Re: [Xen-devel] [PATCH-for-4.13] x86/mm: don't needlessly veto migration

[George Dunlap <george.dunlap@xxxxxxxxxx>]

On 10/1/19 1:32 PM, Paul Durrant wrote:
> On Tue, 1 Oct 2019 at 12:09, George Dunlap <george.dunlap@xxxxxxxxxx> wrote:
>>
>> On 10/1/19 11:40 AM, Paul Durrant wrote:
>>>  -----Original Message-----
>>>> From: George Dunlap <george.dunlap@xxxxxxxxxx>
>>>> Sent: 01 October 2019 11:34
>>>> To: Paul Durrant <Paul.Durrant@xxxxxxxxxx>; 'Jan Beulich' 
>>>> <jbeulich@xxxxxxxx>
>>>> Cc: Andrew Cooper <Andrew.Cooper3@xxxxxxxxxx>; Roger Pau Monne 
>>>> <roger.pau@xxxxxxxxxx>; xen-
>>>> devel@xxxxxxxxxxxxxxxxxxxx; Juergen Gross <jgross@xxxxxxxx>; Wei Liu 
>>>> <wl@xxxxxxx>
>>>> Subject: Re: [Xen-devel] [PATCH-for-4.13] x86/mm: don't needlessly veto 
>>>> migration
>>>>
>>>> On 10/1/19 11:24 AM, Paul Durrant wrote:
>>>>>> -----Original Message-----
>>>>>> From: Jan Beulich <jbeulich@xxxxxxxx>
>>>>>> Sent: 01 October 2019 11:15
>>>>>> To: Paul Durrant <Paul.Durrant@xxxxxxxxxx>
>>>>>> Cc: Andrew Cooper <Andrew.Cooper3@xxxxxxxxxx>; George Dunlap 
>>>>>> <George.Dunlap@xxxxxxxxxx>; Roger Pau
>>>>>> Monne <roger.pau@xxxxxxxxxx>; xen-devel@xxxxxxxxxxxxxxxxxxxx; Juergen 
>>>>>> Gross <jgross@xxxxxxxx>; Wei
>>>> Liu
>>>>>> <wl@xxxxxxx>
>>>>>> Subject: Re: [Xen-devel] [PATCH-for-4.13] x86/mm: don't needlessly veto 
>>>>>> migration
>>>>>>
>>>>>> On 01.10.2019 11:36, Paul Durrant wrote:
>>>>>>>> -----Original Message-----
>>>>>>>> From: Jan Beulich <jbeulich@xxxxxxxx>
>>>>>>>> Sent: 01 October 2019 10:19
>>>>>>>> To: Paul Durrant <Paul.Durrant@xxxxxxxxxx>
>>>>>>>> Cc: Andrew Cooper <Andrew.Cooper3@xxxxxxxxxx>; George Dunlap 
>>>>>>>> <George.Dunlap@xxxxxxxxxx>; Roger
>>>> Pau
>>>>>>>> Monne <roger.pau@xxxxxxxxxx>; xen-devel@xxxxxxxxxxxxxxxxxxxx; Juergen 
>>>>>>>> Gross <jgross@xxxxxxxx>;
>>>> Wei
>>>>>> Liu
>>>>>>>> <wl@xxxxxxx>
>>>>>>>> Subject: Re: [Xen-devel] [PATCH-for-4.13] x86/mm: don't needlessly 
>>>>>>>> veto migration
>>>>>>>>
>>>>>>>> On 01.10.2019 10:52, Paul Durrant wrote:
>>>>>>>>>> -----Original Message-----
>>>>>>>>>> From: Jan Beulich <jbeulich@xxxxxxxx>
>>>>>>>>>> Sent: 01 October 2019 09:46
>>>>>>>>>> To: Paul Durrant <Paul.Durrant@xxxxxxxxxx>
>>>>>>>>>> Cc: xen-devel@xxxxxxxxxxxxxxxxxxxx; Andrew Cooper 
>>>>>>>>>> <Andrew.Cooper3@xxxxxxxxxx>; Roger Pau Monne
>>>>>>>>>> <roger.pau@xxxxxxxxxx>; George Dunlap <George.Dunlap@xxxxxxxxxx>; 
>>>>>>>>>> Juergen Gross
>>>>>> <jgross@xxxxxxxx>;
>>>>>>>> Wei
>>>>>>>>>> Liu <wl@xxxxxxx>
>>>>>>>>>> Subject: Re: [Xen-devel] [PATCH-for-4.13] x86/mm: don't needlessly 
>>>>>>>>>> veto migration
>>>>>>>>>>
>>>>>>>>>> On 01.10.2019 10:28, Paul Durrant wrote:
>>>>>>>>>>> Now that xl.cfg has an option to explicitly enable IOMMU mappings 
>>>>>>>>>>> for a
>>>>>>>>>>> domain, migration may be needlessly vetoed due to the check of
>>>>>>>>>>> is_iommu_enabled() in paging_log_dirty_enable().
>>>>>>>>>>> There is actually no need to prevent logdirty from being enabled 
>>>>>>>>>>> unless
>>>>>>>>>>> devices are assigned to a domain and that domain is sharing HAP 
>>>>>>>>>>> mappings
>>>>>>>>>>> with the IOMMU (in which case disabling write permissions in the 
>>>>>>>>>>> P2M may
>>>>>>>>>>> cause DMA faults).
>>>>>>>>>>
>>>>>>>>>> But that's taking into account only half of the reason of the
>>>>>>>>>> exclusion. The other half is that assigned devices may cause pages
>>>>>>>>>> to be dirtied behind the back of the log-dirty logic.
>>>>>>>>>
>>>>>>>>> But that's no reason to veto logdirty. Some devices have drivers (in 
>>>>>>>>> dom0)
>>>>>>>>> which can extract DMA dirtying information and set dirty tracking
>>>>>>>>> information appropriately.
>>>>>>>>
>>>>>>>> It still needs a positive identification then: Such drivers should tell
>>>>>>>> Xen for which specific devices such information is going to be 
>>>>>>>> provided.
>>>>>>>
>>>>>>> Why does the hypervisor need have the right of veto though? Surely it is
>>>>>>> the toolstack that should decide whether a VM is migratable in the
>>>>>>> presence of assigned h/w. Xen need only be concerned with the integrity
>>>>>>> of the host, which is why the check for ETP sharing remains.
>>>>>>
>>>>>> While the tool stack is to decide, the hypervisor is expected to 
>>>>>> guarantee
>>>>>> correct data coming back from XEN_DOMCTL_SHADOW_OP_{PEEK,CLEAN}.
>>>>>
>>>>> For some definition of 'correct', yes, and I don't think that this change 
>>>>> violates any definition I
>>>> can find in the domctl header.
>>>>>
>>>>> Note: there are already emulators that will be playing with the dirty map 
>>>>> on an arbitrary and
>>>> unsynchronized basis because they are emulating bus mastering h/w.
>>>>
>>>> But the question is, do we want the toolstack to have to become an
>>>> expert in what hardware might have external dirty tracking, and whether
>>>> such tracking is active?  At the moment that would mean either 1)
>>>> putting that information inside of libxc, or 2) duplicating it across
>>>> xapi and libxl, for instance.
>>>
>>> Why not? The toolstack is in charge of migration so why can't it decide 
>>> whether it is 'safe' or not?
>>
>> First of all, it's not about what the toolstack can decide; it's what it
>> knows.  It doesn't currently know anything about the details of devices
>> themselves or how they relate to other functionality, such as migration.
> 
> Doesn't it? Why?
[snip]
> It is, but I don't see that Xen should have any right of veto over
> what a paticular toolstack wishes to do with the domains it has
> created.

I feel like the temperature of this conversation is really high, and I
can't really figure out why.  Could I ask that we try to turn down the
heat a bit, and perhaps help Jan and I figure out where you're coming from?

> It can, in the face of an arbitrary device, use an
> emulator such as QEMU to deal with the pass-through and having so
> decided knows that it can't get dirty page information, and hence the
> domain cannot be safely migrated. In the face of a device it knows
> about though e.g. a GPU, it can run a dedicated emulator from which it
> can get dirty page information and hence (providing shared EPT is not
> in use) it knows the domain can be migrated.

xapi knows what devices *it has asked Xen to pass through*.  Xen knows
*what devices it gave to the guest*.

xapi can *gather* specific information about devices (topology,
characteristics, &c) from Linux and Xen; Xen has it already.

It seems you've encoded in xapi information about how Xen is
implemented.  That could change: More features could begin to interact
with logdirty, or with devices which can implement their own
logdirty-like functionality.  If/when those changes happen, we can
update the rules for when logdirty works within the patch series itself
in Xen.  If you instead encode that knowlegde in xapi, then xapi needs
to be updated to keep in sync with the internal implementation of the
hypervisor.

In any case, having emulators which can handle logdirty externally
report their own capability seems a much better way of doing things than
hard-coding in xapi which emulators know how to do what.

>> Secondly, you haven't answered the question about duplication.  Where do
>> you propose to put this functionality?
>>
> 
> Different toolstacks can have different capabilities. If libxl is
> unaware of a devices capability to provide dirty page information, but
> XAPI is aware, then why is that a problem?

So it sounds like you've already done a lot of this work in xapi.  But
that only benefits XenServer and derivatives: all of Citrix's other
partners who want to do something similar will have to completely
duplicate all of that functionality.  It should be obvious why that's
sub-optimal.

 -George

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel