Re: [PATCH 2/2] KVM: PPC: Book3S HV: Take out virtual core piggybacking code

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



 Hi Paul,

On 15.09.2016 10:09, Paul Mackerras wrote:
> This takes out the code that arranges to run two (or more) virtual
> cores on a single subcore when possible, that is, when both vcores
> are from the same VM, the VM is configured with one CPU thread per
> virtual core, and all the per-subcore registers have the same value
> in each vcore.  Since the VTB (virtual timebase) is a per-subcore
> register, and will almost always differ between vcores, this code
> is disabled on POWER8 machines, meaning that it is only usable on
> POWER7 machines (which don't have VTB).  Given the tiny number of
> POWER7 machines which have firmware that allows them to run HV KVM,
> the benefit of simplifying the code outweighs the loss of this
> feature on POWER7 machines.
> 
> Signed-off-by: Paul Mackerras <paulus@xxxxxxxxxx>
> ---
>  arch/powerpc/kvm/book3s_hv.c | 128 +------------------------------------------
>  1 file changed, 3 insertions(+), 125 deletions(-)

I've built a kernel with your patches and used it for my test where on
of the guests normally generates a soft lockup within some few hours.
The test now ran on two P8 machines over the weekend, and there was no
single soft lockup anymore, all guests survived now - that's great!

Tested-by: Thomas Huth <thuth@xxxxxxxxxx>

--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html



[Index of Archives]     [KVM ARM]     [KVM ia64]     [KVM ppc]     [Virtualization Tools]     [Spice Development]     [Libvirt]     [Libvirt Users]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite Questions]     [Linux Kernel]     [Linux SCSI]     [XFree86]
  Powered by Linux