On Thu, May 2, 2013 at 9:32 PM, Anup Patel <anup.patel@xxxxxxxxxx> wrote: > On 2 May 2013 16:06, Peter Maydell <peter.maydell@xxxxxxxxxx> wrote: >> On 2 May 2013 11:30, Marc Zyngier <marc.zyngier@xxxxxxx> wrote: >>> On 02/05/13 11:02, Anup Patel wrote: >>>> We would like to know KVM ARM/ARM64 plans for handling host implementation >>>> specific registers when underlying host cpu is different from guest cpu. >>> >>> So far, we don't do that. The only case where that could actually be >>> useful is a big.LITTLE-type system. >> >> That's not the only situation where you need it. Consider the case >> where you have a pile of guest VMs on a rack full of A15s. Now >> you'd like to migrate them all over to your new rack full of shiny >> AWhatevers. Unless you can run A15 guests on AWhatever hosts then >> you've just forced all your customers to reboot, maybe reconfigure >> and worst case reinstall their guest VMs. >> >> We haven't implemented it yet, but we absolutely ought to >> at some point. big.LITTLE host support would then build on >> top of that. >> >> -- PMM > > I got more clarifications on non-trappable registers in APM ARMv8 CPU. > > We don't have any host implementation specific non-trappable registers > in APM ARMv8 CPU, so we are good. :) > > I hope we are not missing any thing else for host != guest scenarios. > I don't think there's much more to it than going through the ARM ARM and the TRM and checking which registers are specific to the core and dealing with it it in any possible way, and hope it doesn't require PV or BT if it's a hard requirement. -Christoffer _______________________________________________ kvmarm mailing list kvmarm@xxxxxxxxxxxxxxxxxxxxx https://lists.cs.columbia.edu/cucslists/listinfo/kvmarm
