Repopulate TCDS dual channel SCSI adapter to MI dev/tc, squashing
internals for better layering between TCDS DMA ASIC and ASC SCSI
controller.

Have MI ncr53c9x_attach() the 2nd and 3rd arguments for scsipi_adater
and scsipi_device respectively, with size reduction of ncr53c9x_softc.
Specifying NULL instructs the driver to use default adapter and default
device codes. Every target port has ncr53c9x_attach(sc, NULL, NULL) anyway.

branches: 1.16.2; 1.16.10;
Use the correct definition for the Fast Clock bit in cfg3, and set the
Fast SCSI bit used by the 53C9x chips.

branches: 1.15.4;
Adapt to the new scsipi_adapter interface.

Garbage-collect the open_target_lu and close_target_lu entry points from
struct scsipi_adapter; they were not used.

Add a scsipi_ioctl entry point to struct scsipi_adapter. This will be
used to issue ioctl commands to the host adapters.

Inspired by PR #6090, from Matt Jacob.

Assign my copyright to TNF.

Minor edit.

Assign copyrights to TNF.

- Make this into a proper attachment (i.e. use asc_tcds) and clean up
the namespace accordingly.
- Set FCLK in the CFG3 register if we have a > 25MHz clock.
- Indicate that we support Fast mode to the back-end if we're told we're
fast by the TCDS layer.
- Correctly compute the minimum sync period.

Several changes to the TurboChannel Dual SCSI driver:
- Fix the autoconfiguration model; the TCDS children don't attach to
TurboChannel, so don't make it look like they do.
- Make the driver use bus_space (not bus_dma yet; soon)
- Recognize the PMAZ-FS (baseboard fast SCSI on DEC 3000s) and the
PMAZB-AA (TCDS option card) and PMAZC-AA (TCDS fast SCSI option card).
- Rearrange it a bit, so that it can eventually be made into an "MI"
TurboChannel driver (i.e. work on DECstations).

Adjust for config changes.

Copyright assigned to The NetBSD Foundation.

Nuke the idea of <machine/options.h>. It completely defeats the purpose
of fine-grain option dependencies.

Merge scsipi branch in the mainline. This add support for ATAPI devices
(currently only CD-ROM drives on i386). The sys/dev/scsipi system provides 2
busses to which devices can attach (scsibus and atapibus). This needed to
change some include files and structure names in the low level scsi drivers.

branches: 1.4.2; 1.4.4; 1.4.6;
by default, provide RCS IDs for NetBSD/alpha kernel files in kernel binaries.
This can be disabled (to save a bit of space) with the NO_KERNEL_RCSIDS
options, which is present but commented out in the ALPHA config file.
In ELF-format kernels, these strings are present in the kernel binary but
are not loaded into memory. (In ECOFF-format kernels, there's no easy way
to keep them from being loaded, so they _are_ loaded into memory.)

clean up NetBSD RCS ID strings, include machine/options.h

New ARP system, supports IPv4 over any hardware link.

Some of the stuff (e.g., rarpd, bootpd, dhcpd etc., libsa) still will
only support Ethernet. Tcpdump itself should be ok, but libpcap needs
lot of work.

For the detailed change history, look at the commit log entries for
the is-newarp branch.

branches: 1.1.2;
Add a front-end for the MI 53c9x driver, called "asc", which is the
name used by Digital UNIX (formerly DEC OSF/1).

Repopulate TCDS dual channel SCSI adapter to MI dev/tc, squashing
internals for better layering between TCDS DMA ASIC and ASC SCSI
controller.

branches: 1.6.14;
Make this into a proper attachment (i.e. use asc_tcds) and clean up
the namespace accordingly.

Several changes to the TurboChannel Dual SCSI driver:
- Fix the autoconfiguration model; the TCDS children don't attach to
TurboChannel, so don't make it look like they do.
- Make the driver use bus_space (not bus_dma yet; soon)
- Recognize the PMAZ-FS (baseboard fast SCSI on DEC 3000s) and the
PMAZB-AA (TCDS option card) and PMAZC-AA (TCDS fast SCSI option card).
- Rearrange it a bit, so that it can eventually be made into an "MI"
TurboChannel driver (i.e. work on DECstations).

Copyright assigned to The NetBSD Foundation.

branches: 1.3.2; 1.3.6;
clean up NetBSD RCS ID strings

New ARP system, supports IPv4 over any hardware link.

Some of the stuff (e.g., rarpd, bootpd, dhcpd etc., libsa) still will
only support Ethernet. Tcpdump itself should be ok, but libpcap needs
lot of work.

For the detailed change history, look at the commit log entries for
the is-newarp branch.

branches: 1.1.2;
Add a front-end for the MI 53c9x driver, called "asc", which is the
name used by Digital UNIX (formerly DEC OSF/1).

these have been renamed

MB -> wbflush

update for new clock interrupt handling

print out mode (slow vs. fast)

update from my local tree; better CPU configuration, support for 3000/300.

preliminary Alpha support. note that NOT ALL OF THE MODIFICATIONS TO
THE REST OF THE KERNEL ARE IN THE TREE YET. Also, some of this is
_incredibly_ hack-ish, etc., but it works.

these have been renamed

KERNEL -> _KERNEL

preliminary Alpha support. note that NOT ALL OF THE MODIFICATIONS TO
THE REST OF THE KERNEL ARE IN THE TREE YET. Also, some of this is
_incredibly_ hack-ish, etc., but it works.

- Remove unused bt459reg.h.
- sfbvar.h has gone, placing patch on holes for now. Every TC framebuffer
has a single explosure of xxfb_cnattach() which could be bridged by MI
land.

branches: 1.3.2;
Quote "AS IS" as in the majority of Carnegy Mellon notices.

Fix a pasto in copyright text which has been procreating like rabbits..

branches: 1.1.6;
Preliminary support for both serial consoles and kbd/display wscons
consoles for TC alphas. Based on code from Takuya Koumoto
<takuya-k@is.aist-nara.ac.jp> as well as some parts of Toru Nishimura's
and Matthias Drochner's work on pmax wscons. This relies on the
NEW_SCC_DRIVER.

Add MX (PMAG-A), CX (PMAG-B), TX (PMAG-RO/PMAG[B]-J) TC framebuffer
supports. None of them has any acceleration capability.

branches: 1.25.4;
Change the kernel mmap interface so that the offset to map is an
"off_t" and the return value is a "paddr_t" to allow mappings
at offsets past 2^31 bytes. Somewhat inspired by FreeBSD, which
only changed the offset to a "vm_offset_t".

Includes updates for the i386, pc532 and sh3 mmmmap from Jason Thorpe.

branches: 1.24.4;
- Remove unused bt459reg.h.
- sfbvar.h has gone, placing patch on holes for now. Every TC framebuffer
has a single explosure of xxfb_cnattach() which could be bridged by MI
land.

update for changed struct wsdisplay_accessops:show_screen signature.
no functional changes

branches: 1.22.8; 1.22.14;
remove dummy "load_font" entries

fix problems in many d_mmap routines:
- returned EOPNOTSUPP rather than -1.
- no check for negative offset.
many of these fix potential security problems in these drivers.


XXX XXX XXX
the d_mmap cdev routine should be changed to have a prototype like:
paddr_t (*d_mmap) __P((dev_t, off_t, int));

by someone!

Preliminary support for both serial consoles and kbd/display wscons
consoles for TC alphas. Based on code from Takuya Koumoto
<takuya-k@is.aist-nara.ac.jp> as well as some parts of Toru Nishimura's
and Matthias Drochner's work on pmax wscons. This relies on the
NEW_SCC_DRIVER.

adapt to wscons interface change

adapt to wscons interface change

Simple screen attribute handling for wscons.
The graphics device driver passes a "default attribute" for normal text
output to the wscons framework. If the emulation module needs more
attributes (for different "renditions") it can allocate them via a
callback.
For now, only the "sun" emulation makes use of it.

count allocated screens correctly

convert to mi wscons

Adjust for config changes.

branches: 1.13.2;
Add a "scrollskip" member to wscons_emulfuncs (XXX even though it's not
a function) which specifies how many lines will be skipped when scrolling
up when the bottom of the screen is reached. Dumb framebuffers skip 10
lines (as before) because the copies are s ... l ... o ... w, but it's
silly to skip 10 lines on VGA, since the copies are much faster, so we
only skip one in that case.

Nuke the idea of <machine/options.h>. It completely defeats the purpose
of fine-grain option dependencies.

branches: 1.11.4;
by default, provide RCS IDs for NetBSD/alpha kernel files in kernel binaries.
This can be disabled (to save a bit of space) with the NO_KERNEL_RCSIDS
options, which is present but commented out in the ALPHA config file.
In ELF-format kernels, these strings are present in the kernel binary but
are not loaded into memory. (In ECOFF-format kernels, there's no easy way
to keep them from being loaded, so they _are_ loaded into memory.)

use machine/conf.h instead of sys/conf.h and/or machine/cpuconf.h

clean up NetBSD RCS ID strings, include machine/options.h

remove all traces of __BROKEN_INDIRECT_CONFIG (except in shared drivers)

update these so they compile whether or not __BROKEN_INDIRECT_CONFIG
is defined.

branches: 1.6.2;
catch up with wscons frame buffer attachment, mmap, and ioctl
interface changes.

backout previous kprintf change

printf -> kprintf, sprintf -> ksprintf

change cfprint_t type definition to take a const char *, rather than
a char *, because that's what was really intended, and because
if the print function modifies the string, various things could become
unhappy (so the string should _not_ be modified).

clean and update for new defintions, prototypes, etc.

Preliminary CFB support. Does not work as machine's console (for two
reasons: it won't attach as console, and there's currently no way to do
keyboard input on TC machines), and has no real RAMDAC (colormap, cursor,
etc.) support. Digital UNIX does not support CFB frame buffers in the
Alpha, but they appear to work OK (with an appropriate monitor) in my
3000/300.

Add MX (PMAG-A), CX (PMAG-B), TX (PMAG-RO/PMAG[B]-J) TC framebuffer
supports. None of them has any acceleration capability.

branches: 1.5.26;
Preliminary support for both serial consoles and kbd/display wscons
consoles for TC alphas. Based on code from Takuya Koumoto
<takuya-k@is.aist-nara.ac.jp> as well as some parts of Toru Nishimura's
and Matthias Drochner's work on pmax wscons. This relies on the
NEW_SCC_DRIVER.

vm_offset_t no longer exists. Change to paddr_t, vaddr_t, and psize_t.

convert to mi wscons

clean up NetBSD RCS ID strings

branches: 1.1.2;
Preliminary CFB support. Does not work as machine's console (for two
reasons: it won't attach as console, and there's currently no way to do
keyboard input on TC machines), and has no real RAMDAC (colormap, cursor,
etc.) support. Digital UNIX does not support CFB frame buffers in the
Alpha, but they appear to work OK (with an appropriate monitor) in my
3000/300.

they've moved.

and with 0x1fff, rather than 0x1000

preliminary Alpha support. note that NOT ALL OF THE MODIFICATIONS TO
THE REST OF THE KERNEL ARE IN THE TREE YET. Also, some of this is
_incredibly_ hack-ish, etc., but it works.

they've moved.

preliminary Alpha support. note that NOT ALL OF THE MODIFICATIONS TO
THE REST OF THE KERNEL ARE IN THE TREE YET. Also, some of this is
_incredibly_ hack-ish, etc., but it works.

Remove this driver; it is now obsolete.

branches: 1.28.6;
Fill in sc_link.max_target

kill the last vestiges of __BROKEN_INDIRECT_CONFIG.

update these so they compile whether or not __BROKEN_INDIRECT_CONFIG
is defined.

branches: 1.25.2;
Sync with sparc port:
>Pay attention to DMA errors as reported by DMAINTR() returning -1. If this
>happens reset everything.
>
>Re-schedule a timeout when first attempting to abort an operation.
>Cancel any queued timer events before re-scheduling a timeout, so esp_abort()
>can be called from other places besides the timeout handler.

Sync with sparc revision 1.61:
Add explicit casts to `%l*' formatted arguments.

-Werror -Wall -Wstrict-prototypes -Wmissing-prototypes -Wcast-qual cleanups.

Minor cosmetic change.

backout previous kprintf change

printf -> kprintf, sprintf -> ksprintf

Remove unused variable.

* Bitbucket extra status bytes rather than barfing.
* Add copyright.

Restructure code more like aic6360 driver. Also:
* Check for some errors during reselect.
* Don't recognize MESSAGE PARITY ERROR. A target is not allowed to
send it.
* Check the size of a SDTR.

- Add some missing prototypes and kill an unused label.
- Add some missing SPARC_DRIVER stuff.
- Make SPARC_DRIVER automatic.

This driver now compiles as-is and works on both the SPARC and Alpha.

Fix dangling pointer in debugging message.

If we get an unexpected disconnect while issuing a REQUEST SENSE, do a reset.

Functional changes:
* Handle message retransmissions and partially sent messages correctly.
* Make sure we clear ATN after the last message is sent.
* Do the right thing if the target initiates negotiation for async mode
after having negotiated sync mode.
* Fix some cases where we set ATN with no message queued, or schedule a
message without setting ATN.
* Issue a REQUEST SENSE after an unexpected disconnect, per SCSI spec.
* Fix abort handling in a number of cases.
* Recognize selection timeouts better (to speed up probing).

Sync the 53c94 SCSI driver with the sparc version. Update the DMA code to
match the interfaces required by the driver.

(1) set scsi_link channel to either the appropriate channel (if a
multi-channel driver), or to SCSI_CHANNEL_ONLY_ONE if a
single-channel driver.
(2) use scsiprint() rather than a locally-defined autoconfig print
function, and kill any locally-defined print function.

change cfprint_t type definition to take a const char *, rather than
a char *, because that's what was really intended, and because
if the print function modifies the string, various things could become
unhappy (so the string should _not_ be modified).

interrupt counters for TurboChannel systems.

branches: 1.8.4;
update for latest MI TC and SCSI code, and clean up in the face of
config changes.

New device attachment scheme:

- split softc size and match/attach out from cfdriver into
a new struct cfattach.

- new "attach" directive for files.*. May specify the name of
the cfattach structure, so that devices may be easily attached
to parents with different autoconfiguration semantics.

update to include most of pk's changes to the sparc esp driver.
rework TC configuration code.

MB -> wbflush

update SCSI minphys routines' definitions to match standard minphys()
definition and usage.

deal with renamed header file.

use charles's cleaner message input code

preliminary Alpha support. note that NOT ALL OF THE MODIFICATIONS TO
THE REST OF THE KERNEL ARE IN THE TREE YET. Also, some of this is
_incredibly_ hack-ish, etc., but it works.

Remove this driver; it is now obsolete.

branches: 1.3.2; 1.3.8;
Sync the 53c94 SCSI driver with the sparc version. Update the DMA code to
match the interfaces required by the driver.

branches: 1.2.4;
update to include most of pk's changes to the sparc esp driver.
rework TC configuration code.

preliminary Alpha support. note that NOT ALL OF THE MODIFICATIONS TO
THE REST OF THE KERNEL ARE IN THE TREE YET. Also, some of this is
_incredibly_ hack-ish, etc., but it works.

Remove this driver; it is now obsolete.

branches: 1.12.2; 1.12.8;
sync up with pk's commit to the sparc port:
>One control block per target is insufficient if you have a full complement
>of targets attached and access those simultaneously (like in a ccd(4) array).
>We (now) allocate (somewhat arbitrarily) three per target.
>Noticed by Marshall Midden.

-Werror -Wall -Wstrict-prototypes -Wmissing-prototypes -Wcast-qual cleanups.

Increase the maximum number of ECBs.

backout previous kprintf change

printf -> kprintf, sprintf -> ksprintf

- Add some missing prototypes and kill an unused label.
- Add some missing SPARC_DRIVER stuff.
- Make SPARC_DRIVER automatic.

This driver now compiles as-is and works on both the SPARC and Alpha.

Functional changes:
* Handle message retransmissions and partially sent messages correctly.
* Make sure we clear ATN after the last message is sent.
* Do the right thing if the target initiates negotiation for async mode
after having negotiated sync mode.
* Fix some cases where we set ATN with no message queued, or schedule a
message without setting ATN.
* Issue a REQUEST SENSE after an unexpected disconnect, per SCSI spec.
* Fix abort handling in a number of cases.
* Recognize selection timeouts better (to speed up probing).

Sync the 53c94 SCSI driver with the sparc version. Update the DMA code to
match the interfaces required by the driver.

clean and update for new defintions, prototypes, etc.

branches: 1.3.4;
update to include most of pk's changes to the sparc esp driver.
rework TC configuration code.

MB -> wbflush

preliminary Alpha support. note that NOT ALL OF THE MODIFICATIONS TO
THE REST OF THE KERNEL ARE IN THE TREE YET. Also, some of this is
_incredibly_ hack-ish, etc., but it works.

these are going to be made machine-independent

update for am7990var.h, am7990.c changes

wholesale update from my NetBSD/Alpha source tree. Includes:
Support for AXPpci CPUs,
Support for AlphaStation 600 CPUs,
new boot block structure, which requires an 'installboot'
program and works a lot like the NetBSD/sparc boot blocks.

MB -> wbflush

update to use new paths for MI lance driver

update to use machine-independent lance driver

adjust for networking code changes

make bpf work, by appending ether header before mtap, then discarding it.

update from my local tree; better CPU configuration, support for 3000/300.

preliminary Alpha support. note that NOT ALL OF THE MODIFICATIONS TO
THE REST OF THE KERNEL ARE IN THE TREE YET. Also, some of this is
_incredibly_ hack-ish, etc., but it works.

update to use machine-independent lance driver

preliminary Alpha support. note that NOT ALL OF THE MODIFICATIONS TO
THE REST OF THE KERNEL ARE IN THE TREE YET. Also, some of this is
_incredibly_ hack-ish, etc., but it works.

these are going to be made machine-independent

update to use machine-independent lance driver

Liberally sprinkle static around to get more symbols out of the
global namespace. A small bit of const poisoning in the TC code.

branches: 1.48.4;
malloc(9) -> kmem(9)

branches: 1.47.8;
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.

branches: 1.46.10; 1.46.20; 1.46.30;
kill sprintf

branches: 1.45.6; 1.45.10;
Do a minor cleanup of alpha (this will make applying pullups post branching
easier).
u_int{8,16,32,64}_t -> uint{*}_t
Change all old-style definitions to C89 prototypes.
Whitespace cleanup.
Constification in db_disasm.c

branches: 1.44.2; 1.44.6;
#include <sys/bus.h> instead of <machine/bus.h>.

Split device_t/softc. pmax is tested on GXemul, compile test only for alpha.

branches: 1.42.4; 1.42.6; 1.42.10;
ANSIfy another 1261 function definitions.
The only ones left in sys are beyond by sed script!
(or in sys/dist or sys/external)
Mostly they have function pointer parameters.

Change about 4500 of the K&R function definitions to ANSI ones.
There are still about 1600 left, but they have ',' or /* ... */
in the actual variable definitions - which my awk script doesn't handle.
There are also many that need () -> (void).
(The script does handle misordered arguments.)

Remove all the __P() from sys (excluding sys/dist)
Diff checked with grep and MK1 eyeball.
i386 and amd64 GENERIC and sys still build.

branches: 1.39.4; 1.39.10;
use device_lookup_private to get softc

branches: 1.38.2; 1.38.4;
Remove clause 3 and 4 from TNF licenses

branches: 1.37.108; 1.37.110; 1.37.112;
Use CFATTACH_DECL().

Declare all cfattach structures const.

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

branches: 1.34.2; 1.34.4;
Clear the ISDN TXPTR/RXPTR/OVRN interrupt bits here. Based on
a suggestion by Gregory McGarry.

- Use PMAD-BA name for device match logic. It's not a real TURBOchannel
device and the id string is not found in option ROM string, but
ULTRIX/OSF1 call it so anyway.
- Squash IOASIC DMA buffer allocation into device attach() sequence.

branches: 1.32.2;
Switch to the new `evcnt' mechanism for counting interrupts. Maintain
a per-CPU interrupt counter for clock, device, and interprocessor
interrupts.

make egcs happy: ioasic.c:327: warning: suggest parentheses
around arithmetic in operand of |

IOASIC_INTR_ISDN is used for interrupts on every sample. We use DMA so
need IOASIC_INTR_ISDN_TXLOAD|IOASIC_INTR_ISDN_RXLOAD.

branches: 1.29.2;
Eliminate unused sc_cookie field of struct ioasic_softc.

Merge IOASIC register definitions.

make this compile without DEC_3000_300

Use an MI routine in ioasic_subr.c.

branches: 1.25.2; 1.25.4; 1.25.6;
- Remove defunct codes and arrange consistency between pmax and alpha.
- Reduce 'ioasic_base' expose.

be more consistent about use of 'cputype'. e.g. it's in a header, don't
bother 'externing' it everywhere!

branches: 1.23.4;
g/c proto

XXX rm ioasicprint() & ioasic_submatch() because of <dev/tc/ioasic_subr.c> XXX

MI ioasic softc and child-device structs, gc duplicate struct definitions.
Add MI ioasic functions, but dont enable on alpha until tested.

Fix printf format problems on Alpha. (IRONIC?!)

Define ioasic device offsets in terms of standard ioasic "slot" offsets.

Add dm_mapsize to bus_dmamap_t and rename BUS_DMAMEM_NOSYNC to
BUS_DMA_COHERENT.

Allocate the LANCE DMA area and set it up here. This is still not
optimal (really should be done further down the food chain), but we
can't really do that in a nice way right now, and this is much better
than what we had before (which was to steal memory pages before the
VM system was bootstrapped).

Adjust for config changes.

Pull in DEC_3000_300 option.

Nuke the idea of <machine/options.h>. It completely defeats the purpose
of fine-grain option dependencies.

branches: 1.13.4;
by default, provide RCS IDs for NetBSD/alpha kernel files in kernel binaries.
This can be disabled (to save a bit of space) with the NO_KERNEL_RCSIDS
options, which is present but commented out in the ALPHA config file.
In ELF-format kernels, these strings are present in the kernel binary but
are not loaded into memory. (In ECOFF-format kernels, there's no easy way
to keep them from being loaded, so they _are_ loaded into memory.)

clean up NetBSD RCS ID strings, include machine/options.h

remove all traces of __BROKEN_INDIRECT_CONFIG (except in shared drivers)

update these so they compile whether or not __BROKEN_INDIRECT_CONFIG
is defined.

branches: 1.9.2;
backout previous kprintf change

printf -> kprintf, sprintf -> ksprintf

change cfprint_t type definition to take a const char *, rather than
a char *, because that's what was really intended, and because
if the print function modifies the string, various things could become
unhappy (so the string should _not_ be modified).

clean and update for new defintions, prototypes, etc.

interrupt counters for TurboChannel systems.

branches: 1.4.4;
clean up copyrights and RCS IDs

update for latest MI TC and SCSI code, and clean up in the face of
config changes.

New device attachment scheme:

- split softc size and match/attach out from cfdriver into
a new struct cfattach.

- new "attach" directive for files.*. May specify the name of
the cfattach structure, so that devices may be easily attached
to parents with different autoconfiguration semantics.

update for new turbochannel configuration code.

Merge IOASIC register definitions.

branches: 1.2.22;
clean up NetBSD RCS ID strings

branches: 1.1.2;
update for new turbochannel configuration code.

Wrap the mcclock driver's mc146818_softc inside a new mcclock_softc.
This commit represents no functional change, but forms the basis for
a future functional change.

Liberally sprinkle static around to get more symbols out of the
global namespace. A small bit of const poisoning in the TC code.

branches: 1.17.64;
Do a minor cleanup of alpha (this will make applying pullups post branching
easier).
u_int{8,16,32,64}_t -> uint{*}_t
Change all old-style definitions to C89 prototypes.
Whitespace cleanup.
Constification in db_disasm.c

branches: 1.16.2; 1.16.6;
#include <sys/bus.h> instead of <machine/bus.h>.

Don't forget to initialize sc_dev.

Split device_t and softc for MI mc146818 clock,
and other related misc cosmetics.

branches: 1.13.16;
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: 1.12.6;
Add MI todr(9) support and timercounter(9) support with kern_cctr.c to alpha:
- use todr(9) API with MI mc146818(4) driver and remove homegrown
todr stuff from MD alpha/clock.c and alpha/mcclock.c
- also remove obsolete cc_microtime stuff from MD code
- add ci_pcc_freq member in struct cpu_info for cpu_frequency(), and
calibrate it with mc146818 interval clock in mcclock attachment
- call cc_init() in cpu_initclocks(9) because all alpha cpus have
a pcc counter

Tested on DEC 3000/300 and AlphaPC 164, but not on any SMP machines yet.

branches: 1.11.22; 1.11.64; 1.11.72; 1.11.76; 1.11.78;
Use CFATTACH_DECL().

Declare all cfattach structures const.

branches: 1.9.2; 1.9.4;
Unnecessary to include <dev/tc/tcreg.h>.

branches: 1.8.18;
Nuke the idea of <machine/options.h>. It completely defeats the purpose
of fine-grain option dependencies.

branches: 1.7.2;
Use new <dev/dec/clockvar.h> and and <dev/dec/mcclokvar.h>.
Garbage-collect original headers from sys/arch/alpha.

by default, provide RCS IDs for NetBSD/alpha kernel files in kernel binaries.
This can be disabled (to save a bit of space) with the NO_KERNEL_RCSIDS
options, which is present but commented out in the ALPHA config file.
In ELF-format kernels, these strings are present in the kernel binary but
are not loaded into memory. (In ECOFF-format kernels, there's no easy way
to keep them from being loaded, so they _are_ loaded into memory.)

clean up NetBSD RCS ID strings, include machine/options.h

remove all traces of __BROKEN_INDIRECT_CONFIG (except in shared drivers)

update these so they compile whether or not __BROKEN_INDIRECT_CONFIG
is defined.

branches: 1.2.2;
fix copyrights

clean up clock handling: clock drivers are now seperate from the
chip-independent clock code. 'clock' has been renamed 'mcclock' 'clock'
has been renamed 'mcclock' (since it's a driver for that particular
clock, and since eventually there may be another clock chip driver),
and now attaches via seperate match/attach functions to both TC and ISA.
This removes a whole lot of #ifdefs...

Remove the TCWSCONS config now that zstty can do flow control on IOASIC
machines.

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: 1.59.2;
Eliminate common.

Convert ioctl code to use EPASSTHROUGH instead of -1 or ENOTTY for
indicating an unhandled "command". ERESTART is -1, which can lead to
confusion. ERESTART has been moved to -3 and EPASSTHROUGH has been
placed at -4. No ioctl code should now return -1 anywhere. The
ioctl() system call is now properly restartable.

branches: 1.57.2;
Don't require tp->t_dev to be filled in in cold_sccparam().

Garbage-collect all the (outdated and totally useless) pmax stuff
in this code.

Yanking a pmax-specific header should not have broken Alpha
compiles! Remove all pmax include files, copying 'struct pdma'
from <pmax/dev/pdma.h> into sccvar.h.

XXX: diffs between current pmax and alpha scc.c are almost as large
as the files themselves. Should clean this up...

branches: 1.54.2; 1.54.4;
Add `l_poll' to `struct linesw' and provide an xxxpoll() entry point
in each tty driver to indirect through it.

This allows tty line-disciplines to handle poll(2) system calls.

Remove unnecessary test of tp->t_linesw against NULL; they are results
of confusion while correcting compilation error after t_line is
replaced with t_linesw.

branches: 1.52.2;
/t_line/t_linesw (typo)

Adapt to new line discipline scheme.

- Implement cnbell() -- ring the console bell. The cn_bell entrypoint
is optional.
- Add cn_bell to statically allocated consdevs as appropriate.

Adjust for IOASIC register changes.

Merge IOASIC register definitions.

branches: 1.47.2;
Centralize the declaration and clearing of `cold'.

In the kludges-on-top-of-kludges dept, print SWITCHING TO SERIAL CONSOLE
(but iff we started on the framebuffer) so the would-be installer doesn't
think everything just locked up.

Quote "AS IS" as in the majority of Carnegy Mellon notices.

be more consistent about use of 'cputype'. e.g. it's in a header, don't
bother 'externing' it everywhere!

Fix a pasto in copyright text which has been procreating like rabbits..

branches: 1.42.8;
defopt DDB.

Update #ifdef'ed-out changes from pmax :
* Add more #ifdef pmax/#endif, #ifdef alpha/#endif where appropriate.
Config and heade files need more work (or replacement)
change TK_NOTYET to HAVE_RCONS
change commented-out /* && cn_tab.cn_screen */ to && raster_console()
* Add DDB hooks.
* Note where Alpha console ignores carrier on consoles.
* Add pmax-derived console tty-size code inside HAVE_RCONS
* Fold in gross pmax rcons-input hooks, inside HAVE_RCONS

Untested, but whitespace/ifdef only, cross-compiles OK,
preprocessing shows no significat differences (famous last words)

more TS_WOPEN to tp->t_wopen changes

Centralize cdev_decl() for these MI drivers + a12dc in machine/conf.h (cgd)
(This does not introduce any config or option dependencies.)

Make this compile again after recent config changes.

Adjust for config changes.

Pull in DEC_3000_300 option.

Nuke the idea of <machine/options.h>. It completely defeats the purpose
of fine-grain option dependencies.

branches: 1.34.2;
use locator defines in "locators.h" to index cf_loc[]

- In sccmatch(), check any user-specified offset, allowing scc unit numbers
to be nailed down. From Ken Hornstein <kenh@cmf.nrl.navy.mil>, PR #3626,
slightly modified to allow wildcarded offsets.
- Blow away the cf->cf_unit check in sccmatch(); it's bogus.

by default, provide RCS IDs for NetBSD/alpha kernel files in kernel binaries.
This can be disabled (to save a bit of space) with the NO_KERNEL_RCSIDS
options, which is present but commented out in the ALPHA config file.
In ELF-format kernels, these strings are present in the kernel binary but
are not loaded into memory. (In ECOFF-format kernels, there's no easy way
to keep them from being loaded, so they _are_ loaded into memory.)

use machine/conf.h instead of sys/conf.h and/or machine/cpuconf.h

clean up NetBSD RCS ID strings, include machine/options.h

remove all traces of __BROKEN_INDIRECT_CONFIG (except in shared drivers)

update these so they compile whether or not __BROKEN_INDIRECT_CONFIG
is defined.

branches: 1.27.2;
compile cleanly with:
-Werror -Wall -Wstrict-prototypes -Wmissing-prototypes -Wcast-qual

alpha_mb() -> tc_mb(), for alpha/pmax convergence, even tho the tc_mb()
interface in tc_machdep.h is obsolete.
Change "#if 0" around pmax rcons mouse-redirection to "#ifdef HAVE_RCONS"

spaces -> tabs where appropriate, some (minor) KNF (it needs much more),
kill spaces at EOL.

more updates from jonathan. kill scc_tty array and needs-count

various bits of cleanup, from jonathan stone, to make this and the pmax
driver slowly converge.

backout previous kprintf change

printf -> kprintf, sprintf -> ksprintf

tty stop functions really should return void, not int, and certainly not both.

clean and update for new defintions, prototypes, etc.

only do tty_attach() for first tty of each chip; that's the only real tty.

add tty_attach() calls where appropriate

branches: 1.16.4;
clean up setting of the default tty cflag. don't do the "delay if
high bit set" thing.

clean up copyrights and RCS IDs

update for latest MI TC and SCSI code, and clean up in the face of
config changes.

New device attachment scheme:

- split softc size and match/attach out from cfdriver into
a new struct cfattach.

- new "attach" directive for files.*. May specify the name of
the cfattach structure, so that devices may be easily attached
to parents with different autoconfiguration semantics.

Fix #includes.

update for new turbochannel configuration code.

wholesale update from my NetBSD/Alpha source tree. Includes:
Support for AXPpci CPUs,
Support for AlphaStation 600 CPUs,
new boot block structure, which requires an 'installboot'
program and works a lot like the NetBSD/sparc boot blocks.

MB -> wbflush

note that most of dev/ic's contents have changed names

update for (old) new tty changes

console putc returns void

Use global 8530 defs. Needs more work.

deal reasonably with the 3000/300's console (annoying engineers!)

don't reset the chip when setting paramters

fill up device switches, clean up a couple of mistakes.

preliminary Alpha support. note that NOT ALL OF THE MODIFICATIONS TO
THE REST OF THE KERNEL ARE IN THE TREE YET. Also, some of this is
_incredibly_ hack-ish, etc., but it works.

Remove the TCWSCONS config now that zstty can do flow control on IOASIC
machines.

branches: 1.3.34; 1.3.36;
clean up NetBSD RCS ID strings

branches: 1.2.4;
Use global 8530 defs. Needs more work.

preliminary Alpha support. note that NOT ALL OF THE MODIFICATIONS TO
THE REST OF THE KERNEL ARE IN THE TREE YET. Also, some of this is
_incredibly_ hack-ish, etc., but it works.

Remove the TCWSCONS config now that zstty can do flow control on IOASIC
machines.

Eliminate common.

Yanking a pmax-specific header should not have broken Alpha
compiles! Remove all pmax include files, copying 'struct pdma'
from <pmax/dev/pdma.h> into sccvar.h.

XXX: diffs between current pmax and alpha scc.c are almost as large
as the files themselves. Should clean this up...

branches: 1.6.14; 1.6.16;
In the kludges-on-top-of-kludges dept, print SWITCHING TO SERIAL CONSOLE
(but iff we started on the framebuffer) so the would-be installer doesn't
think everything just locked up.

branches: 1.5.16; 1.5.20;
clean up NetBSD RCS ID strings

branches: 1.4.2;
Add short delays after accessing registers, so that the z8530's settle
times are respected. This solves the duplicated characters problem.

clean and update for new defintions, prototypes, etc.

branches: 1.2.6;
MB -> wbflush

preliminary Alpha support. note that NOT ALL OF THE MODIFICATIONS TO
THE REST OF THE KERNEL ARE IN THE TREE YET. Also, some of this is
_incredibly_ hack-ish, etc., but it works.

Remove local implementation switching to accelerated MI sfb.c.

update for changed struct wsdisplay_accessops:show_screen signature.
no functional changes

branches: 1.22.8; 1.22.14;
remove dummy "load_font" entries

fix problems in many d_mmap routines:
- returned EOPNOTSUPP rather than -1.
- no check for negative offset.
many of these fix potential security problems in these drivers.


XXX XXX XXX
the d_mmap cdev routine should be changed to have a prototype like:
paddr_t (*d_mmap) __P((dev_t, off_t, int));

by someone!

Preliminary support for both serial consoles and kbd/display wscons
consoles for TC alphas. Based on code from Takuya Koumoto
<takuya-k@is.aist-nara.ac.jp> as well as some parts of Toru Nishimura's
and Matthias Drochner's work on pmax wscons. This relies on the
NEW_SCC_DRIVER.

adapt to wscons interface change

adapt to wscons interface change

Simple screen attribute handling for wscons.
The graphics device driver passes a "default attribute" for normal text
output to the wscons framework. If the emulation module needs more
attributes (for different "renditions") it can allocate them via a
callback.
For now, only the "sun" emulation makes use of it.

count allocated screens correctly

convert to mi wscons

Adjust for config changes.

branches: 1.13.2;
Add a "scrollskip" member to wscons_emulfuncs (XXX even though it's not
a function) which specifies how many lines will be skipped when scrolling
up when the bottom of the screen is reached. Dumb framebuffers skip 10
lines (as before) because the copies are s ... l ... o ... w, but it's
silly to skip 10 lines on VGA, since the copies are much faster, so we
only skip one in that case.

Nuke the idea of <machine/options.h>. It completely defeats the purpose
of fine-grain option dependencies.

branches: 1.11.4;
by default, provide RCS IDs for NetBSD/alpha kernel files in kernel binaries.
This can be disabled (to save a bit of space) with the NO_KERNEL_RCSIDS
options, which is present but commented out in the ALPHA config file.
In ELF-format kernels, these strings are present in the kernel binary but
are not loaded into memory. (In ECOFF-format kernels, there's no easy way
to keep them from being loaded, so they _are_ loaded into memory.)

use machine/conf.h instead of sys/conf.h and/or machine/cpuconf.h

clean up NetBSD RCS ID strings, include machine/options.h

remove all traces of __BROKEN_INDIRECT_CONFIG (except in shared drivers)

update these so they compile whether or not __BROKEN_INDIRECT_CONFIG
is defined.

branches: 1.6.2;
catch up with wscons frame buffer attachment, mmap, and ioctl
interface changes.

backout previous kprintf change

printf -> kprintf, sprintf -> ksprintf

change cfprint_t type definition to take a const char *, rather than
a char *, because that's what was really intended, and because
if the print function modifies the string, various things could become
unhappy (so the string should _not_ be modified).

clean and update for new defintions, prototypes, etc.

preliminary SFB support. Does not work as machine's console (for two
reasons: it won't attach as console, and there's currently no way to do
keyboard input on TC machines), and has no real RAMDAC (colormap,
cursor, etc.) support.

Remove local implementation switching to accelerated MI sfb.c.

branches: 1.5.12;
Preliminary support for both serial consoles and kbd/display wscons
consoles for TC alphas. Based on code from Takuya Koumoto
<takuya-k@is.aist-nara.ac.jp> as well as some parts of Toru Nishimura's
and Matthias Drochner's work on pmax wscons. This relies on the
NEW_SCC_DRIVER.

vm_offset_t no longer exists. Change to paddr_t, vaddr_t, and psize_t.

convert to mi wscons

clean up NetBSD RCS ID strings

branches: 1.1.2;
preliminary SFB support. Does not work as machine's console (for two
reasons: it won't attach as console, and there's currently no way to do
keyboard input on TC machines), and has no real RAMDAC (colormap,
cursor, etc.) support.

these are going to be made machine-independent

update from my local tree; better CPU configuration, support for 3000/300.

preliminary Alpha support. note that NOT ALL OF THE MODIFICATIONS TO
THE REST OF THE KERNEL ARE IN THE TREE YET. Also, some of this is
_incredibly_ hack-ish, etc., but it works.

these are going to be made machine-independent

preliminary Alpha support. note that NOT ALL OF THE MODIFICATIONS TO
THE REST OF THE KERNEL ARE IN THE TREE YET. Also, some of this is
_incredibly_ hack-ish, etc., but it works.

Liberally sprinkle static around to get more symbols out of the
global namespace. A small bit of const poisoning in the TC code.

branches: 1.38.4;
malloc(9) -> kmem(9)

branches: 1.37.2;
Changes to make interrupt {,dis}establish MP-safe on Alpha:
- Protect all of the system interrupt linkage with the cpu_lock mutex.
- Re-order some of the stores to the SCB vector table to make it safe
in the face of lockless interrupt dispatch.
- Add a framework for routing interrupts to specific CPUs. Interrupts
are still funneled only to the primary CPU, but that will change for
some systems soon. Ensure that interrupt handler lists are manipulated
only on the CPUs that handle that specific interrupt source. This required
a re-factor of the alpha_shared_intr_*() family of functions.
- Enable __HAVE_INTR_CONTROL, although interrupt redistribution is still
a no-op.
- Reduce code duplication in the Jenson direct-SCB interrupt handlers.

Changes to make MPSAFE interrupts work on Alpha:

- Remove the ipl argument to scb_set() and the associated array of
"mpsafe" booleans initialized based on the ipl. It was bogus
anyway; all IPL_{BIO,NET,TTY}, etc. values are aliases of IPL_VM,
and for all practical purposes, there is really only one device
interrrupt level on Alpha anyway. Intead, we now treat all dispatches
from the SCB vector table as MP-safe, and it is now the handler for
that vector who is responsible for acquiring the KERNEL_LOCK if needed.

- Update the direct interrupt vector handlers in jensenio and TURBOchannel
to acquire the KERNEL_LOCK.

- Introduce a new ALPHA_INTR_MPSAFE flag, and add a flags argument to
alpha_shared_intr_establish(). When it is set, indicate that the
handler is MP-safe. Update alpha_shared_intr_dispatch() to pay
attention and acquire the KERNEL_LOCK (or not) as indicated.

- Re-factor all of the PCI interrupt handling, providing "generic PCI"
"PCI interrupts through ISA IRQs" implementations to significantly
reduce code duplication. Supplement the PCI chipset tag with more
info to facilitate this, and make the PCI interrupt-related routines
take a pci_chipset_tag_t argument rather than a void * argument.

- Because PCI interrupts on KN8AE are dispatched directly from the
SCB, provide a wrapper for non-MPSAFE interrupt handlers that
acquires the KERNEL_LOCK.

- Change the pci_intr_handle_t type to be a struct rather than an
integer type in order to catch any direct use of it as a value.
Add a set of functions to interact with pci_intr_handle_t, including
setting interrupt flags.

- Implement pci_intr_setattr() so that the PCI_INTR_MPSAFE attribute
can be set on a pci_intr_handle_t.

- While I'm here, make all of the MI PCI back-end operations call
through real functions rather than hopping directly through function
pointers in the chipset tag.

This change looks a lot bigger than it really is because of the re-factor
in the plethora of model-specific PCI interrupt back-ends. The KN8AE,
KN300, and T2/T3/T4 (Sable) are largely un-changed.

- Document all of the various interrupt levels in the Processor Stataus
register, and provide symbolic names for them as well.
- Use ALPHA_PSL_IPL_* values directly for IPL_*.

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.

branches: 1.33.30;
kill sprintf

branches: 1.32.6; 1.32.10;
Do a minor cleanup of alpha (this will make applying pullups post branching
easier).
u_int{8,16,32,64}_t -> uint{*}_t
Change all old-style definitions to C89 prototypes.
Whitespace cleanup.
Constification in db_disasm.c

branches: 1.31.2; 1.31.6;
struct device * -> device_t

branches: 1.30.10;
ANSIfy functions with function-pointer arguments

Change about 4500 of the K&R function definitions to ANSI ones.
There are still about 1600 left, but they have ',' or /* ... */
in the actual variable definitions - which my awk script doesn't handle.
There are also many that need () -> (void).
(The script does handle misordered arguments.)

Remove all the __P() from sys (excluding sys/dist)
Diff checked with grep and MK1 eyeball.
i386 and amd64 GENERIC and sys still build.

branches: 1.27.112; 1.27.120; 1.27.126;
remove trailing \n in panic(). approved perry.

branches: 1.26.2;
Rework the interrupt code, shaving some cycles off in the process.
Rather than an "iointr" routine that decomposes a vector into an
IRQ, we maintain a vector table directly, hooking up each "iointr"
routine at the correct vector. This also allows us to hook device
interrupts up to specific vectors (c.f. Jensen).

We can shave even more cycles off, here, and I will, but it requires
some changes to the alpha_shared_intr stuff.

branches: 1.25.6;
Switch to the new `evcnt' mechanism for counting interrupts. Maintain
a per-CPU interrupt counter for clock, device, and interprocessor
interrupts.

Implement the more flexiable `evcnt' interface as discussed (briefly) on
tech-kern and now documented in evcnt(9).

branches: 1.23.2;
- Remove unused bt459reg.h.
- sfbvar.h has gone, placing patch on holes for now. Every TC framebuffer
has a single explosure of xxfb_cnattach() which could be bridged by MI
land.

Fix DEC3000/300 TC slot numbering. There is an address space hole
between two TC option cards and 3 builtin TC devices. Was uncovered by
mrg's change made at Nov 1999 and pointed by tsutsui.

Merge IOASIC register definitions.

branches: 1.20.2;
clean up error handling in TC graphics console attachment: all functions
return 0 on success and an errno on failure

Leave out display console related bits if wscons is not compiled in.
Should fix link error reported by Kevin <kev@drule.org>.

Fix printf format problems on Alpha. (IRONIC?!)

DTRT if sfb is not configured in the kernel.

Preliminary support for both serial consoles and kbd/display wscons
consoles for TC alphas. Based on code from Takuya Koumoto
<takuya-k@is.aist-nara.ac.jp> as well as some parts of Toru Nishimura's
and Matthias Drochner's work on pmax wscons. This relies on the
NEW_SCC_DRIVER.

Nuke the idea of <machine/options.h>. It completely defeats the purpose
of fine-grain option dependencies.

branches: 1.14.4;
by default, provide RCS IDs for NetBSD/alpha kernel files in kernel binaries.
This can be disabled (to save a bit of space) with the NO_KERNEL_RCSIDS
options, which is present but commented out in the ALPHA config file.
In ELF-format kernels, these strings are present in the kernel binary but
are not loaded into memory. (In ECOFF-format kernels, there's no easy way
to keep them from being loaded, so they _are_ loaded into memory.)

clean up NetBSD RCS ID strings, include machine/options.h

branches: 1.12.2;
backout previous kprintf change

printf -> kprintf, sprintf -> ksprintf

(1) Clean up interrupt handling slightly.

clean and update for new defintions, prototypes, etc.

interrupt counters for TurboChannel systems.

branches: 1.7.4;
fix the test for option slot interrupts. If a given interrupt is
disabled (and therefore can't have caused the actual I/O interrupt),
don't try to call its handler even if its bit is set in the interrupt
register.

clean up copyrights and RCS IDs

update for latest MI TC and SCSI code, and clean up in the face of
config changes.

update for new turbochannel configuration code.

MB -> wbflush

cleanup, add some comments, and (oh yeah!) look at interrupt bits

update from my local tree; better CPU configuration, support for 3000/300.

spelling (DEC called it "TURBOchannel")

branches: 1.6.6; 1.6.24;
Do a minor cleanup of alpha (this will make applying pullups post branching
easier).
u_int{8,16,32,64}_t -> uint{*}_t
Change all old-style definitions to C89 prototypes.
Whitespace cleanup.
Constification in db_disasm.c

branches: 1.5.12; 1.5.16;
Remove all the __P() from sys (excluding sys/dist)
Diff checked with grep and MK1 eyeball.
i386 and amd64 GENERIC and sys still build.

branches: 1.4.152; 1.4.160; 1.4.166;
Preliminary support for both serial consoles and kbd/display wscons
consoles for TC alphas. Based on code from Takuya Koumoto
<takuya-k@is.aist-nara.ac.jp> as well as some parts of Toru Nishimura's
and Matthias Drochner's work on pmax wscons. This relies on the
NEW_SCC_DRIVER.

clean up NetBSD RCS ID strings

branches: 1.2.2;
update for new turbochannel configuration code.

update from my local tree; better CPU configuration, support for 3000/300.

Liberally sprinkle static around to get more symbols out of the
global namespace. A small bit of const poisoning in the TC code.

branches: 1.37.4;
malloc(9) -> kmem(9)

branches: 1.36.2;
Changes to make interrupt {,dis}establish MP-safe on Alpha:
- Protect all of the system interrupt linkage with the cpu_lock mutex.
- Re-order some of the stores to the SCB vector table to make it safe
in the face of lockless interrupt dispatch.
- Add a framework for routing interrupts to specific CPUs. Interrupts
are still funneled only to the primary CPU, but that will change for
some systems soon. Ensure that interrupt handler lists are manipulated
only on the CPUs that handle that specific interrupt source. This required
a re-factor of the alpha_shared_intr_*() family of functions.
- Enable __HAVE_INTR_CONTROL, although interrupt redistribution is still
a no-op.
- Reduce code duplication in the Jenson direct-SCB interrupt handlers.

Changes to make MPSAFE interrupts work on Alpha:

- Remove the ipl argument to scb_set() and the associated array of
"mpsafe" booleans initialized based on the ipl. It was bogus
anyway; all IPL_{BIO,NET,TTY}, etc. values are aliases of IPL_VM,
and for all practical purposes, there is really only one device
interrrupt level on Alpha anyway. Intead, we now treat all dispatches
from the SCB vector table as MP-safe, and it is now the handler for
that vector who is responsible for acquiring the KERNEL_LOCK if needed.

- Update the direct interrupt vector handlers in jensenio and TURBOchannel
to acquire the KERNEL_LOCK.

- Introduce a new ALPHA_INTR_MPSAFE flag, and add a flags argument to
alpha_shared_intr_establish(). When it is set, indicate that the
handler is MP-safe. Update alpha_shared_intr_dispatch() to pay
attention and acquire the KERNEL_LOCK (or not) as indicated.

- Re-factor all of the PCI interrupt handling, providing "generic PCI"
"PCI interrupts through ISA IRQs" implementations to significantly
reduce code duplication. Supplement the PCI chipset tag with more
info to facilitate this, and make the PCI interrupt-related routines
take a pci_chipset_tag_t argument rather than a void * argument.

- Because PCI interrupts on KN8AE are dispatched directly from the
SCB, provide a wrapper for non-MPSAFE interrupt handlers that
acquires the KERNEL_LOCK.

- Change the pci_intr_handle_t type to be a struct rather than an
integer type in order to catch any direct use of it as a value.
Add a set of functions to interact with pci_intr_handle_t, including
setting interrupt flags.

- Implement pci_intr_setattr() so that the PCI_INTR_MPSAFE attribute
can be set on a pci_intr_handle_t.

- While I'm here, make all of the MI PCI back-end operations call
through real functions rather than hopping directly through function
pointers in the chipset tag.

This change looks a lot bigger than it really is because of the re-factor
in the plethora of model-specific PCI interrupt back-ends. The KN8AE,
KN300, and T2/T3/T4 (Sable) are largely un-changed.

- Document all of the various interrupt levels in the Processor Stataus
register, and provide symbolic names for them as well.
- Use ALPHA_PSL_IPL_* values directly for IPL_*.

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.

branches: 1.32.30;
kill sprintf

branches: 1.31.6; 1.31.10;
Do a minor cleanup of alpha (this will make applying pullups post branching
easier).
u_int{8,16,32,64}_t -> uint{*}_t
Change all old-style definitions to C89 prototypes.
Whitespace cleanup.
Constification in db_disasm.c

branches: 1.30.2; 1.30.6;
struct device * -> device_t

branches: 1.29.10;
ANSIfy functions with function-pointer arguments

ANSIfy another 1261 function definitions.
The only ones left in sys are beyond by sed script!
(or in sys/dist or sys/external)
Mostly they have function pointer parameters.

Change about 4500 of the K&R function definitions to ANSI ones.
There are still about 1600 left, but they have ',' or /* ... */
in the actual variable definitions - which my awk script doesn't handle.
There are also many that need () -> (void).
(The script does handle misordered arguments.)

Remove all the __P() from sys (excluding sys/dist)
Diff checked with grep and MK1 eyeball.
i386 and amd64 GENERIC and sys still build.

branches: 1.25.112; 1.25.120; 1.25.126;
remove trailing \n in panic(). approved perry.

branches: 1.24.2;
Rework the interrupt code, shaving some cycles off in the process.
Rather than an "iointr" routine that decomposes a vector into an
IRQ, we maintain a vector table directly, hooking up each "iointr"
routine at the correct vector. This also allows us to hook device
interrupts up to specific vectors (c.f. Jensen).

We can shave even more cycles off, here, and I will, but it requires
some changes to the alpha_shared_intr stuff.

branches: 1.23.6;
Switch to the new `evcnt' mechanism for counting interrupts. Maintain
a per-CPU interrupt counter for clock, device, and interprocessor
interrupts.

Implement the more flexiable `evcnt' interface as discussed (briefly) on
tech-kern and now documented in evcnt(9).

branches: 1.21.2;
- Remove unused bt459reg.h.
- sfbvar.h has gone, placing patch on holes for now. Every TC framebuffer
has a single explosure of xxfb_cnattach() which could be bridged by MI
land.

branches: 1.20.2;
clean up error handling in TC graphics console attachment: all functions
return 0 on success and an errno on failure

Leave out display console related bits if wscons is not compiled in.
Should fix link error reported by Kevin <kev@drule.org>.

Fix printf format problems on Alpha. (IRONIC?!)

DTRT if sfb is not configured in the kernel.

Preliminary support for both serial consoles and kbd/display wscons
consoles for TC alphas. Based on code from Takuya Koumoto
<takuya-k@is.aist-nara.ac.jp> as well as some parts of Toru Nishimura's
and Matthias Drochner's work on pmax wscons. This relies on the
NEW_SCC_DRIVER.

Nuke the idea of <machine/options.h>. It completely defeats the purpose
of fine-grain option dependencies.

branches: 1.14.4;
by default, provide RCS IDs for NetBSD/alpha kernel files in kernel binaries.
This can be disabled (to save a bit of space) with the NO_KERNEL_RCSIDS
options, which is present but commented out in the ALPHA config file.
In ELF-format kernels, these strings are present in the kernel binary but
are not loaded into memory. (In ECOFF-format kernels, there's no easy way
to keep them from being loaded, so they _are_ loaded into memory.)

clean up NetBSD RCS ID strings, include machine/options.h

branches: 1.12.2;
Include a PMAGB-BA in the list of built-in devices only if SV_GRAPHICS
is set in the RPB's rpb_variation field. This fixes a bug where machines
(e.g. the 3000/900) would see that they could touch memory where a built-in
PMAGB-BA, assume that it was there, and panic later because the memory they
were accessing didn't look like PMAGB-BA registers (because it wasn't).

backout previous kprintf change

printf -> kprintf, sprintf -> ksprintf

(1) Clean up interrupt handling slightly.

clean and update for new defintions, prototypes, etc.

clean up copyright notices

interrupt counters for TurboChannel systems.

don't depend on being able to read the IMR. Apparently, it doesn't
work (as documented in the DEC 3000/[3-9]00 System Programmer's Manual)
in some of the older 3000/500 systems.

branches: 1.4.4;
add a bit of clarafication in a comment, and change the interrupt
handling code so that if a given interrupt is disabled (and therefore
can't have caused the actual I/O interrupt), its handler won't be
called even if its bit is set in the interrupt register.

update for new turbochannel configuration code.

MB -> wbflush

preliminary Alpha support. note that NOT ALL OF THE MODIFICATIONS TO
THE REST OF THE KERNEL ARE IN THE TREE YET. Also, some of this is
_incredibly_ hack-ish, etc., but it works.

spelling (DEC called it "TURBOchannel")

branches: 1.6.6; 1.6.24;
Do a minor cleanup of alpha (this will make applying pullups post branching
easier).
u_int{8,16,32,64}_t -> uint{*}_t
Change all old-style definitions to C89 prototypes.
Whitespace cleanup.
Constification in db_disasm.c

branches: 1.5.12; 1.5.16;
Remove all the __P() from sys (excluding sys/dist)
Diff checked with grep and MK1 eyeball.
i386 and amd64 GENERIC and sys still build.

branches: 1.4.152; 1.4.160; 1.4.166;
Preliminary support for both serial consoles and kbd/display wscons
consoles for TC alphas. Based on code from Takuya Koumoto
<takuya-k@is.aist-nara.ac.jp> as well as some parts of Toru Nishimura's
and Matthias Drochner's work on pmax wscons. This relies on the
NEW_SCC_DRIVER.

clean up NetBSD RCS ID strings

branches: 1.2.2;
update for new turbochannel configuration code.

preliminary Alpha support. note that NOT ALL OF THE MODIFICATIONS TO
THE REST OF THE KERNEL ARE IN THE TREE YET. Also, some of this is
_incredibly_ hack-ish, etc., but it works.

branches: 1.1.4;
Re-factor the boot device and consinit routines into common functions
per platform type (PCI vs TC), eliminating a bunch of copy-pasta.

there is no TC I/O space, so in the new bus.h world order there
is no tc_bus_io.c.

clean and update for new defintions, prototypes, etc.

implement bus_mem_subregion() and bus_io_subregion().

branches: 1.1.4;
bus_io_* functions. These have no meaning on the TurboChannel, since
it's a memory-mapped bus and has no I/O space. They all are
fancy wrappers around panic().

Remove unnecessary #include <sys/malloc.h>

Moar static.

branches: 1.38.22;
spelling (DEC called it "TURBOchannel")

More cleanups from Felix Deichmann (code) and me (panics).

From Felix Deichmann:
While writing "slhci at tc" support, I noticed TC bus_space functions
for alpha don't work as they should (tc_bus_mem.c). Consideration of
dense vs. sparse space seems problematic, especially for widths < 4.

I found information mainly in

[1] https://web-docs.gsi.de/~kraemer/COLLECTION/DEC/d3syspmb.pdf
[2]
http://h20565.www2.hpe.com/hpsc/doc/public/display?docId=emr_na-c04623255

and a dirty (but tested) hack based on this is attached as an example.
It would be great if someone could have a look and do a cleaner version
preferably.

Especially the end of ([2] section 12.2.2) provides a summary of which
"instruction" to use for which data width in which space, and possible
side effects (unintentional double reads/writes) in dense space...

branches: 1.35.6; 1.35.10;
remove unused code

branches: 1.34.6; 1.34.10;
Do a minor cleanup of alpha (this will make applying pullups post branching
easier).
u_int{8,16,32,64}_t -> uint{*}_t
Change all old-style definitions to C89 prototypes.
Whitespace cleanup.
Constification in db_disasm.c

branches: 1.33.2; 1.33.6;
change inline to static inline to appease gcc 4.5.

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

rework <uvm/uvm_extern.h> includes.

branches: 1.30.4;
ANSIfy another 1261 function definitions.
The only ones left in sys are beyond by sed script!
(or in sys/dist or sys/external)
Mostly they have function pointer parameters.

Change about 4500 of the K&R function definitions to ANSI ones.
There are still about 1600 left, but they have ',' or /* ... */
in the actual variable definitions - which my awk script doesn't handle.
There are also many that need () -> (void).
(The script does handle misordered arguments.)

Remove all the __P() from sys (excluding sys/dist)
Diff checked with grep and MK1 eyeball.
i386 and amd64 GENERIC and sys still build.

branches: 1.27.78; 1.27.86; 1.27.92;
merge ktrace-lwp.

Rename an inner variable shadowing the 'v' parameter to appease -Wshadow.

branches: 1.25.20;
Implement bus_space_mmap().

branches: 1.24.2; 1.24.4;
remove include of <vm/vm.h>. <vm/vm.h> -> <uvm/uvm_extern.h>

Fix typo in last commit.

implement bus_space_vaddr()

- Add a bus space method for getting the translation for a window.
- Add sysarch methods for "get bus window count", "get bus window",
and "pci conf read/write".

These are a hack, but they're what's necessary in order to make
XFree86 work in its current state.

Add an internal bus space method alpha_bus_space_translate(), which
provides a method to translate an address on an I/O bus into a sysBus
address, along with acccess method information.

branches: 1.19.8;
ovbcopy -> memmove

Provide a hook for bypassing space accounting, needed to support ISA PnP
for now.

branches: 1.17.6;
Nuke the idea of <machine/options.h>. It completely defeats the purpose
of fine-grain option dependencies.

branches: 1.16.2;
when i was documenting the bus_space interfaces (the document
was sent to developers, but will eventually become a man page or
something), i noticed a few in various names:
(1) _map and _alloc should take a flags argument, rather than a
'cacheable' boolean.
(2) BUS_BARRIER_* flags should be BUS_SPACE_BARRIER_*.
(3) bus_space_copy_* should be bus_space_copy_region_* for consistency
with other region ops.
Fix all of these (in a backward-compatible way, at least for now). Redefine
internal usees of those names to use the new names. Also, while at it,
clean up the copy functions (remove unnecessary variables) and make sure
that they and other functions conform to the spec.

by default, provide RCS IDs for NetBSD/alpha kernel files in kernel binaries.
This can be disabled (to save a bit of space) with the NO_KERNEL_RCSIDS
options, which is present but commented out in the ALPHA config file.
In ELF-format kernels, these strings are present in the kernel binary but
are not loaded into memory. (In ECOFF-format kernels, there's no easy way
to keep them from being loaded, so they _are_ loaded into memory.)

clean up NetBSD RCS ID strings, include machine/options.h

branches: 1.13.2;
implement bus_space_copy_*().

move barrier operation definition closer to the top of the bus space ops
structure. In the implementations, allow the barrier op to be inlined.

implement bus_space_set_multi_* and bus_space_set_region_* operations.

declare the single-datum read and write methods as inline, so that
the multi and region methods can inline them.

update for new bus.h macros. bus_io_* and bus_mem_* integrated into
single bus_space_* framework. Unfortunately, bus_space_{read,write}_*
operations still imply barriers. That will change soon.

update for new bus.h

clean and update for new defintions, prototypes, etc.

fix pasto in last commit.

since bus_mem_subregion() can return failure, don't even bother to try
to handle the case where subregion offset isn't a multiple of 8.

implement bus_mem_subregion() and bus_io_subregion().

gross hack to get around the fact that there are currently devices with
common back-ends that live on multiple very-different busses (e.g. PCI and
TC), which need bus-specific DMA mapping support. As a nice side effect,
this will allow the especially nasty (vtophys(va) | 0x40000000) expressions
to go away in favor of less nasty bus-specific function calls.

branches: 1.2.4;
add wbflush() calls in strategic places. These really shouldn't be
here, and there should be a generic 'bus' interface to do memory
read/write barriers.

TurboChannel bus_mem_* functions. bus_mem_{read,write}_8, and
bus_mem_read_{1,2} are not yet supported for sparse space.

Liberally sprinkle static around to get more symbols out of the
global namespace. A small bit of const poisoning in the TC code.

branches: 1.14.22;
spelling (DEC called it "TURBOchannel")

branches: 1.13.6; 1.13.24;
Do a minor cleanup of alpha (this will make applying pullups post branching
easier).
u_int{8,16,32,64}_t -> uint{*}_t
Change all old-style definitions to C89 prototypes.
Whitespace cleanup.
Constification in db_disasm.c

branches: 1.12.2; 1.12.6;
struct device * -> device_t

branches: 1.11.10;
Remove all the __P() from sys (excluding sys/dist)
Diff checked with grep and MK1 eyeball.
i386 and amd64 GENERIC and sys still build.

branches: 1.10.132; 1.10.140; 1.10.146;
Implement the more flexiable `evcnt' interface as discussed (briefly) on
tech-kern and now documented in evcnt(9).

branches: 1.9.12; 1.9.20;
Preliminary support for both serial consoles and kbd/display wscons
consoles for TC alphas. Based on code from Takuya Koumoto
<takuya-k@is.aist-nara.ac.jp> as well as some parts of Toru Nishimura's
and Matthias Drochner's work on pmax wscons. This relies on the
NEW_SCC_DRIVER.

Pull in DEC_3000_300 and DEC_3000_500 options.

branches: 1.7.4;
Pull thorpej-bus-dma branch into mainline.

branches: 1.6.2;
Oops, last commit was meant for the thorpej-bus-dma branch, not the
trunk.

Split DEC 3000/300 and 3000/500 DMA functions into separate files,
suggested by Chris Demetriou <cgd@netbsd.org>.

clean up NetBSD RCS ID strings

branches: 1.3.2;
Include a PMAGB-BA in the list of built-in devices only if SV_GRAPHICS
is set in the RPB's rpb_variation field. This fixes a bug where machines
(e.g. the 3000/900) would see that they could touch memory where a built-in
PMAGB-BA, assume that it was there, and panic later because the memory they
were accessing didn't look like PMAGB-BA registers (because it wasn't).

(1) Clean up interrupt handling slightly.

branches: 1.1.4;
update for new turbochannel configuration code.

Allow for the SGMAP implementation to specify a minimum alignment for
SGMAP DMA segments. If not specified, PAGE_SIZE will be used, as before.

branches: 1.14.6;
G/C alpha_XXX_dmamap() / alpha_XXX_dmamap_or. They haven't been needed
for a long time.

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

Ansify function definitions w/o arguments. Generated with sed.

branches: 1.11.8; 1.11.14;
Remove clause 3 and 4 from TNF licenses

branches: 1.10.2; 1.10.124; 1.10.126; 1.10.128;
Add support for mbufs to the Alpha SGMAP DMA maps.

branches: 1.9.26;
vm_offset_t -> {paddr_t,vaddr_t}, vm_size_t -> vsize_t

Allow the DMA tag to specify a boundary contraint. If the device has a
more strict boundary, the map will use it, otherwise the map will inherit
the tag's, unless the tag's constraint is 0 (no boundary constraint).

Add support for chaining DMA windows together, for falling back on
SGMAPs if a direct-mapped window fails.

Simplify the direct-mapped DMA case somewhat by adding a window base
member to the DMA tag, and calling the direct-mapped back-ends directly,
rather than through chipset-specific front-ends which pass the window
base as an additional argument.

Use the common _bus_dmamap_sync() as the _dmamap_sync method in the
bus_dma_tag_t.

G/c an unused function pointer.

Nuke the idea of <machine/options.h>. It completely defeats the purpose
of fine-grain option dependencies.

branches: 1.2.2; 1.2.6;
Pull thorpej-bus-dma branch into mainline.

branches: 1.1.2;
file tc_dma.c was initially added on branch thorpej-bus-dma.

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

Change about 4500 of the K&R function definitions to ANSI ones.
There are still about 1600 left, but they have ',' or /* ... */
in the actual variable definitions - which my awk script doesn't handle.
There are also many that need () -> (void).
(The script does handle misordered arguments.)

branches: 1.4.8; 1.4.14;
Remove clause 3 and 4 from TNF licenses

branches: 1.3.154; 1.3.156; 1.3.158;
Nuke the idea of <machine/options.h>. It completely defeats the purpose
of fine-grain option dependencies.

branches: 1.2.2; 1.2.6;
Pull thorpej-bus-dma branch into mainline.

branches: 1.1.2;
file tc_dma_3000_300.c was initially added on branch thorpej-bus-dma.

Remove all the __P() from sys (excluding sys/dist)
Diff checked with grep and MK1 eyeball.
i386 and amd64 GENERIC and sys still build.

branches: 1.3.8; 1.3.14;
Remove clause 3 and 4 from TNF licenses

branches: 1.2.2; 1.2.160; 1.2.162; 1.2.164;
Pull thorpej-bus-dma branch into mainline.

branches: 1.1.2;
file tc_dma_3000_300.h was initially added on branch thorpej-bus-dma.

Allow for the SGMAP implementation to specify a minimum alignment for
SGMAP DMA segments. If not specified, PAGE_SIZE will be used, as before.

branches: 1.23.4;
malloc(9) -> kmem(9)

branches: 1.22.2;
Add some bus_dma instrumentation.

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.

branches: 1.20.6;
spelling (DEC called it "TURBOchannel")

branches: 1.19.6;
remove unused variable

branches: 1.18.6; 1.18.10;
Do a minor cleanup of alpha (this will make applying pullups post branching
easier).
u_int{8,16,32,64}_t -> uint{*}_t
Change all old-style definitions to C89 prototypes.
Whitespace cleanup.
Constification in db_disasm.c

branches: 1.17.2; 1.17.6;
#include <sys/bus.h> instead of <machine/bus.h>.

rework <uvm/uvm_extern.h> includes.

branches: 1.15.4;
Change about 4500 of the K&R function definitions to ANSI ones.
There are still about 1600 left, but they have ',' or /* ... */
in the actual variable definitions - which my awk script doesn't handle.
There are also many that need () -> (void).
(The script does handle misordered arguments.)

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

branches: 1.13.2; 1.13.124; 1.13.126; 1.13.128;
Add support for mbufs to the Alpha SGMAP DMA maps.

bcopy -> memcpy

bzero -> memset

branches: 1.10.4;
remove include of <vm/vm.h>. <vm/vm.h> -> <uvm/uvm_extern.h>

Include sys/device.h so that this file compiles again; file dev/tc/tcvar.h
now contains the definition of tc_softc.

branches: 1.8.14; 1.8.20;
Allow the DMA tag to specify a boundary contraint. If the device has a
more strict boundary, the map will use it, otherwise the map will inherit
the tag's, unless the tag's constraint is 0 (no boundary constraint).

Add support for chaining DMA windows together, for falling back on
SGMAPs if a direct-mapped window fails.

Simplify the direct-mapped DMA case somewhat by adding a window base
member to the DMA tag, and calling the direct-mapped back-ends directly,
rather than through chipset-specific front-ends which pass the window
base as an additional argument.

Use the common _bus_dmamap_sync() as the _dmamap_sync method in the
bus_dma_tag_t.

Put SGMAP-related stuff in the DMA map structure directly, rather than
indirecting through a pointer.

Nuke the idea of <machine/options.h>. It completely defeats the purpose
of fine-grain option dependencies.

branches: 1.2.2; 1.2.6;
Pull thorpej-bus-dma branch into mainline.

branches: 1.1.2;
file tc_dma_3000_500.c was initially added on branch thorpej-bus-dma.

Remove all the __P() from sys (excluding sys/dist)
Diff checked with grep and MK1 eyeball.
i386 and amd64 GENERIC and sys still build.

branches: 1.3.8; 1.3.14;
Remove clause 3 and 4 from TNF licenses

branches: 1.2.2; 1.2.160; 1.2.162; 1.2.164;
Pull thorpej-bus-dma branch into mainline.

branches: 1.1.2;
file tc_dma_3000_500.h was initially added on branch thorpej-bus-dma.

move tc/tc_machdep.h to include/tc_machdep.h. (can't do it via
repository copy because alpha already had an include/tc_machdep.h in
the attic.)

Get the per-slot DMA tag a bit differently.

Pull thorpej-bus-dma branch into mainline.

branches: 1.4.2;
clean up NetBSD RCS ID strings

branches: 1.3.2;
update for new bus.h

clean and update for new defintions, prototypes, etc.

portability definitions for machine-independent TC configuration code.

Remove unnecessary #include <sys/malloc.h>

branches: 1.9.6;
Add some bus_dma instrumentation.

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

rework <uvm/uvm_extern.h> includes.

branches: 1.6.22;
Remove clause 3 and 4 from TNF licenses

branches: 1.5.2; 1.5.124; 1.5.126; 1.5.128;
Add support for mbufs to the Alpha SGMAP DMA maps.

branches: 1.4.4;
remove include of <vm/vm.h>. <vm/vm.h> -> <uvm/uvm_extern.h>

branches: 1.3.18;
Nuke the idea of <machine/options.h>. It completely defeats the purpose
of fine-grain option dependencies.

branches: 1.2.2; 1.2.6;
Pull thorpej-bus-dma branch into mainline.

branches: 1.1.2;
file tc_sgmap.c was initially added on branch thorpej-bus-dma.

spelling (DEC called it "TURBOchannel")

branches: 1.4.6; 1.4.24;
Do a minor cleanup of alpha (this will make applying pullups post branching
easier).
u_int{8,16,32,64}_t -> uint{*}_t
Change all old-style definitions to C89 prototypes.
Whitespace cleanup.
Constification in db_disasm.c

branches: 1.3.34; 1.3.38;
Remove clause 3 and 4 from TNF licenses

branches: 1.2.2; 1.2.160; 1.2.162; 1.2.164;
Pull thorpej-bus-dma branch into mainline.

branches: 1.1.2;
file tc_sgmap.h was initially added on branch thorpej-bus-dma.

Merge thorpej-cfargs2.

branches: 1.51.6;
Static'ify more symbols.

branches: 1.50.2;
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: 1.49.4;
Changes to make interrupt {,dis}establish MP-safe on Alpha:
- Protect all of the system interrupt linkage with the cpu_lock mutex.
- Re-order some of the stores to the SCB vector table to make it safe
in the face of lockless interrupt dispatch.
- Add a framework for routing interrupts to specific CPUs. Interrupts
are still funneled only to the primary CPU, but that will change for
some systems soon. Ensure that interrupt handler lists are manipulated
only on the CPUs that handle that specific interrupt source. This required
a re-factor of the alpha_shared_intr_*() family of functions.
- Enable __HAVE_INTR_CONTROL, although interrupt redistribution is still
a no-op.
- Reduce code duplication in the Jenson direct-SCB interrupt handlers.

Changes to make MPSAFE interrupts work on Alpha:

- Remove the ipl argument to scb_set() and the associated array of
"mpsafe" booleans initialized based on the ipl. It was bogus
anyway; all IPL_{BIO,NET,TTY}, etc. values are aliases of IPL_VM,
and for all practical purposes, there is really only one device
interrrupt level on Alpha anyway. Intead, we now treat all dispatches
from the SCB vector table as MP-safe, and it is now the handler for
that vector who is responsible for acquiring the KERNEL_LOCK if needed.

- Update the direct interrupt vector handlers in jensenio and TURBOchannel
to acquire the KERNEL_LOCK.

- Introduce a new ALPHA_INTR_MPSAFE flag, and add a flags argument to
alpha_shared_intr_establish(). When it is set, indicate that the
handler is MP-safe. Update alpha_shared_intr_dispatch() to pay
attention and acquire the KERNEL_LOCK (or not) as indicated.

- Re-factor all of the PCI interrupt handling, providing "generic PCI"
"PCI interrupts through ISA IRQs" implementations to significantly
reduce code duplication. Supplement the PCI chipset tag with more
info to facilitate this, and make the PCI interrupt-related routines
take a pci_chipset_tag_t argument rather than a void * argument.

- Because PCI interrupts on KN8AE are dispatched directly from the
SCB, provide a wrapper for non-MPSAFE interrupt handlers that
acquires the KERNEL_LOCK.

- Change the pci_intr_handle_t type to be a struct rather than an
integer type in order to catch any direct use of it as a value.
Add a set of functions to interact with pci_intr_handle_t, including
setting interrupt flags.

- Implement pci_intr_setattr() so that the PCI_INTR_MPSAFE attribute
can be set on a pci_intr_handle_t.

- While I'm here, make all of the MI PCI back-end operations call
through real functions rather than hopping directly through function
pointers in the chipset tag.

This change looks a lot bigger than it really is because of the re-factor
in the plethora of model-specific PCI interrupt back-ends. The KN8AE,
KN300, and T2/T3/T4 (Sable) are largely un-changed.

spelling (DEC called it "TURBOchannel")

Make VAXstation 4000 TURBOchannel adaptor work.
PIO and IRQ tested on VS4000/60 and /96 with TC-USB (slhci) and PMAD-AA (le).
DMA is untested.
vax needs space to be mapped out, so memory space is pre-mapped for TC option
drivers not bus_space'ified yet, also providing custom bus_space methods for
this kludge.

branches: 1.45.8;
Clean-up (remove redundant declarations, concentrate checks).

branches: 1.44.6; 1.44.24; 1.44.28;
Do a minor cleanup of alpha (this will make applying pullups post branching
easier).
u_int{8,16,32,64}_t -> uint{*}_t
Change all old-style definitions to C89 prototypes.
Whitespace cleanup.
Constification in db_disasm.c

branches: 1.43.2; 1.43.6;
Split device_t/softc. pmax is tested on GXemul.
Compile test only for alpha and vax.

XXX: alpha/tc/tcasic.c calls tcattach() via config_found(9), but
XXX: pmax/tc/tcbus.c and vax/vsa/tc_vsbus.c calls tcattach() directly.
XXX: It looks horrible...

branches: 1.42.4; 1.42.6; 1.42.10;
Change about 4500 of the K&R function definitions to ANSI ones.
There are still about 1600 left, but they have ',' or /* ... */
in the actual variable definitions - which my awk script doesn't handle.
There are also many that need () -> (void).
(The script does handle misordered arguments.)

Remove all the __P() from sys (excluding sys/dist)
Diff checked with grep and MK1 eyeball.
i386 and amd64 GENERIC and sys still build.

branches: 1.40.18; 1.40.26; 1.40.32;
Interrupt handling changes, in discussion since February:

- Reduce available SPL levels for hardware devices to none, vm, sched, high.
- Acquire kernel_lock only for interrupts at IPL_VM.
- Implement threaded soft interrupts.

branches: 1.39.18; 1.39.60; 1.39.78; 1.39.80; 1.39.86;
Use aprint_normal() for cfprint routines.

Use CFATTACH_DECL().

Declare all cfattach structures const.

branches: 1.36.2;
Add MX (PMAG-A), CX (PMAG-B), TX (PMAG-RO/PMAG[B]-J) TC framebuffer
supports. None of them has any acceleration capability.

Rework the interrupt code, shaving some cycles off in the process.
Rather than an "iointr" routine that decomposes a vector into an
IRQ, we maintain a vector table directly, hooking up each "iointr"
routine at the correct vector. This also allows us to hook device
interrupts up to specific vectors (c.f. Jensen).

We can shave even more cycles off, here, and I will, but it requires
some changes to the alpha_shared_intr stuff.