]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
24 years agoArgh, fix cut/paste mistake. This contributed to LINT not building.
peter [Mon, 20 Mar 2000 10:17:13 +0000 (10:17 +0000)]
Argh, fix cut/paste mistake.  This contributed to LINT not building.

24 years agoA fairly simple newbusification of if_de
peter [Mon, 20 Mar 2000 09:18:46 +0000 (09:18 +0000)]
A fairly simple newbusification of if_de

24 years agoUnused definitions.
peter [Mon, 20 Mar 2000 08:21:52 +0000 (08:21 +0000)]
Unused definitions.

24 years agoWhite-space-only change. The file originally had screwed indentation.
yokota [Mon, 20 Mar 2000 08:18:00 +0000 (08:18 +0000)]
White-space-only change.  The file originally had screwed indentation.

24 years agoMake this compile. This is a hack.
peter [Mon, 20 Mar 2000 08:17:34 +0000 (08:17 +0000)]
Make this compile. This is a hack.

24 years agoAdd Mitsumi Wireless Scroll Mouse PnP ID.
yokota [Mon, 20 Mar 2000 08:11:11 +0000 (08:11 +0000)]
Add Mitsumi Wireless Scroll Mouse PnP ID.

24 years agoThis file was repo copied and I forgot to cvs rm this here.
peter [Mon, 20 Mar 2000 07:21:40 +0000 (07:21 +0000)]
This file was repo copied and I forgot to cvs rm this here.

24 years agoNoriyoshi Kawano-san reported in nomads another OUI for linksys. 4 is
imp [Mon, 20 Mar 2000 07:09:32 +0000 (07:09 +0000)]
Noriyoshi Kawano-san reported in nomads another OUI for linksys.  4 is
too many, so I removed the checks for the valid OUIs.  We already do a
checksum of the entire ethernet address, so extra checking against the
OUI shouldn't be needed.

24 years agoOnly character devices exsist these days. Make savecore understand
ps [Mon, 20 Mar 2000 06:54:06 +0000 (06:54 +0000)]
Only character devices exsist these days.  Make savecore understand
this.

24 years agoTag a warning on the isa compat shims at config time.
peter [Mon, 20 Mar 2000 04:47:40 +0000 (04:47 +0000)]
Tag a warning on the isa compat shims at config time.

24 years agoIncrement number of retries before giving up. This may fix the
kato [Mon, 20 Mar 2000 03:55:29 +0000 (03:55 +0000)]
Increment number of retries before giving up.  This may fix the
problem that some CD-ROM drives are not recognized if they are empty.

Submitted by: Tomokazu HARADA <tkhara@osk4.3web.ne.jp>

24 years agoUse an intermediate file containing the module binary before converting
peter [Mon, 20 Mar 2000 01:29:59 +0000 (01:29 +0000)]
Use an intermediate file containing the module binary before converting
it into a ``shared'' .ko file.  This intermediate file can be directly
linked into a static kernel.  This isn't all that useful yet but will
become much more interesting shortly.

24 years agoBandaid for src/sys/modules which broke world
peter [Mon, 20 Mar 2000 01:22:27 +0000 (01:22 +0000)]
Bandaid for src/sys/modules which broke world

24 years agoUpdate the copyright. Remove the advertising clause.
paul [Mon, 20 Mar 2000 00:53:46 +0000 (00:53 +0000)]
Update the copyright. Remove the advertising clause.

24 years agoMake the SUBDIR list more granular to assist in my GCC hacking.
obrien [Sun, 19 Mar 2000 23:47:50 +0000 (23:47 +0000)]
Make the SUBDIR list more granular to assist in my GCC hacking.

24 years agoAdd missing depend file
ache [Sun, 19 Mar 2000 23:07:39 +0000 (23:07 +0000)]
Add missing depend file

24 years agoInstall readline API info
ache [Sun, 19 Mar 2000 23:04:23 +0000 (23:04 +0000)]
Install readline API info

24 years agoUpgrade to 4.1
ache [Sun, 19 Mar 2000 22:26:49 +0000 (22:26 +0000)]
Upgrade to 4.1
Minor bumped because new function added

24 years agoMerge conflicts
ache [Sun, 19 Mar 2000 22:00:57 +0000 (22:00 +0000)]
Merge conflicts

24 years agoo Replace most magic numbers related to token ring with #defines
lile [Sun, 19 Mar 2000 21:34:39 +0000 (21:34 +0000)]
o Replace most magic numbers related to token ring with #defines
  from iso88025.h.

o Add minimal llc support to iso88025_input.

o Clean up most of the source routing code.

* Submitted by: Nikolai Saoukh <nms@otdel-1.org>

24 years agoThis commit was generated by cvs2svn to compensate for changes in r58310,
ache [Sun, 19 Mar 2000 21:18:09 +0000 (21:18 +0000)]
This commit was generated by cvs2svn to compensate for changes in r58310,
which included commits to RCS files with non-trunk default branches.

24 years agoInitial import of v4.1
ache [Sun, 19 Mar 2000 21:18:09 +0000 (21:18 +0000)]
Initial import of v4.1

24 years agoAdd a new flag: -E enables "extended" regular expressions.
green [Sun, 19 Mar 2000 19:41:53 +0000 (19:41 +0000)]
Add a new flag: -E enables "extended" regular expressions.

24 years agoAdd $FreeBSD$.
green [Sun, 19 Mar 2000 19:39:41 +0000 (19:39 +0000)]
Add $FreeBSD$.

24 years agoEnable the K6-2 MTRR driver again, since there are reports of it working
green [Sun, 19 Mar 2000 19:22:24 +0000 (19:22 +0000)]
Enable the K6-2 MTRR driver again, since there are reports of it working
with the applied change.

Submitted by:  Coleman Kane <cokane@one.net>

24 years agoSet the bits in the mask from the left to the right, not backwards.
green [Sun, 19 Mar 2000 19:21:37 +0000 (19:21 +0000)]
Set the bits in the mask from the left to the right, not backwards.

Submitted by:  Coleman Kane <cokane@one.net>

24 years agoFix typo.
asmodai [Sun, 19 Mar 2000 16:52:59 +0000 (16:52 +0000)]
Fix typo.

24 years agoTrim trailing whitespace.
asmodai [Sun, 19 Mar 2000 15:55:38 +0000 (15:55 +0000)]
Trim trailing whitespace.

24 years agoDocument the sysctl knob for setting the severity of the SPLASSERT
asmodai [Sun, 19 Mar 2000 15:47:33 +0000 (15:47 +0000)]
Document the sysctl knob for setting the severity of the SPLASSERT
macros.

Reviewed by: ps

24 years agoMerged from sys/isa/fd.c.
kato [Sun, 19 Mar 2000 15:18:22 +0000 (15:18 +0000)]
Merged from sys/isa/fd.c.

24 years agoo Fix typo in the RapidFire 3540 adapter name
lile [Sun, 19 Mar 2000 15:08:39 +0000 (15:08 +0000)]
o Fix typo in the RapidFire 3540 adapter name
  4/16/100 -> 100/16/4 so that it matches all of the others.

o Since we went to all the trouble of getting the correct
  working memory size actually use it.

* Submitted by: Nikolai Saoukh <nms@otdel-1.org>

24 years agoDefine PC98 when MACHINE == pc98, because msdos_vfsops.c contains
kato [Sun, 19 Mar 2000 15:06:18 +0000 (15:06 +0000)]
Define PC98 when MACHINE == pc98, because msdos_vfsops.c contains
`#ifdef PC98'.

24 years agoMake SPLASSERT sysctl and boot time tunable with kern.splassertmode.
ps [Sun, 19 Mar 2000 14:55:42 +0000 (14:55 +0000)]
Make SPLASSERT sysctl and boot time tunable with kern.splassertmode.
The following values are understood:  0 (ignore), 1 (log), and 2
(panic).

The default value is 1.

Reviewed by: peter

24 years agoSpell Jonathan's name correctly in the copyright. Which is funny,
asmodai [Sun, 19 Mar 2000 14:46:56 +0000 (14:46 +0000)]
Spell Jonathan's name correctly in the copyright.  Which is funny,
considering the fact he wrote it himself. =)

Also add his email address in the AUTHORS section.

24 years agoMerged from sys/i386/i386/userconfig.c rev 1.178.
kato [Sun, 19 Mar 2000 14:46:22 +0000 (14:46 +0000)]
Merged from sys/i386/i386/userconfig.c rev 1.178.

24 years agoAdd the manpage for CONDSPLASSERT().
asmodai [Sun, 19 Mar 2000 14:43:03 +0000 (14:43 +0000)]
Add the manpage for CONDSPLASSERT().
This was based upon the manpage written by jmb for SPLASSERT().

24 years agoAdded COMPAT_OLDPCI and COMPAT_OLDISA options.
kato [Sun, 19 Mar 2000 14:42:52 +0000 (14:42 +0000)]
Added COMPAT_OLDPCI and COMPAT_OLDISA options.

24 years agoAdd the COMPAT_OLDPCI option for the alpha so GENERIC compiles.
peter [Sun, 19 Mar 2000 13:57:09 +0000 (13:57 +0000)]
Add the COMPAT_OLDPCI option for the alpha so GENERIC compiles.

24 years agoDocument and supply COMPAT_OLDPCI and COMPAT_OLDISA so 'make release'
peter [Sun, 19 Mar 2000 13:33:26 +0000 (13:33 +0000)]
Document and supply COMPAT_OLDPCI and COMPAT_OLDISA so 'make release'
still works.

24 years agoConnect the ISA and PCI compatability shims to an option. In this case
peter [Sun, 19 Mar 2000 13:07:12 +0000 (13:07 +0000)]
Connect the ISA and PCI compatability shims to an option.  In this case
it's options COMPAT_OLDISA and COMPAT_OLDPCI.  This is meant to be a
fairly strong incentive to update the older drivers to newbus, but doesn't
(quite) leave anybody hanging with no hardware support.  I was talking with
a few folks and I was encouraged to simply break or disable the shims but
that was a bit too drastic for my liking.

24 years agoCompletely decouple userconfig from 'struct isa_device' and friends.
peter [Sun, 19 Mar 2000 12:57:49 +0000 (12:57 +0000)]
Completely decouple userconfig from 'struct isa_device' and friends.
Userconfig was only using this for convenience since newbus, and even
then it was rather inconvenient.

24 years agoAdd some diagnostics to prove that incoming IP fragments are
brian [Sun, 19 Mar 2000 12:37:27 +0000 (12:37 +0000)]
Add some diagnostics to prove that incoming IP fragments are
being dealt with correctly.

24 years agoRemove the old a.out-only boot loader, we don't use it any more.
peter [Sun, 19 Mar 2000 11:42:34 +0000 (11:42 +0000)]
Remove the old a.out-only boot loader, we don't use it any more.
The new boot1/boot2 can load a.out and elf kernels directly.
I think the sys/pc98 version can go too as the sys/boot/pc98 code
appears to be functional, but I'll leave that for the pc98 folks.

"There can be only one."

24 years agoAdd conditional splassert.
ps [Sun, 19 Mar 2000 11:39:26 +0000 (11:39 +0000)]
Add conditional splassert.

Reviewed by: peter

24 years agoAdd device isic to the ASUSCOM_IPAC entry.
asmodai [Sun, 19 Mar 2000 10:19:23 +0000 (10:19 +0000)]
Add device isic to the ASUSCOM_IPAC entry.

Remove quotes around some i4b options to be consistent with the rest.

24 years agoAdd and document a NO_X knob to force disabling of X support in doscmd
kris [Sun, 19 Mar 2000 09:49:03 +0000 (09:49 +0000)]
Add and document a NO_X knob to force disabling of X support in doscmd

Requested by: sysop@silver.komanda.com.ua

24 years agoMake _FindLinkIn() static and only define GetDestPort when
brian [Sun, 19 Mar 2000 09:11:05 +0000 (09:11 +0000)]
Make _FindLinkIn() static and only define GetDestPort when
NO_FW_PUNCH isn't defined.

24 years agoModify if_wi.c to call bpfdetach(ifp) before if_detach(ifp), so as to
rwatson [Sun, 19 Mar 2000 05:46:50 +0000 (05:46 +0000)]
Modify if_wi.c to call bpfdetach(ifp) before if_detach(ifp), so as to
prevent a page fault if the card is ejected while BPF is in use.  This
could happen if DHCP or tcpdump was in use on that interface during
ejection.  Other drivers may also require this modification.

Reviewed by: wes

24 years agoThe advent of if_detach, allowing interface removal at runtime, makes it
rwatson [Sun, 19 Mar 2000 05:42:34 +0000 (05:42 +0000)]
The advent of if_detach, allowing interface removal at runtime, makes it
possible for a panic to occur if BPF is in use on the interface at the
time of the call to if_detach.  This happens because BPF maintains pointers
to the struct ifnet describing the interface, which is freed by if_detach.

To correct this problem, a new call, bpfdetach, is introduced.  bpfdetach
locates BPF descriptor references to the interface, and NULLs them.  Other
BPF code is modified so that discovery of a NULL interface results in
ENXIO (already implemented for some calls).  Processes blocked on a BPF
call will also be woken up so that they can receive ENXIO.

Interface drivers that invoke bpfattach and if_detach must be modified to
also call bpfattach(ifp) before calling if_detach(ifp).  This is relevant
for buses that support hot removal, such as pccard and usb.  Patches to
all effected devices will not be committed, only to if_wi.c, due to
testing limitations.  To reproduce the crash, load up tcpdump on you
favorite pccard ethernet card, and then eject the card.  As some pccard
drivers do not invoke if_detach(ifp), this bug will not manifest itself
for those drivers.

Reviewed by: wes

24 years agoMissing pieces of the last commit ;-(
yokota [Sun, 19 Mar 2000 04:37:18 +0000 (04:37 +0000)]
Missing pieces of the last commit ;-(

24 years ago- Properly keep track of I/O port resources.
yokota [Sun, 19 Mar 2000 03:25:13 +0000 (03:25 +0000)]
- Properly keep track of I/O port resources.
- Use bus_space_read/write() to access the ports.

24 years agoo Replace the old "oltr" driver with the completely re-written
lile [Sat, 18 Mar 2000 23:51:54 +0000 (23:51 +0000)]
o Replace the old "oltr" driver with the completely re-written
  new-bus Olicom driver, previously known as "ol".  The new
  driver unfortunately does not support ISA cards yet.

o Update the microcode files, interface library and include files
  to the latest PowerMACH works version.  Force even byte alignment
  of adapter microcode.

o Roll in some of the patches from Nikolai Saoukh <nms@ethereal.ru>.

24 years agoRemove the old DSC code, it doesn't work anymore.
sos [Sat, 18 Mar 2000 22:42:45 +0000 (22:42 +0000)]
Remove the old DSC code, it doesn't work anymore.

24 years agoExplicitly include sys/cdefs.h to get the definition of __strong_reference(),
jasone [Sat, 18 Mar 2000 22:36:46 +0000 (22:36 +0000)]
Explicitly include sys/cdefs.h to get the definition of __strong_reference(),
rather than getting lucky due to header dependencies.

24 years agoo Correct the offsets into the syncrate table for paritcular
gibbs [Sat, 18 Mar 2000 22:28:20 +0000 (22:28 +0000)]
o Correct the offsets into the syncrate table for paritcular
  negotiation features (DT, ULTRA2, ULTRA, FAST).  The offsets
  where not properly updated when the DT entry was added and so
  the driver could attempt to negotiate a speed faster than that
  supported by the target device or even requested by the user
  via SCSI-Select settings. *

o Update the target mode incoming command queue kernel index value
  ever 128 commands instead of 32.  This means that the kernel will
  always try to keep its index (as seen on the card - the kernel may
  actually have cleared more space) 128 commands ahead of where the
  sequencer is adding entries.

o Use the HS_MAILBOX register instead of the KERNEL_TQINPOS location
  in SRAM to indicate the kernel's target queue possition on Ultra2
  cards.  This avoids the "pause bug" on these cards and also turns
  out to be much more efficient.

o When enabling or disabling a particular target id for target mode,
  make sure that the taret id in the SCSIID register does not
  reference an ID that is not to receive target selections.  This
  is only an issue on chips that support the multiple target id
  feature where the value in SCSIID will still affect selection
  behavior regardless of the values in the target id bit field
  registers.

o Remove some target mode debugging printfs.

o Make sure that the sense length reported in ATIO commands is
  always zero.  This driver does not, yet, report HBA generated
  sense information for accepted commands.

o Honor the CAM_TIME_INFINITY and CAM_TIME_DEFAULT values for
  the CCB timeout field.

o Make the driver compile with AHC_DEBUG again.

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

24 years agoUse __alias__ instead of alias in the definition of __strong_reference().
jasone [Sat, 18 Mar 2000 22:26:54 +0000 (22:26 +0000)]
Use __alias__ instead of alias in the definition of __strong_reference().

Submitted by: jdp

24 years agoKNR -> ANSI function definition in two places. This matches the rest of
gibbs [Sat, 18 Mar 2000 22:15:00 +0000 (22:15 +0000)]
KNR -> ANSI function definition in two places.  This matches the rest of
the code in this driver.

24 years agoo clear the ccb_flags field before sending an accept target I/O ccb to
gibbs [Sat, 18 Mar 2000 22:13:27 +0000 (22:13 +0000)]
o clear the ccb_flags field before sending an accept target I/O ccb to
  an HBA.  Garbage in this field confuses the driver in targdone().

o When completing a CCB on behalf of a user process, we need to
  *de-queue* the ccb from our pending ccb list, not queue it again.

o All continue target I/O operations need to have a timeout set.
  We use 5 seconds throughout this driver.

o Remove some logging printfs.

o During abort processing, remove ccbs that are on the pending queue
  from the pending queue, not the work queue.

24 years agoPer Justin's request- remove his name from Copyright.
mjacob [Sat, 18 Mar 2000 22:03:41 +0000 (22:03 +0000)]
Per Justin's request- remove his name from Copyright.

24 years agoAfter about two months I am the first to notice that SPLASERT.9 never
asmodai [Sat, 18 Mar 2000 21:51:16 +0000 (21:51 +0000)]
After about two months I am the first to notice that SPLASERT.9 never
got installed.

Also, correct the copyright to a more realistic date (1996->2000).
Fix some wordings and omissions.

24 years agoSigh, it would have been nice but I just can't make the rsaref
jkh [Sat, 18 Mar 2000 20:14:58 +0000 (20:14 +0000)]
Sigh, it would have been nice but I just can't make the rsaref
hack work until we figure out some better way of handling package
interaction.

24 years agoPut ASUSCOM_IPAC in the section where it belongs, namely i4b.
asmodai [Sat, 18 Mar 2000 19:30:58 +0000 (19:30 +0000)]
Put ASUSCOM_IPAC in the section where it belongs, namely i4b.

24 years agoPrint the PCI resources even if they are disabled. This shows more clearly
n_hibma [Sat, 18 Mar 2000 19:18:36 +0000 (19:18 +0000)]
Print the PCI resources even if they are disabled. This shows more clearly
when the BIOS is forgetful about initialising the USB controllers.

24 years agoAdd the entries for urio.
n_hibma [Sat, 18 Mar 2000 18:45:48 +0000 (18:45 +0000)]
Add the entries for urio.

24 years agoPut the undocumented options back at the bottom as per old
asmodai [Sat, 18 Mar 2000 18:39:02 +0000 (18:39 +0000)]
Put the undocumented options back at the bottom as per old
practice.

Place the debug options above the undocumented options.

24 years agoMake the Y-E Data PCMCIA floppy of the Toshiba Libretto work under
joerg [Sat, 18 Mar 2000 18:27:01 +0000 (18:27 +0000)]
Make the Y-E Data PCMCIA floppy of the Toshiba Libretto work under
-current.  It doesn't work yet as stable as the 3.x/PAO version of the
driver does, however, i get occasional `FDC direction bit not set' and
other weird messages, but it basically works at least.

The old (defunct) #ifdef FDC_YE stuff has been eliminated completely
now, PCMCIA-FDC specific functions have been implemented differently
where needed.

Unfortunately, due to the fact that the traditional PeeCee FDC with
its funny non-contiguous register space (one register for WD1003
harddisk controllers is interleaved into the FDC register set), and
Peter's subsequent changes involving two different bus space handles
for normal FDCs, the changes required for the Y-E stuff are more
complex than i'd love them to be.  I've done my best to keep the logic
for normal FDCs intact.

Since the Y-E FDC seems to lose interrupts after a FDC reset
sometimes, i've also replaced the timeout logic in fd_turnoff() to
generate an artificial pseudo interrupt in case of a timeout while the
drive has still outstanding transfers waiting.  This avoids the total
starvation of the driver that could be observed with highly damaged
media under 3.x/PAO.  This part of the patch has been revied by bde
previously.

I've fixed a number of occasions where previous commits have been
missing the encapuslation of ISA DMA related functions inside
FDC_NODMA checks.

I've added one call to SET_BCDR() during preparation of the format
floppy operation.  Floppy formatting has been totally broken before in
3.x/PAO (garbage ID fields have been written to the medium, causing
`wrong cylinder' errors upon media reading).  This is just black
magic, i don't have the slightes idea _why_ this needs to be but just
copied over the hack that has been used by the PAO folks in the normal
read/write case anyway.

The entired device_busy() stuff seems to be pointless to me.  In any
case, i had to add device_unbusy() calls symmetrical to the
device_busy() calls, otherwise the PCMCIA floppy driver could never be
deactivated.  (As it used to be, it caused a `mark the device busier
and busier' situation.)  IMHO, all block device drivers should be
marked busy based on active buffers still waiting for the driver, so
the device_unbusy() calls should probably go to biodone().  Only one
other driver (whose name escapes me at the moment) uses device_busy()
calls at all, so i question the value of all this...

I think this entire `device busy' logic simply doesn't fit for PCMCIA
&al.  It cannot be the decision of some piece of kernel software to
declare a device `busy by now, you can't remove it', when the actual
physical power of removing it is the user pulling the card.  The
kernel simply has to cope with the removal, however busy the device
might have been by the time of the removal, period.  Perhaps a force
flag needs to be added?

Upon inserting the card a second time, i get:

WARNING: "fd" is usurping "fd"'s cdevsw[]
WARNING: "fd" is usurping "fd"'s bmaj

I suspect this is related to the XXX comment at the call to
cdevsw_add().  Does anybody know what the correct way is to cleanup
this?

24 years agoReflect the fact that we supported the Y-E Data floppy as well.
n_hibma [Sat, 18 Mar 2000 17:59:27 +0000 (17:59 +0000)]
Reflect the fact that we supported the Y-E Data floppy as well.

24 years agoThe completely rewritten USB Mass Storage driver. It supports the following
n_hibma [Sat, 18 Mar 2000 17:33:41 +0000 (17:33 +0000)]
The completely rewritten USB Mass Storage driver. It supports the following
devices:

Tested:
   Zip 100 USB drive
   Y-E Data USB floppy drive

Untested: (let me know whether or not it works!)
   TEAC USB floppy drive
   Zip 250 USB drive

24 years agoDocument three debug options: npx, bus and vfs locks debugging.
asmodai [Sat, 18 Mar 2000 17:30:12 +0000 (17:30 +0000)]
Document three debug options: npx, bus and vfs locks debugging.

24 years agoActivate the pc98 directory.
kato [Sat, 18 Mar 2000 16:49:22 +0000 (16:49 +0000)]
Activate the pc98 directory.

24 years agoAdded PC98 version of fdisk.
kato [Sat, 18 Mar 2000 16:48:41 +0000 (16:48 +0000)]
Added PC98 version of fdisk.

24 years ago- `Newbus'ified the mse driver.
yokota [Sat, 18 Mar 2000 15:43:13 +0000 (15:43 +0000)]
- `Newbus'ified the mse driver.

24 years ago- Fix bugs when painting border and the mouse cursor in the raster
yokota [Sat, 18 Mar 2000 15:28:48 +0000 (15:28 +0000)]
- Fix bugs when painting border and the mouse cursor in the raster
  text mode.

24 years ago- Add support for IntelliMouse Explorer, NetScroll Optical, SmartScroll,
yokota [Sat, 18 Mar 2000 15:23:21 +0000 (15:23 +0000)]
- Add support for IntelliMouse Explorer, NetScroll Optical, SmartScroll,
  Scroll Point, and 4D/4D+ mice.
- Add a couple of serial mouse PnP IDs.
- Extend the `-z' option so that the second wheel (or the horizontal
  movement of the `scroll' device) can be mapped to buttons.

24 years ago- Add Support for the following PS/2 mice:
yokota [Sat, 18 Mar 2000 15:21:40 +0000 (15:21 +0000)]
- Add Support for the following PS/2 mice:
  - Microsoft IntelliMouse Explorer: 2 buttons on top, 2 side buttons
    and a wheel which also acts as the middle button.  The mouse is
    recognized as "IntelliMouse Explorer".
  - Genius NetScroll Optical: 2 buttons on top, 2 side buttons and a
    wheel which also acts as the middle button.  The mouse is recognized
    as "NetMouse/NetScroll Optical".
  - MouseSystems SmartScroll Mouse (OEM from Genius?): 3 buttons on top,
    1 side button and a wheel.  The mouse is recognized as Genius
    "NetScroll".
  - IBM ScrollPoint: 2 buttons on top and a stick between the buttons.
    The stick can perform "horizontal scroll" in W*ndows environment.
    The horizontal movement of the stick is detected.  It is currently
    mapped to the Z axis movement in the same way as the first wheel.
    The mouse is recognized as "MouseMan+", as it is considered to be
    a variation of MouseMan.
  - A4 Tech 4D and 4D+ mice. These mice have two wheels! The movement
    of the second wheel is reported as the Z axis movement in the
    same way as the first wheel. These mice are recognized as "4D
    Mouse" and "4D+ Mouse".
  - Tweak IntelliMouse support code a bit so that less-than-compatible
    wheel mice can work properly with the psm driver.
- Add driver configuration flags which correspond to the kernel
  options PSM_HOOKRESUME and PSM_RESETAFTERSUSPEND, so that we don't
  need to recompile the kernel when we need these functions.
- Properly keep track of the irq resource.
- Add a watchdog timer in case interrupts are lost (experimental).
- Add `detach' function (experimental).

24 years ago- `Newbus'ified the driver.
yokota [Sat, 18 Mar 2000 15:13:30 +0000 (15:13 +0000)]
- `Newbus'ified the driver.
- Properly keep track of resources (I/O ports and irq).
- Use bus_space_read/write() to access the ports.
- Add PnP IDs.
- Add a watchdog timer in case interrupts are lost (experimental).
- Add `detach' function (experimental).

24 years agoGet rid of the last remnants of powerhook abuse.
n_hibma [Sat, 18 Mar 2000 14:54:54 +0000 (14:54 +0000)]
Get rid of the last remnants of powerhook abuse.

24 years agoMake sure to free the socket in soabort() if the protocol couldn't
fenner [Sat, 18 Mar 2000 08:56:56 +0000 (08:56 +0000)]
Make sure to free the socket in soabort() if the protocol couldn't
 free it (this could happen if the protocol already freed its part
 and we just kept the socket around to make sure accept(2) didn't block)

24 years agoChange the output of truss to more closely resemble SysV's. (Yes, it
sef [Sat, 18 Mar 2000 08:49:41 +0000 (08:49 +0000)]
Change the output of truss to more closely resemble SysV's.  (Yes, it
really is much nicer looking.)

Submitted by: "Matthew N. Dodd" <winter@jurai.net>

24 years agoWe still need a hack for rsaref package; add it. This doesn't
jkh [Sat, 18 Mar 2000 08:46:23 +0000 (08:46 +0000)]
We still need a hack for rsaref package; add it.  This doesn't
affect CDs since they don't contain the rsaref package anyway.

24 years agoSigh, when the DIST_LOCAL thing was added, it unfortunately wasn't added
jkh [Sat, 18 Mar 2000 08:40:39 +0000 (08:40 +0000)]
Sigh, when the DIST_LOCAL thing was added, it unfortunately wasn't added
completely.  Take out another gratuitous warning which is omitted when
All is selected and local dist does not exist.

24 years agoUpdate to match current reality (not that anyone reads these anyway :)
jkh [Sat, 18 Mar 2000 08:06:16 +0000 (08:06 +0000)]
Update to match current reality (not that anyone reads these anyway :)

24 years agoAlpha 8200: redo files arrangement for 8200 support. There sort of
mjacob [Sat, 18 Mar 2000 08:05:27 +0000 (08:05 +0000)]
Alpha 8200: redo files arrangement for 8200 support. There sort of
still *is* a problem about how to get an mcclock attached w/o an
ISA declaration though.

24 years agoAlpha 8200: remove the SIMOS usage of zs_cnattach- we have a real
mjacob [Sat, 18 Mar 2000 08:04:23 +0000 (08:04 +0000)]
Alpha 8200: remove the SIMOS usage of zs_cnattach- we have a real
zs_cnattach now.

24 years agoAlpha 8200: Redo console attachment to be more 'normal'. Make sure the
mjacob [Sat, 18 Mar 2000 08:03:48 +0000 (08:03 +0000)]
Alpha 8200: Redo console attachment to be more 'normal'. Make sure the
halt function *stays* halted.

24 years agoAlpha 8200: redo prom console code to be more up to date, etc. This
mjacob [Sat, 18 Mar 2000 08:02:35 +0000 (08:02 +0000)]
Alpha 8200: redo prom console code to be more up to date, etc. This
is now a functional PROM console (previously broken). Do the correct
promcnatach && CONS_DRIVER dance.

24 years agoAlpha 8200: add DEC_KN8AE (TurboLaser) platform option.
mjacob [Sat, 18 Mar 2000 08:01:34 +0000 (08:01 +0000)]
Alpha 8200: add DEC_KN8AE (TurboLaser) platform option.

24 years agoAlpha 8200: Add prototype from promcnattach.
mjacob [Sat, 18 Mar 2000 08:00:51 +0000 (08:00 +0000)]
Alpha 8200: Add prototype from promcnattach.

24 years agoAlpha 8200: Add in a dummy child of dwlpx so that we can make the leap
mjacob [Sat, 18 Mar 2000 08:00:13 +0000 (08:00 +0000)]
Alpha 8200: Add in a dummy child of dwlpx so that we can make the leap
from DWLPX to PCI space. Just a methods holder such that we have a parent
which is a "pcib" and we create a child which is a "pci". Add the appropriate
ivar code (which is for a hose #).

24 years agoAlpha 8200: Do some compilation cleanup. Conditionalize some IDE stuff
mjacob [Sat, 18 Mar 2000 07:57:58 +0000 (07:57 +0000)]
Alpha 8200: Do some compilation cleanup. Conditionalize some IDE stuff
based upon presence/absence of ISA (there is no ISA bus on an 8200- okay,
well, there *could* be one in a DWLPX tray, but we don't support it)).

Most importantly change the interrupt resource map to cover a whole 16
bits. The 8200 uses 16 bit interrupt vectors which we construct that
contain the I/O-board, hose, an pci slot in them, and then we write these
vectors into the appropriate DWLPX registers. At any rate, a flat array
of 64 'IRQs' isn't enough.

24 years agoAlpha 8200: Reinsert licence from NetBSD that should have been there
mjacob [Sat, 18 Mar 2000 07:54:11 +0000 (07:54 +0000)]
Alpha 8200: Reinsert licence from NetBSD that should have been there
to begin with. Redo newbus attachment code so that all the DMA mapping
and further pci attachment is done right. Insert config space functions
(jeez- how do you do type 1 cycles?). Do the interrupt setups, etc.
Basically, this is the core I/O module for 8200s, even though logically
it's the 3rd level down from the nominal principle backplane bus
(turbolaser). Still to be done here: S/G code isn't done yet, so we
better live with 2GB or less primary memory.

24 years agoAlpha 8200: Remove clause 3 of licence. Clean up code slightly.
mjacob [Sat, 18 Mar 2000 07:50:58 +0000 (07:50 +0000)]
Alpha 8200: Remove clause 3 of licence. Clean up code slightly.

24 years agoAlpha 8200: Remove clause 3 of licence. Redo KFTXX bus attachment
mjacob [Sat, 18 Mar 2000 07:49:53 +0000 (07:49 +0000)]
Alpha 8200: Remove clause 3 of licence. Redo KFTXX bus attachment
code (merge in progress made in NetBSD since the initial import to
FreeBSD). Create dwlpx as the child device. Of course, if we had
more h/w and time, we could find out whether the child device was
a FutureBus module, etc...Anyone ever actually seen one?

24 years agoAlpha 8200: Remove clause 3 of licence and and some KFTXX defines.
mjacob [Sat, 18 Mar 2000 07:48:15 +0000 (07:48 +0000)]
Alpha 8200: Remove clause 3 of licence and and some KFTXX defines.

24 years agoAlpha 8200: minor formatting tweaks; removal of clause 3 of licence.
mjacob [Sat, 18 Mar 2000 07:46:34 +0000 (07:46 +0000)]
Alpha 8200: minor formatting tweaks; removal of clause 3 of licence.

24 years agoAlpha 8200: Rework to get working properly with real h/w. This isn't a
mjacob [Sat, 18 Mar 2000 07:46:05 +0000 (07:46 +0000)]
Alpha 8200: Rework to get working properly with real h/w. This isn't a
clean MI/MD driver, but it *does* actually work at this time. Updated
to use new make_dev stuff. A CONS_DRIVER declaration is also put in
so that this can be the real console for the 8200s.

24 years agoAlpha 8200: Ooops on last delta- real changes include redoing bus attachment
mjacob [Sat, 18 Mar 2000 07:44:18 +0000 (07:44 +0000)]
Alpha 8200: Ooops on last delta- real changes include redoing bus attachment
code and interrupt mapping code.

24 years agoAlpha 8200: Some minor formatting tweaks and removal of clause 3 of licence.
mjacob [Sat, 18 Mar 2000 07:42:13 +0000 (07:42 +0000)]
Alpha 8200: Some minor formatting tweaks and removal of clause 3 of licence.

24 years agoAlpha 8200: add a dummy driver for TLSB CPU nodes. Not so dummy, actually,
mjacob [Sat, 18 Mar 2000 07:40:17 +0000 (07:40 +0000)]
Alpha 8200: add a dummy driver for TLSB CPU nodes. Not so dummy, actually,
as you have to hook a GBus into it.

24 years agoAlpha 8200: Add in variables and defines for DWLPX PCI controllers.
mjacob [Sat, 18 Mar 2000 07:30:06 +0000 (07:30 +0000)]
Alpha 8200: Add in variables and defines for DWLPX PCI controllers.