> > Reinoud's patch is somewhat reasonable...  
> > subr_disk_mbr.c:readdisklabel() is the only version of that function  
> > that does not return "no disk label" when no label is found, as far  
> > as I can tell.
> > 
> > What, exactly, depends on the semantics of it NOT returning that  
> > message?
> 
> Since the message gets written to the console, you get an annoying message
> every time you access a 'vnd' disk that contains a single filesystem,
> a raidframe partition that isn't subdivided, or a CD or ....

if you want to suppress the message, readdisklabel is not
a right place to do it, IMO.

YAMAMOTO Takashi