Re: [PATCH v3 0/5] Handle monitor trap flag during instruction emulation

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

 



On Wed, Feb 26, 2020 at 10:38 PM Paolo Bonzini <pbonzini@xxxxxxxxxx> wrote:
>
> On 27/02/20 01:22, Oliver Upton wrote:
> > Are there any strong opinions about how the newly introduced nested
> > state should be handled across live migrations? When applying this
> > patch set internally I realized live migration would be busted in the
> > case of a kernel rollback (i.e. a kernel with this patchset emits the
> > nested state, kernel w/o receives it + refuses).
>
> Only if you use MTF + emulation.  In this case it's a pure bugfix so
> it's okay to break backwards migration.  If it's really a new feature,
> it should support KVM_ENABLE_CAP to enable it.
>
> Paolo
>

True. I suppose I've conflated the pure bugfix here with the fact that
MTF is new to us.

Thanks Paolo!

--
Best,
Oliver

> > Easy fix is to only turn on the feature once it is rollback-proof, but
> > I wonder if there is any room for improvement on this topic..
>



[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