Re: Questions on gdbsx

Re: Questions on gdbsx

[Mukesh Rathor]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=x-gbk; format=flowed, Size: 1118 bytes --]



yujiageng734 wrote:
 > Hi,
 >
 >     I am trying to use your gdbsx.  But I don't know how to make a 32bit
 > dom0 running on 64bit hypervisor. Can you explain this for me?  I'm
 > extremely grateful to you.

Hmm.. good question. I don't know, it's always done for me here. One
option would be to just download OVM from edelivery.oracle.com, it's
free, and should ahve both binary and sources. You could look at spec
file to see how they do it.


 >     By the way, could you give me some tools to debug hypervisor and
 > monitor the process of hypervisor start. Any thought will be thankful.

To debug hypervisor, you'll need kdb. You could get that from OVM (xen
3.1.4) also, or http://xenbits.xensource.com/ext/debuggers.hg  for
unstable. See README-dbg. If you put earlykdb in xen kernel line in
grub.conf, it would break into kdb during boot very early. Then you can
single step, set breakpoints, etc...


Mukesh


 > Yu Jiageng,
 > China.
 >
 > ------------------------------------------------------------------------
 > ÍøÒ×ÓÊÏ䣬ÖйúµÚÒ»´óµç×ÓÓʼþ·þÎñÉÌ <http://www.yeah.net>

How expensive is a VM EXIT?

[Emre Can Sezer]

Hi,

I've been trying to get a ballpark number on how expensive a VM EXIT is 
in terms of clock cycles. I'm also interested in how slow a TLB miss is. 
The only reference I was able to find was 13000 cycles for VM EXIT and 
around several hundred cycles for a TLB miss and a complete walk of page 
tables. Does anyone know from the top of their head or is there a 
reference you can share?

Thanks in advance.

John

Re: How expensive is a VM EXIT?

[Keir Fraser]

For just the VMEXIT itself, it's probably about 1000-1500 cycles on the
latest processors. I haven't measured it very recently though.

 -- Keir

On 03/03/2009 15:35, "Emre Can Sezer" <ecsezer@ncsu.edu> wrote:

> Hi,
> 
> I've been trying to get a ballpark number on how expensive a VM EXIT is
> in terms of clock cycles. I'm also interested in how slow a TLB miss is.
> The only reference I was able to find was 13000 cycles for VM EXIT and
> around several hundred cycles for a TLB miss and a complete walk of page
> tables. Does anyone know from the top of their head or is there a
> reference you can share?
> 
> Thanks in advance.
> 
> John
> 
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel

Re: How expensive is a VM EXIT?

[NAHieu]

On Wed, Mar 4, 2009 at 1:34 AM, Keir Fraser <keir.fraser@eu.citrix.com> wrote:
> For just the VMEXIT itself, it's probably about 1000-1500 cycles on the
> latest processors. I haven't measured it very recently though.
>

Could you explain how do you measure the cycles for VMEXIT?

Thanks,
H


>  -- Keir
> - Show quoted text -
> On 03/03/2009 15:35, "Emre Can Sezer" <ecsezer@ncsu.edu> wrote:
>
>> Hi,
>>
>> I've been trying to get a ballpark number on how expensive a VM EXIT is
>> in terms of clock cycles. I'm also interested in how slow a TLB miss is.
>> The only reference I was able to find was 13000 cycles for VM EXIT and
>> around several hundred cycles for a TLB miss and a complete walk of page
>> tables. Does anyone know from the top of their head or is there a
>> reference you can share?
>>
>> Thanks in advance.
>>
>> John
>>
>>
>> _______________________________________________
>> Xen-devel mailing list
>> Xen-devel@lists.xensource.com
>> http://lists.xensource.com/xen-devel
>
>
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel
>

Re: How expensive is a VM EXIT?

[Keir Fraser]

On 04/03/2009 00:15, "NAHieu" <nahieu@gmail.com> wrote:

> On Wed, Mar 4, 2009 at 1:34 AM, Keir Fraser <keir.fraser@eu.citrix.com> wrote:
>> For just the VMEXIT itself, it's probably about 1000-1500 cycles on the
>> latest processors. I haven't measured it very recently though.
>> 
> 
> Could you explain how do you measure the cycles for VMEXIT?

RDTSC in the guest immediately before triggering a VMEXIT and RDTSC in the
hypervisor on entry.

 -- Keir

RE: How expensive is a VM EXIT?

[Nadolski, Ed]

Is there a good tool to look at for this -- xentrace, xenoprofile, other?

Thanks,
Ed

-----Original Message-----
From: xen-devel-bounces@lists.xensource.com [mailto:xen-devel-bounces@lists.xensource.com] On Behalf Of Keir Fraser
Sent: Wednesday, March 04, 2009 1:46 AM
To: NAHieu
Cc: xen-devel@lists.xensource.com
Subject: Re: [Xen-devel] How expensive is a VM EXIT?

On 04/03/2009 00:15, "NAHieu" <nahieu@gmail.com> wrote:

> On Wed, Mar 4, 2009 at 1:34 AM, Keir Fraser <keir.fraser@eu.citrix.com> wrote:
>> For just the VMEXIT itself, it's probably about 1000-1500 cycles on the
>> latest processors. I haven't measured it very recently though.
>> 
> 
> Could you explain how do you measure the cycles for VMEXIT?

RDTSC in the guest immediately before triggering a VMEXIT and RDTSC in the
hypervisor on entry.

 -- Keir



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

Re: How expensive is a VM EXIT?

[Keir Fraser]

Either of those can give you profiling info about common paths through the
hypervisor, and time spent in the hypervisor. For the specific
microbenchmark of time taken for a processor to execute a vmexit, I'd hack
in custom tracing to get a precise measurement.

 -- Keir

On 04/03/2009 18:39, "Nadolski, Ed" <Ed.Nadolski@lsi.com> wrote:

> Is there a good tool to look at for this -- xentrace, xenoprofile, other?
> 
> Thanks,
> Ed
> 
> -----Original Message-----
> From: xen-devel-bounces@lists.xensource.com
> [mailto:xen-devel-bounces@lists.xensource.com] On Behalf Of Keir Fraser
> Sent: Wednesday, March 04, 2009 1:46 AM
> To: NAHieu
> Cc: xen-devel@lists.xensource.com
> Subject: Re: [Xen-devel] How expensive is a VM EXIT?
> 
> On 04/03/2009 00:15, "NAHieu" <nahieu@gmail.com> wrote:
> 
>> On Wed, Mar 4, 2009 at 1:34 AM, Keir Fraser <keir.fraser@eu.citrix.com>
>> wrote:
>>> For just the VMEXIT itself, it's probably about 1000-1500 cycles on the
>>> latest processors. I haven't measured it very recently though.
>>> 
>> 
>> Could you explain how do you measure the cycles for VMEXIT?
> 
> RDTSC in the guest immediately before triggering a VMEXIT and RDTSC in the
> hypervisor on entry.
> 
>  -- Keir
> 
> 
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel