Mail Archives: djgpp/1999/10/08/00:10:44
Eli Zaretskii wrote:
> > crash.log always ends with:
> > FPU unavailable at eip=4f19; flags=3246
> > eax=00000000 ebx=00000007 ecx=37fcb98d edx=37fcb98d
> > esi=00098f3b edi=00001eca ebp=00098920 esp=000988e8 cs=a7
> > ds=af es=af fs=8f gs=bf ss=af error=0006
>
> This message comes from CWSDPMI, not from Make or the emulator. I have
>never seen anything like this before. If this problem persists after you
>start using the patched emulator, you will have to investigate some
>more. No one of the development team has access to an FPU-less machine,
>so all we can do is set 387=n and assume this produces the same behavior
>as on a machine without an FPU.
Just in case it's helpful to know..
I've seen something very similar to this (variously "FPU unavailable"
and a dump like the above, or "FP exception" with no dump) from CWSDPMI
on a system with a known-buggy CPU (Intel P24T "Overdrive" which is an
83MHz 486DX. This CPU also caused OS/2 to commit seppuku with an
"internal processing error at location [long address]".)
What hardware do you need for testing, and what needs to be run to get a
good test? I have these boards here that I could set up with some
reasonable amt of memory (up to 16mb) if you've got something smallish
that would work to test against 'em:
386SX16 w/ 387SX16
386SX40, no mathco
386DX40 w/ 33MHz mathco
386DX40, no mathco
486SX25, no mathco.
Yes, I do have a weird collexion of old hardware :)
~REZ~
- Raw text -