Define missing unifycr_gfid replacements for various unifycr_fid functions

Nobody has claimed this yet.

In CRUISE, the fid was used as both a file identification number and a reference into data structures holding the file's metadata and storage. However, each process "owned" its file and files were not shared across processes.

In UnifyCR, the fid was carried over from CRUISE, and also a new global file id was created to uniquely identify a file across all processes. The fid is still useful given the context of a particular client, as a reference into data structures for the file, e.g., the list of storage blocks allocated to a file.

However, many of the old unifycr_fid_* functions, like unifycr_fid_truncate() and unifycr_fid_unlink() need to be updated to account for global ids. This can be accomplished by replacing many of these calls with gfid equivalents or by keeping with fids. Alternatively, we could keep using fid values here, but we need to add a function so a process can allocate a fid for a file if it does not already have one assigned to that file.

Start by tracing the mentioned unifycr_fid_truncate() and unifycr_fid_unlink() functions and locating their gfid counterparts. Review the other unifycr_fid_* callers to identify which require global IDs and which remain client-local. Done means the missing replacements are defined consistently, or the required fid-allocation behavior is specified and implemented.

Written by the indexing model from the issue text.