Xen performance kernel patches backported to 3.2 for Precise
Stefan Bader
stefan.bader at canonical.com
Wed Nov 30 16:13:10 UTC 2011
On 30.11.2011 12:38, Mike McClurg wrote:
> Hello Kernel Team,
>
> I have just created bug report #898112 [1]. This bug is related to a
> kernel issue which negatively affects Xen performance on Ubuntu 11.04
> and later. This issue is caused by the kernel incorrectly reporting the
> cpufreq capabilities of the CPU to Xen, which prevents the hypervisor
> from setting the CPU to the maximum frequency scaling.
>
> A blog post describing the bug in detail can be found here [2]. The
> patch series, which has been backported to 3.2 (and I believe 3.0 as
> well) can be found here [3].
>
> I ask the kernel team to consider including the 3.2 backport of this
> branch in the Precise kernel, and possibly the Oneiric kernel as well,
> if this is possible. Please feel free to contact me or Konrad (the patch
> author) with any questions. Thanks,
>
> Mike
>
Darn, why is reply-all sometimes the default and then again not...
> Hi Mike,
>
> if Konrad would/will manage to upstream those changes into 3.2, those would be
> picked up automatically. Just looking at the number of patches I would be quite
> careful to pick them from a non-upstream branch, even more because those all
> touch the acpi subsystem. And even being useful in Xen does not guarantee there
> is no side effect.
> I will look at them, but still would be careful and cannot make any promise
> right now.
>
> Thanks,
> Stefan
The set is changing a considerable amount of code (including the addition of xen
specific processor driver). I tried to find this being discussed on either the
xen or the acpi mailing list but at least I cannot find any of the patches on
any of those.
-Stefan
>
> [1] https://bugs.launchpad.net/ubuntu/+source/linux/+bug/898112
> [2] http://blog.xen.org/index.php/2011/11/29/baremetal-vs-xen-vs-kvm-redux/
> [3]
> http://git.kernel.org/?p=linux/kernel/git/konrad/xen.git;a=shortlog;h=refs/heads/devel/acpi-cpufreq.v4
>
More information about the kernel-team
mailing list