On Fri, Feb 02, 2007 at 10:24:28PM +0000, Thiemo Seufer wrote:
> Juergen Lock wrote:
> > > > Ok Jung-uk Kim found the following fix: (Thanx!)
> > > >
> > > > --- qemu/cpu-exec.c.orig Wed Jan 31 16:58:03 2007
> > > > +++ qemu/cpu-exec.c Wed Jan 31 17:08:11 2007
> > > > @@ -226,9 +226,9 @@
> > > >
> > > > int cpu_exec(CPUState *env1)
> > > > {
> > > > - int saved_T0, saved_T1;
> > > > + long saved_T0, saved_T1;
> > > > #if defined(reg_T2)
> > > > - int saved_T2;
> > > > + long saved_T2;
> > >
> > > I used target_ulong instead.
> >
> > Isn't that 32 bit for 32 bit targets? Then it wouldn't fix the bug,
> > the problem happened for 32 bit guests (i386-softmmu) on 64 bit hosts
> > (FreeBSD-current/amd64), there the upper half of rbx which held a
> > pointer in the calling function was lost...
>
> Apparently I miss something here. If T0 is a (32bit) target_ulong,
> why does saving it to a (32bit) target_ulong and restoring it from
> there cause any trouble?
Hmm. All I can say is the upper half of rbx (which holds T0) gets
spilled on FreeBSD-current/amd64 hosts unless saving and restoring
the full 64 bit of it...
_______________________________________________
Qemu-devel mailing list
Qemu-devel@nongnu.org
http://lists.nongnu.org/mailman/listinfo/qemu-devel
