Hmmm.. I'll have to think about this some. 

No... the code I added to st_interpret_sense only modifies
whether st_interpret_sense will call scsi_print_sense.
The return values from st_interpret_sense still stay the same:
0 for no error, -1 for 'continue processing' or errno of some kind.

In scsi_interpret_sense the case statement I modified is already
set to return zero (error = 0). I'm returning instead of doing
a break only to avoid the call to scsi_print_sense (or the
non-SCSIVERBOSE version).

So, am I still not getting your point? Let me know, and I'll
go ponder for a while (during a 50 mile drive down to Los Gatos
to pick up a test system I have to work on,......)