Ok thanks for the guidance, I'll look into ARMv7 first to get some benchmarks. - Mario -----Original Message----- From: Marc Zyngier [mailto:marc.zyngier@xxxxxxx] Sent: Thursday, March 06, 2014 6:50 PM To: Mario Smarduch Cc: kvmarm@xxxxxxxxxxxxxxxxxxxxx Subject: Re: Asynchronous Page Fault Handling On Thu, Mar 06 2014 at 6:04:22 pm GMT, Mario Smarduch <m.smarduch@xxxxxxxxxxx> wrote: > Hi Marc, > I don't want to put this on the mailing to. > > Do you think ARMv8 is a better target then ARMv7 or start with ARMv7 > since there is silicon available? It shouldn't make any difference. Furthermore, we'd probably request feature parity between ARMv7 and ARMv8 as we share all the MMU code. Cheers, M. > Thanks. > > -----Original Message----- > From: Marc Zyngier [mailto:marc.zyngier@xxxxxxx] > Sent: Wednesday, March 05, 2014 6:54 PM > To: Mario Smarduch > Cc: kvmarm@xxxxxxxxxxxxxxxxxxxxx > Subject: Re: Asynchronous Page Fault Handling > > Hi Mario, > > On 2014-03-04 17:40, Mario Smarduch wrote: >> Is this feature on the road map? Memory over committed VM vCPUs spend >> a lot of time waiting on IO. > > As it doesn't seem to support non-PV guests, it has never been considered. > > Patches welcomed. > > M. > -- > Fast, cheap, reliable. Pick two. -- AAAFNRAA _______________________________________________ kvmarm mailing list kvmarm@xxxxxxxxxxxxxxxxxxxxx https://lists.cs.columbia.edu/cucslists/listinfo/kvmarm
