-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512

Marek Marczykowski-Górecki:
> On Tue, Jan 23, 2018 at 02:39:00PM +0000, Simon Gaiser wrote:
>> Marek Marczykowski-Górecki:
>>> Hi,
>>>
>>> The "8.4.3pre-shim-comet" branch on xenbits.xen.org/xen.git contains a
>>> backport of PVH to Xen 4.8 (and comet shim as the top 3-4 patches). In
>>> practice this is mostly about libxl changes (type="pvh" instead of
>>> device_model_version="none"), but there are also few fixes for other
>>> code, like ACPI tables builder. Those do fix some existing minor bugs.
>>> For example wrongly advertised i8042 controller, causing ~1s delay on
>>> Linux boot.
>>>
>>> It seems like the type="pvh" is what Xen is going to support and
>>> device_model_version="none" is abandoned in new versions. This means
>>> that it may be harder to backport further PVH-related patches, and also
>>> upstream our patches to Xen. There is not much changes on libxl side,
>>> but still.
>>> This also means that our libvirt patch in its current shape will be
>>> rejected, and we'll need to maintain it separately.
>>>
>>> Changing this that late in the release cycle may not be a good idea, but
>>> changing it later, or having to maintain duplicated set of patches
>>> and having additional difficulty with backporting other patches may also
>>> be a pain.
>>>
>>> Simon, can you take a look at this? What do you think about this? And if
>>> you think it worth changing it now, please create appropriate pull
>>> requests.
> 
>> I think the deciding factor is whether future 4.8.z releases will
>> include them (so that we can simply update instead of applying single
>> patches). But unfortunately this seems to be simply not decided yet [1].
> 
>> If they change it, updating now would be much better because to install
>> the update libxenlight.so, libvirt and core-admin need to be
>> synchronized.
> 
> Yes, this is exactly why I bring this idea right now. Having
> synchronized update is always a pain (in some corner cases may lead to
> removing packages).
> 
>> Since the mail reads like it's more likely that it will be backported and
>> AFAICS the change to type=pvh is rather small (so not so hard to
>> backport if it not ends in the official release) I lean towards
>> including the backports now. What do you think?
> 
> Well, change to type=pvh is not exactly small. It's 15-something patches
> touching different libxl files... But it looks it wasn't that hard to
> backport it from 4.10 (mostly differences because of libxl.c split),
> generally the code structure is very similar.

Ah, yeah, you're right. I seem to have missed some "preparation" patches
when looking at it yesterday.

> So, I also lean towards including it now.

Ok. Then let's hope it will be included in future releases.

>> [1]: 
>> https://lists.xenproject.org/archives/html/xen-devel/2018-01/msg02004.html
-----BEGIN PGP SIGNATURE-----
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=UCXZ
-----END PGP SIGNATURE-----

-- 
You received this message because you are subscribed to the Google Groups 
"qubes-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/qubes-devel/6aa7dd0f-04eb-f980-ae70-bfc63c346c74%40invisiblethingslab.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to