[jcifs] Multi-fragment PDUs

[Michael B Allen]

Elrond,

Maybe you can help me with an RPC question? Our RPC code does not properly
handle multi-fragment PDUs. It handles PDUs that fit into a single
transaction just fine but when I try to read an additional fragment I get
ACCESS_DENIED.

Using a Win32 test program I have observed the following buffering for
multi-fragment PDUs:

frag 1: 5680
   trans: 1024 @ 0
    read: 4293 @ 0
    read: 363  @ 4293
frag 2: 5680
    read: 4293 @ 4656 (5680 - 1024)
    read: 1387 @ 8949
    ...

with Jarapac/JCIFS I'm trying to do the following:

frag 1: 4280
   trans: 4280 @ 0
frag 2: 4280
    read: 4280 @ 0  <- ACCESS DENIED

So clearly one thing that is quite different here is that the fragment
fits entirely within the transaction. So when I try to read the next
fragment I use offset 0.

Any advice?

Thanks,
Mike