Re: Important for fs devs: rcu-walk merged upstream

[Miklos Szeredi]

On Fri, 14 Jan 2011, Nick Piggin wrote:
> On Fri, Jan 14, 2011 at 7:45 PM, Miklos Szeredi <miklos@xxxxxxxxxx> wrote:
> > On Fri, 14 Jan 2011, Nick Piggin wrote:
> >> On Fri, Jan 14, 2011 at 3:18 AM, Miklos Szeredi <miklos@xxxxxxxxxx> wrote:
> 
> >> Surely you'd need some filtering anyway? I don't think any function
> >> involving path lookup could sanely return -ECHILD.
> >
> > No, but not filtering doesn't normally hurt. ÂAnd it's not quite
> > trivial deciding what should be allowed and what shoudln't, and the
> > filter would have to be updated for each addition of a new errno. ÂSo
> > I'm not sure I want to go there.
> 
> Well if you allow untrusted filesystems it is possible that
> -ECHILD return will do something a bit silly. So it would be
> good to filter it I guess.

What could it do (other than resulting in a silly error printout)?

If we'd want errno filtering for fuse how would you define "sane"?

> >> That said, it probably is a good idea to have a new errno.
> >
> > Yeah, that makes the fitering much easier.
> 
> How so? Would -ECHILD ever be sane to return? I'm not
> arguing against changing it but I just want to know what
> the issue is there.

All kernel private errnos are >=512 (see <linux/errno.h), filtering
out those is quite easy and clearly desirable.

Thanks,
Miklos
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/