more patches: arlad/messages.c, fs.c

[Love]

Nickolai Zeldovich <kolya at mit.edu> writes:

> Looks like I've ran into another arlad bug tonight -- in viocgetacl,
> the opaque struct is never initialized, and thus contains random
> stack garbage. If the getacl call fails, opaque is passed without
> initialization to xfs_send_message_wakeup_data() which tries to
> memcpy from it..

The initization really need to be inside the finst_first_fs/find_next_fs
loop, since the RPC might modify the data-structure. I made it so.
 
> The second patch makes fs behave more like the Transarc fs, in that
> it will not stop doing listacl if it encounters EACCES.

Ok, thanks.

Love