|Re: [hatari-devel] FPU update|
[ Thread Index |
| More lists.tuxfamily.org/hatari-devel Archives
On 01/23/2017 10:10 PM, Douglas Little wrote:
Yes I'm aware of this problem (diagnosing the cause) - but I didn't really
want to do anything about it unless/until somebody raised a reason for it.
I wrote it mainly because two of my programs were flaking out on some
Falcons, and it turned out to be faulty FPUs.
I think it wouldn't be too hard to record all of the hashes, instead of
just the final hash on each test - and stop on the first error. It doesn't
necessarily generate too much data. However all it would do is halt near
the fault - you'd still need to debug the instruction before the fault.
Hatari debugger history command can be used to check what
instructions were called before the error and registers
command lists all the register values (including FP ones).
On next run, one can break on the suspicious instruction,
set debugger to show regs after every instruction:
and then just step forward.