Package: linux-image-amd64
Severity: wishlist


Dear Maintainers

With the release of the 3.2 kernel (yet to happen), some of the
modules improving the glue for Linux VMs running on a MS Hyper-V
hypervisor have left the staging area (drivers/hv).

Could this be taken as an opportunity to enable them in the default
kernel image in experimental (i386, amd64) allowing testing with Debian?
Currently they are not unloadable, if that is a problem, I know there is
a patch floating in the upstream.

What would they enable?
- add virtual SMP support for VMs
- time provider support from the hypervisor

Best regards
Mathieu



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 20111214061021.17898.26204.reportbug@cassini.gymna sium.koeniz">http://lists.debian.org/20111214061021.17898.26204.reportbug@cassini.gymna sium.koeniz

On Wed, 2011-12-14 at 07:10 +0100, Mathieu Simon wrote:
> Package: linux-image-amd64
> Severity: wishlist
>
>
> Dear Maintainers
>
> With the release of the 3.2 kernel (yet to happen), some of the
> modules improving the glue for Linux VMs running on a MS Hyper-V
> hypervisor have left the staging area (drivers/hv).

There is no need to wait for drivers to leave staging, in general.

I think there was some problem with them earlier that meant we didn't
enable them (possibly limitations on building as modules?). Also, MS
abandoned them for a while. But it seems that they are being properly
maintained now.

> Could this be taken as an opportunity to enable them in the default
> kernel image in experimental (i386, amd64) allowing testing with Debian?
> Currently they are not unloadable, if that is a problem, I know there is
> a patch floating in the upstream.

Do you have a reference for that? (Commit hash, message ID, URL...)

> What would they enable?
> - add virtual SMP support for VMs
> - time provider support from the hypervisor

Yes, all this paravirt stuff is useful.

Ben.

--
Ben Hutchings
Computers are not intelligent. They only think they are.

Hi (forwarding to the bugtracker this time)

> There is no need to wait for drivers to leave staging, in general.
Great. In this case please consider enabling the following modules too:

Have both left staging-next, not made into 3.2:
Microsoft Hyper-V virtual network driver
Microsoft Hyper-V virtual mouse driver

Still in staging also in staging-next:
Microsoft Hyper-V virtual storage driver

> I think there was some problem with them earlier that meant we didn't
> enable them (possibly limitations on building as modules?). *Also, MS
> abandoned them for a while. *But it seems that they are being properly
> maintained now.
True, there were quite unstable at least until 2.6.38. I wouldn't
favour to enable them on earlier kernels (especially the storave driver)
(This is based on my experience driver-related crashes using
self-built mainline kernel)

>> Currently they are not unloadable, if that is a problem, I know there is
>> a patch floating in the upstream.
>
> Do you have a reference for that? *(Commit hash, message ID, URL...)
The patch seems not to be integrated yet anywhere, but the discussion mentions
the patch here

http://www.spinics.net/lists/linux-driver-devel/msg21785.html

Thanks for your positive answer.

- Mathieu



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: CANCJCZbO20zR=8HPGCCP3KoKSxO=vQ-jqvj74C-Z0quqTRGGSQ@mail.gmail.com">http://lists.debian.org/CANCJCZbO20zR=8HPGCCP3KoKSxO=vQ-jqvj74C-Z0quqTRGGSQ@mail.gmail.com

Ben Hutchings <ben@decadent.org.uk> schrieb:

>> Could this be taken as an opportunity to enable them in the default
>> kernel image in experimental (i386, amd64) allowing testing with Debian?=
>=20
>> Currently they are not unloadable, if that is a problem, I know there is=
>=20
>> a patch floating in the upstream.
>
> Do you have a reference for that? (Commit hash, message ID, URL...)

http://permalink.gmane.org/gmane.linux.kernel/1227920

Cheers,
Moritz


--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: slrnjeiecf.3vd.jmm@inutil.org">http://lists.debian.org/slrnjeiecf.3vd.jmm@inutil.org