This is only available with the UII(+) version of C128DM and is intended for
debugging your own software as well as helping out with debugging possible
issues in DM.

When the 'extended break handler' in dm config is enabled, a brk instruction
will trigger writing a complete ram dump to /temp/crashdump*

This can be used to inspect ram as it was when the brk instruction was encountered
 
The crashdump consists of a number of files:
crashdump.prg  : this file contains metadata and register information
crashdump0.prg : ram block 0, $0200 - $ffff
crashdump1.prg : ram block 1, $0200 - $ffff
crashdump2.prg : color ram, zeropage, stack, $0000-$0200 of ram block 2 (see below)
crashdumpv.prg : first 16k VDC ram

If VDC was configured for 64k ram when brk was encountered, there will be 3 more
files: crashdumpw.prg, crashdumpx,prg and crashdumpy.prg, each containing 16k
VDC ram


crashdump2.prg format:

The default load address for this file is $2000.

$2000-$20ff : zeropage
$2100-$21ff : stack
$2200-$22ff : page 0 of ram block 1
$2300-$23ff : page 1 of ram block 1
$2400-$27ff : color ram block 0
$2800-$2bff : color ram block 1

The crashdump format is identical to the snapshot format, the difference between
a snapshot and crashdump is detected by DM based on the metadata

Documentation on the metadata format is for now not available as this is still
a work in progress.

The crashdump.prg file can be loaded and run, this will get you the ml monitor
as if the brk was just encountered.
