Re: [PATCH 1/3] btrfs: return errno instead of -1 from compression

[David Sterba <dsterba@suse.cz>]

On Fri, May 09, 2014 at 01:40:15PM -0700, Zach Brown wrote:
> > > @@ -335,7 +335,7 @@ cont:
> > >  					break;
> > >  
> > >  				if (page_in_index + 1 >= total_pages_in) {
> > > -					ret = -1;
> > > +					ret = -EIO;
> > 
> > That looks like an internal error, we should never ask for more pages
> > than is in the input, so the buffer offset calculations are wrong.
> 
> Yeah, but EIO is still arguably the right thing.  There's nothing
> userspace can do about broken kernel code.  We don't want to give them
> an error that could be misinterpreted as them having used an interface
> incorrectly.  We could wrap WARN_ON_ONCE() around it, I suppose.  I'm
> inclined to leave it as is.

Ok. An EIO is better than an ASSERT or BUG_ON, the error paths will
handle it. I think adding the WARN_ON_ONCE is a good compromise, I'm not
expecting to see this error but the stack trace will help.