Peter Wemm [Thu, 20 May 1999 00:00:58 +0000 (00:00 +0000)]
Oops, set module->file..

PR: 1179
Submitted-by: lha@stacken.kth.se
25 years ago* Remove "'s in the synopsis as config(8) does not require them any more.
David E. O'Brien [Wed, 19 May 1999 22:08:01 +0000 (22:08 +0000)]
* Remove "'s in the synopsis as config(8) does not require them any more.
* Optional bits now shown as such.

Submitted by: Philippe Charnier & bde

25 years agoUpdated Australian channel frequencies
Roger Hardiman [Wed, 19 May 1999 22:04:21 +0000 (22:04 +0000)]
Updated Australian channel frequencies

Submitted by: Ivan Brawley <brawley@internode.com.au>