branches: 1.1.4; 1.1.6; 1.1.8; 1.1.12; 1.1.14;
Don't be lazy and let one directory's Makefile install an include
file from another directory, as that only causes problems with
mtree later on when encountering the /.. path component. Instead,
descend into dev/ and use the Makefile there to install dev/crmfbreg.h
in the appropriate directory.

Following the example of the hpc/, gio/, and ioc/ directories, move the
mace devices to their own mace/ directory. Alter conf/files.sgimips to
reflect this change in a sane manner (i.e., pull in dev/files.dev and
mace/files.mace when appropriate).

At the same time, allow crime_intr_establish() to fall through to
mace_intr_establish(). mace devices now call cpu_intr_establish().

www.netbsd.org -> www.NetBSD.org

Apply a bunch of patches written by Christopher SEKIYA
(with several cosmetic changes by me) which fixes O2 (IP32) support.

Now my R5000 O2 works fine in multiuser with on-board AIC7880 SCSIs
and several PCI network cards (but only on serial console yet).
L2 cache on R5000/Rm5200 is still disabled for now, but it will be
fixed later, hopefully.

See recent discussion on port-sgimips for details.

__KERNEL_RCSID()

branches: 1.8.2;
Improve the interrupt code somewhat by having callers of xxx_intr_establish
pass in an interrupt handle (which is currently to the CRIME interrupt the
device is attached to) so the interrupt handlers know which device was the
one looking for attention.

While here, fix up PCI interrupt routing for both the on-board devices and
the PCI slots -- even though there is only one PCI slot in the chasis, the
hardware can accomodate up to three and provides an interrupt mapping for
all the PCI interrupt pins for both the internal SCSI & PCI slot and the
two "extra" slots.

Checkpoint of O2 work by Chris Sekiya and myself. This is the sgimips bit;
still needs some arch/mips support code before it will fully work.

Add trailing ; to CFATTACH_DECL.

Use CFATTACH_DECL().

Declare all cfattach structures const.

Replace lots of 8x<space> with <tabs> and other miscellaneous indentation
fixes.
Wrap a couple of long lines.
Use <return-type>\n<function name> as per KNF in a few places.

branches: 1.2.4;
read dbaud from ARCS and use that for com rate, not hardcoded 38400

branches: 1.1.4; 1.1.6; 1.1.8; 1.1.10;
A start at O2 support..

if ddb is enabled set cpu_reset_address to crime_reboot.
'mach reset' works from ddb now.

branches: 1.38.34;
switch sgimips to common bus_dma and bus_space in arch/mips/
Tested on O2 and Indy.
Things that are compile-tested only for lack of hardware:
- GIO ethernet cards with PCI bridges
- IP2x hardware not found on Indy
- IP1x

branches: 1.37.4;
split crime_bus_reset() into crime_{cpu,mem}_reset() and only reset
the cpu or the memory upon either error, not both.

seems to make my O2 slightly less likely to hang, and matches what
the linux crime driver does.

branches: 1.36.12; 1.36.16; 1.36.26;
device_t-ify

#include <sys/bus.h> instead of <machine/bus.h>.

Merge forward matt-nb5-mips64
Adapt to new interrupt/spl framework
Deal with arcbios calls being O32 and kernels being N32/N64.
Need to save/restore T8 across arcbios calls.

branches: 1.33.12; 1.33.16; 1.33.22; 1.33.24;
Misc cosmetics:
- some KNF and ANSIfy
- use uintNN_t types
- remove a redundant declaration
- use __func__ to print a function name
- use PRIu64 to print uint64_t
- wrap long lines

In crime_attach(), call crime_watchdog_disable() rather than
writing zero to the CRIME registers implicitly.

Fix a typo so that the crimedog is actually shut up.

- replace some mysterious magic numbers in CPU clock measurement code
with some more meaningful expressions
- compare counter values in 100ms rather than 2ms
- check CPU_MIPS_DOUBLE_COUNT in mips_cpu_flags explicitly

branches: 1.29.4;
Remove all initialization of obsolete ci_divisor_recip in
mips struct cpu_info and related macroes.
The member was prepared for a hack in MD microtime(9) implementation
but it has been superseded by MI timecounter(9).

branches: 1.28.6; 1.28.8; 1.28.10; 1.28.12;
Fix typo in macro name and comments.

branches: 1.27.2; 1.27.8;
Merge the ppcoea-renovation branch to HEAD.

This branch was a major cleanup and rototill of many of the various OEA
cpu based PPC ports that focused on sharing as much code as possible
between the various ports to eliminate near-identical copies of files in
every tree. Additionally there is a new PIC system that unifies the
interface to interrupt code for all different OEA ppc arches. The work
for this branch was done by a variety of people, too long to list here.

TODO:
bebox still needs work to complete the transition to -renovation.
ofppc still needs a bunch of work, which I will be looking at.
ev64260 still needs to be renovated
amigappc was not attempted.

NOTES:
pmppc was removed as an arch, and moved to a evbppc target.

add crime_reboot()

branches: 1.25.8; 1.25.10; 1.25.12;
merge yamt-idlelwp branch. asked by core@. some ports still needs work.

from doc/BRANCHES:

idle lwp, and some changes depending on it.

1. separate context switching and thread scheduling.
(cf. gmcgarry_ctxsw)
2. implement idle lwp.
3. clean up related MD/MI interfaces.
4. make scheduler(s) modular.

branches: 1.24.2;
Fix crime_intr so intr handlers registers w/ crime_intr_establish actually
get called. "seems more right to me" mrg@

Disable the CRIME watchdog for now; looks like we can't keep up with it
under load. With this patch, my O2 is finally running stable.

branches: 1.22.26; 1.22.30; 1.22.32;
Reset and disable watchdog explicitly in crime_attach().

TAB/space cleanup.

branches: 1.20.2;
Avoid shadow warnings

Use the mace interrupt handler for PCI interrupts.

From KIYOHARA Takashi

Following the example of the hpc/, gio/, and ioc/ directories, move the
mace devices to their own mace/ directory. Alter conf/files.sgimips to
reflect this change in a sane manner (i.e., pull in dev/files.dev and
mace/files.mace when appropriate).

At the same time, allow crime_intr_establish() to fall through to
mace_intr_establish(). mace devices now call cpu_intr_establish().

Finish moving code between ip2x.c and ip3x.c to imc.c and crime.c.

Assign memory controller-dependent watchdog reset functions to
platform.watchdog_reset.

CRIME revision 1.1 has a bug that affects PIO operations to/from the MACE.
The recommended workaround is a 5-10ms delay before and after accesses.
Therefore, move the affected bus_space_* operations from bus.h to bus.c
and special-case MACE accesses.

CRIME accesses are not affected, so introduce SGIMIPS_BUS_SPACE_CRIME and
use it as the CRIME tag.

My ip32 seems a little bit happier with this change, and my ip22 didn't
notice the change.

www.netbsd.org -> www.NetBSD.org

Apply a bunch of patches written by Christopher SEKIYA
(with several cosmetic changes by me) which fixes O2 (IP32) support.

Now my R5000 O2 works fine in multiuser with on-board AIC7880 SCSIs
and several PCI network cards (but only on serial console yet).
L2 cache on R5000/Rm5200 is still disabled for now, but it will be
fixed later, hopefully.

See recent discussion on port-sgimips for details.

__KERNEL_RCSID()

branches: 1.11.2;
Simplify CRIME rev. calculation (we don't appear to need the low nibble).

Improve the interrupt code somewhat by having callers of xxx_intr_establish
pass in an interrupt handle (which is currently to the CRIME interrupt the
device is attached to) so the interrupt handlers know which device was the
one looking for attention.

While here, fix up PCI interrupt routing for both the on-board devices and
the PCI slots -- even though there is only one PCI slot in the chasis, the
hardware can accomodate up to three and provides an interrupt mapping for
all the PCI interrupt pins for both the internal SCSI & PCI slot and the
two "extra" slots.

Checkpoint of O2 work by Chris Sekiya and myself. This is the sgimips bit;
still needs some arch/mips support code before it will fully work.

Add trailing ; to CFATTACH_DECL.

Use CFATTACH_DECL().

Declare all cfattach structures const.

Replace lots of 8x<space> with <tabs> and other miscellaneous indentation
fixes.
Wrap a couple of long lines.
Use <return-type>\n<function name> as per KNF in a few places.

branches: 1.4.2; 1.4.8;
- Use the identifier strings that are fetched by the MI ARC BIOS
code.
- Garbage-collect some non-useful stuff from the mainbus_attach_args.

Get the major number of the CRIME properly.

From Rafal K. Boni.

branches: 1.2.2; 1.2.4;
Move a few things from ip32.c.

branches: 1.1.4;
A start at O2 support..

s/endianess/endianness/

merge ktrace-lwp.

www.netbsd.org -> www.NetBSD.org

Apply a bunch of patches written by Christopher SEKIYA
(with several cosmetic changes by me) which fixes O2 (IP32) support.

Now my R5000 O2 works fine in multiuser with on-board AIC7880 SCSIs
and several PCI network cards (but only on serial console yet).
L2 cache on R5000/Rm5200 is still disabled for now, but it will be
fixed later, hopefully.

See recent discussion on port-sgimips for details.

TAB/space cleanup.

branches: 1.6.2;
Simplify CRIME rev. calculation (we don't appear to need the low nibble).

add timer calibration

from Chris Sekiya

CRIME register definitions

from Chris Sekiya

Replace lots of 8x<space> with <tabs> and other miscellaneous indentation
fixes.
Wrap a couple of long lines.
Use <return-type>\n<function name> as per KNF in a few places.

branches: 1.2.2; 1.2.6; 1.2.10;
Some CRIME definitions.

branches: 1.1.4;
A start at O2 support..

device_t-ify

#include <sys/bus.h> instead of <machine/bus.h>.

Merge the ppcoea-renovation branch to HEAD.

This branch was a major cleanup and rototill of many of the various OEA
cpu based PPC ports that focused on sharing as much code as possible
between the various ports to eliminate near-identical copies of files in
every tree. Additionally there is a new PIC system that unifies the
interface to interrupt code for all different OEA ppc arches. The work
for this branch was done by a variety of people, too long to list here.

TODO:
bebox still needs work to complete the transition to -renovation.
ofppc still needs a bunch of work, which I will be looking at.
ev64260 still needs to be renovated
amigappc was not attempted.

NOTES:
pmppc was removed as an arch, and moved to a evbppc target.

add crime_reboot()

branches: 1.5.30; 1.5.38; 1.5.48; 1.5.50; 1.5.52;
merge ktrace-lwp.

branches: 1.4.12;
Use the mace interrupt handler for PCI interrupts.

From KIYOHARA Takashi

branches: 1.3.4;
Finish moving code between ip2x.c and ip3x.c to imc.c and crime.c.

www.netbsd.org -> www.NetBSD.org

Apply a bunch of patches written by Christopher SEKIYA
(with several cosmetic changes by me) which fixes O2 (IP32) support.

Now my R5000 O2 works fine in multiuser with on-board AIC7880 SCSIs
and several PCI network cards (but only on serial console yet).
L2 cache on R5000/Rm5200 is still disabled for now, but it will be
fixed later, hopefully.

See recent discussion on port-sgimips for details.

Remove unnecessary <sys/malloc.h>.

Merge thorpej-cfargs2.

branches: 1.48.8;
Merge thorpej-cfargs branch:

Simplify and make extensible the config_search() / config_found() /
config_attach() interfaces: rather than having different variants for
which arguments you want pass along, just have a single call that
takes a variadic list of tag-value arguments.

Adjust all call sites:
- Simplify wherever possible; don't pass along arguments that aren't
actually needed.
- Don't be explicit about what interface attribute is attaching if
the device only has one. (More simplification.)
- Add a config_probe() function to be used in indirect configuiration
situations, making is visibly easier to see when indirect config is
in play, and allowing for future change in semantics. (As of now,
this is just a wrapper around config_match(), but that is an
implementation detail.)

Remove unnecessary or redundant interface attributes where they're not
needed.

There are currently 5 "cfargs" defined:
- CFARG_SUBMATCH (submatch function for direct config)
- CFARG_SEARCH (search function for indirect config)
- CFARG_IATTR (interface attribte)
- CFARG_LOCATORS (locators array)
- CFARG_DEVHANDLE (devhandle_t - wraps OFW, ACPI, etc. handles)

...and a sentinel value CFARG_EOL.

Add some extra sanity checking to ensure that interface attributes
aren't ambiguous.

Use CFARG_DEVHANDLE in MI FDT, OFW, and ACPI code, and macppc and shark
ports to associate those device handles with device_t instance. This
will trickle trough to more places over time (need back-end for pre-OFW
Sun OBP; any others?).

branches: 1.47.10;
s/suport/support/

Cleanup i2c bus acquire / release, centralizing all of the logic into
iic_acquire_bus() / iic_release_bus(). "acquire" and "release" hooks
no longer need to be provided by back-end controller drivers (only if
they need special handling, e.g. powering on the i2c controller).
This results in the removal of a bunch of rendundant code from each
back-end controller driver.

Assert that we are not in hard interrupt context in iic_acquire_bus(),
iic_exec(), and iic_release_bus().

branches: 1.45.2;
Add support for overriding the video mode by setting an ARCS environment
variable.

If the "crmfb_mode" environment variable is set, treat it as a mode
string in the form WIDTHxHEIGHT[@REFRESH] and generate monitor timings
using the VESA GTF formula. If set, this overrides the EDID preferred
mode.

branches: 1.44.2; 1.44.8;
enable font loading / screen resizing

Remove extra stray break missed by Christos when committing PR 50783.

PR/50783: David Binderman: Indent switch properly, add missing break.

no need to set RI_CLEAR - we do our own clearing and rasops assumes a linear
framebuffer anyway. Now we can get away with leaving ri_bits at NULL where it
belongs on this hardware.

switch sgimips to common bus_dma and bus_space in arch/mips/
Tested on O2 and Indy.
Things that are compile-tested only for lack of hardware:
- GIO ethernet cards with PCI bridges
- IP2x hardware not found on Indy
- IP1x

- remove scratch area, it's been unused since we're drawing characters by
hardware
- don't sync the rendering engine unless we have to, watch FIFO levels
instead
- support anti-aliased fonts and let the drawing engine do all the work

branches: 1.38.2;
support ioctl() WSDISPLAYIO_SVIDEO, _GVIDEO and _GET_EDID
now xf86-video-crime can actually turn the monitor off

- remove unused variables
- use (void) when we don't care about the return value

branches: 1.36.6; 1.36.10;
use rasops_init(0,0)

branches: 1.35.2; 1.35.6;
#include <sys/bus.h> instead of <machine/bus.h>.

add support for running the console in more than 8 bit
Now modes like 1680x1050 work. The hardware requires a stride of a multiple
of 32 bytes, with 1680 this requires 16bit colour.

use PRIx64 for printing 64-bit values

add mode setting support
If we find a monitor via DDC we switch to its preferred mode if it reports one,
otherwise stick with whatever the firmware set up.

remove some old debug code

add DDC2 support. Not too useful yet.

device_t-ify, use aprint_*_dev() where appropriate

use hardware to draw characters, while there use vcons_replay_msgbuf()

Merge forward matt-nb5-mips64
Adapt to new interrupt/spl framework
Deal with arcbios calls being O32 and kernels being N32/N64.
Need to save/restore T8 across arcbios calls.

branches: 1.26.16; 1.26.22; 1.26.24;
Add a BUS_DMASYNC_PREWRITE call against tile's DMA address descriptors.
Fixes mangled screen on recent kernels.
Tested on my R5K IP32 and on R10K IP32 by Jorge Acereda Macia on port-sgimips.

branches: 1.25.2; 1.25.4;
Use 2-clause license.

branches: 1.24.2; 1.24.4; 1.24.6;
Remove debug printf: ``so we need 10 x 8 tiles -> 00500000''

Swap red and blue for 8bpp paletted mode.

allow userland to map all video memory, not just the visible part

allocate a linear buffer, don't smoke crack when setting up the engine's TLBs

branches: 1.20.2; 1.20.6;
correct accidentially committed garbage (what the hell... )

correct typo

avoid unnecessary writes to the MTE's direction registers

use MTE for rectangle filling as well
TODO: figure out why the MTE freaks out when used for character drawing

adjust some register names, also use MTE for scrolling down

use the memory transfer engine for scrolling when possible
currently that's only for scrolling upwards but the speedup should be very
noticeable
TODO: use MTE for erase rasops, figure out how to control the MTEs direction

bunch of minor fixes:
- use default colour when clearing the screen on attach
- sprinkle some break; so we no longer panic in 32bit colour
- allow userland to mmap() engine registers but not TLBs
- get rid of a delay() in crmfb_wait_idle()
- use aprint_*()

get rid of some leftover debugging cruft

update (c) sections

add hardware acceleration for scrolling and clearing.
The framebuffer is now run in tiled mode so we can use the rendering engine
which means X is broken for now.
This needs more testing.

Allow sync-on-green to be controlled by ARCS environment variable "SyncOnGreen".

Set to anything starting with the letter 'n' to disable sync-on-green, set to
anything else or leave undefined to keep default behavior.

CRMFB_VT_VCMAP and CRMFB_VT_HCMAP register definitions were transposed.
From jsing@openbsd in private mail.

branches: 1.8.2;
Merge the ppcoea-renovation branch to HEAD.

This branch was a major cleanup and rototill of many of the various OEA
cpu based PPC ports that focused on sharing as much code as possible
between the various ports to eliminate near-identical copies of files in
every tree. Additionally there is a new PIC system that unifies the
interface to interrupt code for all different OEA ppc arches. The work
for this branch was done by a variety of people, too long to list here.

TODO:
bebox still needs work to complete the transition to -renovation.
ofppc still needs a bunch of work, which I will be looking at.
ev64260 still needs to be renovated
amigappc was not attempted.

NOTES:
pmppc was removed as an arch, and moved to a evbppc target.

branches: 1.7.4; 1.7.6; 1.7.8;
- move register definitions to crmfbreg.h
- identify as WSDISPLAY_TYPE_CRIME
- always claim we're 32bit
- switch to 32bit colour in WSDISPLAYIO_MODE_MAPPED, switch back to 8bit
in WSDISPLAYIO_MODE_EMUL
- restore palette, turn off the cursor when switching back to console mode
- add a few delay()s, hopefully that will deal with occasional messed up
video timings *fingers crossed*

branches: 1.6.2; 1.6.4;
add hardware cursor support

branches: 1.5.2; 1.5.4; 1.5.6;
We don't support WSDISPLAYIO_[SG]VIDEO, so don't try to hide it.

branches: 1.4.2;
Avoid crashing if booting via serial console

Add shadow framebuffer support; scrolling the screen still isn't great,
but it's noticeably better now than it was before.

Don't claim to be the console if we're supposed to be using a serial
console. While we're here, save a few MB of RAM by properly calculating
the framebuffer size based on the selected video mode.

Add crmfb, a framebuffer console driver for the onboard O2 video (SGI-CRM)

- remove scratch area, it's been unused since we're drawing characters by
hardware
- don't sync the rendering engine unless we have to, watch FIFO levels
instead
- support anti-aliased fonts and let the drawing engine do all the work

branches: 1.15.14; 1.15.32;
add some constants, clarify some hw limitations

add mode setting support
If we find a monitor via DDC we switch to its preferred mode if it reports one,
otherwise stick with whatever the firmware set up.

add DDC2 support. Not too useful yet.

add a bunch of GBE registers, no functional change (yet)

use hardware to draw characters, while there use vcons_replay_msgbuf()

branches: 1.10.4; 1.10.6; 1.10.8;
add a few more register definitions, needed by the latest crime_drv.

branches: 1.9.4;
add a couple more registers

branches: 1.8.4; 1.8.6; 1.8.8;
Use 2-clause license.

branches: 1.7.6; 1.7.8; 1.7.10; 1.7.12;
adjust some register names, also use MTE for scrolling down

use the memory transfer engine for scrolling when possible
currently that's only for scrolling upwards but the speedup should be very
noticeable
TODO: use MTE for erase rasops, figure out how to control the MTEs direction

update (c) sections

add hardware acceleration for scrolling and clearing.
The framebuffer is now run in tiled mode so we can use the rendering engine
which means X is broken for now.
This needs more testing.

Define relevent register data so that we can turn off sync-on-green when needed.

CRMFB_VT_VCMAP and CRMFB_VT_HCMAP register definitions were transposed.
From jsing@openbsd in private mail.

branches: 1.1.4; 1.1.6; 1.1.8; 1.1.10; 1.1.12; 1.1.14; 1.1.20;
- move register definitions to crmfbreg.h
- identify as WSDISPLAY_TYPE_CRIME
- always claim we're 32bit
- switch to 32bit colour in WSDISPLAYIO_MODE_MAPPED, switch back to 8bit
in WSDISPLAYIO_MODE_EMUL
- restore palette, turn off the cursor when switching back to console mode
- add a few delay()s, hopefully that will deal with occasional messed up
video timings *fingers crossed*

switch sgimips to common bus_dma and bus_space in arch/mips/
Tested on O2 and Indy.
Things that are compile-tested only for lack of hardware:
- GIO ethernet cards with PCI bridges
- IP2x hardware not found on Indy
- IP1x

branches: 1.3.12; 1.3.30;
#include <sys/bus.h> instead of <machine/bus.h>.

Merge forward matt-nb5-mips64
Adapt to new interrupt/spl framework
Deal with arcbios calls being O32 and kernels being N32/N64.
Need to save/restore T8 across arcbios calls.

branches: 1.1.2; 1.1.4; 1.1.6; 1.1.10; 1.1.12; 1.1.14;
Check in work in progress of Personal IRIS 4D/20 and 4D/25 support (IP6
and IP10, respectively).

branches: 1.1.2; 1.1.4; 1.1.6;
Check in work in progress of Personal IRIS 4D/20 and 4D/25 support (IP6
and IP10, respectively).

Fix typo again... s/Comparision/Comparison/

s/comparsion/comparision/ in comment.

branches: 1.1.2; 1.1.4; 1.1.6;
Enable Personal IRIS 4D/20 and 4D/25 support:
- Adapt int(4) to handle the INT1 chip
- Move generic rtc clocks out of hpc/ and into dev/
- Handle the very strangely wired eeprom and other bits in arcemu
- Sprinkle MACH_SGI_IP6 as necessary
- Enable IP6/IP10 devices in GENERIC32_IP12. Yes, the naming is poor but
there's no winning with kernel/hw compatibility on sgimips...

Tested on my 4D/25. Doesn't (appear to) break macallan@'s IP22.

Change todr_chip_handle::cookie -> todr_chip_handle::todr_dev, and
make it a device_t. Upcoming functional changes will require the
device_t associated with a TODR device.

Change todr_chip_handle::bus_cookie -> todr_chip_handle::todr_devaux.
Nothing was using the old field, but I decided to keep it around just
in cause something needs it in the future.

And with these largely mechanical yet semantically meaningful changes,
thus spake the Oracle: "Welcome to NetBSD 11.99.2."

Remove unnecessary NULL-initialization of TODR handle fields.

Use todr_gettime_ymdhms / todr_settime_ymdhms.

branches: 1.6.18;
switch sgimips to common bus_dma and bus_space in arch/mips/
Tested on O2 and Indy.
Things that are compile-tested only for lack of hardware:
- GIO ethernet cards with PCI bridges
- IP2x hardware not found on Indy
- IP1x

branches: 1.5.2;
use the inline bcdtobin and bintobcd directly instead through a macro.

split device_t/softc for all remaining drivers.
replace "struct device *" with "device_t".
use device_xname(), device_unit(), etc.

branches: 1.3.2; 1.3.12;
#include <sys/bus.h> instead of <machine/bus.h>.

Remove `volatile' qualifier from argument types of
struct timeval passed to todr_gettime(9) and todr_settime(9).
We no longer have an ancient and volatile struct timeval `time'
global since we have switched to MI timercounter(9) on all port.

XXX1: some of these RTC drivers still assume 32bit time_t
XXX2: some of these should be rewritten to use todr_[gs]ettime_ymdhms()
XXX3: todr(9) man page doesn't mention todr_[gs]ettime_ymdhms()

branches: 1.1.2; 1.1.4; 1.1.6;
Enable Personal IRIS 4D/20 and 4D/25 support:
- Adapt int(4) to handle the INT1 chip
- Move generic rtc clocks out of hpc/ and into dev/
- Handle the very strangely wired eeprom and other bits in arcemu
- Sprinkle MACH_SGI_IP6 as necessary
- Enable IP6/IP10 devices in GENERIC32_IP12. Yes, the naming is poor but
there's no winning with kernel/hw compatibility on sgimips...

Tested on my 4D/25. Doesn't (appear to) break macallan@'s IP22.

Change todr_chip_handle::cookie -> todr_chip_handle::todr_dev, and
make it a device_t. Upcoming functional changes will require the
device_t associated with a TODR device.

Change todr_chip_handle::bus_cookie -> todr_chip_handle::todr_devaux.
Nothing was using the old field, but I decided to keep it around just
in cause something needs it in the future.

And with these largely mechanical yet semantically meaningful changes,
thus spake the Oracle: "Welcome to NetBSD 11.99.2."

Remove unnecessary NULL-initialization of TODR handle fields.

switch sgimips to common bus_dma and bus_space in arch/mips/
Tested on O2 and Indy.
Things that are compile-tested only for lack of hardware:
- GIO ethernet cards with PCI bridges
- IP2x hardware not found on Indy
- IP1x

branches: 1.6.2;
use the inline bcdtobin and bintobcd directly instead through a macro.

branches: 1.5.12;
#include <sys/bus.h> instead of <machine/bus.h>.

Split device_t/softc, some KNF.

Switch to using todr_[gs]ettime_ymdhms() format
which doesn't require struct timeval details.

Remove `volatile' qualifier from argument types of
struct timeval passed to todr_gettime(9) and todr_settime(9).
We no longer have an ancient and volatile struct timeval `time'
global since we have switched to MI timercounter(9) on all port.

XXX1: some of these RTC drivers still assume 32bit time_t
XXX2: some of these should be rewritten to use todr_[gs]ettime_ymdhms()
XXX3: todr(9) man page doesn't mention todr_[gs]ettime_ymdhms()

branches: 1.1.2; 1.1.4; 1.1.6;
Enable Personal IRIS 4D/20 and 4D/25 support:
- Adapt int(4) to handle the INT1 chip
- Move generic rtc clocks out of hpc/ and into dev/
- Handle the very strangely wired eeprom and other bits in arcemu
- Sprinkle MACH_SGI_IP6 as necessary
- Enable IP6/IP10 devices in GENERIC32_IP12. Yes, the naming is poor but
there's no winning with kernel/hw compatibility on sgimips...

Tested on my 4D/25. Doesn't (appear to) break macallan@'s IP22.

The crmfb device does not depend on iic; i2cbus and i2c_bitbang is
sufficient, says jmcneill@
Fixes a build problem for (at least) the GENERIC32_IP3x kernel.

crmfb may run in more than 8 bit now so we need the other rasops* as well

defflag CRMFB_DEBUG

crmfb needs edid and friends now

branches: 1.10.4; 1.10.6;
Enable Personal IRIS 4D/20 and 4D/25 support:
- Adapt int(4) to handle the INT1 chip
- Move generic rtc clocks out of hpc/ and into dev/
- Handle the very strangely wired eeprom and other bits in arcemu
- Sprinkle MACH_SGI_IP6 as necessary
- Enable IP6/IP10 devices in GENERIC32_IP12. Yes, the naming is poor but
there's no winning with kernel/hw compatibility on sgimips...

Tested on my 4D/25. Doesn't (appear to) break macallan@'s IP22.

branches: 1.9.20; 1.9.28; 1.9.34;
Merge the ppcoea-renovation branch to HEAD.

This branch was a major cleanup and rototill of many of the various OEA
cpu based PPC ports that focused on sharing as much code as possible
between the various ports to eliminate near-identical copies of files in
every tree. Additionally there is a new PIC system that unifies the
interface to interrupt code for all different OEA ppc arches. The work
for this branch was done by a variety of people, too long to list here.

TODO:
bebox still needs work to complete the transition to -renovation.
ofppc still needs a bunch of work, which I will be looking at.
ev64260 still needs to be renovated
amigappc was not attempted.

NOTES:
pmppc was removed as an arch, and moved to a evbppc target.

crime needs-flag

branches: 1.7.2; 1.7.10; 1.7.12; 1.7.14;
Add crmfb, a framebuffer console driver for the onboard O2 video (SGI-CRM)

branches: 1.6.8; 1.6.12; 1.6.14;
Add pic and imc-specific functions to alter the GIO bus arbiter
configuration. These are to be used exclusively by a higher interface in
gio/gio.c.

Also add functions to enable, disable and query the state of SysAd bus
parity detection since certain gio(4) cards need to be able to disable
it in order to work around their own buggy implementations.

branches: 1.5.4; 1.5.8; 1.5.18;
merge ktrace-lwp.

TAB/space cleanup.

branches: 1.3.2; 1.3.14;
Add support for keyboard and mouse attachments at zs*, used on IP12 and IP20.
Code written by Steve Rumble; committed with his blessing.

Using a glass TTY as a console on IP20 (IP12 untested) should now be possible.

Attach int0 at mainbus (precedent being the ip32 CRIME device, which handled
interrupts and timer calibration yet is also attached at mainbus).

Introduce the INDY_R4600_CACHE config option, which more accurately describes
the code enabled by this option.

This renders #ifdef IPxx completely obsolete, thus we theoretically can unify
the GENERIC files if a workable load address relocation scheme can be found.

Add files.dev and files.mace

Following the example of the hpc/, gio/, and ioc/ directories, move the
mace devices to their own mace/ directory. Alter conf/files.sgimips to
reflect this change in a sane manner (i.e., pull in dev/files.dev and
mace/files.mace when appropriate).

At the same time, allow crime_intr_establish() to fall through to
mace_intr_establish(). mace devices now call cpu_intr_establish().

Checkpoint if_mec work (just in case I get hit by a bus). This driver is still
not operational. The MII code mostly works, DMA data structures are defined
and allocated, but tx/rx logic is missing.

www.netbsd.org -> www.NetBSD.org

__KERNEL_RCSID()

branches: 1.10.2;
sync with mecreg change, ie. make this compile

Add trailing ; to CFATTACH_DECL.

Use CFATTACH_DECL().

Declare all cfattach structures const.

Replace lots of 8x<space> with <tabs> and other miscellaneous indentation
fixes.
Wrap a couple of long lines.
Use <return-type>\n<function name> as per KNF in a few places.

branches: 1.5.2; 1.5.6; 1.5.10;
Use ETHER_ADDR_LEN.

Move bpfattach()/bpfdetach() calls into ether_ifattach()/ether_ifdetach().

A few tweaks. Still not a real driver.

remove include of <vm/vm.h>. <vm/vm.h> -> <uvm/uvm_extern.h>

branches: 1.1.4;
A start at O2 support..

Following the example of the hpc/, gio/, and ioc/ directories, move the
mace devices to their own mace/ directory. Alter conf/files.sgimips to
reflect this change in a sane manner (i.e., pull in dev/files.dev and
mace/files.mace when appropriate).

At the same time, allow crime_intr_establish() to fall through to
mace_intr_establish(). mace devices now call cpu_intr_establish().

Checkpoint if_mec work (just in case I get hit by a bus). This driver is still
not operational. The MII code mostly works, DMA data structures are defined
and allocated, but tx/rx logic is missing.

www.netbsd.org -> www.NetBSD.org

branches: 1.3.2;
brush up situation with mec register definitions

from Chris Sekiya

Replace lots of 8x<space> with <tabs> and other miscellaneous indentation
fixes.
Wrap a couple of long lines.
Use <return-type>\n<function name> as per KNF in a few places.

branches: 1.1.4; 1.1.6; 1.1.10; 1.1.14;
A start at O2 support..

Use {,e}isabus_attach().

Merge thorpej-cfargs2.

branches: 1.36.8;
Merge thorpej-cfargs branch:

Simplify and make extensible the config_search() / config_found() /
config_attach() interfaces: rather than having different variants for
which arguments you want pass along, just have a single call that
takes a variadic list of tag-value arguments.

Adjust all call sites:
- Simplify wherever possible; don't pass along arguments that aren't
actually needed.
- Don't be explicit about what interface attribute is attaching if
the device only has one. (More simplification.)
- Add a config_probe() function to be used in indirect configuiration
situations, making is visibly easier to see when indirect config is
in play, and allowing for future change in semantics. (As of now,
this is just a wrapper around config_match(), but that is an
implementation detail.)

Remove unnecessary or redundant interface attributes where they're not
needed.

There are currently 5 "cfargs" defined:
- CFARG_SUBMATCH (submatch function for direct config)
- CFARG_SEARCH (search function for indirect config)
- CFARG_IATTR (interface attribte)
- CFARG_LOCATORS (locators array)
- CFARG_DEVHANDLE (devhandle_t - wraps OFW, ACPI, etc. handles)

...and a sentinel value CFARG_EOL.

Add some extra sanity checking to ensure that interface attributes
aren't ambiguous.

Use CFARG_DEVHANDLE in MI FDT, OFW, and ACPI code, and macppc and shark
ports to associate those device handles with device_t instance. This
will trickle trough to more places over time (need back-end for pre-OFW
Sun OBP; any others?).

branches: 1.35.12;
fix misleading error message

branches: 1.34.18;
switch sgimips to common bus_dma and bus_space in arch/mips/
Tested on O2 and Indy.
Things that are compile-tested only for lack of hardware:
- GIO ethernet cards with PCI bridges
- IP2x hardware not found on Indy
- IP1x

branches: 1.33.14;
split device_t/softc for all remaining drivers.
replace "struct device *" with "device_t".
use device_xname(), device_unit(), etc.

branches: 1.32.2; 1.32.12;
#include <sys/bus.h> instead of <machine/bus.h>.

dependant -> dependent

Merge forward matt-nb5-mips64
Adapt to new interrupt/spl framework
Deal with arcbios calls being O32 and kernels being N32/N64.
Need to save/restore T8 across arcbios calls.

branches: 1.29.12; 1.29.16; 1.29.22; 1.29.24;
KNF and misc cosmetics.

branches: 1.28.42; 1.28.46; 1.28.48; 1.28.52;
Support the sq(4) interface on the Challenge S's IOPLUS mezzanine. This
requires us to configure the IOPLUS to use the appropriate DMA channel,
depending on what other expansion cards may be installed. Further, we need
to do a bit more configuration of the sq(4) DMA engine, which apparently
was formerly being configured properly by the PROM. Use the same values,
which appear to work fine.

Also, extend the gio_arb_config interface as needed.

NB: We're currently setting the IOPLUS board to long burst, but it may be
better to use realtime. This will require some experimentation.

branches: 1.27.2;
Guiness -> Guinness.

branches: 1.26.2;
Add pic and imc-specific functions to alter the GIO bus arbiter
configuration. These are to be used exclusively by a higher interface in
gio/gio.c.

Also add functions to enable, disable and query the state of SysAd bus
parity detection since certain gio(4) cards need to be able to disable
it in order to work around their own buggy implementations.

branches: 1.25.4; 1.25.8; 1.25.18;
merge ktrace-lwp.

TAB/space cleanup.

branches: 1.23.12;
Various fixes:

* clean up EISA detection logic,
* really fix config_found_ia() usage (from drochner@),
* sprinkle a bit of KNF.

Oooops, iaa.iaa_name must be initialized.

Use config_found_ia() rather than config_found().

Implement several changes that get my E++ (HPC1.5 GIO ethernet
adapter) to work on IP24:
o Force HPC and GIO slot endianness.
o Allow slots one and two to be masters and configure for
realtime mode.

Also, while we're here move IP22 invariants into the IP22 case.

I suspect that IP20 will work as well since it exhibited the same
broken DMA that IP24 did before I flipped the master bit.

Properly detect the EISA bus for IP22-class machines.

sekiya-san's platform abstraction allows us to keep these
functions static.

Hook the IMC bus error handler into the ISR, otherwise bus error conditions
won't be properly dealt with.

branches: 1.16.2;
Keep chip revision and EISA information on same line as attach information.

Properly disable watchdog timer.

Finish moving code between ip2x.c and ip3x.c to imc.c and crime.c.

Assign memory controller-dependent watchdog reset functions to
platform.watchdog_reset.

As the platform-specific bus reset routine varies based on the memory
controller used, set it in the memory controller driver rather than the
platform init code.

Add watchdog and bus error handling functions (these features live in
IMC space, thus should be handled here rather than in ip2x.c)

Bus-space'ify the memory controller driver, and add IP20-specific bits.

__KERNEL_RCSID()

branches: 1.8.2;
Use aprint_normal() for cfprint routines.

Add trailing ; to CFATTACH_DECL.

Use CFATTACH_DECL().

Declare all cfattach structures const.

Replace lots of 8x<space> with <tabs> and other miscellaneous indentation
fixes.
Wrap a couple of long lines.
Use <return-type>\n<function name> as per KNF in a few places.

branches: 1.3.4;
Use the newly committed imcreg.h to get rid of the magic numbers polluting
this file; while I'm here, rework the GIO arbitrator parameter setup to
match more closely what is done by lunix on the same hardware... Finally,
on a tip from Wilbern Cobb, preserve the graphics-related bits of the
GIO arbitrator parameter register from the PROM-initialized settings so
ARCS console on the graphics display works (tested on both my Indy with
Newport and my Indigo2 with XL gfx).

branches: 1.2.2; 1.2.4;
- Use the identifier strings that are fetched by the MI ARC BIOS
code.
- Garbage-collect some non-useful stuff from the mainbus_attach_args.

branches: 1.1.2;
Support for the Indigo2/Indy IMC (this is that to which all other
devices are attached).

Improve spelling: regsiter -> register

This change is non-functional.

Approved by <riastradh>, <pgoyette>

branches: 1.4.122; 1.4.142;
merge ktrace-lwp.

Bus-space'ify the memory controller driver, and add IP20-specific bits.

branches: 1.2.12;
Replace lots of 8x<space> with <tabs> and other miscellaneous indentation
fixes.
Wrap a couple of long lines.
Use <return-type>\n<function name> as per KNF in a few places.

branches: 1.1.2; 1.1.4; 1.1.8;
Register defines for the SGI MC memory controller. Not 100% complete, but
has enough register/bit definitions to get rid of all the magic numbers that
currently pollute imc.c

branches: 1.1.2; 1.1.6; 1.1.8; 1.1.14;
Add pic and imc-specific functions to alter the GIO bus arbiter
configuration. These are to be used exclusively by a higher interface in
gio/gio.c.

Also add functions to enable, disable and query the state of SysAd bus
parity detection since certain gio(4) cards need to be able to disable
it in order to work around their own buggy implementations.

malloc(9) -> kmem(9)

branches: 1.31.2;
For struct timecounter, use C99 initializers.
Compile tested. No functional changes intended.

in many device attach paths, allocate memory with M_WAITOK instead of M_NOWAIT
and remove code to handle failures that can no longer happen.

add event counters for int0 at mainbus

branches: 1.28.16; 1.28.18;
switch sgimips to common bus_dma and bus_space in arch/mips/
Tested on O2 and Indy.
Things that are compile-tested only for lack of hardware:
- GIO ethernet cards with PCI bridges
- IP2x hardware not found on Indy
- IP1x

branches: 1.27.4;
actually get rid of actually unused variables, still works on my Indy
thanks mrg

sprinkle __unused()

branches: 1.25.10;
split device_t/softc for all remaining drivers.
replace "struct device *" with "device_t".
use device_xname(), device_unit(), etc.

branches: 1.24.2; 1.24.12;
#include <sys/bus.h> instead of <machine/bus.h>.

Set correct struct clockframe .intr value for hardclock(9).

Merge forward matt-nb5-mips64
Adapt to new interrupt/spl framework
Deal with arcbios calls being O32 and kernels being N32/N64.
Need to save/restore T8 across arcbios calls.

branches: 1.21.4; 1.21.6; 1.21.8;
Merge from matt-nb5-mips64
Merge mips-specific arch files.

Enable Personal IRIS 4D/20 and 4D/25 support:
- Adapt int(4) to handle the INT1 chip
- Move generic rtc clocks out of hpc/ and into dev/
- Handle the very strangely wired eeprom and other bits in arcemu
- Sprinkle MACH_SGI_IP6 as necessary
- Enable IP6/IP10 devices in GENERIC32_IP12. Yes, the naming is poor but
there's no winning with kernel/hw compatibility on sgimips...

Tested on my 4D/25. Doesn't (appear to) break macallan@'s IP22.

branches: 1.19.2; 1.19.8; 1.19.12;
KNF and misc cosmetics.

branches: 1.18.4;
Remove all initialization of obsolete ci_divisor_recip in
mips struct cpu_info and related macroes.
The member was prepared for a hack in MD microtime(9) implementation
but it has been superseded by MI timecounter(9).

branches: 1.17.6; 1.17.8; 1.17.10; 1.17.12;
Fix typo in macro name and comments.

branches: 1.16.24; 1.16.30; 1.16.36;
Change printing such that our 'timecounter:' message gets its own line on
IP12.

Move the mips1_clock_intr handling into int.c where it belongs.

Add i8254 timecounter support for use on IP12.

Now pings aren't 0.000 ms anymore.

Sprinkle static.

branches: 1.12.2;
Remove the 'FIFO Full' hack for interrupt vector 0.

It is unclear that this interrupt fails to latch as int2_wait_fifo()
depends on it doing so and appears to work. Furthermore, we were previously
unconditionally unmasking the interrupt 0 vector, which led to more
'unexpected interrupt' messages on the console, rather than avoiding them.

Approved by sekiya@.

branches: 1.11.4; 1.11.8; 1.11.18;
merge ktrace-lwp.

TAB/space cleanup.

branches: 1.9.2; 1.9.12; 1.9.14;
Fake a FIFO interrupt iff the FIFO interrupt is unmasked. Should eliminate
the "spurious interrupt" messages on machines without grtwo displays.

Add the following to the int driver:

* if the driver hasn't been properly initialized, spin for a bit in wait_fifo()
rather than try to access the local0 status register.
* allow interrupt sharing (from Steve Rumble; necessary for IP20 glass TTY
support).

Fix base address for i8254 calibration code. This has the effect of
us writing to the correct addresses and the thing actually working.

branches: 1.6.2;
KNF

replace terms ip1x and ip2x in comment

Sprinkle some #ifdef MIPS3 to be able to compile the port without
including MIPS3 code.

for upcoming IP12

Add int2_wait_fifo(), which will idle while awaiting a flag in the interrupt
controller to deassert. Needed for the more interesting features of various
SGI framebuffers.

opt_machtype.h no longer exists.

Attach int0 at mainbus (precedent being the ip32 CRIME device, which handled
interrupts and timer calibration yet is also attached at mainbus).

Introduce the INDY_R4600_CACHE config option, which more accurately describes
the code enabled by this option.

This renders #ifdef IPxx completely obsolete, thus we theoretically can unify
the GENERIC files if a workable load address relocation scheme can be found.

branches: 1.2.2; 1.2.4; 1.2.6;
Enable Personal IRIS 4D/20 and 4D/25 support:
- Adapt int(4) to handle the INT1 chip
- Move generic rtc clocks out of hpc/ and into dev/
- Handle the very strangely wired eeprom and other bits in arcemu
- Sprinkle MACH_SGI_IP6 as necessary
- Enable IP6/IP10 devices in GENERIC32_IP12. Yes, the naming is poor but
there's no winning with kernel/hw compatibility on sgimips...

Tested on my 4D/25. Doesn't (appear to) break macallan@'s IP22.

Check in work in progress of Personal IRIS 4D/20 and 4D/25 support (IP6
and IP10, respectively).

Enable Personal IRIS 4D/20 and 4D/25 support:
- Adapt int(4) to handle the INT1 chip
- Move generic rtc clocks out of hpc/ and into dev/
- Handle the very strangely wired eeprom and other bits in arcemu
- Sprinkle MACH_SGI_IP6 as necessary
- Enable IP6/IP10 devices in GENERIC32_IP12. Yes, the naming is poor but
there's no winning with kernel/hw compatibility on sgimips...

Tested on my 4D/25. Doesn't (appear to) break macallan@'s IP22.

branches: 1.4.78; 1.4.86; 1.4.92;
merge ktrace-lwp.

branches: 1.3.2;
Add the following to the int driver:

* if the driver hasn't been properly initialized, spin for a bit in wait_fifo()
rather than try to access the local0 status register.
* allow interrupt sharing (from Steve Rumble; necessary for IP20 glass TTY
support).

Finish moving code between ip2x.c and ip3x.c to imc.c and crime.c.

Add register definitions for INT2 interrupt controller.

KNF and misc cosmetics.

branches: 1.2.74; 1.2.78; 1.2.80; 1.2.84;
merge ktrace-lwp.

branches: 1.1.4;
Add int2_wait_fifo(), which will idle while awaiting a flag in the interrupt
controller to deassert. Needed for the more interesting features of various
SGI framebuffers.

Following the example of the hpc/, gio/, and ioc/ directories, move the
mace devices to their own mace/ directory. Alter conf/files.sgimips to
reflect this change in a sane manner (i.e., pull in dev/files.dev and
mace/files.mace when appropriate).

At the same time, allow crime_intr_establish() to fall through to
mace_intr_establish(). mace devices now call cpu_intr_establish().

www.netbsd.org -> www.NetBSD.org

Apply a bunch of patches written by Christopher SEKIYA
(with several cosmetic changes by me) which fixes O2 (IP32) support.

Now my R5000 O2 works fine in multiuser with on-board AIC7880 SCSIs
and several PCI network cards (but only on serial console yet).
L2 cache on R5000/Rm5200 is still disabled for now, but it will be
fixed later, hopefully.

See recent discussion on port-sgimips for details.

__KERNEL_RCSID()

branches: 1.5.6;
Add trailing ; to CFATTACH_DECL.

Use CFATTACH_DECL().

Declare all cfattach structures const.

Replace lots of 8x<space> with <tabs> and other miscellaneous indentation
fixes.
Wrap a couple of long lines.
Use <return-type>\n<function name> as per KNF in a few places.

branches: 1.1.4; 1.1.6; 1.1.10; 1.1.14;
A start at O2 support..

Following the example of the hpc/, gio/, and ioc/ directories, move the
mace devices to their own mace/ directory. Alter conf/files.sgimips to
reflect this change in a sane manner (i.e., pull in dev/files.dev and
mace/files.mace when appropriate).

At the same time, allow crime_intr_establish() to fall through to
mace_intr_establish(). mace devices now call cpu_intr_establish().

www.netbsd.org -> www.NetBSD.org

branches: 1.2.12;
Replace lots of 8x<space> with <tabs> and other miscellaneous indentation
fixes.
Wrap a couple of long lines.
Use <return-type>\n<function name> as per KNF in a few places.

branches: 1.1.4; 1.1.6; 1.1.10; 1.1.14;
A start at O2 support..

Following the example of the hpc/, gio/, and ioc/ directories, move the
mace devices to their own mace/ directory. Alter conf/files.sgimips to
reflect this change in a sane manner (i.e., pull in dev/files.dev and
mace/files.mace when appropriate).

At the same time, allow crime_intr_establish() to fall through to
mace_intr_establish(). mace devices now call cpu_intr_establish().

www.netbsd.org -> www.NetBSD.org

NULL -> 0

Apply a bunch of patches written by Christopher SEKIYA
(with several cosmetic changes by me) which fixes O2 (IP32) support.

Now my R5000 O2 works fine in multiuser with on-board AIC7880 SCSIs
and several PCI network cards (but only on serial console yet).
L2 cache on R5000/Rm5200 is still disabled for now, but it will be
fixed later, hopefully.

See recent discussion on port-sgimips for details.

__KERNEL_RCSID()

branches: 1.11.2;
Improve the interrupt code somewhat by having callers of xxx_intr_establish
pass in an interrupt handle (which is currently to the CRIME interrupt the
device is attached to) so the interrupt handlers know which device was the
one looking for attention.

While here, fix up PCI interrupt routing for both the on-board devices and
the PCI slots -- even though there is only one PCI slot in the chasis, the
hardware can accomodate up to three and provides an interrupt mapping for
all the PCI interrupt pins for both the internal SCSI & PCI slot and the
two "extra" slots.

Checkpoint of O2 work by Chris Sekiya and myself. This is the sgimips bit;
still needs some arch/mips support code before it will fully work.

Use aprint_normal() for cfprint routines.

Add trailing ; to CFATTACH_DECL.

Use CFATTACH_DECL().

Declare all cfattach structures const.

Introduce a new routine, config_match(), which invokes the
cfattach->ca_match function in behalf of the caller. Use it
rather than invoking cfattach->ca_match directly.

Replace lots of 8x<space> with <tabs> and other miscellaneous indentation
fixes.
Wrap a couple of long lines.
Use <return-type>\n<function name> as per KNF in a few places.

branches: 1.3.4;
* fix thinko in mace_print() to prevent intr with value MACECF_INTR_DEFAULT
from getting printed
* fetch device name from dv_xname instead of using hardcoded value

branches: 1.2.2;
- Use the identifier strings that are fetched by the MI ARC BIOS
code.
- Garbage-collect some non-useful stuff from the mainbus_attach_args.

branches: 1.1.4; 1.1.6;
A start at O2 support..

Following the example of the hpc/, gio/, and ioc/ directories, move the
mace devices to their own mace/ directory. Alter conf/files.sgimips to
reflect this change in a sane manner (i.e., pull in dev/files.dev and
mace/files.mace when appropriate).

At the same time, allow crime_intr_establish() to fall through to
mace_intr_establish(). mace devices now call cpu_intr_establish().

www.netbsd.org -> www.NetBSD.org

Apply a bunch of patches written by Christopher SEKIYA
(with several cosmetic changes by me) which fixes O2 (IP32) support.

Now my R5000 O2 works fine in multiuser with on-board AIC7880 SCSIs
and several PCI network cards (but only on serial console yet).
L2 cache on R5000/Rm5200 is still disabled for now, but it will be
fixed later, hopefully.

See recent discussion on port-sgimips for details.

TAB/space cleanup.

branches: 1.3.2;
MACE register definitions

from Chris Sekiya

Replace lots of 8x<space> with <tabs> and other miscellaneous indentation
fixes.
Wrap a couple of long lines.
Use <return-type>\n<function name> as per KNF in a few places.

branches: 1.1.4; 1.1.6; 1.1.10; 1.1.14;
A start at O2 support..

Following the example of the hpc/, gio/, and ioc/ directories, move the
mace devices to their own mace/ directory. Alter conf/files.sgimips to
reflect this change in a sane manner (i.e., pull in dev/files.dev and
mace/files.mace when appropriate).

At the same time, allow crime_intr_establish() to fall through to
mace_intr_establish(). mace devices now call cpu_intr_establish().

www.netbsd.org -> www.NetBSD.org

Apply a bunch of patches written by Christopher SEKIYA
(with several cosmetic changes by me) which fixes O2 (IP32) support.

Now my R5000 O2 works fine in multiuser with on-board AIC7880 SCSIs
and several PCI network cards (but only on serial console yet).
L2 cache on R5000/Rm5200 is still disabled for now, but it will be
fixed later, hopefully.

See recent discussion on port-sgimips for details.

branches: 1.3.2;
Checkpoint of O2 work by Chris Sekiya and myself. This is the sgimips bit;
still needs some arch/mips support code before it will fully work.

Replace lots of 8x<space> with <tabs> and other miscellaneous indentation
fixes.
Wrap a couple of long lines.
Use <return-type>\n<function name> as per KNF in a few places.

branches: 1.1.4; 1.1.6; 1.1.10; 1.1.14;
A start at O2 support..

Following the example of the hpc/, gio/, and ioc/ directories, move the
mace devices to their own mace/ directory. Alter conf/files.sgimips to
reflect this change in a sane manner (i.e., pull in dev/files.dev and
mace/files.mace when appropriate).

At the same time, allow crime_intr_establish() to fall through to
mace_intr_establish(). mace devices now call cpu_intr_establish().

NULL -> 0

Apply a bunch of patches written by Christopher SEKIYA
(with several cosmetic changes by me) which fixes O2 (IP32) support.

Now my R5000 O2 works fine in multiuser with on-board AIC7880 SCSIs
and several PCI network cards (but only on serial console yet).
L2 cache on R5000/Rm5200 is still disabled for now, but it will be
fixed later, hopefully.

See recent discussion on port-sgimips for details.

TAB/space cleanup.

__KERNEL_RCSID()

branches: 1.8.2;
Update to use the full DS1687 register set, therefore allowing the clock to
store absolute year rather than an offset -- this means the clock is now
consitent across the ARCS PROM, IRIX and NetBSD.

XXX: This attachment is now a mismoner, since it's a Dallas Semi RTC, not
a Motorola RTC. Should be renamed.

Add trailing ; to CFATTACH_DECL.

Use CFATTACH_DECL().

Declare all cfattach structures const.

slight last-minute typo here

make RTC year IRIX semantics adjustments a macro

branches: 1.2.2; 1.2.4;
Forgot pooka's license header.

O2 RTC attachment, based on code from Antti Kantee.

Following the example of the hpc/, gio/, and ioc/ directories, move the
mace devices to their own mace/ directory. Alter conf/files.sgimips to
reflect this change in a sane manner (i.e., pull in dev/files.dev and
mace/files.mace when appropriate).

At the same time, allow crime_intr_establish() to fall through to
mace_intr_establish(). mace devices now call cpu_intr_establish().

www.netbsd.org -> www.NetBSD.org

Mouse port offset was incorrect.

Apply a bunch of patches written by Christopher SEKIYA
(with several cosmetic changes by me) which fixes O2 (IP32) support.

Now my R5000 O2 works fine in multiuser with on-board AIC7880 SCSIs
and several PCI network cards (but only on serial console yet).
L2 cache on R5000/Rm5200 is still disabled for now, but it will be
fixed later, hopefully.

See recent discussion on port-sgimips for details.

__KERNEL_RCSID()

branches: 1.5.6;
Add trailing ; to CFATTACH_DECL.

Use CFATTACH_DECL().

Declare all cfattach structures const.

Replace lots of 8x<space> with <tabs> and other miscellaneous indentation
fixes.
Wrap a couple of long lines.
Use <return-type>\n<function name> as per KNF in a few places.

branches: 1.1.2; 1.1.6; 1.1.10;
Just a stub. Needs different register accesses than other 8-bit MACE
deviecs..

Merge thorpej-cfargs2.

branches: 1.18.8;
Merge thorpej-cfargs branch:

Simplify and make extensible the config_search() / config_found() /
config_attach() interfaces: rather than having different variants for
which arguments you want pass along, just have a single call that
takes a variadic list of tag-value arguments.

Adjust all call sites:
- Simplify wherever possible; don't pass along arguments that aren't
actually needed.
- Don't be explicit about what interface attribute is attaching if
the device only has one. (More simplification.)
- Add a config_probe() function to be used in indirect configuiration
situations, making is visibly easier to see when indirect config is
in play, and allowing for future change in semantics. (As of now,
this is just a wrapper around config_match(), but that is an
implementation detail.)

Remove unnecessary or redundant interface attributes where they're not
needed.

There are currently 5 "cfargs" defined:
- CFARG_SUBMATCH (submatch function for direct config)
- CFARG_SEARCH (search function for indirect config)
- CFARG_IATTR (interface attribte)
- CFARG_LOCATORS (locators array)
- CFARG_DEVHANDLE (devhandle_t - wraps OFW, ACPI, etc. handles)

...and a sentinel value CFARG_EOL.

Add some extra sanity checking to ensure that interface attributes
aren't ambiguous.

Use CFARG_DEVHANDLE in MI FDT, OFW, and ACPI code, and macppc and shark
ports to associate those device handles with device_t instance. This
will trickle trough to more places over time (need back-end for pre-OFW
Sun OBP; any others?).

branches: 1.17.32;
switch sgimips to common bus_dma and bus_space in arch/mips/
Tested on O2 and Indy.
Things that are compile-tested only for lack of hardware:
- GIO ethernet cards with PCI bridges
- IP2x hardware not found on Indy
- IP1x

branches: 1.16.14;
split device_t/softc for all remaining drivers.
replace "struct device *" with "device_t".
use device_xname(), device_unit(), etc.

branches: 1.15.2; 1.15.12;
#include <sys/bus.h> instead of <machine/bus.h>.

Merge forward matt-nb5-mips64
Adapt to new interrupt/spl framework
Deal with arcbios calls being O32 and kernels being N32/N64.
Need to save/restore T8 across arcbios calls.

branches: 1.13.12; 1.13.16; 1.13.22; 1.13.24;
KNF and misc cosmetics.

branches: 1.12.44; 1.12.48; 1.12.50; 1.12.54;
Catch and report bus errors.

branches: 1.11.2;
Add pic and imc-specific functions to alter the GIO bus arbiter
configuration. These are to be used exclusively by a higher interface in
gio/gio.c.

Also add functions to enable, disable and query the state of SysAd bus
parity detection since certain gio(4) cards need to be able to disable
it in order to work around their own buggy implementations.

branches: 1.10.4; 1.10.8; 1.10.18;
merge ktrace-lwp.

branches: 1.9.12;
Print board identification, same as we do on the IP2x machines.

branches: 1.8.2;
General cleanup to improve style consistency and KNF adherence.
Updated comments and removed superfluous code. No functional
change intended.

ok'd by pooka and sekiya

RCS Id police

and finally, the big moment you've all been waiting for:
make it actually compile

fix off-by-one in revision print

add watchdog enable and disable

from Steve Rumble

branches: 1.3.2;
Assign memory controller-dependent watchdog reset functions to
platform.watchdog_reset.

As the platform-specific bus reset routine varies based on the memory
controller used, set it in the memory controller driver rather than the
platform init code.

Add pic (memory controller for IP12) driver. Originally written by
Steve Rumble, with mostly stylistic changes by myself.

Use the SI capitalization for "Hz", "kHz", and "MHz" in comments and strings.
Add a space between numbers and Hz unit.

branches: 1.4.4; 1.4.6; 1.4.8; 1.4.10;
merge ktrace-lwp.

TAB/space cleanup.

branches: 1.2.2; 1.2.14;
Add physical offsets for memory configuration registers to be used
outside bus_space.

branches: 1.1.2;
Add pic (memory controller for IP12) driver. Originally written by
Steve Rumble, with mostly stylistic changes by myself.

Remove unnecessary <sys/malloc.h>.

sgimips/scn(4): Convert to ttylock/ttyunlock.

Remove unused extern declaration of constty.

don't malloc a tiny, fixed size buffer to scribble into, then not use it
and never free it either

found by Brainy

branches: 1.7.2; 1.7.4;
Add d_discard to all struct cdevsw instances I could find.

All have been set to "nodiscard"; some should get a real implementation.

branches: 1.6.2;
- use cpu_{g,s}etmodel
- fix unused

Change (mostly mechanically) every cdevsw/bdevsw I can find to use
designated initializers.

I have not built every extant kernel so I have probably broken at
least one build; however I've also found and fixed some wrong
cdevsw/bdevsw entries so even if so I think we come out ahead.

branches: 1.4.2;
split device_t/softc for all remaining drivers.
replace "struct device *" with "device_t".
use device_xname(), device_unit(), etc.

branches: 1.3.4; 1.3.14;
Rename ttymalloc() to tty_alloc(), and ttyfree() to tty_free() for
consistency. Remove some unnecessary malloc.h inclusions as well.

branches: 1.2.4; 1.2.6;
Use lwp_getpcb() on mips, powerpc and sh3, clean from struct user usage.

branches: 1.1.2; 1.1.4; 1.1.6;
Check in work in progress of Personal IRIS 4D/20 and 4D/25 support (IP6
and IP10, respectively).

branches: 1.1.2; 1.1.4; 1.1.6;
Check in work in progress of Personal IRIS 4D/20 and 4D/25 support (IP6
and IP10, respectively).

split device_t/softc for all remaining drivers.
replace "struct device *" with "device_t".
use device_xname(), device_unit(), etc.

branches: 1.1.2; 1.1.4; 1.1.6; 1.1.18; 1.1.28;
Check in work in progress of Personal IRIS 4D/20 and 4D/25 support (IP6
and IP10, respectively).

Make backspace and control keys work.

XXX: Why does KS_BackSpace not do what we want? We're copying pckbd now...

branches: 1.4.20;
merge ktrace-lwp.

TAB/space cleanup.

branches: 1.2.12;
Use C-style comments.

branches: 1.1.2;
Add support for keyboard and mouse attachments at zs*, used on IP12 and IP20.
Code written by Steve Rumble; committed with his blessing.

Using a glass TTY as a console on IP20 (IP12 untested) should now be possible.

declare zskgdb() in zs.c, if KGBD option is enabled.

Fixes compilation issue with the KGBD option enabled for sgimips.

Add missing double p and d for stopped and overriden accordingly.
Fix few more typos along the way, mainly in copy-pasted comments.

Merge thorpej-cfargs2.

branches: 1.40.8;
Merge thorpej-cfargs branch:

Simplify and make extensible the config_search() / config_found() /
config_attach() interfaces: rather than having different variants for
which arguments you want pass along, just have a single call that
takes a variadic list of tag-value arguments.

Adjust all call sites:
- Simplify wherever possible; don't pass along arguments that aren't
actually needed.
- Don't be explicit about what interface attribute is attaching if
the device only has one. (More simplification.)
- Add a config_probe() function to be used in indirect configuiration
situations, making is visibly easier to see when indirect config is
in play, and allowing for future change in semantics. (As of now,
this is just a wrapper around config_match(), but that is an
implementation detail.)

Remove unnecessary or redundant interface attributes where they're not
needed.

There are currently 5 "cfargs" defined:
- CFARG_SUBMATCH (submatch function for direct config)
- CFARG_SEARCH (search function for indirect config)
- CFARG_IATTR (interface attribte)
- CFARG_LOCATORS (locators array)
- CFARG_DEVHANDLE (devhandle_t - wraps OFW, ACPI, etc. handles)

...and a sentinel value CFARG_EOL.

Add some extra sanity checking to ensure that interface attributes
aren't ambiguous.

Use CFARG_DEVHANDLE in MI FDT, OFW, and ACPI code, and macppc and shark
ports to associate those device handles with device_t instance. This
will trickle trough to more places over time (need back-end for pre-OFW
Sun OBP; any others?).

branches: 1.39.32;
switch sgimips to common bus_dma and bus_space in arch/mips/
Tested on O2 and Indy.
Things that are compile-tested only for lack of hardware:
- GIO ethernet cards with PCI bridges
- IP2x hardware not found on Indy
- IP1x

branches: 1.38.6;
- use cpu_{g,s}etmodel
- fix unused

branches: 1.37.4; 1.37.14; 1.37.18;
Merge forward matt-nb5-mips64
Adapt to new interrupt/spl framework
Deal with arcbios calls being O32 and kernels being N32/N64.
Need to save/restore T8 across arcbios calls.

branches: 1.36.4; 1.36.6; 1.36.8;
Enable Personal IRIS 4D/20 and 4D/25 support:
- Adapt int(4) to handle the INT1 chip
- Move generic rtc clocks out of hpc/ and into dev/
- Handle the very strangely wired eeprom and other bits in arcemu
- Sprinkle MACH_SGI_IP6 as necessary
- Enable IP6/IP10 devices in GENERIC32_IP12. Yes, the naming is poor but
there's no winning with kernel/hw compatibility on sgimips...

Tested on my 4D/25. Doesn't (appear to) break macallan@'s IP22.

branches: 1.35.4; 1.35.10;
use device_lookup_private to get softc

branches: 1.34.2; 1.34.4;
Remove clause 3 and 4 from TNF licenses

branches: 1.33.2; 1.33.4;
Split softc and device_t for zsc(4) and its children.

XXX we should restructure MI APIs and make it really machine independent.

branches: 1.32.14;
Use the softint API.

Call zs_lock_init() to set up the chanstate's lock.

branches: 1.30.18; 1.30.36; 1.30.38; 1.30.42; 1.30.44;
integrate kauth.

Use device_unit().

branches: 1.28.4; 1.28.6; 1.28.8; 1.28.10; 1.28.12;
merge ktrace-lwp.

TAB/space cleanup.

branches: 1.26.2;
Constify

Use ANSI function declarations.

Add support for keyboard and mouse attachments at zs*, used on IP12 and IP20.
Code written by Steve Rumble; committed with his blessing.

Using a glass TTY as a console on IP20 (IP12 untested) should now be possible.

At least my IP12 works better without the port-flipping glue,
so un-glueify IP12 for the time being.

branches: 1.22.2;
ANSIfy, de-__P(), add a little glue for IP12.

The serial controller on IP20 lives in a different place. Modify
zs_get_chan_addr() to reflect this.

XXX hardcoded addresses are quite distasteful. This should be passed in
through consinit() -- to avoid conflicts with Ilpo's soon-to-be-committed
framebuffer driver, I'll hold off until his code hits the tree.

TAB/space cleanup.

__KERNEL_RCSID()

branches: 1.18.2;
Provide locking required by the interrupt handlers running at IPL_SERIAL.

Use aprint_normal() for cfprint routines.

Add trailing ; to CFATTACH_DECL.

Use CFATTACH_DECL().

Declare all cfattach structures const.

remove trailing \n in panic(). approved perry.

Rather than referencing the cfdriver directly in the cfdata entries,
instead use a string naming the driver. The cfdriver is then looked
up in a list which is built at run-time.

Merge the gehenna-devsw branch into the trunk.

This merge changes the device switch tables from static array to
dynamically generated by config(8).

- All device switches is defined as a constant structure in device drivers.

- The new grammer ``device-major'' is introduced to ``files''.

device-major <prefix> char <num> [block <num>] [<rules>]

- All device major numbers must be listed up in port dependent majors.<arch>
by using this grammer.

- Added the new naming convention.
The name of the device switch must be <prefix>_[bc]devsw for auto-generation
of device switch tables.

- The backward compatibility of loading block/character device
switch by LKM framework is broken. This is necessary to convert
from block/character device major to device name in runtime and vice versa.

- The restriction to assign device major by LKM is completely removed.
We don't need to reserve LKM entries for dynamic loading of device switch.

- In compile time, device major numbers list is packed into the kernel and
the LKM framework will refer it to assign device major number dynamically.

branches: 1.10.2;
If we handle multiple events from the zs chip in one interrupt, don't count
that as multiple interrupts.

Add event counter for zs interrupts and attach and use it; also, get the
zs major number out of the cdevsw table rather than hardcoding it in zs.c.

Replace lots of 8x<space> with <tabs> and other miscellaneous indentation
fixes.
Wrap a couple of long lines.
Use <return-type>\n<function name> as per KNF in a few places.

- pull in opt_kgdb.h where necessary
- replace opt_kgdb_machdep.h with opt_kgdb.h
- defparam opt_kgdb.h:
KGDB_DEV KGDB_DEVNAME KGDB_DEVADDR KGDB_DEVRATE KGDB_DEVMODE
- move from opt_ddbparam.h to opt_ddb.h:
DDB_FROMCONSOLE DDB_ONPANIC DDB_HISTORY_SIZE DDB_BREAK_CHAR SYMTAB_SPACE
- replace KGDBDEV with KGDB_DEV
- replace KGDBADDR with KGDB_DEVADDR
- replace KGDBMODE with KGDB_DEVMODE
- replace KGDBRATE with KGDB_DEVRATE
- use `9600' instead of `0x2580' for 9600 baud rate
- use correct quotes for options KGDB_DEVNAME="\"com\""
- use correct quotes for options KGDB_DEV="17*256+0"
- remove unnecessary dependancy on Makefile for kgdb_stub.o
- minor whitespace cleanup

* Clean up HPC autoconfiguration a little. Pass device register and
DMA register offsets, as well as IRQ, to children. Use direct
config. Use machine type/subtype to determine which devices are
present.
* Add support for the second SCSI controller on the Indigo2.

branches: 1.5.2; 1.5.8;
bcopy -> memcpy, bzero -> memset

Use sys/dev/arcbios.

- Finish off KGDB kernel debugger support for sgimips

branches: 1.2.2;
Don't use PCLK as the baud rate generator, as the Indy doens't
like that, and it has no effect on the Indigo2.
From Christopher SEKIYA <wileyc@rezrov.net>.

Also, query the "ConsoleOut" ARCS variable (i.e. "what the console
really is") rather than "console" ("what the console is configured
as").

Support for the SCC serial ports on Indigo2/Indy.

From Rafal K. Boni.

Remove unnecessary (and unimplemented) WSKBD_*BELL ioctls in kbd drivers.

It looks the only ioctl(4) to be implemented for WSKBDIO_*BELL in
the backend keyboard driver is WSKBDIO_COMPLEXBELL.
All other BELL ioctls (WSKBDIO_BELL, WSKBDIO_SETBELL, WSKBDIO_GETBELL,
WSKBDIO_SETDEFAULTBELL, and WSKBDIO_GETDEFAULTBELL) are handled in the
MI wskbd driver (in wskbd_displayioctl() in src/sys/dev/wscons/wskbd.c).

Merge thorpej-cfargs2.

branches: 1.12.8;
Merge thorpej-cfargs branch:

Simplify and make extensible the config_search() / config_found() /
config_attach() interfaces: rather than having different variants for
which arguments you want pass along, just have a single call that
takes a variadic list of tag-value arguments.

Adjust all call sites:
- Simplify wherever possible; don't pass along arguments that aren't
actually needed.
- Don't be explicit about what interface attribute is attaching if
the device only has one. (More simplification.)
- Add a config_probe() function to be used in indirect configuiration
situations, making is visibly easier to see when indirect config is
in play, and allowing for future change in semantics. (As of now,
this is just a wrapper around config_match(), but that is an
implementation detail.)

Remove unnecessary or redundant interface attributes where they're not
needed.

There are currently 5 "cfargs" defined:
- CFARG_SUBMATCH (submatch function for direct config)
- CFARG_SEARCH (search function for indirect config)
- CFARG_IATTR (interface attribte)
- CFARG_LOCATORS (locators array)
- CFARG_DEVHANDLE (devhandle_t - wraps OFW, ACPI, etc. handles)

...and a sentinel value CFARG_EOL.

Add some extra sanity checking to ensure that interface attributes
aren't ambiguous.

Use CFARG_DEVHANDLE in MI FDT, OFW, and ACPI code, and macppc and shark
ports to associate those device handles with device_t instance. This
will trickle trough to more places over time (need back-end for pre-OFW
Sun OBP; any others?).

branches: 1.11.2;
malloc(9) -> kmem(9)

branches: 1.10.50;
fix device/softc split errors.

split device_t/softc for all remaining drivers.
replace "struct device *" with "device_t".
use device_xname(), device_unit(), etc.

branches: 1.8.38; 1.8.48;
Split softc and device_t for zsc(4) and its children.

XXX we should restructure MI APIs and make it really machine independent.

branches: 1.7.40;
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.

branches: 1.6.26;
merge ktrace-lwp.

TAB/space cleanup.

branches: 1.4.2; 1.4.14;
KNF and consistency

Allow for console attachment by statically allocating a softc
extension. Tweak cs_softreq when appending the transmit queue.
Finally, add a bit of debugging.

Properly initialise the tx and rx queue head and tail positions.

Add support for keyboard and mouse attachments at zs*, used on IP12 and IP20.
Code written by Steve Rumble; committed with his blessing.

Using a glass TTY as a console on IP20 (IP12 untested) should now be possible.

cast kgdb_dev to int as sun3, sgimips and sgimips64 have different types,
same %lld format specifier doesn't work for both.

apply similar changes between different archs/machines for zs_kgdb.c:
1) adjust zs_kgdb_init tty attachment message for kgdb_dev type change.
(one arch uses cast to int instead, but kept one approach for consistency)
2) remove unused rr0 declaration and assignment in zs_kgdb_txint()
3) adjust kgdb_dev format specifier to %llx/d if needed.
4) add const for zs_kgdb_devname declaration (macppc only).
should fix zs_kgdb.c build for all these ports.

branches: 1.15.44; 1.15.54;
Ansify function definitions w/o arguments. Generated with sed.

branches: 1.14.8; 1.14.14;
Remove clause 3 and 4 from TNF licenses

branches: 1.13.6; 1.13.8; 1.13.10;
Fix a typo: s/Kgdb_dev/kgdb_dev/

branches: 1.12.2; 1.12.8;
Merge the ppcoea-renovation branch to HEAD.

This branch was a major cleanup and rototill of many of the various OEA
cpu based PPC ports that focused on sharing as much code as possible
between the various ports to eliminate near-identical copies of files in
every tree. Additionally there is a new PIC system that unifies the
interface to interrupt code for all different OEA ppc arches. The work
for this branch was done by a variety of people, too long to list here.

TODO:
bebox still needs work to complete the transition to -renovation.
ofppc still needs a bunch of work, which I will be looking at.
ev64260 still needs to be renovated
amigappc was not attempted.

NOTES:
pmppc was removed as an arch, and moved to a evbppc target.

branches: 1.11.10;
Fix a warning from gcc 4 about stripping the volatile qualifier by a cast
by doing the proper __UNVOLATILE() dance.

branches: 1.10.30; 1.10.32; 1.10.38;
merge ktrace-lwp.

branches: 1.9.12;
Use ANSI function declarations.

Add support for keyboard and mouse attachments at zs*, used on IP12 and IP20.
Code written by Steve Rumble; committed with his blessing.

Using a glass TTY as a console on IP20 (IP12 untested) should now be possible.

de-__P.

__KERNEL_RCSID()

branches: 1.5.6;
Merge the gehenna-devsw branch into the trunk.

This merge changes the device switch tables from static array to
dynamically generated by config(8).

- All device switches is defined as a constant structure in device drivers.

- The new grammer ``device-major'' is introduced to ``files''.

device-major <prefix> char <num> [block <num>] [<rules>]

- All device major numbers must be listed up in port dependent majors.<arch>
by using this grammer.

- Added the new naming convention.
The name of the device switch must be <prefix>_[bc]devsw for auto-generation
of device switch tables.

- The backward compatibility of loading block/character device
switch by LKM framework is broken. This is necessary to convert
from block/character device major to device name in runtime and vice versa.

- The restriction to assign device major by LKM is completely removed.
We don't need to reserve LKM entries for dynamic loading of device switch.

- In compile time, device major numbers list is packed into the kernel and
the LKM framework will refer it to assign device major number dynamically.

branches: 1.4.4;
Replace lots of 8x<space> with <tabs> and other miscellaneous indentation
fixes.
Wrap a couple of long lines.
Use <return-type>\n<function name> as per KNF in a few places.

branches: 1.3.2; 1.3.8;
bcopy -> memcpy, bzero -> memset

- Finish off KGDB kernel debugger support for sgimips

branches: 1.1.2;
Support for the SCC serial ports on Indigo2/Indy.

From Rafal K. Boni.

Merge thorpej-cfargs2.

branches: 1.9.8;
Merge thorpej-cfargs branch:

Simplify and make extensible the config_search() / config_found() /
config_attach() interfaces: rather than having different variants for
which arguments you want pass along, just have a single call that
takes a variadic list of tag-value arguments.

Adjust all call sites:
- Simplify wherever possible; don't pass along arguments that aren't
actually needed.
- Don't be explicit about what interface attribute is attaching if
the device only has one. (More simplification.)
- Add a config_probe() function to be used in indirect configuiration
situations, making is visibly easier to see when indirect config is
in play, and allowing for future change in semantics. (As of now,
this is just a wrapper around config_match(), but that is an
implementation detail.)

Remove unnecessary or redundant interface attributes where they're not
needed.

There are currently 5 "cfargs" defined:
- CFARG_SUBMATCH (submatch function for direct config)
- CFARG_SEARCH (search function for indirect config)
- CFARG_IATTR (interface attribte)
- CFARG_LOCATORS (locators array)
- CFARG_DEVHANDLE (devhandle_t - wraps OFW, ACPI, etc. handles)

...and a sentinel value CFARG_EOL.

Add some extra sanity checking to ensure that interface attributes
aren't ambiguous.

Use CFARG_DEVHANDLE in MI FDT, OFW, and ACPI code, and macppc and shark
ports to associate those device handles with device_t instance. This
will trickle trough to more places over time (need back-end for pre-OFW
Sun OBP; any others?).

branches: 1.8.52;
split device_t/softc for all remaining drivers.
replace "struct device *" with "device_t".
use device_xname(), device_unit(), etc.

branches: 1.7.38; 1.7.48;
Split softc and device_t for zsc(4) and its children.

XXX we should restructure MI APIs and make it really machine independent.

branches: 1.6.40;
Kill caddr_t; there will be some MI fallout, but it will be fixed shortly.

branches: 1.5.4;
Tidy away wsmouse_input() abstractions and update
documentation to include the W direction.

branches: 1.4.20; 1.4.22;
merge ktrace-lwp.

TAB/space cleanup.

Pass proper values to MI wsmouse_input(9):
http://mail-index.netbsd.org/port-sgimips/2005/10/18/0000.html

Ok'ed by Steve Rumble.

branches: 1.1.2; 1.1.14;
Add support for keyboard and mouse attachments at zs*, used on IP12 and IP20.
Code written by Steve Rumble; committed with his blessing.

Using a glass TTY as a console on IP20 (IP12 untested) should now be possible.