Alexander Graf <ag...@suse.de> writes: > On 19.06.2013, at 22:40, Anthony Liguori wrote: > >> Today if an invalid hcall is made to an otherwise valid device we >> abort due to QOM casting. This adds a parameter to find_by_reg() >> to indicate which type of device we're looking for. > > This should also adjust the hack to only match reg=0 for > TYPE_VIO_SPAPR_VTY_DEVICE, no?
Yup, that'll result in a bad cast otherwise. Nice catch. I'll respin. Regards, Anthony Liguori > > > Alex