]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
24 years agoHopefully wave the unbreak-release magic wand at dokern.sh. Use GENERIC's
peter [Wed, 14 Jun 2000 17:53:38 +0000 (17:53 +0000)]
Hopefully wave the unbreak-release magic wand at dokern.sh.  Use GENERIC's
hints strings with BOOTMFS.

24 years agoFix typo (accessable --> accessible) and add period.
alex [Wed, 14 Jun 2000 17:53:11 +0000 (17:53 +0000)]
Fix typo (accessable --> accessible) and add period.

PR: 18588
Submitted by: Anatoly Vorobey <mellon@pobox.com>
Reviewed by: asmodai

24 years ago/etc/weekly --> /etc/periodic/weekly/310.locate
alex [Wed, 14 Jun 2000 17:41:31 +0000 (17:41 +0000)]
/etc/weekly --> /etc/periodic/weekly/310.locate

PR: 19268
Submitted by: Uwe Pierau <uwe.pierau@tu-clausthal.de>
Reviewed by: asmodai

24 years agopthread_mutex_lock(), pthread_cond_trywait(), and pthread_cond_wait() are
jasone [Wed, 14 Jun 2000 17:17:41 +0000 (17:17 +0000)]
pthread_mutex_lock(), pthread_cond_trywait(), and pthread_cond_wait() are
not allowed to return EINTR, but use of pthread_suspend_np() could cause
EINTR to be returned.  To fix this, restructure pthread_suspend_np() so that
it does not interrupt a thread that is waiting on a mutex or condition, and
keep enough state around that pthread_resume_np() can fix things up
afterwards.

Reviewed by: deischen

24 years agoRemoved support for generating inline code for MALLOC() and FREE()
bde [Wed, 14 Jun 2000 17:11:47 +0000 (17:11 +0000)]
Removed support for generating inline code for MALLOC() and FREE()
in the dysfunctional !KMEMSTATS case.  This hasn't compiled since
rev.1.31 of kern_malloc.c quietly removed the core of the support
for the !KMEMSTATS case.  I fixed it to see if it was worth saving
and found that (as usual) inlining just wasted space and increased
complexity without significantly affecting time, at least for the
lmbench2 micro-benchmark on a Celeron.  The space bloat was
surprisingly large - the text size increased from 1700K to 1840K
for a version with the entire malloc() family inlined.

Removed even older garbage (kmemxtob() and btokmemx() macros).

Attempt to deprecate MALLOC() and FREE().  Given current compilers
(gcc-2.x or C99), they don't do anything that (safe) function-like
macros or inline functions named malloc() and free() couldn't do.

Fixed missing casts of macro args in MALLOC() and FREE().

24 years ago- Added support for passive mode FTP by aliasing 227 replies.
ru [Wed, 14 Jun 2000 16:09:35 +0000 (16:09 +0000)]
- Added support for passive mode FTP by aliasing 227 replies.
  It does mean that it is now possible to run passive-mode FTP
  server behind NAT.

- SECURITY: FTP aliasing engine now ensures that:
  o the segment preceding a PORT/227 segment terminates with a \r\n;
  o the IP address in the PORT/227 matches the source IP address of
    the packet;
  o the port number in the PORT command or 277 reply is greater than
    or equal to 1024.

Submitted by: Erik Salander <erik@whistle.com>
Reviewed by: ru

24 years agoWarn about config's bout with Peter's big axe
imp [Wed, 14 Jun 2000 15:42:50 +0000 (15:42 +0000)]
Warn about config's bout with Peter's big axe

24 years agoMake sure to use native IPv4 addrerss even if getaddrinfo()
ume [Wed, 14 Jun 2000 15:26:58 +0000 (15:26 +0000)]
Make sure to use native IPv4 addrerss even if getaddrinfo()
returns IPv4 mapped IPv6 address.  FTP is nervous about address
family.
Submitted by itojun and slightly modified to fit our ftp(1).

24 years agoFix "op" for xterm-color
ache [Wed, 14 Jun 2000 15:08:24 +0000 (15:08 +0000)]
Fix "op" for xterm-color

24 years agoadd libxpg4
ache [Wed, 14 Jun 2000 14:38:11 +0000 (14:38 +0000)]
add libxpg4

24 years agoYou need options USER_LDT in your kernel to use these functions.
alex [Wed, 14 Jun 2000 13:38:21 +0000 (13:38 +0000)]
You need options USER_LDT in your kernel to use these functions.

PR: 18943
Submitted by: Ben Smithurst <ben@scientia.demon.co.uk>
Reviewed by: asmodai

24 years agoDelay calling the device cleanup routines until the absolute last
ps [Wed, 14 Jun 2000 10:34:29 +0000 (10:34 +0000)]
Delay calling the device cleanup routines until the absolute last
moment.  We were cleaning up after PXE too early and the module
dependancy code would not be able to load any files if it needed
too.

24 years agoFix behaviour of "ipfw pipe show" -- previous code gave
luigi [Wed, 14 Jun 2000 10:07:22 +0000 (10:07 +0000)]
Fix behaviour of "ipfw pipe show" -- previous code gave
ambiguous data to the userland program (kernel operation was
safe, anyways).

24 years agos/iomem/maddr/
peter [Wed, 14 Jun 2000 10:04:06 +0000 (10:04 +0000)]
s/iomem/maddr/
s/iosiz/msize/

24 years agos/iomem/maddr/ - these were generated from an older verion of the
peter [Wed, 14 Jun 2000 10:01:39 +0000 (10:01 +0000)]
s/iomem/maddr/ - these were generated from an older verion of the
gethints script. :-(

24 years agoCatch up with Peter's config(8) changes.
kato [Wed, 14 Jun 2000 09:20:43 +0000 (09:20 +0000)]
Catch up with Peter's config(8) changes.

24 years agoMake typing 'q' or 'Q' work for the show command as the pager prompt
ps [Wed, 14 Jun 2000 09:05:03 +0000 (09:05 +0000)]
Make typing 'q' or 'Q' work for the show command as the pager prompt
says it should.

24 years agoArgh! I broke the static hints parser at the last minute on freefall when
peter [Wed, 14 Jun 2000 07:43:56 +0000 (07:43 +0000)]
Argh!  I broke the static hints parser at the last minute on freefall when
I added the $FreeBSD$ (commented) line.
Fix:
 1: s/break/continue/
 2: will somebody please shoot me! :-]

24 years agoMake find -Wall -Wredundant-decls clean.
roberto [Wed, 14 Jun 2000 07:43:52 +0000 (07:43 +0000)]
Make find -Wall -Wredundant-decls clean.

Submitted by: nrahlstr

24 years agoWith apologies to dcs, temporarily comment out the version check code. It
peter [Wed, 14 Jun 2000 07:18:18 +0000 (07:18 +0000)]
With apologies to dcs, temporarily comment out the version check code.  It
is failing for everybody that I have spoken with that has tried it.

FreeBSD/i386 bootstrap loader, Revision 0.8
(root@outback.netplex.com.au, Tue Jun 13 23:26:49 PDT 2000)
Loader version 0.3+ required
Aborted!
start not found

Note that the 0.3+ message is from inside the arch-alpha block, not the
i386 block of code.  And even then, 0.8 is higher than 0.3.

This prevents the rest of the loader.conf stuff working. :-/

24 years agoAdd option ALT_BREAK_TO_DEBUGGER.
ps [Wed, 14 Jun 2000 06:41:33 +0000 (06:41 +0000)]
Add option ALT_BREAK_TO_DEBUGGER.

Implement the Solaris way to break into DDB over a serial console
instead of sending a break.  Sending the character sequence
CR ~ ^b will break the kernel into DDB (if DDB is enabled).

Reviewed by: peter

24 years agoDo not perform any opeartion with mbuf after it placed into
bp [Wed, 14 Jun 2000 05:56:53 +0000 (05:56 +0000)]
Do not perform any opeartion with mbuf after it placed into
interface queue.

Tested by: Bosko Milekic <bmilekic@dsuper.net>

24 years agoMention a bug in the Mylex driver's geometry handling.
msmith [Wed, 14 Jun 2000 04:13:01 +0000 (04:13 +0000)]
Mention a bug in the Mylex driver's geometry handling.

24 years agoPrint error messages to stderr, not stdout.
peter [Wed, 14 Jun 2000 02:45:11 +0000 (02:45 +0000)]
Print error messages to stderr, not stdout.

24 years agoadd alpha-quality recording code and handle pci error interrupts - this may
cg [Tue, 13 Jun 2000 23:24:40 +0000 (23:24 +0000)]
add alpha-quality recording code and handle pci error interrupts - this may
prevent the card generating an nmi on ecc systems.  for now a message is
printed on every pci error and it seems every time we start playng we get one

24 years agohandle closing differently - should fix the end-of-sample cutoff bug
cg [Tue, 13 Jun 2000 23:18:43 +0000 (23:18 +0000)]
handle closing differently - should fix the end-of-sample cutoff bug

24 years agoBorrow phk's axe and apply the next stage of config(8)'s evolution.
peter [Tue, 13 Jun 2000 22:28:50 +0000 (22:28 +0000)]
Borrow phk's axe and apply the next stage of config(8)'s evolution.

Use Warner Losh's "hint" driver to decode ascii strings to fill the
resource table at boot time.

config(8) no longer generates an ioconf.c table - ie: the configuration
no longer has to be compiled into the kernel.  You can reconfigure your
isa devices with the likes of this at loader(8) time:
  set hint.ed.0.port=0x320

userconfig will be rewritten to use this style interface one day and will
move to /boot/userconfig.4th or something like that.

It is still possible to statically compile in a set of hints into a kernel
if you do not wish to use loader(8).  See the "hints" directive in GENERIC
as an example.

All device wiring has been moved out of config(8).  There is a set of
helper scripts (see i386/conf/gethints.pl, and the same for alpha and pc98)
that extract the 'at isa? port foo irq bar' from the old files and produces
a hints file.  If you install this file as /boot/device.hints (and update
/boot/defaults/loader.conf - You can do a build/install in sys/boot) then
loader will load it automatically for you.  You can also compile in the
hints directly with:  hints "device.hints"  as well.

There are a few things that I'm not too happy with yet.  Under this scheme,
things like LINT would no longer be useful as "documentation" of settings.
I have renamed this file to 'NOTES' and stored the example hints strings
in it.  However... this is not something that config(8) understands, so
there is a script that extracts the build-specific data from the
documentation file (NOTES) to produce a LINT that can be config'ed and
built.  A stack of man4 pages will need updating. :-/

Also, since there is no longer a difference between 'device' and
'pseudo-device' I collapsed the two together, and the resulting 'device'
takes a 'number of units' for devices that still have it statically
allocated.  eg:  'device fe 4' will compile the fe driver with NFE set
to 4.  You can then set hints for 4 units (0 - 3).  Also note that
'device fe0' will be interpreted as "zero units of 'fe'" which would be
bad, so there is a config warning for this.  This is only needed for
old drivers that still have static limits on numbers of units.
All the statically limited drivers that I could find were marked.

Please exercise EXTREME CAUTION when transitioning!

Moral support by: phk, msmith, dfr, asmodai, imp, and others

24 years agoFix a problem of user settings from TEKRAM NVRAM
groudier [Tue, 13 Jun 2000 20:17:41 +0000 (20:17 +0000)]
Fix a problem of user settings from TEKRAM NVRAM
layout introduced in driver 1.5.3. The driver was
confused by the bogus TEKRAM table used to translate
user sync. setting to SCSI sync. factor.
Btw, the new TEKRAM DC-390 U3D and U3W Ultra-160
controllers seem to be using BIOS from SYMBIOS/LSI
and thus SYMBIOS NVRAM layout.
If that means that TEKRAM will now offer real
SYMBIOS software compatible SCSI controllers, then
it is a *GREAT NEWS*.

24 years agoAdd libxpg4 as the current version had it's SO version number bumped and
obrien [Tue, 13 Jun 2000 18:09:45 +0000 (18:09 +0000)]
Add libxpg4 as the current version had it's SO version number bumped and
it no longer contains the locale functions.

24 years agoFix panic by moving the prp == 0 check up the order of sanity checks.
asmodai [Tue, 13 Jun 2000 15:44:04 +0000 (15:44 +0000)]
Fix panic by moving the prp == 0 check up the order of sanity checks.

Submitted by: Bart Thate <freebsd@1st.dudi.org> on -current
Approved by: rwatson

24 years agoMerged from sys/conf/options.i386 rev 1.137.
kato [Tue, 13 Jun 2000 13:08:30 +0000 (13:08 +0000)]
Merged from sys/conf/options.i386 rev 1.137.

24 years agoTreat \t and \n inside /boot.config as whitespaces.
ru [Tue, 13 Jun 2000 13:07:53 +0000 (13:07 +0000)]
Treat \t and \n inside /boot.config as whitespaces.

PR: 19215

24 years agoMerged from sys/i386/i386/machdep.c rev 1.395.
kato [Tue, 13 Jun 2000 13:05:51 +0000 (13:05 +0000)]
Merged from sys/i386/i386/machdep.c rev 1.395.

24 years agoFix typo: turn of -> turn off.
alex [Tue, 13 Jun 2000 12:50:47 +0000 (12:50 +0000)]
Fix typo: turn of -> turn off.

PR: 18805
Submitted by: Yoshihiro Ota <ota@mail.drexel.edu>
Kind of Reviewed by: asmodai ("sure")

24 years agoRecognize Coppermine Celeron processors whose CPU ID = 0x68?. They
kato [Tue, 13 Jun 2000 12:33:45 +0000 (12:33 +0000)]
Recognize Coppermine Celeron processors whose CPU ID = 0x68?.  They
were recognized as "Pentium III/Pentium III Xeon."

24 years agomdoc improvements and better readability of the text.
alex [Tue, 13 Jun 2000 12:25:20 +0000 (12:25 +0000)]
mdoc improvements and better readability of the text.

PR: 18811
Submitted by: Christian Weisgerber <naddy@unix-ag.uni-kl.de>
Reviewed by: asmodai

24 years agoA few more hard-sentence breaks.
brian [Tue, 13 Jun 2000 10:03:28 +0000 (10:03 +0000)]
A few more hard-sentence breaks.

24 years agoCorrectly set the Maximum DHCP Message Size. bootpd now works
ps [Tue, 13 Jun 2000 09:32:09 +0000 (09:32 +0000)]
Correctly set the Maximum DHCP Message Size.  bootpd now works
again as well as ISC dhcpd.

24 years agoAdded new options CPU_PPRO2CELERON and CPU_L2_LATENCY to support
kato [Tue, 13 Jun 2000 09:10:37 +0000 (09:10 +0000)]
Added new options CPU_PPRO2CELERON and CPU_L2_LATENCY to support
Socket 8 to 370 converters.  When (1) CPU_PPRO2CELERON option is
defined, (2) Intel CPU is found and (3) CPU ID is 0x66?, L2 cache is
enabled through MSR 0x11e.  The L2 cache latency value can be
specified by CPU_L2_LATENCY option.  Default value of L2 cache latency
is 5.

These options are useful if you use Socket 8 to Socket 370 converter
(e.g. Power Leap's PL-Pro/II.)  Most PentiumPro BIOSs don't enable L2
cache of Mendocino Celeron CPUs because they don't know Celeron CPUs.
These options are needles if you use a Coppermine (FCPGA) Celeron or
PentiumIII, becuase the L2 cache enable bit is hard wired and L2 cache
is always enabled.

24 years agoRemove the "any" and "pio" stuff from PAO. We don't (yet) support
imp [Tue, 13 Jun 2000 06:08:58 +0000 (06:08 +0000)]
Remove the "any" and "pio" stuff from PAO.  We don't (yet) support
that notation.

Reported by: jkh on the train from Tokyo to Nagoya.

24 years agoAdd 'tools' to the list of tarballs to be extracted when 'all' is
jwd [Tue, 13 Jun 2000 03:05:56 +0000 (03:05 +0000)]
Add 'tools' to the list of tarballs to be extracted when 'all' is
specified as the distribution.

PR: 18927
Submitted by: Stephen J. Roznowski <sjr@home.net>

24 years agoBump major to preserve 3.x binaries compatibility
ache [Tue, 13 Jun 2000 00:41:38 +0000 (00:41 +0000)]
Bump major to preserve 3.x binaries compatibility

Found by: "Akinori -Aki- MUSHA" <knu@idaemons.org>

24 years agoFix breakage to target mode support.
mjacob [Mon, 12 Jun 2000 23:08:31 +0000 (23:08 +0000)]
Fix breakage to target mode support.

What we'd like to know is whether or not we have a listener
upstream that really hasn't configured yet. If we do, then
we can give a more sensible reply here. If not, then we can
reject this out of hand.

Choices for what to send were
Not Ready, Unit Not Self-Configured Yet
(0x2,0x3e,0x00)
for the former and
Illegal Request, Logical Unit Not Supported
(0x5,0x25,0x00)
for the latter.

We used to decide whether there was at least one listener
based upon whether the black hole driver was configured.

However, recent config(8) changes have made this hard to do
at this time.

Actually, we didn't use the above quite yet, but were sure considering it.

24 years agoThe buslogic driver doesn't have static unit limits (ie: no bt.h or
peter [Mon, 12 Jun 2000 20:53:39 +0000 (20:53 +0000)]
The buslogic driver doesn't have static unit limits (ie: no bt.h or
arbitary NBT limits)

24 years agoUse objcopy to strip the gcc2_compiled. symbol from kernels.
peter [Mon, 12 Jun 2000 20:22:07 +0000 (20:22 +0000)]
Use objcopy to strip the gcc2_compiled. symbol from kernels.
ld(1) doesn't seem to have the ability to delete an arbitary symbol.

24 years agoMention about -4 and -6 options.
ume [Mon, 12 Jun 2000 18:30:16 +0000 (18:30 +0000)]
Mention about -4 and -6 options.

24 years agounstatic getfp() so that other subsystems can use it.
alfred [Mon, 12 Jun 2000 18:06:12 +0000 (18:06 +0000)]
unstatic getfp() so that other subsystems can use it.

make sendfile() use it.

Approved by: dg

24 years agoSometimes there isn't an ISA bus configured.
mjacob [Mon, 12 Jun 2000 17:07:57 +0000 (17:07 +0000)]
Sometimes there isn't an ISA bus configured.

24 years agoFix REFILL. It must throw RESTART instead of OUTOFTEXT so that
dcs [Mon, 12 Jun 2000 16:46:28 +0000 (16:46 +0000)]
Fix REFILL. It must throw RESTART instead of OUTOFTEXT so that
execution can take place at the point where it stopped after the
input buffer has been refilled.

Add ANS Forth CORE EXT and FILE word SOURCE-ID.

24 years agoThe word environment? returns a flag indicating whether the variable
dcs [Mon, 12 Jun 2000 16:45:01 +0000 (16:45 +0000)]
The word environment? returns a flag indicating whether the variable
was found or not. Fix it's usage. Alas, it caused no problem before,
besides leaving garbage in the stack, because refill, used by [if]
[else] [then], was broken.

24 years agoMake abort" functional in interpret mode. This behavior is undefined
dcs [Mon, 12 Jun 2000 16:42:02 +0000 (16:42 +0000)]
Make abort" functional in interpret mode. This behavior is undefined
by ANS Forth standard, but it's useful.

Also, define the constant true in a more strict way. C might garantee
2-complement math, but Forth doesn't.

24 years agoMake comment reflect reality.
dcs [Mon, 12 Jun 2000 16:40:00 +0000 (16:40 +0000)]
Make comment reflect reality.

24 years agoAdd MLINKS for queue(3) FOREACH_REVERSE and HEAD_INITIALIZER macros.
jake [Mon, 12 Jun 2000 15:28:08 +0000 (15:28 +0000)]
Add MLINKS for queue(3) FOREACH_REVERSE and HEAD_INITIALIZER macros.

Forgotten by: jake, archie

24 years ago- Eliminate rpread(). Call generic ttyread(). (cf rev 1.33)
tanimura [Mon, 12 Jun 2000 15:21:59 +0000 (15:21 +0000)]
- Eliminate rpread(). Call generic ttyread(). (cf rev 1.33)
- Comment out deftermios. Termioschars() will give the default
  value.

Pointed out by: bde

24 years agoDocument possible values for the waitfor arg.
alex [Mon, 12 Jun 2000 13:53:04 +0000 (13:53 +0000)]
Document possible values for the waitfor arg.

PR: 18627, 18628
Reviewed by: asmodai

24 years ago This patch adds the -mindepth and -maxdepth options to find(1), which
roberto [Mon, 12 Jun 2000 11:12:41 +0000 (11:12 +0000)]
This patch adds the -mindepth and -maxdepth options to find(1), which
    behave as in GNU find (and of course as described in the manual page
    diff included).  I think these options would be useful for some people.

    Some missing $FreeBSD$ tags are also added.

    The patch was slightly modified (send-pr mangling of TABS).

PR: bin/18941
Submitted by: Ben Smithurst <ben@scientia.demon.co.uk>

24 years agoDocument SVR4 branding is supported.
asmodai [Mon, 12 Jun 2000 11:12:19 +0000 (11:12 +0000)]
Document SVR4 branding is supported.

PR: 19131
Submitted by: Mike Barcroft <mike@q9media.com>

24 years ago The find -perm option currently supports an exact match,
roberto [Mon, 12 Jun 2000 10:36:52 +0000 (10:36 +0000)]
    The find -perm option currently supports an exact match,
        or if the mode is preceded by a '-', it checks for a match
        in at least the bits specified on the command line.  It is
        often desirable to find things with any execute or setuid or
        setgid bits set.

PR: bin/10169
Submitted by: Monte Mitzelfelt <monte@gonefishing.org>

24 years ago before this commit, specfs reported disk partitions
jmb [Mon, 12 Jun 2000 10:20:18 +0000 (10:20 +0000)]
before this commit, specfs reported disk partitions
using decimal major and minor numbers.  "ls -l" reports
disk partitions using decimal major numbers and hex
minor numbers.

make specfs use decimal major numbers and hex minor numbers,
just like "ls -l"

24 years agoFixed style bugs of rev 1.66.
ru [Mon, 12 Jun 2000 09:43:00 +0000 (09:43 +0000)]
Fixed style bugs of rev 1.66.

24 years agoAdd (another) PnP entry for the ESS ES1869 (Compaq OEM)
brian [Mon, 12 Jun 2000 09:09:37 +0000 (09:09 +0000)]
Add (another) PnP entry for the ESS ES1869 (Compaq OEM)

Submitted by: mdharnois@home.com
PR: 19206

24 years ago0x39 == plan9
brian [Mon, 12 Jun 2000 09:08:40 +0000 (09:08 +0000)]
0x39 == plan9

Obtained from: OpenBSD

24 years agoReformat and update sound section a bit. Still need to add more
asmodai [Sun, 11 Jun 2000 21:53:38 +0000 (21:53 +0000)]
Reformat and update sound section a bit.  Still need to add more
soundcards.

24 years agoFix syntax error in previous commit.
kris [Sun, 11 Jun 2000 21:41:25 +0000 (21:41 +0000)]
Fix syntax error in previous commit.

Submitted by: Udo Schweigert <ust@cert.siemens.de>

24 years agoThe very feeble beginnings of a driver for the LanMedia LMC1504 card.
phk [Sun, 11 Jun 2000 19:09:47 +0000 (19:09 +0000)]
The very feeble beginnings of a driver for the LanMedia LMC1504 card.

New-Bus wizards are encouraged to look at this, I think it poses a
challenge for the current newbus design.

24 years agoThe USB library from NetBSD by Lennart Augustsson <lennart@augustsson.net>.
n_hibma [Sun, 11 Jun 2000 18:19:14 +0000 (18:19 +0000)]
The USB library from NetBSD by Lennart Augustsson <lennart@augustsson.net>.

24 years agoDNS IPv6 transport support.
ume [Sun, 11 Jun 2000 15:43:34 +0000 (15:43 +0000)]
DNS IPv6 transport support.
It is nessesary for IPv6 only life.

Obtained from: KAME

24 years agoHard sentence breaks and trailing space tidy-ups
brian [Sun, 11 Jun 2000 14:58:32 +0000 (14:58 +0000)]
Hard sentence breaks and trailing space tidy-ups

Obtained from: OpenBSD

24 years agoSuck out all of the current and desired n/w parameters.
dmlb [Sun, 11 Jun 2000 13:57:59 +0000 (13:57 +0000)]
Suck out all of the current and desired n/w parameters.

24 years agoGenerated a new macor, RAY_RECERR for reporting errors with. Verbosity set with IFF_D...
dmlb [Sun, 11 Jun 2000 13:56:11 +0000 (13:56 +0000)]
Generated a new macor, RAY_RECERR for reporting errors with. Verbosity set with IFF_DEBUG (recommended at present).

Move promisc flag into the nw parameter structure.

24 years agoGenerated a new macor, RAY_RECERR for reporting errors with. Verbosity set with IFF_D...
dmlb [Sun, 11 Jun 2000 13:54:59 +0000 (13:54 +0000)]
Generated a new macor, RAY_RECERR for reporting errors with. Verbosity set with IFF_DEBUG (recommended at present).

24 years agoTry and get this list a bit more consistent in style.
asmodai [Sun, 11 Jun 2000 13:51:29 +0000 (13:51 +0000)]
Try and get this list a bit more consistent in style.
Also, clarify a comment with proper comma placement.

Also tried to get the supported soundcard section more fleshed out.  This
needs more work.

24 years agoA bunch of misc. tidy ups really.
dmlb [Sun, 11 Jun 2000 13:32:07 +0000 (13:32 +0000)]
A bunch of misc. tidy ups really.

Generated a new macor, RAY_RECERR for reporting errors with. Verbosity set with IFF_DEBUG (recommended at present).

Add PRIBIO to tsleeps.

Catch detach on ray_ccs_alloc a little better.

Move sc_promisc into desired and current n/w parameters.

Remove IFQ_PEEK, we know the driver runs okay without it.

Drain the output queue in ray_stop.

Only use ray_mcast for ADD/DEL multi ioctls. ray_init_multi resets the
multicast list on startup. Simplifies ray_init a little.

Tidy some old comments.

ray_download_done now copies the whole desired n/w parameter set into the
current set. This is because I was missing soem parameters - like the
net type!

24 years agoDon't oan about padding characters
brian [Sun, 11 Jun 2000 12:22:06 +0000 (12:22 +0000)]
Don't oan about padding characters

24 years agoAdd Adico and Accton NICs to the list.
asmodai [Sun, 11 Jun 2000 12:04:08 +0000 (12:04 +0000)]
Add Adico and Accton NICs to the list.
Move NDC a bit lower so that it is alphabetical.

Inspired by PR: 18735
Submitted by: Adoal Xu <adoal@iname.com>

24 years agoAdd support for the Accton EN1217.
asmodai [Sun, 11 Jun 2000 11:54:52 +0000 (11:54 +0000)]
Add support for the Accton EN1217.

PR: 18735
Submitted by: Adoal Xu <adoal@iname.com>

24 years agoFixed allocation of unit numbers. Allocate the amount of space actually
bde [Sun, 11 Jun 2000 07:19:20 +0000 (07:19 +0000)]
Fixed allocation of unit numbers.  Allocate the amount of space actually
required (rounded up a little) instead of twice the previous amount (or
a fixed amount for the first allocation).

The bug caused memory corruption when a new unit number for a devclass
was more than about twice the previous maximum one (or more than 3 for
the first one), so it corrupted memory (which happened to be the atkbdc
port resource list) in the reporter's configuration with sio unit
numbers { 0, 25, 1, 2, ... }.

Reviewed by: dfr
Reported by: Leonid Lukiyanets <stalwar78@hotmail.com>

24 years ago1. Update Comtrol RocketPort driver(rp) to version 3.02.
tanimura [Sun, 11 Jun 2000 06:43:16 +0000 (06:43 +0000)]
1. Update Comtrol RocketPort driver(rp) to version 3.02.
2. Newbusify the driver.
3. Build as a module.

4. Use correct minor numbers when creating device files.
5. Correctly lock control characters.
6. Return ENXIO when device not configured.
Submitted by: Tor Egge <Tor.Egge@fast.no>

7. Fix the baud_table.
Submitted by: Elliot Dierksen <ebd@oau.org>

Note:
- the old driver still lives in src/sys/i386/isa, so that you can
  revert to it if something goes wrong.
- The module does not detach very well. Attaching works fine.

24 years agocpu_fork(): Check "flags" before dereferencing "p2". Otherwise,
alc [Sun, 11 Jun 2000 06:22:01 +0000 (06:22 +0000)]
cpu_fork(): Check "flags" before dereferencing "p2".  Otherwise,
the call "vm_fork(p1, 0, flags);" early in fork1 can cause a kernel
panic.

24 years agoLink explicitly against -lmd. I'm not sure what was pulling this in
kris [Sun, 11 Jun 2000 05:30:52 +0000 (05:30 +0000)]
Link explicitly against -lmd. I'm not sure what was pulling this in
on -current, but it doesnt do it on -stable.

24 years ago Fix memory leak in mount_nfs's background (-b) mode which occurs when
dillon [Sun, 11 Jun 2000 05:19:58 +0000 (05:19 +0000)]
Fix memory leak in mount_nfs's background (-b) mode which occurs when
    the mount is not available.

Submitted-by: Jonathan Hanna <pangolin@home.com>
24 years agoActivate Japanese NLS and KANJI support
ache [Sun, 11 Jun 2000 03:26:08 +0000 (03:26 +0000)]
Activate Japanese NLS and KANJI support

24 years agoAdd ``set ifqueue'' to control the size of the outgoing packet
brian [Sun, 11 Jun 2000 02:30:51 +0000 (02:30 +0000)]
Add ``set ifqueue'' to control the size of the outgoing packet
queue.  Doing ``set ifqueue 0'' and ``set urgent none'' will allow
full use of luigi's WF2Q code.

Requested by: luigi

24 years agoDon't include opt_smp.h - we don't use anything defined in it.
msmith [Sat, 10 Jun 2000 22:59:50 +0000 (22:59 +0000)]
Don't include opt_smp.h - we don't use anything defined in it.

24 years agoCorrect the tests for ISA PIC/APIC so that they actually work.
msmith [Sat, 10 Jun 2000 22:56:09 +0000 (22:56 +0000)]
Correct the tests for ISA PIC/APIC so that they actually work.

24 years agoFix security botch in "UseLogin Yes" case: commands are executed with
kris [Sat, 10 Jun 2000 22:32:57 +0000 (22:32 +0000)]
Fix security botch in "UseLogin Yes" case: commands are executed with
uid 0.

Obtained from: OpenBSD

24 years agoThis commit was generated by cvs2svn to compensate for changes in r61524,
obrien [Sat, 10 Jun 2000 22:20:53 +0000 (22:20 +0000)]
This commit was generated by cvs2svn to compensate for changes in r61524,
which included commits to RCS files with non-trunk default branches.

24 years agoOffical fixes to allow longer lines in the history and tweak expand_lex()
obrien [Sat, 10 Jun 2000 22:20:53 +0000 (22:20 +0000)]
Offical fixes to allow longer lines in the history and tweak expand_lex()
useage.

24 years agoA checkpoint of a part of a work-in-progress. Some more cleanups for
peter [Sat, 10 Jun 2000 22:13:40 +0000 (22:13 +0000)]
A checkpoint of a part of a work-in-progress.  Some more cleanups for
config(8).  This commit allows control of the creation of the
#include "foo.h" files.  We now only create them explicitly when needed.
BTW; these are mostly bad because they usually imply static limits on
numbers of units for devices.  eg: struct mysoftc sc[NFOO];
These static limits have Got To Go.

24 years agoOffical fixes to quiet compiler warnings.
obrien [Sat, 10 Jun 2000 22:07:54 +0000 (22:07 +0000)]
Offical fixes to quiet compiler warnings.

24 years agoThis commit was generated by cvs2svn to compensate for changes in r61521,
obrien [Sat, 10 Jun 2000 22:07:54 +0000 (22:07 +0000)]
This commit was generated by cvs2svn to compensate for changes in r61521,
which included commits to RCS files with non-trunk default branches.

24 years agoOffical fixes to the history and I18N bugs ache found.
obrien [Sat, 10 Jun 2000 21:38:51 +0000 (21:38 +0000)]
Offical fixes to the history and I18N bugs ache found.

24 years agoThis commit was generated by cvs2svn to compensate for changes in r61519,
obrien [Sat, 10 Jun 2000 21:38:51 +0000 (21:38 +0000)]
This commit was generated by cvs2svn to compensate for changes in r61519,
which included commits to RCS files with non-trunk default branches.

24 years agoRemoved a few RAY_MAP_CMs that were not needed.
dmlb [Sat, 10 Jun 2000 21:24:36 +0000 (21:24 +0000)]
Removed a few RAY_MAP_CMs that were not needed.

Made checking sc->gone a lot safer by checking to see if sc is NULL first.

Made return from tsleep in the ccs allocator detach safe.

24 years agoIt is not necessary to wrap usage of $f28 with ".set noat/at" pairs.
obrien [Sat, 10 Jun 2000 20:42:27 +0000 (20:42 +0000)]
It is not necessary to wrap usage of $f28 with ".set noat/at" pairs.
A very bogus gas warning gave people the impression it was.  noat/at had
no useful affect here.

24 years agoOffical fixes to the history and I18N bugs ache found.
obrien [Sat, 10 Jun 2000 20:14:21 +0000 (20:14 +0000)]
Offical fixes to the history and I18N bugs ache found.

24 years agoThis commit was generated by cvs2svn to compensate for changes in r61515,
obrien [Sat, 10 Jun 2000 20:14:21 +0000 (20:14 +0000)]
This commit was generated by cvs2svn to compensate for changes in r61515,
which included commits to RCS files with non-trunk default branches.

24 years agoHandle minors like ls(1) does it.
phk [Sat, 10 Jun 2000 19:55:39 +0000 (19:55 +0000)]
Handle minors like ls(1) does it.

PR: 19179
Submitted by: Kelly Yancey <kbyanc@posi.net>

24 years agoAdd an autologin example.
phk [Sat, 10 Jun 2000 19:49:16 +0000 (19:49 +0000)]
Add an autologin example.

PR: 18082
Submitted by: C. Stephen Gunn <csg@dustdevil.waterspout.com>

24 years agoAvoid useless check against NULL, since we assign the checked value
alex [Sat, 10 Jun 2000 19:48:19 +0000 (19:48 +0000)]
Avoid useless check against NULL, since we assign the checked value
to the var anyways.

Reviewed by: bde