I am converting some code, written for DOS, that accesses a VME interface,
using huge pointers. There is a lot of bitshifting and masking
going on that I do not quite understand. Is there a safe and clean procedure
to convert this code and get it to run correctly under Linux?
Right now my approach is
1) to get io permission on the specified ports to do the command
passing (which works fine), and then to
2) mmap the physical addresses of the dual ported ram to the virtual
address space of my (suid) Linux process.
Getting the addresses right is the main problem.
Does anyone have a clean recipe of how to solve this problem?
Any help is much appreciated.