On 17.12.2012, at 18:21, Andreas Färber wrote: > Am 17.12.2012 15:49, schrieb Alexander Graf: >> >> On 14.12.2012, at 17:46, Jens Freimann wrote: >> >>> @@ -37,12 +42,30 @@ static void s390_cpu_reset(CPUState *s) >>> log_cpu_state(env, 0); >>> } >>> >>> - scc->parent_reset(s); >>> + s390_del_running_cpu(env); >>> >>> + scc->parent_reset(s); >>> memset(env, 0, offsetof(CPUS390XState, breakpoints)); >> >> Shouldn't parent_reset already do the memset? > > No, because "env" location and size are specific to S390CPU. > > And yes, it is ugly boilerplate code, but it cannot be solved with my > CPU_COMMON field movements alone (which partially add explicit reset > code based on the field location), there's quite a large number of > per-target fields that get reset that way, some intentionally, some > accidentally. ;-)
I see :) Alex
