[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: single software TLB vs. multiple software TLBs (was RE: [Xen-ia64-devel] [PATCH] [Resend]Enable hash vtlb)


  • To: "Tristan Gingold" <Tristan.Gingold@xxxxxxxx>, "Xu, Anthony" <anthony.xu@xxxxxxxxx>, <xen-ia64-devel@xxxxxxxxxxxxxxxxxxx>
  • From: "Dong, Eddie" <eddie.dong@xxxxxxxxx>
  • Date: Wed, 12 Apr 2006 22:51:09 +0800
  • Delivery-date: Wed, 12 Apr 2006 07:51:26 -0700
  • List-id: Discussion of the ia64 port of Xen <xen-ia64-devel.lists.xensource.com>
  • Thread-index: AcZeIOmsV1lih27VSzi/Kix2Hgw3mQAHwG0Q
  • Thread-topic: single software TLB vs. multiple software TLBs (was RE: [Xen-ia64-devel] [PATCH] [Resend]Enable hash vtlb)

Tristan Gingold wrote:
>> 
>> Just make sure I am understanding correctly. With the help of
>> software TLB for guest huge translation, you have the flexibility to
>> either breakdown it into smaller ones to VHPT for all entries at a
>> time, or you can do one at a time. Defenitely several at a time is
>> also OK if people want to :-) 
> I don't understand how it can work without writing all entries.  Xen
> doesn't know which entry will be used.
At every tlb miss time, you can get guest translation from software TLB
(not from VHPT).
You actually don't need to care about VHPT entries no matter it is all
there or nothing there.
Then using p2m information, Xen can generate machine translation for the
fault address and then insert to VHPT or TLB base on other policy.

Thx,eddie

_______________________________________________
Xen-ia64-devel mailing list
Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-ia64-devel


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.