Re: [RFC 0/3] netlink: extended error reporting

[Pablo Neira Ayuso <pablo-Cap9r6Oaw4JrovVCs/uTlw@public.gmane.org>]

On Fri, Apr 07, 2017 at 12:22:23PM -0700, David Miller wrote:
> From: Johannes Berg <johannes-cdvu00un1VgdHxzADdlk8Q@public.gmane.org>
> Date: Fri, 07 Apr 2017 21:09:45 +0200
> 
> > On Fri, 2017-04-07 at 21:06 +0200, Pablo Neira Ayuso wrote:
> >> On Fri, Apr 07, 2017 at 08:59:12PM +0200, Johannes Berg wrote:
> >> [...]
> >> > Heh. I think I really want to solve - at least partially -
> >> > nla_parse()
> >> > to see that it can be done this way. It'd be nice to even transform
> >> > all
> >> > the callers (I generated half of these patches with spatch anyway)
> >> > to
> >> > have at least that.
> >> 
> >> We can just have a modified version of nla_parse that deals with
> >> this.
> > 
> > Yes, but we need to figure out a good way to have the offset.
> > 
> > We also need to see if we want to *force* having the offset. In some
> > sense that'd be useful, in another it might be very complicated to fill
> > it in at all times, if for example errors come from lower layers like
> > drivers.
> 
> It has to be optional, some kinds of errors don't have an exact
> context per-se.
> 
> Also another way to look at this is that we're providing a lot of
> new power and expressability.  So even if only one aspect of the
> new error reporting is used it's a positive step forward.
> 
> So allow offset "0" meaning "unspecified".

Instead, we can just not send the offset attribute to userspace if
it's not specified. So missing attribute means "unspecified".

I'm always a bit worried this "0 means something" semantics :)