All,

OK, last question for today (except maybe follow ups... ;) ).

I'm in the process of updating the DayStar '030 accelerator patch from
1.3 to 1.4.  For the most part, this has just been a matter of
updating the patch diffs for different offsets in the source files.
There is one exception however.  In mac68k/pmap.c, a function is
defined to map the daystar hardware.  This function calls pmap_enter()
to do this.  In 1.3 the interface to pmap_enter() was:

    void
    pmap_enter(pmap, va, pa, prot, wired)
            register pmap_t pmap;
            vm_offset_t va;
            register vm_offset_t pa;
            vm_prot_t prot;
            boolean_t wired;

However, in 1.4, pmap_enter now has an additional parameter, access_type: 

    void
    pmap_enter(pmap, va, pa, prot, wired, access_type)
            pmap_t pmap;
            vaddr_t va;
            paddr_t pa;
            vm_prot_t prot;
            boolean_t wired;
            vm_prot_t access_type;

What is the purpose of "access_type", and how is it different from
"prot"?  Or perhaps more to the point, the map_daystar() function in
1.3 called pmap_enter with, eg:

        pmap_enter (pmap_kernel(), 0x52040000, 0x52040000,
                    VM_PROT_READ | VM_PROT_WRITE, TRUE);

What should be used for access_type?  I tried copying the flags passed
to prot:

        pmap_enter (pmap_kernel(), 0x52040000, 0x52040000,
                    VM_PROT_READ | VM_PROT_WRITE, TRUE,
                    VM_PROT_READ | VM_PROT_WRITE);

and that *seems* to work OK.  The addresses mapped w/ pmap_enter() are
only tested (read), eg:

    asm ("tstb      0x52040000 | disable writeback cache");

Does that mean access_type should only be VM_PROT_READ (or does it not
really matter in this case)?

Thanks,

John