Xen devel - Feb 2010 - Memory hot-add and c/s 20892: bad interaction?

Jan, Yunhong,

I was just thinking about xen-unstable:20892, which exposes real current
max_mfn to guests, so that they can more accurately clamp their m2p address
translations.

I was wondering whether this changeset is actually a bad idea in light of
memory hot-add, as now implemented by Yunhong? I would imagine this can mean
that max_mfn is now dynamic, and can increase in value after boot. So would
20892 thus leave all existing guests (e.g., dom0!) broken after a hot-add
which adds new highest RAM addresses?

 -- Keir



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

>>> Keir Fraser <keir.fraser@eu.citrix.com> 08.02.10 11:32
>>>
>I was just thinking about xen-unstable:20892, which exposes real current
>max_mfn to guests, so that they can more accurately clamp their m2p address
>translations.
>
>I was wondering whether this changeset is actually a bad idea in light of
>memory hot-add, as now implemented by Yunhong? I would imagine this can mean
>that max_mfn is now dynamic, and can increase in value after boot. So would
>20892 thus leave all existing guests (e.g., dom0!) broken after a hot-add
>which adds new highest RAM addresses?
You probably overlooked the

+        if ( !mem_hotplug )

in that patch? I was intending to return some sort of boundary for the
hot-add case too, but that needs propagation from the SRAT parsing
code, and I didn''t think that would be urgent (i.e. for 4.0).

Jan


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

On 08/02/2010 11:09, "Jan Beulich" <JBeulich@novell.com> wrote:
>> I was wondering whether this changeset is actually a bad idea in light
of
>> memory hot-add, as now implemented by Yunhong? I would imagine this can
mean
>> that max_mfn is now dynamic, and can increase in value after boot. So
would
>> 20892 thus leave all existing guests (e.g., dom0!) broken after a
hot-add
>> which adds new highest RAM addresses?
> 
> You probably overlooked the
> 
> +        if ( !mem_hotplug )
Yeah, I looked at the changeset comment and not the patch itself! Sorry
about that.

 -- Keir



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

Keir/Jan, thanks for your notice.

Thanks
--jyh
>-----Original Message-----
>From: Keir Fraser [mailto:keir.fraser@eu.citrix.com]
>Sent: Monday, February 08, 2010 7:15 PM
>To: Jan Beulich; Jiang, Yunhong
>Cc: xen-devel@lists.xensource.com
>Subject: Re: Memory hot-add and c/s 20892: bad interaction?
>
>On 08/02/2010 11:09, "Jan Beulich" <JBeulich@novell.com>
wrote:
>
>>> I was wondering whether this changeset is actually a bad idea in
light of
>>> memory hot-add, as now implemented by Yunhong? I would imagine this
can
>mean
>>> that max_mfn is now dynamic, and can increase in value after boot.
So would
>>> 20892 thus leave all existing guests (e.g., dom0!) broken after a
hot-add
>>> which adds new highest RAM addresses?
>>
>> You probably overlooked the
>>
>> +        if ( !mem_hotplug )
>
>Yeah, I looked at the changeset comment and not the patch itself! Sorry
>about that.
>
> -- Keir
>

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