Mikolaj,

Unfortunately this is an enormous undertaking.  The problem is that we 
need to be interoperable both at the disk level and network level with 
32 bit platforms. 

The first task is to get 32bit types for the Volume/Vnode etc
definitions into coda.h and compile kernel code and potemkin (under
util).  Potemkin is a Coda simulator and speaks the same language with
the kernel as Venus but mounts a directory from the disk filesystem.

Potemkin on AXP would be a great step forward.  Then lwp and rpc2 need
to be done - but those are easier.  BTW, the cutting and pasting from
ARLA may give us copyright problems.

If you start doing this keep me posted -- since you'll be making large 
amounts of little changes.

- Peter -




Mikolaj J. Habryn writes:
 >   I've decided to spend some time toying with coda on my alpha. It seems
 > that this might be an unpleasant undertaking. I have dropped the alpha
 > lwp assembly routines from arla into the 4.7.2 release, and fixed a
 > bunch of pointer conversion warning while compiling, but this seems
 > entirely secondary next to the requirement of sticking 64 bit pointers
 > into the PARM structure which is commented as being (needing to be?) 
 > four bytes in size. Who should I be talking to for help?
 > 
 > m.