RE: [Xen-devel] PSE in guests

["Ian Pratt" <m+Ian.Pratt@xxxxxxxxxxxx>]

> Looking at   the 2.0 code, it looks like the guest OSs are made to
> believe that PSE is not available. Is this a design choice? 
> Has any work been done on supporting 4M pages in guest kernels?

We did some experiments using 4M pages to map the kernel, but didn't get
any speedup as recorded by lmbench.

We'd like to get hugetlbfs working so that databases can use it at some
point, but it's a bit subtle: Xen uses linear (recursive) pagetables
internally for performance reasons, and the dumb x86 page table format
means that its not possible to cause a fault to occur if an attempt is
made to use a super page entry in the PGD as a PTE (as a result of the
recursive map).  We'll need to add secondary checks :-(

Ian


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_ide95&alloc_id396&op=click
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxxx
https://lists.sourceforge.net/lists/listinfo/xen-devel