> > Of course we could just tell FS authors to be carefull and copy byte by 
> > byte ;-)
> 
> Actually this option is not as bad as it initially sounded - creating the
> real fid on the stack and memcpy to the pointer passed should work in this
> case.
> 
> Guess I should convert all vptofh() in tree... - any objections? Example
> (uncompiled, untested) patch appended.

btw, will you take a look at fhtovp as well?

YAMAMOTO Takashi