Add GCC macros to detect x86 compilation and add x86 compatible code

serialcons_putc is not available on ARM platforms.

Update hw for benchmarking and api changes

The hw platform is maintained in order to run benchmark comparisons with
the seL4 platform. As we develop the seL4 platform we change some of
the library apis. This commit adds:
- Extra interrupt numbers to the PIC
- A serial input interrupt handler for saving, resetting and printing
the tsc for measuring idle time
- Changes default console output to use COM port instead of VGA
- Changes for changed library functions
- A rumprun-bake configuration for only using ethernet modules

space -> tab indent

no functional change (one would at least hope ...)

Treat routed and non-routed interrupts separately.

For systems with routed interrupts, we don't know exactly
how they're routed (since we don't parse acpi tables or
mpbios or any of that rubbish). However, for bog-standard
ISA interrupts we do know where they'll appear. So avoid
calling the non-routed interrupt handler in case of a routed
interrupt and vice versa.

also voidify intr_init()

p-p-p-p-panic: make bmk_isr_rumpkernel() type void

In other words: just panic if something doesn't work out.
It's ~100% likely that the resulting state would not be what
the user wanted in case we'd return an error.

Rename bmk_isr_init -> bmk_isr_rumpkernel

The established handler will always call the rump kernel, so
reflect that in the name. Also, give it a flags parameter
(always 0 for now).

Disable "stray interrupt" print

Doesn't convey anything useful in its current state.

<bmk-rumpuser/core_types.h> needed here too

make stray interrupt print "work" under BMK_INTRLEVS == 1

Make isr_todo a volatile.

It's set in the interrupt handler and examined in thread context.

Flag interrupts we tried to handle as processed for cpu_intr_ack().

If we mark only ones which were acknowledged by devices, buggy
drivers may disable interrupts unwillingly. Of course, the flipside
is that buggy drivers with level-triggered interrupts will perpetually
raise the interrupt. Not quite what to do about that. But at least
ARM/integratorcp interrupts work now (or, will work after the next, MD
commit).

pass info on which interrupts where handled to cpu_intr_ack()

Pass interrupt as a mask instead of shift to isr()

On some platforms (e.g. ARM) we get the mask but not the shift(s),
so this makes like a bit easier there.

actually commit the code changes too (for previous commit)

remove bmk_ prefix from routines private to platform/hw

hw: Consolidate clock-related code

This commit consolidates the common clock-related functions for the x86
and amd64 hw platform into platform/hw/arch/x86/clock.c.

No functional change intended.

if compiled with BMK_SCREW_INTERRUPT_ROUTING, use only one level

Start introducing memory allocation arenas to bmk_memalloc.

Step 1: include information on who is allocation memory
in the memory allocation requests.

use assert() for invariants

judge stray interrupts per wakeup, not per loop

Rename baremetal -> hw, as discussed on rumpkernel-users

Migration: use "hw" where you would have previously used "baremetal".