read/write should use libnfs's fh offset

[jsgf]

libnfs maintains its own per-FH offset for use with read/write/seek, so the wrapper should use that rather than always using an offset of 0. Or alternatively not provide read/write and just expose pread/pwrite.

[cholcombe973]

Interesting. How do you access that?

[jsgf]

Oh, I meant using the nfs_read/write libnfs API calls.