On Fri, Apr 26, 2013 at 7:14 AM, Andre Przywara <andre.przywara@xxxxxxxxxx> wrote: > On 04/26/2013 03:42 PM, Peter Maydell wrote: >> On 26 April 2013 14:24, Andre Przywara <andre.przywara@xxxxxxxxxx> wrote: >>> On 04/26/2013 03:18 PM, Peter Maydell wrote: >>>> The obvious question here is "why do we need a new command?". >>>> The kernel booting specification says "boot the kernel in >>>> Hyp mode" so we should just always do that for booting Linux, >>>> surely? >> >>> Because it avoids regressions. I kind of feel uneasy to do a lot of >>> tinkering with secure state and the GIC unconditionally, especially if >>> enabled on many boards with virt-capable CPUs. >> >> There aren't exactly very many of those out there, so if we >> default to "boot in HYP mode" then (a) KVM will just work >> out of the box and (b) people doing u-boot ports to their >> board will find any issues and be able to submit fixes. >> If we don't turn it on by default then we'll end up stuck >> in a world where KVM doesn't work on half the virt capable >> boards out there. > > OK, that's a point. I changed the code already and will include it in > the next revision. > I strongly agree with Peter, and in fact I think it should only be a compile time option, not a run-time option, only for those debugging u-boot and a kernel on some new platform. -Christoffer _______________________________________________ kvmarm mailing list kvmarm@xxxxxxxxxxxxxxxxxxxxx https://lists.cs.columbia.edu/cucslists/listinfo/kvmarm
