All of lore.kernel.org
 help / color / mirror / Atom feed
* Re: [Xen-devel] Re: Xen 4 - Page Sharing HowTo?
@ 2010-04-09 16:28 Nick Couchman
  2010-04-09 16:40 ` Re: [Xen-users] " Keir Fraser
  0 siblings, 1 reply; 6+ messages in thread
From: Nick Couchman @ 2010-04-09 16:28 UTC (permalink / raw)
  To: Keir Fraser, Andrew Lyon, pasik; +Cc: xen-devel, XEN Mailing List

> 
> Should be unconditionally printed during Xen boot. Not when memory sharing
> is first used.

Yep, I see it!  What's the best way to tell how much memory or how many pages are being shared?  Will Xen's memory status (e.g. with xentop or xm list) show lower memory usage per machine?  Or is there somewhere else I need to dig to find how much memory is shared?

> 
>>> Also the 4.0 release announcement mentioned page to disk, but again I
>>> can find no reference to that... has anybody tested these new
>>> features?
> 
> I think you run the xenpaging daemon manually on each domain you want to
> have paged to disc. The daemon is probably under /usr/sbin/xenpaging. It
> takes two arguments: domain-id and memory allocated to the domain, in 4kB
> pages (e.g., memory in MB, multiplied by 256).
> 
>  -- Keir
> 

I think the RPMs I have are built without this - I cannot seem to locate it.  I'll have to see if I can compile it myself.

Thanks!

-Nick



--------
This e-mail may contain confidential and privileged material for the sole use of the intended recipient.  If this email is not intended for you, or you are not responsible for the delivery of this message to the intended recipient, please note that this message may contain SEAKR Engineering (SEAKR) Privileged/Proprietary Information.  In such a case, you are strictly prohibited from downloading, photocopying, distributing or otherwise using this message, its contents or attachments in any way.  If you have received this message in error, please notify us immediately by replying to this e-mail and delete the message from your mailbox.  Information contained in this message that does not relate to the business of SEAKR is neither endorsed by nor attributable to SEAKR.

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: Re: [Xen-users] Xen 4 - Page Sharing HowTo?
  2010-04-09 16:28 [Xen-devel] Re: Xen 4 - Page Sharing HowTo? Nick Couchman
@ 2010-04-09 16:40 ` Keir Fraser
  2010-04-09 20:45   ` Nick Couchman
  0 siblings, 1 reply; 6+ messages in thread
From: Keir Fraser @ 2010-04-09 16:40 UTC (permalink / raw)
  To: Nick Couchman, Andrew Lyon, pasik; +Cc: xen-devel, XEN Mailing List

On 09/04/2010 17:28, "Nick Couchman" <Nick.Couchman@seakr.com> wrote:

>> Should be unconditionally printed during Xen boot. Not when memory sharing
>> is first used.
> 
> Yep, I see it!  What's the best way to tell how much memory or how many pages
> are being shared?  Will Xen's memory status (e.g. with xentop or xm list) show
> lower memory usage per machine?  Or is there somewhere else I need to dig to
> find how much memory is shared?

Sharing opportunities should result in memory being freed back to Xen. So
'xm info', or whatever else you use for monitoring Xen's free memory, should
work.

 -- Keir

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: Re: [Xen-users] Xen 4 - Page Sharing HowTo?
  2010-04-09 16:40 ` Re: [Xen-users] " Keir Fraser
@ 2010-04-09 20:45   ` Nick Couchman
  2010-04-10  6:27     ` [Xen-devel] " Keir Fraser
  0 siblings, 1 reply; 6+ messages in thread
From: Nick Couchman @ 2010-04-09 20:45 UTC (permalink / raw)
  To: Keir Fraser, Andrew Lyon, pasik; +Cc: xen-devel, XEN Mailing List

> 
> Sharing opportunities should result in memory being freed back to Xen. So
> 'xm info', or whatever else you use for monitoring Xen's free memory, should
> work.
> 
>  -- Keir

Hmmm...two XP HVM domUs running, and a total of 7 M freed to Xen.  Kind of hard to tell if memory sharing is working, or if this is normal rounding behavior...

-Nick



--------
This e-mail may contain confidential and privileged material for the sole use of the intended recipient.  If this email is not intended for you, or you are not responsible for the delivery of this message to the intended recipient, please note that this message may contain SEAKR Engineering (SEAKR) Privileged/Proprietary Information.  In such a case, you are strictly prohibited from downloading, photocopying, distributing or otherwise using this message, its contents or attachments in any way.  If you have received this message in error, please notify us immediately by replying to this e-mail and delete the message from your mailbox.  Information contained in this message that does not relate to the business of SEAKR is neither endorsed by nor attributable to SEAKR.

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [Xen-devel] Re: Xen 4 - Page Sharing HowTo?
  2010-04-09 20:45   ` Nick Couchman
@ 2010-04-10  6:27     ` Keir Fraser
  0 siblings, 0 replies; 6+ messages in thread
From: Keir Fraser @ 2010-04-10  6:27 UTC (permalink / raw)
  To: Nick Couchman, Andrew Lyon, pasik; +Cc: xen-devel, XEN Mailing List

On 09/04/2010 21:45, "Nick Couchman" <Nick.Couchman@seakr.com> wrote:

>> Sharing opportunities should result in memory being freed back to Xen. So
>> 'xm info', or whatever else you use for monitoring Xen's free memory, should
>> work.
>> 
>>  -- Keir
> 
> Hmmm...two XP HVM domUs running, and a total of 7 M freed to Xen.  Kind of
> hard to tell if memory sharing is working, or if this is normal rounding
> behavior...

The two images have to be using a common disc image (or a delta of a common
disc image) exported by a memshr-aware block-device backend: blktap1 or
blktap2. In fact, would have to be the former, since blktap2 has memshr
support disabled as it was crashing. The sharing opportunities aren't
discovered by page scanning, as in some other schemes.

 -- Keir

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: Re: [Xen-users] Xen 4 - Page Sharing HowTo?
  2010-04-09 14:59   ` Re: [Xen-users] " Andrew Lyon
@ 2010-04-09 15:22     ` Keir Fraser
  0 siblings, 0 replies; 6+ messages in thread
From: Keir Fraser @ 2010-04-09 15:22 UTC (permalink / raw)
  To: Andrew Lyon; +Cc: xen-devel, XEN Mailing List

On 09/04/2010 15:59, "Andrew Lyon" <andrew.lyon@gmail.com> wrote:

> ERROR Internal error: Error flushing ioemu cache
> ERROR Internal error: Error mapping page
> 
> Perhaps there are hardware requirements for page to disk?
> 
> Any idea what I need to do to have memory sharing support compiled
> into blktap2? The combination of memory sharing + blktap2 results in
> this error when I try to start a vm:
> 
> using config file "./jos-ts3.cfg".
> Error: Failed to create device.
>     stdout:
>     stderr: MEMSHR support not compiled in.
> Check that target "/mnt/tb1_store/Xen_VMs/WinXP/WinXP.hda" exists and
> that blktap2 driver installed in dom0.

Tbh I think these features are barely even tech preview in the 4.0.0
release. They got little testing and the devs who implemented them are also
off doing other things at least part time. I hope they will get some more
love before 4.0.1.

 -- Keir

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: Re: [Xen-users] Xen 4 - Page Sharing HowTo?
  2010-04-09 14:36 ` [Xen-devel] " Keir Fraser
@ 2010-04-09 14:59   ` Andrew Lyon
  2010-04-09 15:22     ` Keir Fraser
  0 siblings, 1 reply; 6+ messages in thread
From: Andrew Lyon @ 2010-04-09 14:59 UTC (permalink / raw)
  To: Keir Fraser; +Cc: xen-devel, XEN Mailing List

On Fri, Apr 9, 2010 at 3:36 PM, Keir Fraser <keir.fraser@eu.citrix.com> wrote:
> On 09/04/2010 15:09, "Pasi Kärkkäinen" <pasik@iki.fi> wrote:
>
>>>> void mem_sharing_init(void)
>>>> {
>>>>    printk("Initing memory sharing.\n");
>>>>
>>>>
>>>> But I've not seen that at all on my 4.x system.
>
> Should be unconditionally printed during Xen boot. Not when memory sharing
> is first used.
>
>>> Also the 4.0 release announcement mentioned page to disk, but again I
>>> can find no reference to that... has anybody tested these new
>>> features?
>
> I think you run the xenpaging daemon manually on each domain you want to
> have paged to disc. The daemon is probably under /usr/sbin/xenpaging. It
> takes two arguments: domain-id and memory allocated to the domain, in 4kB
> pages (e.g., memory in MB, multiplied by 256).
>
>  -- Keir

Hmm, I tried that but as soon as I run xenpaging it outputs lots of errors:


ERROR Internal error: Error mapping page
ERROR Internal error: Error mapping page
ERROR Internal error: Error mapping page
ERROR Internal error: Error mapping page
ERROR Internal error: Error flushing ioemu cache
ERROR Internal error: Error mapping page
ERROR Internal error: Error mapping page
ERROR Internal error: Error mapping page
ERROR Internal error: Error mapping page
ERROR Internal error: Error mapping page
ERROR Internal error: Error mapping page
ERROR Internal error: Error mapping page

Perhaps there are hardware requirements for page to disk?

Any idea what I need to do to have memory sharing support compiled
into blktap2? The combination of memory sharing + blktap2 results in
this error when I try to start a vm:

using config file "./jos-ts3.cfg".
Error: Failed to create device.
    stdout:
    stderr: MEMSHR support not compiled in.
Check that target "/mnt/tb1_store/Xen_VMs/WinXP/WinXP.hda" exists and
that blktap2 driver installed in dom0.

Andy

^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2010-04-10  6:27 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2010-04-09 16:28 [Xen-devel] Re: Xen 4 - Page Sharing HowTo? Nick Couchman
2010-04-09 16:40 ` Re: [Xen-users] " Keir Fraser
2010-04-09 20:45   ` Nick Couchman
2010-04-10  6:27     ` [Xen-devel] " Keir Fraser
  -- strict thread matches above, loose matches on Subject: below --
2010-04-09 14:09 [Xen-users] " Pasi Kärkkäinen
2010-04-09 14:36 ` [Xen-devel] " Keir Fraser
2010-04-09 14:59   ` Re: [Xen-users] " Andrew Lyon
2010-04-09 15:22     ` Keir Fraser

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.