On Thu, Nov 22, 2018 at 08:59:32AM +0100, Cédric Le Goater wrote:
> On 11/22/18 7:50 AM, Benjamin Herrenschmidt wrote:
> > On Thu, 2018-11-22 at 15:44 +1100, David Gibson wrote:
> >>
> >> Sorry, didn't think of this in my first reply.
> >>
> >> 1) Does the hardware ever actually write back to the EAS?  I know it
> >> does for the END, but it's not clear why it would need to for the
> >> EAS.  If not, we don't need the setter.
> > 
> > Nope, though the PAPR model will via hcalls
> 
> Indeed. The H_INT_SET_SOURCE_CONFIG hcall updates the EAT.
> 
> >> 2) The signatures are a bit odd here.  For the setter, a value would
> >> make sense than a (XiveEAS *), since it's just a word.  For the getter
> >> you could return the EAS value directly rather than using a pointer -
> >> there's already a valid bit in the EAS so you can construct a value
> >> with that cleared if the lisn is out of bounds.
> 
> Yes we could. I think I made it that way to be consistent with the 
> other XIVE internal structures which are bigger : END, NVT

Yeah, but as noted elsewhere I don't really like the get/set model for
the bigger-than-word-size structures.  It gives the impression that
they're atomic updates when they can't be, as well as unnecessarily
copying a bunch of stuff, sometimes on hot paths

> There might be other reasons in Pnv. One was to use generic accessors 
> to the guest RAM but I didn't do it finally. Take a look at the Pnv
> model and we might decide to change the prototype then. I don't 
> think it's a major change.

Hmmm.

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson