> Just allowing misaligned accesses shouldn't cause a m68k to panic.
> Perhaps there is another problem which was masked by the previous
> problem, that is, now that gdb gets through reading the registers,
> it does something else which causes your kernel to panic.

Right, basically I haven't had alot of time to investigate the problems
however, it seems that things get pretty wacky when ptrace is used.
I noticed that (if ddb trace is to be counted on..) a call to 
uiomove durring a PT_READ_I (caused panic) was for the address
0x0e000254 (or something similar, while the address requested was
0x00002194 (or something similar.)  This is obviously bogus.  Note
too (may/may not be important) 0x0e000000 is VM_MAXUSER_ADDRESS and the
base of the amiga user stack.  Anyway it doesn't seem to happen
right away.  When I was trying to catch it I set some breakpoints
however it hit them many times before crapping out (i.e. I finally gave
up and deleted the breakpoint.. it then panic'd after..of course :/)

Note that the alignment change was just part of a larger overhaul
(complete re-write I believe) of ptrace.  It now uses procfs_xxx()
calls.  Basically there is probably a bug in these somewhere no of
this stuff has been tested yet (well I guess its being tested now)

I think that ptrace() was overhauled for poltical reasons beyond
our control.

Chris.