[Jeremy Fitzhardinge <jeremy@zip.com.au> writes:]
>> There was some PCI work quite a while back, which got far enough to map in
>> the card and run some of its code. I recall that there was a real problem
>> that the only rules for what BIOS-type code was allowed to do was basically
>> "if it works for Windoze...".
>Yeah, I was forced to stop working on that approach because it broke completel
>y
>when I changed motherboards/bioses. The BIOS32 mechanism which is supposed to
>allow a 32-bit kernel make use of BIOS code is flawed, because there's no way
>of telling what memory it wants access to, so you can't do the appropriate
>mmaps.
Is there any possibility we could install a SEGV handler and manipulate the
environment on the fly?
>There's also the secondary problem that you can't identity map all
>parts of the physical address space. I also needed to change the mmap()
>syscall itself to allow the use of MAP_FIXED and MAP_PHYS together, but I
>haven't looked at recent kernels to see if this is still the case.
Well, changing the kernel to allow this (especially for a privileged
process) would be the least of the problem.
>I suspect that using direct IO to the PCI chipset is more likely to work in th
>e
>VSTa environment.
If there's a modest set of chip types to support, I'm sure this is workable.
It seems like the PC video arena is an example where there's just barely a
contained enough set for this approach to work.
Andy
Received on Fri Jun 26 08:12:53 1998
This archive was generated by hypermail 2.1.8 : Thu Sep 22 2005 - 15:12:43 PDT