Re: [PATCH 2/2] arm/xen: Add misuse warning to virt_to_gfn

[Stefano Stabellini <sstabellini@xxxxxxxxxx>]

On Thu, 27 Aug 2020, Jürgen Groß wrote:
> On 27.08.20 10:24, Julien Grall wrote:
> > 
> > 
> > On 27/08/2020 06:21, Jürgen Groß wrote:
> > > On 26.08.20 20:37, Julien Grall wrote:
> > > "Usually" is a bit gross here. The only generic call site I could find
> > > is xenbus_grant_ring(). All other instances (I counted 22) are not
> > > generic at all.
> > > 
> > > > will only catch one instance and it means we would have to fix the first
> > > > instance and then re-run to catch the others.
> > > > 
> > > > So I think we want to switch to WARN_ON() here.
> > > 
> > > No, please don't. In case there would be a frequent path the result
> > > would be a basically unusable system due to massive console clobbering.
> > 
> > Right, but if that's really happenning then you have a much bigger problem
> > on your platform because the address returned will be invalid.
> > 
> > So I still don't see the advantage of WARN_ON_ONCE() here.
> 
> Depends of the (potential) source of the warnings. I think we can agree
> that e.g. a problem in the pv network stack is rather improbable, as it
> would have been detected long ago.
> 
> If, however, the problem is being introduced by one of the rather new
> pv-drivers (like sound, pvcalls, 9pfs) it is perfectly fine to assume
> the overall system is still functional even without those drivers
> working correctly. Having a message storm from those sources is still
> quite undesirable IMO and doesn't really help.

I agree with this.