Handle loop reconfiguration in a kernel thread, instead of doing it from
interrupt context.

Postpone the hil console detection logic to the first detection of a keyboard.

On hp300, hil would claim console against dnkbd if no dnkbd was found at
the time the loop is probed, even if the loop is empty. Because of this,
plugging dnkbd later would not select it as console keyboard, which is
really annoying on kernels without wsmux, such as hp300 RAMDISK.
Now the first keyboard plugged will become the console keyboard, whatever
its type.

No functional change on hppa, since the console path gives a definite console
device setting.

Reliability fixes:
- Let the loop initialize completely before attempting to probe its devices.
Fixes the "no answer from device 1" problem.
- Handle ``loop unplugged'' events and force detach of all children in this
case.

Allow send_hil{,dev}_cmd to return failure, and handle this where
applicable.

During device probe, if a device does not answer commands, display a warning
message. This apparently happens on hp300 when the console is configured
as remote (i.e. serial console). Unplugging and replugging the device works
fine afterwards...

Remove the advertising clause in the UCB license which Berkeley
rescinded 22 July 1999. Proofed by myself and Theo.

branches: 1.5.4;
New hil child devices attachment framework, that allow them to be detached
at runtime.

Handle reconfiguration notices from the loop, and do the necessary
detach/attach work so that our vision of the loop is in sync with reality.

Adapt all hil child devices to the above changes.

"This is not as plug'n'play as usb, but you get the same feeling anyways..."

Real polled mode console support, enough to play hangman in ddb.
Still a minor issue left for tomorrow.

Rework the console management on hppa:
- only attach a keyboard as a console if it matches the PDC keyboard path
- on hil, as there can be multiple keyboards on the loop, attach only the
first hilkbd device configured as console keyboard. Right now this means
the one with the lowest hil code, which was the existing behaviour so far.
- do not try to switch to the wscons consdev structure early at all in
wscons_machdep, but rather wait for the console to be completely
configured (i.e. both wskbd and wsdisplay are attached) to switch.

With feedback and help from mickey@

- Better type usage in the various hil code: prefer u_int8_t to u_char, but
do not stick to u_int8_t when native word size can do the job better.
- Allow send_hildev_cmd() to return the command response buffer to its
caller, rather than forcing it to look at the guts of its parent device
softc... this will be needed shortly.

Machine independent HP-HIL loop support code.

Derived from the hp300 HIL code, and some information found in XFree86
HP-UX specific parts.

However, this code does not provide an HP-UX compatible /dev/hil* interface,
but will rather attach real BSD drivers to the hil driver glue.

Currently, only a driver for the HP-HIL keyboards is provided. More to come
as resources permit.

The international layout tables for hilkbd are derived from the ite tables
found in the hp300 code, but only the US layout could be tested.

Sample dmesg output on a heavily charged hil loop:
hil0 at gsc0 offset 21000 irq 1
hilkbd0 at hil0 code 1: 109-key keyboard, layout 1b
wskbd0 at hilkbd0: console keyboard
hilkbd1 at hil0 code 2: 109-key keyboard, layout 1f
wskbd1 at hilkbd1
"ID module" at hil0 id 34 code 3 not configured
"ID module" at hil0 id 34 code 4 not configured
"Tablet" at hil0 id 94 code 5 not configured
"Mouse" at hil0 id 68 code 6 not configured

Some feedback from and ok mickey@