On 07.12.2011 01:12, Jan Kiszka wrote: > On 2011-12-06 21:58, Jan Kiszka wrote: [] >> How embarrassing. Messed something up, maybe during rebase, and forgot >> to retest non-irqchip mode. Will fix and repost. > > Was even worse, a simple -ENOTTESTED case. Pushed an update at the same > location.
Ok, with the updated series everything works correctly. Guests boots, qemu does not segfault at startup anymore, WinXP with StdPC HAL works, even coroutines on mixed 32/64 environment breaks exactly the same way as with qemu-kvm/master ;) Tested both -enable-kvm and -machine pc,accel=kvm,kernel_irqchip=on - both behaves equally (I wasnt' able to spot a difference). What is the "target" syntax -- I assume it is the more verbose one, with -machine, so that -enable-kvm will be dropped at some point, right? Thanks! /mjt
