Hi Christos,
     Just a quick one...

     The patches I submitted in the pr are in a terrible format. I will
repost them as a diff -ruN from inside sys/, to the pr tonight.
     Please note I have still not quite got my head all around the amd64
pmap so I am not sure if this change is even worthwhile.... However,
i've been running these patches on my development/main pc at home and
I haven't noticed any problems. Also, the patches only do the pool
thing. I have not ported any of openbsd's other pmap changes.
    On another note, I am in the process of rewriting the page table code
on amd64 to make use of the long mode 2M page size feature and the
change to the page table structure for this page size. (I have
rewritten pte.h but the changes emmanate through other code as well,
eg. pmap.{h,c} & locore.S) Are you someone I can ask a question of
here and there with respect to the amd64 implementation? PS: This is
only a hobby for me so don't hold your breath :-) I'm also thinking of
going through locore.S as well...

Take care,
    murray

> On Sep 26,  2:59pm, murray@river-styx.org (murray@river-styx.org) wrote:
> -- Subject: Re: amd64 pmap & using a pool for pv_entry
>
> | Hi,
> |     My code changes have been submitted as PR port-amd64/34618
>
> Thanks, I don't have an amd64 machine yet, so it will be a while before
> I get a chance to test them. Hopefully someone else will.
>
> christos
>