On Tue, Jan 20, 2015 at 09:34:39AM -0600, Greg Bellows wrote:
> On Tue, Jan 20, 2015 at 9:22 AM, Igor Mammedov <imamm...@redhat.com> wrote:
>
> > On Mon, 19 Jan 2015 16:30:17 -0600
> > Greg Bellows <greg.bell...@linaro.org> wrote:
> >
> > > Adds a CPU feature parsing function and assigns to the CPU class. The
> > only
> > > feature added was "-aarch64" which disabled the AArch64 execution state
> > on a
> > > 64-bit ARM CPU.
> > >
> > > Also adds stripping of features from CPU model string in acquiring the
> > ARM CPU
> > > by name.
> > >
> > > Signed-off-by: Greg Bellows <greg.bell...@linaro.org>
> > > ---
> > > target-arm/cpu.c | 45 ++++++++++++++++++++++++++++++++++++++++++++-
> > > 1 file changed, 44 insertions(+), 1 deletion(-)
> > >
> > > diff --git a/target-arm/cpu.c b/target-arm/cpu.c
> > > index 285947f..f327dd7 100644
> > > --- a/target-arm/cpu.c
> > > +++ b/target-arm/cpu.c
> > > @@ -514,13 +514,17 @@ static ObjectClass *arm_cpu_class_by_name(const
> > char *cpu_model)
> > > {
> > > ObjectClass *oc;
> > > char *typename;
> > > + char *cpuname;
> > >
> > > if (!cpu_model) {
> > > return NULL;
> > > }
> > >
> > > - typename = g_strdup_printf("%s-" TYPE_ARM_CPU, cpu_model);
> > > + cpuname = g_strdup(cpu_model);
> > > + cpuname = strtok(cpuname, ",");
> > > + typename = g_strdup_printf("%s-" TYPE_ARM_CPU, cpuname);
> > > oc = object_class_by_name(typename);
> > > + g_free(cpuname);
> > > g_free(typename);
> > > if (!oc || !object_class_dynamic_cast(oc, TYPE_ARM_CPU) ||
> > > object_class_is_abstract(oc)) {
> > > @@ -1163,6 +1167,44 @@ static Property arm_cpu_properties[] = {
> > > DEFINE_PROP_END_OF_LIST()
> > > };
> > >
> > > +static void arm_cpu_parse_features(CPUState *cs, char *features,
> > > + Error **errp)
> > > +{
> > > + ARMCPU *cpu = ARM_CPU(cs);
> > > + char *featurestr;
> > > +
> > > + featurestr = features ? strtok(features, ",") : NULL;
> > > + while (featurestr) {
> > > + if (featurestr[0] == '-') {
> > ...
> > > + } else if (featurestr[0] == '+') {
> > Please do not use legacy +-feature format and support only foo=val format.
> > Other targets have it only for to being able support legacy setups
> > which use +- format.
> >
> >
> Thanks Igor. I was under the impression that the +/- notation was still
> relevant. Perhaps it makes the most sense to convert to using object
> properties similar to how machine options are specified? What do you think
> Peter?
This is what we have been working on, on x86. "+FOO" should be
translated to setting QOM property "feat-FOO=on" (note that you
shouldn't need to use "feat-FOO" for the proeprty names, as you don't
need to support the legacy +-feature format).
--
Eduardo
