_C_LABEL() and _ASM_LABEL() are no longer useful in the "everything
is ELF" world. Eliminate use of them in m88k code.

ok aoyama@

Unnecessary include

Switch m88k ports to ELF.

Drop the pm_cpus bitmask field from struct pmap, and instead remember the
currently active userland pmap in each processors struct cpu_info.

This thus skips the complete tlb flush if idle switches b

Drop the pm_cpus bitmask field from struct pmap, and instead remember the
currently active userland pmap in each processors struct cpu_info.

This thus skips the complete tlb flush if idle switches back to
the proc previously running on this processor.

show more ...

When switching processes, do not reenable interrupts until pmap_activate()
has been invoked on the new process.

In cpu_switchto(), always pmap_deactivate() the process being switched out,
even in non-MP kernels, to avoid unnecessary tlb flushes later when
pmap operates on shared pages.

Be sure to pmap_deactivate() a process during context switches, so that
the cpu which runs it is accounted correctly in MP kernels.

Disable interrupts around changing curproc and curpcb so these always match.

Rely on 16 byte pcb alignment, and use double loads and stores during
context switches. Should have been commited ages ago (when pcb alignment
was fixed). No functional change.

It is no longer necessary to fiddle with spl in cpu_idle_{enter,leave} now
that proc_trampoline has been fixed.

Make context switching much more MI:
- Move the functionality of choosing a process from cpu_switch into
a much simpler function: cpu_switchto. Instead of having the locore
code walk the run q

Make context switching much more MI:
- Move the functionality of choosing a process from cpu_switch into
a much simpler function: cpu_switchto. Instead of having the locore
code walk the run queues, let the MI code choose the process we
want to run and only implement the context switching itself in MD
code.
- Let MD context switching run without worrying about spls or locks.
- Instead of having the idle loop implemented with special contexts
in MD code, implement one idle proc for each cpu. make the idle
loop MI with MD hooks.
- Change the proc lists from the old style vax queues to TAILQs.
- Change the sleep queue from vax queues to TAILQs. This makes
wakeup() go from O(n^2) to O(n)

there will be some MD fallout, but it will be fixed shortly.
There's also a few cleanups to be done after this.

deraadt@, kettenis@ ok

show more ...

On multiprocessor kernels, update p->p_pcu on exit from the scheduler.

Be more generous with interrupt disabling in the scheduler.

Stop saving and restoring the current ipl in the pcb when switching processes,
since we know we are at splsched().

Work in progress SMP code; mvme88k boards can spin up secondary CPUs,
kernel boots single user. Still a lot of polishing and bugfixing to do.

Better use of the scheduler lock; no functional change on UP kernels.

Switch m88k ports to __HAVE_CPUINFO. Current cpu pointer is held in SR0
on all running processors.
Tested aoyama@ and I

In the core of the idle loop, be sure to invoke spl0 more than once to
process soft interrupts; fixes luna88k hanging after the rootdev line.
Tested aoyama@ martin@ and I

Stop mapping the u area at fixed UADDR in addition to its actual va.
While there, attempt to clean and comment stack usage in the kernel.
No functional change.

From the m88k SMP tree; help&test mart

Stop mapping the u area at fixed UADDR in addition to its actual va.
While there, attempt to clean and comment stack usage in the kernel.
No functional change.

From the m88k SMP tree; help&test martin@

show more ...

General scheduler cleanup:
- at the end of switch_exit(), jump in a better location inside cpu_switch()
- skip unnecessary curproc == NULL tests, and remove other dead code
- simplify the idle loop
-

General scheduler cleanup:
- at the end of switch_exit(), jump in a better location inside cpu_switch()
- skip unnecessary curproc == NULL tests, and remove other dead code
- simplify the idle loop
- optimize whichqs bit operations
- and add comments at strategic places

help&test martin@

show more ...

Take advantage of the recent ABI change to use less stack space and do
less stack pointer manipulation; the kernel needs to be compiled by an
up-to-date compiler now, or a tree will fall on your hous

Take advantage of the recent ABI change to use less stack space and do
less stack pointer manipulation; the kernel needs to be compiled by an
up-to-date compiler now, or a tree will fall on your house.

show more ...

No need to make __savectx visible and 8-byte aligned.

When restoring a process' ipl, invoke setipl() on behalf on the kernel stack,
rather than on the process stack.

Do not trash r14 by mistake in __savectx.

Fix alignment problem between savectx and __savectx. How could this have
ever worked?