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

Re: [Xen-devel] [PATCH] Xenperf Bug Fix


  • To: "Wilson, Stephen" <Stephen.Wilson@xxxxxxx>, <xen-devel@xxxxxxxxxxxxxxxxxxx>
  • From: Keir Fraser <Keir.Fraser@xxxxxxxxxxxx>
  • Date: Wed, 31 Oct 2007 16:19:12 +0000
  • Delivery-date: Wed, 31 Oct 2007 09:20:02 -0700
  • List-id: Xen developer discussion <xen-devel.lists.xensource.com>
  • Thread-index: Acgb2HSwkSVjGOLcSleXzHkFaQkbsQAAVEW4
  • Thread-topic: [Xen-devel] [PATCH] Xenperf Bug Fix

If the VMEXIT reasons aren’t densely encoded (which they actually are, except for NPT), then rather than making the array an order of magnitude bigger it’s probably better either to define your own dense encoding (e.g., remap NPT==1024 to a lower number) or give NPT exits their own singleton counter which you increment in the existing ‘case VMEXIT_NPF’ in svm_vmexit_handler()).

 -- Keir

On 31/10/07 16:09, "Wilson, Stephen" <Stephen.Wilson@xxxxxxx> wrote:


This patch fixes/enables the following:
 
* Fix typo bug in xenperf tool.
* Increase the size of percpu data area from 4K to 8K to accommodate larger PERFCOUNTER_ARRAY.
* Increase the size of svmexits PERFCOUNTER_ARRAY to include nested paging perf counter.
 
Signed-off-by: Stephen Wilson <stephen.wilson@xxxxxxx>
 
--Stephen




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

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

 


Rackspace

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