Re: [Qemu-devel] [PATCH v1 3/5] xlnx-zynqmp: Make the RPU subsystem optional

["Edgar E. Iglesias" <edgar.iglesias@gmail.com>]

On Tue, May 24, 2016 at 05:26:54PM +0100, Peter Maydell wrote:
> On 19 May 2016 at 23:54, Edgar E. Iglesias <edgar.iglesias@gmail.com> wrote:
> > From: "Edgar E. Iglesias" <edgar.iglesias@xilinx.com>
> >
> > The way we currently model the RPU subsystem is of quite
> > limited use. In addition to that, it causes problems for
> > KVM and for GDB debugging.
> >
> > Make the RPU optional by adding a has_rpu property and
> > default to having it disabled.
> >
> > Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
> > ---
> >  hw/arm/xlnx-zynqmp.c         | 50 +++++++++++++++++++++++++++++++++++++-------
> >  include/hw/arm/xlnx-zynqmp.h |  2 ++
> >  2 files changed, 44 insertions(+), 8 deletions(-)
> >
> > diff --git a/hw/arm/xlnx-zynqmp.c b/hw/arm/xlnx-zynqmp.c
> > index 250ecc4..c180206 100644
> > --- a/hw/arm/xlnx-zynqmp.c
> > +++ b/hw/arm/xlnx-zynqmp.c
> > @@ -83,6 +83,40 @@ static inline int arm_gic_ppi_index(int cpu_nr, int ppi_index)
> >      return GIC_NUM_SPI_INTR + cpu_nr * GIC_INTERNAL + ppi_index;
> >  }
> >
> > +static bool xlnx_zynqmp_get_has_rpu(Object *obj, Error **errp)
> > +{
> > +    XlnxZynqMPState *s = XLNX_ZYNQMP(obj);
> > +
> > +    return s->has_rpu;
> > +}
> > +
> > +static void xlnx_zynqmp_set_has_rpu(Object *obj, bool value, Error **errp)
> > +{
> > +    XlnxZynqMPState *s = XLNX_ZYNQMP(obj);
> > +    int i;
> > +
> > +    if (s->has_rpu == value) {
> > +        /* Nothing to do.  */
> > +        return;
> > +    }
> > +
> > +    /* We don't support clearing the flag.  */
> > +    if (s->has_rpu) {
> > +        error_setg(errp, "has_rpu is already set to %u",
> > +                   s->has_rpu);
> > +        return;
> > +    }
> > +
> > +    /* Create the Cortex R5s.  */
> > +    for (i = 0; i < XLNX_ZYNQMP_NUM_RPU_CPUS; i++) {
> > +        object_initialize(&s->rpu_cpu[i], sizeof(s->rpu_cpu[i]),
> > +                          "cortex-r5-" TYPE_ARM_CPU);
> > +        object_property_add_child(obj, "rpu-cpu[*]", OBJECT(&s->rpu_cpu[i]),
> > +                                  &error_abort);
> > +    }
> 
> Do we have to create them in the set function so we can
> set properties before realize, or could this be deferred
> to realize time?

Yes, I thought it was recommended to avoid object creation
in realize. But creating the PRU in realize works too.

Cheers,
Edgar