]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
24 years agoinstall(1) lookalike that can be used during a build. Any options
marcel [Wed, 8 Dec 1999 13:40:44 +0000 (13:40 +0000)]
install(1) lookalike that can be used during a build. Any options
are ignored and the remaining is passed on to cp(1). This allows
a build to be run as non-root without forcing any user/group
setting and also prevents setting any file flags.

This may not be the right place for buildtools.

24 years agoAdd note about `mountd -r'
phantom [Wed, 8 Dec 1999 13:12:49 +0000 (13:12 +0000)]
Add note about `mountd -r'

PR: docs/15067
Submitted by: Nick Hibma <hibma@skylink.it>

24 years agoAdd support for compiling SVR4 as a static module
newton [Wed, 8 Dec 1999 12:13:13 +0000 (12:13 +0000)]
Add support for compiling SVR4 as a static module

("AND THE CROWD GOES... uh.")

Tested by: Joerg Wunsch <joerg_wunsch@interface-business.de>

24 years agoChange the devstat priorities to the std values now in devicestats.h
sos [Wed, 8 Dec 1999 12:08:47 +0000 (12:08 +0000)]
Change the devstat priorities to the std values now in devicestats.h

24 years agoRemove unnecessary includes
newton [Wed, 8 Dec 1999 12:01:59 +0000 (12:01 +0000)]
Remove unnecessary includes

Prodded by: phk

24 years agoMake this buildable with MROUTING defined.
shin [Wed, 8 Dec 1999 11:57:36 +0000 (11:57 +0000)]
Make this buildable with MROUTING defined.

Specified by: eivind, phk

24 years agoForced commit for correct version ID on files dependent on
newton [Wed, 8 Dec 1999 11:50:48 +0000 (11:50 +0000)]
Forced commit for correct version ID on files dependent on
syscalls.master

Noticed by: marcel

24 years agoMake this actually build with current sources.
dfr [Wed, 8 Dec 1999 11:47:48 +0000 (11:47 +0000)]
Make this actually build with current sources.

24 years agoAccept 12 for %l, because it's logical to expect "%l:%M" to work for
sheldonh [Wed, 8 Dec 1999 11:11:40 +0000 (11:11 +0000)]
Accept 12 for %l, because it's logical to expect "%l:%M" to work for
"12:00" and because strftime(3) does the same.

24 years agoSVR4 emulator source files now take their compilation options from
newton [Wed, 8 Dec 1999 10:51:17 +0000 (10:51 +0000)]
SVR4 emulator source files now take their compilation options from
opt_global.h and opt_svr4.h, instead of from the command line.  This
brings them in-line with most of the rest of the kernel.

svr4_ioctl.c has also failed to compile with debugging for a while
now;  fixed by adding systm.h and socketvar.

Some svr4 source files are automatically generated from syscalls.master;
these have been committed as consequential changes, otherwise everyone
will have to "make svr4_sysent.c".

Changes:

sys/svr4/svr4.h include opt_global.h and opt_svr4.h
sys/svr4/svr4_ioctl.c include svr4.h, sys/systm.h and sys/socketvar.h
sys/svr4/svr4_ipc.c include svr4.h
sys/svr4/svr4_resource.c  include svr4.h
sys/svr4/svr4_socket.c include svr4.h
sys/svr4/svr4_ttold.c include svr4.h
sys/svr4/syscalls.master include svr4.h
sys/svr4/svr4_syscallnames.c dependent on syscalls.master
sys/svr4/svr4_sysent.c dependent on syscalls.master
sys/svr4/svr4_syscall.h dependent on syscalls.master
sys/svr4/svr4_proto.h dependent on syscalls.master
sys/modules/svr4/Makefile create opt_global.h and opt_svr4.h

24 years agoMake adjtime(2) adjust boottime so it doesn't cause non-monotonous
phk [Wed, 8 Dec 1999 10:02:12 +0000 (10:02 +0000)]
Make adjtime(2) adjust boottime so it doesn't cause non-monotonous
uptime.

24 years agoScan cdevs for potential root devices, rather than bdevs.
phk [Wed, 8 Dec 1999 10:01:18 +0000 (10:01 +0000)]
Scan cdevs for potential root devices, rather than bdevs.

24 years agoRemove sysinstall knowledge of the wd based devices..
sos [Wed, 8 Dec 1999 09:51:00 +0000 (09:51 +0000)]
Remove sysinstall knowledge of the wd based devices..

24 years agoRemove BAD144 support, it has already been disabled for some time.
phk [Wed, 8 Dec 1999 09:33:00 +0000 (09:33 +0000)]
Remove BAD144 support, it has already been disabled for some time.

24 years agoFinally use the new ata driver.
sos [Wed, 8 Dec 1999 09:19:37 +0000 (09:19 +0000)]
Finally use the new ata driver.

24 years agoRemove unused #includes.
phk [Wed, 8 Dec 1999 08:59:40 +0000 (08:59 +0000)]
Remove unused #includes.

Obtained from: http://bogon.freebsd.dk/include

24 years agoSwitch to using suspend/resume methods rather than the APM hooks.
imp [Wed, 8 Dec 1999 07:55:20 +0000 (07:55 +0000)]
Switch to using suspend/resume methods rather than the APM hooks.
Should have no effect, but does make things a little cleaner.  I
thought this was the race that was causing problems, but it turned out
to be in pccardd waking up after the empty/insert events had happened
and being confused.

Minor cleanup:
o Remove isahd from pccard_devinfo
o remove extra from controller
o formatting nits
o use PCCARD_DEVINFO(d) rather than a bare device_get_ivars(d)

24 years agoIf last state was supended, remove the card. This is a kludge, there
imp [Wed, 8 Dec 1999 07:47:41 +0000 (07:47 +0000)]
If last state was supended, remove the card.  This is a kludge, there
is a race here that the old code didn't deal with, and I'm not
completely sure this is the right way to solve it, but it works here.

Should get rid of the dreaded "No free configuration for card" message.

24 years agoStyle nit.
obrien [Wed, 8 Dec 1999 06:42:50 +0000 (06:42 +0000)]
Style nit.

24 years agoDocument what "X" in the table means.
obrien [Wed, 8 Dec 1999 06:42:08 +0000 (06:42 +0000)]
Document what "X" in the table means.

24 years agoWhen erroring out that there is not enough space to write the corefile,
obrien [Wed, 8 Dec 1999 06:22:58 +0000 (06:22 +0000)]
When erroring out that there is not enough space to write the corefile,
tell the user how much space is avaible for writing the corefile, and how
much space we wanted.

24 years agoFix alloc_driver() so that calls devclass_get_device() with slt->slotnum
wpaul [Wed, 8 Dec 1999 05:34:17 +0000 (05:34 +0000)]
Fix alloc_driver() so that calls devclass_get_device() with slt->slotnum
as the unit argument instead of 0. disable_slot() calls
devclass_get_device() correctly, however because alloc_driver() does
it wrong, disable_slot() is unable to locate the child devices
attached to the pccard bus and thus fails to call device_delete_child()
on them. The end result is that when a card is removed, its detach
routine is never called, and re-insertion always fails.

With this fix (and the previous one to if_wi.c), I can now insert,
remove and re-insert my WaveLAN/IEEE card and things behave correctly.
kldunloading the if_wi.ko module also seems to work properly now.

Ok'ed by: imp

24 years agoRevamp the devstat priority system. All disks now have the same priority.
ken [Wed, 8 Dec 1999 04:45:23 +0000 (04:45 +0000)]
Revamp the devstat priority system.  All disks now have the same priority.
The same goes for CD drivers and tape drivers.  In systems with mixed IDE
and SCSI, devices in the same priority class will be sorted in attach
order.

Also, the 'CCD' priority is now the 'ARRAY' priority, and a number of
drivers have been modified to use that priority.

This includes the necessary changes to all drivers, except the ATA drivers.
Soren will modify those separately.

This does not include and does not require any change in the devstat
version number, since no known userland applications use the priority
enumerations.

Reviewed by: msmith, sos, phk, jlemon, mjacob, bde

24 years agoI've been Brucified! I did evil things with typedefs, but I'll do it
green [Wed, 8 Dec 1999 02:44:46 +0000 (02:44 +0000)]
I've been Brucified!  I did evil things with typedefs, but I'll do it
the correct (but less aesthetic) way, now.  New lesson: correctness
and aestheticity may be mutually exclusive at times.

Brucified by: bde

24 years agoFix invocation of bus_teardown_intr(). The last arg is supposed to be
wpaul [Wed, 8 Dec 1999 02:00:10 +0000 (02:00 +0000)]
Fix invocation of bus_teardown_intr(). The last arg is supposed to be
sc->wi_intrhand, not &sc->wi_intrhand.

24 years agoo Implement IRQ activation. We do this in bus_{setup,teardown}_intr,
imp [Wed, 8 Dec 1999 01:42:44 +0000 (01:42 +0000)]
o Implement IRQ activation.  We do this in bus_{setup,teardown}_intr,
  per discussions on -arch.  This is different than -current's pccard
  code, but shouldn't have any real difference.
o More code cleanup.

Next step: convert bus_space_* mapping stuff to newbus.

24 years agoAdd a cross-reference to fabs(3) man page.
chris [Tue, 7 Dec 1999 23:09:58 +0000 (23:09 +0000)]
Add a cross-reference to fabs(3) man page.

PR: docs/15337
Submitted by: Bruce A. Mah <bmah@ca.sandia.gov>

24 years agoCommit the kernel part of our DVD support. Nothing much to say really,
sos [Tue, 7 Dec 1999 22:25:28 +0000 (22:25 +0000)]
Commit the kernel part of our DVD support. Nothing much to say really,
its just a number of new ioctl's, the rest is done in userland.

24 years agoFix typo errors, remove repeated paragraphs, update incorrect (i.e. mixed up)
chris [Tue, 7 Dec 1999 22:10:35 +0000 (22:10 +0000)]
Fix typo errors, remove repeated paragraphs, update incorrect (i.e. mixed up)
information.

PR: docs/15139
Submitted by: Kazuo Horikawa <horikawa@jp.FreeBSD.org>

24 years agoTry a little harder at the VIA DMA support.
sos [Tue, 7 Dec 1999 22:07:18 +0000 (22:07 +0000)]
Try a little harder at the VIA DMA support.
Only use UDMA66 on 82C686 and with prober cable.

24 years agoCorrect the synopsis for the wi(4) driver: it's just "device wi0" now, not
wpaul [Tue, 7 Dec 1999 21:52:41 +0000 (21:52 +0000)]
Correct the synopsis for the wi(4) driver: it's just "device wi0" now, not
"device wi0 at isa? port? irq?".

24 years agoSpruce up the Sundance ST201 driver:
wpaul [Tue, 7 Dec 1999 20:14:42 +0000 (20:14 +0000)]
Spruce up the Sundance ST201 driver:

- Convert to using TX descritor polling similar to the xl driver (the
  ST201 is a clone of the 3c90xB chipset and offers the same transmit
  polling scheme). This should reduce TX overhad a little.

- Make sure to reset PHY when switching mode, as in the starfire driver.

- Fix instances of free() that should be contigfree().

- Remove dead code.

24 years agoBump CIS_MAXSTR from 30 to 254. pccard appears to define the entire
imp [Tue, 7 Dec 1999 19:23:56 +0000 (19:23 +0000)]
Bump CIS_MAXSTR from 30 to 254.  pccard appears to define the entire
section we take them from to be up to 255 bytes long, so that's the
max size for the string.  They can't all be this big, but I don't have
a better number and better to be a little long than a little short.

Also only consume len characters of the cis buffer so we don't run off
the end into the next buffer and get garbage.  This second patch
shouldn't impact anything, but I'll hold off back porting this to
-stable until I get more reports on the stability before/after this
fix.

24 years agoTweak the DC_REDUCED_MII_POLL code in dc_mii_tick() for the DC_IS_INTEL()
wpaul [Tue, 7 Dec 1999 19:18:41 +0000 (19:18 +0000)]
Tweak the DC_REDUCED_MII_POLL code in dc_mii_tick() for the DC_IS_INTEL()
case. The idea is to reduce how often we call mii_tick(), however currently
it may not be called often enough, which prevents autonegotiation from
being driven correctly.

This should improve the chances of successfully autonegotiating media
settings on non-MII 21143 NICs. (Still waiting for confirmation from
some testers, but the code is clearly wrong in any case.)

24 years agoUpdate the sample for $init_path to reflect the kernel default.
msmith [Tue, 7 Dec 1999 18:35:58 +0000 (18:35 +0000)]
Update the sample for $init_path to reflect the kernel default.

24 years agoUpdate to reflect use of : not ; as separator in paths passed to the
msmith [Tue, 7 Dec 1999 18:31:43 +0000 (18:31 +0000)]
Update to reflect use of : not ; as separator in paths passed to the
kernel.

24 years agoudp IPv6 support, IPv6/IPv4 tunneling support in kernel,
shin [Tue, 7 Dec 1999 17:39:16 +0000 (17:39 +0000)]
udp IPv6 support, IPv6/IPv4 tunneling support in kernel,
packet divert at kernel for IPv6/IPv4 translater daemon

This includes queue related patch submitted by jburkhol@home.com.

Submitted by: queue related patch from jburkhol@home.com
Reviewed by: freebsd-arch, cvs-committers
Obtained from: KAME project

24 years agoFix bug where tuner i2c address was not set for IMS Turbo cards
roger [Tue, 7 Dec 1999 17:23:39 +0000 (17:23 +0000)]
Fix bug where tuner i2c address was not set for IMS Turbo cards
Put back Hauppauge Tuner types 0x5, 0x0a, 0x1a which I deleted by mistake.

24 years agoBuild gasp (GAS Preprocessor). This is needed for building Glide.
dfr [Tue, 7 Dec 1999 17:01:27 +0000 (17:01 +0000)]
Build gasp (GAS Preprocessor). This is needed for building Glide.

24 years agoFix support for disk with > 64K cyls.
sos [Tue, 7 Dec 1999 14:49:00 +0000 (14:49 +0000)]
Fix support for disk with > 64K cyls.

24 years agoTemporary adjustment to build after pnp changes. There is a new version
peter [Tue, 7 Dec 1999 12:06:31 +0000 (12:06 +0000)]
Temporary adjustment to build after pnp changes.  There is a new version
in the pipeline apparently.

OK'ed by: hm

24 years agoFix the ioctl CONS_FINDMODE and its underlying subroutine
yokota [Tue, 7 Dec 1999 11:23:58 +0000 (11:23 +0000)]
Fix the ioctl CONS_FINDMODE and its underlying subroutine
xxx_query_mode() in the vga and vesa drivers.

- xxx_query_mode() returns 0 (success) and a positive error number.
- Copy mode information on success.
- Remove redundant structure copy.

The bug first found in -STABLE by jmg.

24 years agoMerge from sys/isa/fd.c rev 1.170.
kato [Tue, 7 Dec 1999 09:32:59 +0000 (09:32 +0000)]
Merge from sys/isa/fd.c rev 1.170.

24 years agoMerge from sys/isa/sio.c rev 1.277 & 1.278.
kato [Tue, 7 Dec 1999 09:29:15 +0000 (09:29 +0000)]
Merge from sys/isa/sio.c rev 1.277 & 1.278.

24 years agoMerge from sys/i386/conf/files.i386 rev 1.283.
kato [Tue, 7 Dec 1999 09:23:44 +0000 (09:23 +0000)]
Merge from sys/i386/conf/files.i386 rev 1.283.

24 years agoMerge from sys/dev/syscons/syscons.c rev 1.328.
kato [Tue, 7 Dec 1999 09:21:19 +0000 (09:21 +0000)]
Merge from sys/dev/syscons/syscons.c rev 1.328.

24 years agoMerge from sys/i386/i386/userconfig.c rev 1.162 & 1.163.
kato [Tue, 7 Dec 1999 09:16:52 +0000 (09:16 +0000)]
Merge from sys/i386/i386/userconfig.c rev 1.162 & 1.163.

24 years agoTonight's cleanups.
imp [Tue, 7 Dec 1999 06:44:38 +0000 (06:44 +0000)]
Tonight's cleanups.
o Implement memory and I/O activation/deactivation.  irq not handled.
o switch pcic_chip functions around to use more convenient types.
o kill __P and most of the old K&R prototypes just to be mean.
o minor other nits

24 years agoRemove a bunch of un-needed includes.
julian [Tue, 7 Dec 1999 05:50:48 +0000 (05:50 +0000)]
Remove a bunch of un-needed includes.
Submitted by: phk@freebsd.org

24 years agoChange the default poweroff delay from 0 to 5 seconds. This seems to be
msmith [Tue, 7 Dec 1999 04:35:37 +0000 (04:35 +0000)]
Change the default poweroff delay from 0 to 5 seconds.  This seems to be
adequate for the IDE disks that I have available for testing.  Most seem
to wait between 1 and 3 seconds before flushing their caches.

Add the ability to override the delay at compile time via the
undocumented option POWEROFF_DELAY.  The delay can still be set via
sysctl as it was originally implemented.

24 years agoMake some examples reflect defaults.
dcs [Tue, 7 Dec 1999 04:24:05 +0000 (04:24 +0000)]
Make some examples reflect defaults.

24 years agoRecognize the Pioneer CD-ROM DRM-6324X as a changer.
ken [Tue, 7 Dec 1999 04:14:56 +0000 (04:14 +0000)]
Recognize the Pioneer CD-ROM DRM-6324X as a changer.

PR: kern/15198
Submitted by: Don Lewis <truckman@FreeBSD.org>

24 years agoDo proper constification in args.c. This shuts up -Wcast-qual (thanks,
green [Tue, 7 Dec 1999 03:32:37 +0000 (03:32 +0000)]
Do proper constification in args.c.  This shuts up -Wcast-qual (thanks,
bfumerola for that pointer!) in GCC complaining about losing a const.

While I'm here, might as well mark in the Makefile that I'm the
${MAINTAINER}.  It seems like that's what everyone's doing these days.

24 years agoThe typo was actually that ``if'' was misspelled as ``it.''
chris [Tue, 7 Dec 1999 03:16:18 +0000 (03:16 +0000)]
The typo was actually that ``if'' was misspelled as ``it.''

24 years agoFix a typo (specifically, the word ``it'' was used twice in a row where it
chris [Tue, 7 Dec 1999 03:13:17 +0000 (03:13 +0000)]
Fix a typo (specifically, the word ``it'' was used twice in a row where it
did not need to be).

PR: 15207
Submitted by: Kazuo Seki <seki@is.utsunomiya-u.ac.jp>

24 years ago-Wall cleanup / variable correctness.
billf [Tue, 7 Dec 1999 03:00:55 +0000 (03:00 +0000)]
-Wall cleanup / variable correctness.

Reviewed by: green

24 years agoChange ``.Dq Fl o'' to ``.Dq -o''
chris [Tue, 7 Dec 1999 02:47:59 +0000 (02:47 +0000)]
Change ``.Dq Fl o'' to ``.Dq -o''

24 years ago-Wall cleanup.
billf [Tue, 7 Dec 1999 02:01:29 +0000 (02:01 +0000)]
-Wall cleanup.

XXX: The signal handling "fix" for worm(6) is wrong. However, the functions
themselves are wrong as well: See sigaction(2).

24 years agoReckognize the 'Avance Asound 100'
billf [Tue, 7 Dec 1999 01:53:24 +0000 (01:53 +0000)]
Reckognize the 'Avance Asound 100'

Submitted by: Oliver Fromme <olli@dorifer.heim3.tu-clausthal.de>

24 years agoAdd vendor NetChip with its device NET1080 (host-to-host adapter)
n_hibma [Tue, 7 Dec 1999 01:19:51 +0000 (01:19 +0000)]
Add vendor NetChip with its device NET1080 (host-to-host adapter)

24 years agoAdd reference to netgraph(4) in the 'see also' section.
archie [Mon, 6 Dec 1999 23:35:40 +0000 (23:35 +0000)]
Add reference to netgraph(4) in the 'see also' section.

24 years ago"raw" -> "block" Block devices have been what are now gone, not "raw"
imp [Mon, 6 Dec 1999 22:42:01 +0000 (22:42 +0000)]
"raw" -> "block" Block devices have been what are now gone, not "raw"
devices.

Submitted by: David O'Brian

24 years agoI always forget to check before I reboot a system, and while it
phk [Mon, 6 Dec 1999 22:35:51 +0000 (22:35 +0000)]
I always forget to check before I reboot a system, and while it
boots I try in vain to remember which month or even year this system
was last booted in.

Print out the uptime before rebooting, and give people like me
less (or more as it may be) to think about while the systems boots.

24 years agoInclude BSD-license.
wosch [Mon, 6 Dec 1999 22:33:40 +0000 (22:33 +0000)]
Include BSD-license.

PR: misc/15262
Submitted by:  alex@big.endian.de

24 years agoMention the '6' flag in the flags summary. (Forgotten in the last commit).
joe [Mon, 6 Dec 1999 21:49:07 +0000 (21:49 +0000)]
Mention the '6' flag in the flags summary.  (Forgotten in the last commit).

24 years agoLast minute patch that I forgot to apply: check return code of iplattach()
guido [Mon, 6 Dec 1999 21:21:47 +0000 (21:21 +0000)]
Last minute patch that I forgot to apply: check return code of iplattach()

24 years agoAdd braces ({}) to all the places that they're missing.
joe [Mon, 6 Dec 1999 21:18:32 +0000 (21:18 +0000)]
Add braces ({}) to all the places that they're missing.

24 years agoMove basic ifilter utils to sbin where they shold have been committed by
guido [Mon, 6 Dec 1999 20:50:04 +0000 (20:50 +0000)]
Move basic ifilter utils to sbin where they shold have been committed by
me in the first place. While we're at it: add MAINTAINER line

24 years agoRevive mlfk_ipl here. This version is slightly changed from
guido [Mon, 6 Dec 1999 20:36:50 +0000 (20:36 +0000)]
Revive mlfk_ipl here. This version is slightly changed from
the old one: an unnecessary define (KLD_MODULE) has been deleted and
the initialisation of the module is  done after domaininit was called
to be sure inet is running.

Some slight changed were made to ip_auth.c and ip_state.c in order
to assure including of sys/systm.h in case we make a kld

Make sure ip_fil does nmot include osreldate in kernel mode

Remove mlfk_ipl.c from here: no sources allowed in these directories!

24 years agoAdd forgotten ip_fil.c and fil.c to kld version
guido [Mon, 6 Dec 1999 20:33:38 +0000 (20:33 +0000)]
Add forgotten ip_fil.c and fil.c to kld version

Remove mlfk_ipl.c from here: no sources allowed in these directories!

24 years agoAdd ifilter klld module to ipfilter based kernels
guido [Mon, 6 Dec 1999 20:32:53 +0000 (20:32 +0000)]
Add ifilter klld module to ipfilter based kernels
Allow kld based kernel with option IPFILTER_LKM.

24 years agoRewrite of the manual page to make it clear that NSI and InterNIC
joe [Mon, 6 Dec 1999 19:58:57 +0000 (19:58 +0000)]
Rewrite of the manual page to make it clear that NSI and InterNIC
are no longer the same organisation.

Submitted by: Szilveszter Adam <sziszi@petra.hos.u-szeged.hu>

24 years agoPrototypes should either have explicit parameter names, or not,
archie [Mon, 6 Dec 1999 19:08:31 +0000 (19:08 +0000)]
Prototypes should either have explicit parameter names, or not,
but some combination of the two.

24 years agoChange definition of NG_BPF_HOOKPROG_SIZE() so as not to require
archie [Mon, 6 Dec 1999 18:43:32 +0000 (18:43 +0000)]
Change definition of NG_BPF_HOOKPROG_SIZE() so as not to require
a structure pointer, just the number of BPF instructions.

24 years agoSimplify my license.
gibbs [Mon, 6 Dec 1999 18:29:03 +0000 (18:29 +0000)]
Simplify my license.

Don't arbitrarily limit the initiator ID of the card to something <= 7.

Fix a bug in the checksum code that would incorrectly prevent a valid
checksum of zero. (cp)

Don't touch rely on seeprom data when configuring termination.  We may
not have seeprom data. (cp)

Treat all ULTRA2 capable adapters the same way when reading or writing
the BRDCTL register.  We previously only did this correctly for aic7890/91
chips.  This should correct some problems with termination settings on
aic7896/97 adapters. (cp)

Changes marked with "(cp)"
Pointed out by: Chuck Paterson <cp@bsdi.com>

24 years agoUpdate for pnp adjustments regarding NPNP.
peter [Mon, 6 Dec 1999 18:26:33 +0000 (18:26 +0000)]
Update for pnp adjustments regarding NPNP.

Also, optimize out a mess of #if's that were duplicating work already
done by config(8).  For example, if a file is marked as
"dev/sound/pci/foo.c optional pcm pci" then it's only added if pcm *and*
pci are present, so #if NPCM > 0 and #if NPCI > 0 are totally redundant.
A bit more work is still needed.

Discussed with: cg (a few weeks ago)

24 years agoSimplify my copyright license terms.
gibbs [Mon, 6 Dec 1999 18:23:31 +0000 (18:23 +0000)]
Simplify my copyright license terms.

aic7xxx.c:
Add a function for sucking firmware out of the controller
prior to reset.

Remove some inline bloat from functions that should not have
been inlined.

During initialization, wait 1ms after the chip reset before
touching any registers.  You can get machine checks on certain
architectures (Atari I think?) without the delay.

Return CAM_REQ_CMP for external BDR requests instead
of CAM_BDR_SENT.

Bump some messages to bootverbose levels above 1.

Don't clear any negotiated sync rate if the target rejects
a WDTR message.  The sync rate is only cleared if the target
accepts a WDTR message.

Fix a small bug in the mesgin handling code that could cause
us to believe that we had recieved a message that was actually
received by another target.  This could only confuse us in
some very rare transmission negotiation scenarios.

Remove some unecessary cleanup of residual information after
a residual is reported.  The sequencer does this when the
command is queued now.

24 years agoUpdate for loss of pnp.h
peter [Mon, 6 Dec 1999 18:22:19 +0000 (18:22 +0000)]
Update for loss of pnp.h

24 years agoMerge pnp change from i386/conf/GENERIC (rev 1.211)
peter [Mon, 6 Dec 1999 18:20:31 +0000 (18:20 +0000)]
Merge pnp change from i386/conf/GENERIC (rev 1.211)

24 years agoFold the pnp code into the base isa system to pave the way for PNPBIOS.
peter [Mon, 6 Dec 1999 18:17:42 +0000 (18:17 +0000)]
Fold the pnp code into the base isa system to pave the way for PNPBIOS.

Reviewed by: dfr (a few weeks ago)

24 years agoMake this compile again. (missing #include for RFPROC)
peter [Mon, 6 Dec 1999 18:12:29 +0000 (18:12 +0000)]
Make this compile again.  (missing #include for RFPROC)

24 years agoFix a pair of silly warnings that I introduced (that would have been
peter [Mon, 6 Dec 1999 17:52:20 +0000 (17:52 +0000)]
Fix a pair of silly warnings that I introduced (that would have been
nasty on an Alpha or some other sizeof(int) != sizeof(long) machine).

24 years agoFix another quirk in the unknown device detection, and also deal with
peter [Mon, 6 Dec 1999 15:52:18 +0000 (15:52 +0000)]
Fix another quirk in the unknown device detection, and also deal with
unknown 'controller' lines.

24 years agoComplain if we get an error that is not ENXIO while opening /dev/usb\d+
n_hibma [Mon, 6 Dec 1999 14:50:41 +0000 (14:50 +0000)]
Complain if we get an error that is not ENXIO while opening /dev/usb\d+
devices. For example, starting 'usbd -e' would give a 'No USB
controllers found' message instead of a '/dev/usb0: Permission denied'.

Submitted-By: Dirk-Willem van Gulik <dirkx@webweaving.org>
24 years agoPut on my asbestos underwear and commit the patch that I posted to -arch
peter [Mon, 6 Dec 1999 11:13:50 +0000 (11:13 +0000)]
Put on my asbestos underwear and commit the patch that I posted to -arch
some time ago that changes kern.randompid from a boolean to a randomness
range for the next pid assigment.  Too high causes a lot of extra work
to scan for free pids, and too low merely wastes randomness entropy.  It's
still possible to select a completely random range by using PID_MAX (100k)
or -1 as a shortcut to mean "the whole range".
Also, don't waste randomness when doing a wraparound.

24 years agoRemove DSO_BAD144 from wd driver(s) so people with bad144'ed disks get
phk [Mon, 6 Dec 1999 10:19:03 +0000 (10:19 +0000)]
Remove DSO_BAD144 from wd driver(s) so people with bad144'ed disks get
a bit of warning.

24 years agoAdd defines for additional EEPROM locations and register definitions.
mdodd [Mon, 6 Dec 1999 09:29:34 +0000 (09:29 +0000)]
Add defines for additional EEPROM locations and register definitions.

24 years agoWhitespace cleanups.
mdodd [Mon, 6 Dec 1999 09:16:13 +0000 (09:16 +0000)]
Whitespace cleanups.

24 years agoCorrect the ttys.5 and init.8 manpages with respect to the incorrect
sheldonh [Mon, 6 Dec 1999 09:07:14 +0000 (09:07 +0000)]
Correct the ttys.5 and init.8 manpages with respect to the incorrect
assumption that only getty processes can be managed.  Describe the
SysV-like ability to keep arbitrary long-running processes alive
using a non-device first field in /etc/ttys.

PR: 12767
Submitted by: Peter Jeremy <peter.jeremy@alcatel.com.au>

24 years agoAdd a delay (per the databook) to the get_eeprom_data() loop so
mdodd [Mon, 6 Dec 1999 08:59:52 +0000 (08:59 +0000)]
Add a delay (per the databook) to the get_eeprom_data() loop so
that the read EEPROM command has time to execute.

I didn't observe any difference in behavior on my test system but
this is the documented "correct behavior".

24 years agoStandardize device descriptions so that there is parity between PnP
mdodd [Mon, 6 Dec 1999 08:40:59 +0000 (08:40 +0000)]
Standardize device descriptions so that there is parity between PnP
descriptions and non-PnP descriptions.

24 years agoCorrectly implement the non-PnP enumeration of ISA board.
mdodd [Mon, 6 Dec 1999 08:31:47 +0000 (08:31 +0000)]
Correctly implement the non-PnP enumeration of ISA board.

We now correctly skip boards that have PnP support enabled, or are in
test mode.  The 3c509s support a number of combinations of device
probing, as per the databook.

- ISA only
- PnP only
- ISA or PnP

We will allow cards that can be dealt with by PnP to be attached by the
PnP enumerator.

This fixes the bogus detection of boards at weird ioports.

24 years agoAdd a truely evil workaround (hack!) for some unfortunate BIOS
peter [Mon, 6 Dec 1999 06:59:09 +0000 (06:59 +0000)]
Add a truely evil workaround (hack!) for some unfortunate BIOS
programming practices.  It seems that newer fdc chips have an
alternative way of setting the transfer speed (including high speed
modes for floppy tape) that doesn't use the control register (which
we don't support - we use the old way only).  So, they (the BIOS
programmers) sometimes leave out the 0x3f6 control register from
the PnP ports descriptor(!!).  "Hey, it works with windows, so
what's the problem?" :-(  Anyway, this hack tries to compensate
for that.  This was discussed with dfr (who did the pnp attachment).

24 years agoUse bus_get_resource() instead of bus_get_resource_start() as the latter
peter [Mon, 6 Dec 1999 06:47:02 +0000 (06:47 +0000)]
Use bus_get_resource() instead of bus_get_resource_start() as the latter
returns "0" on failure, which is indistinguishable from (say) irq 0.  This
should stop a couple of stray messages that turn up.

Also, if a BUS_SETUP_INTR() fails with INTR_TYPE_FAST, try falling back to
a normal interrupt.  This might help pccard folks with a shared slot
interrupt.  This whole thing needs to be revisited.

24 years agoMake this compile again after the device_add_child change.
imp [Mon, 6 Dec 1999 05:22:47 +0000 (05:22 +0000)]
Make this compile again after the device_add_child change.

24 years agoNeed header <machine/smp.h> for prototype declaration of smp_rendezvous()
luoqi [Mon, 6 Dec 1999 05:12:26 +0000 (05:12 +0000)]
Need header <machine/smp.h> for prototype declaration of smp_rendezvous()
in my previous commit.

24 years agoAdd "-v" to usage().
obrien [Mon, 6 Dec 1999 05:02:32 +0000 (05:02 +0000)]
Add "-v" to usage().

24 years agoDocument the "-v" flag. [It was like that when I got here.. honestly.]
obrien [Mon, 6 Dec 1999 05:01:14 +0000 (05:01 +0000)]
Document the "-v" flag.  [It was like that when I got here.. honestly.]

24 years agoUser ldt sharing.
luoqi [Mon, 6 Dec 1999 04:53:08 +0000 (04:53 +0000)]
User ldt sharing.

24 years agobad144 elimination
imp [Mon, 6 Dec 1999 04:21:50 +0000 (04:21 +0000)]
bad144 elimination
bdev removal and its required actions (MAKEDEV)
sd -> da