evbarm/generic_v5: fix setstatclockrate crash
setstatclockrate is an optional function for timers that currently cannot be multiplexed through device trees. We don't provide an implementation in the am18 code, but we need to make sure the imx23 driver doesn't crash us.
The 68020 does not have an on-chip data cache, so:
- Make the comments reflect that.
- no-op PCIA(), as well, since DC_CLEAR doesn't do anything on a 68020.
We should not be intermingling ATC/TLB and I/D cache operations, but
for now ensure that the TBI*() functions are only made available for
Moto and/or HP MMU configurations.
Consolidate a bunch of sub-includes of various config definitions
that were extremely-but-not-universally-common across the various
m68k platforms into files.m68k.
Apparently we have kgdb_machdep.c and kgdb_m68k.c, the latter is already
in files.m68k, and the former everyone uses the same one, so move it to
files.m68k, too.
Obviously, more cleanup in this area is going to be needed.
sun2 and sun3 don't *really* need their own trimmed-down copy of
sys_machdep.c; the common version was already perfectly suitable
for sun3, and can be easily conditionally-trimmed-down for the
68010 case.
(Sharing is caring!)
Move the do-nothing(-except-DIAGNOSTICly-panic) m68881_save/restore
stubs to switch_subr.s so that future 68010 ports don't have to
declare them on their own.