On Wed, Jan 08, 2020 at 11:58:45AM +0100, Greg Kurz wrote: > On Wed, 8 Jan 2020 11:54:53 +1100 > David Gibson <da...@gibson.dropbear.id.au> wrote: > > > On Tue, Jan 07, 2020 at 07:32:03PM +0100, Philippe Mathieu-Daudé wrote: > > > Hi Greg, > > > > > > On 1/7/20 5:32 PM, Greg Kurz wrote: > > > > The proper way to do that would be to use > > > > device_class_set_parent_realize(), > > > > but defining a Pnv8PsiClass and a Pnv9PsiClass types with a > > > > parent_realize > > > > pointer adds a fair amount of code. Calling pnv_psi_realize() > > > > explicitely > > > > is fine for now. > > > > > > > > This should probably be achieved with a device realize hook in the > > > > PSI base class and device_class_set_parent_realize() in the children > > > > classes. > > > > > I realize that this last paragraph is a leftover. First paragraph already > mentions device_class_set_parent_realize() as being the "proper way". > > David, > > Can you remove it in your tree ? No big deal if you can't.
Sorry, not any more - I've already sent the PR.
>
> > > Can you add a note explaining why the POWER10 PSI doesn't need it?
> >
> > For now, POWER10 uses the Pnv9PsiClass, I believe, so the question
> > doesn't arise.
> >
>
> This is correct and also a bit confusing, as proves Philippe's remark.
> Maybe we should come up with a PnvXivePsiClass and specialize it for
> POWER9 and POWER10.
>
> > >
> > > >
> > > > Signed-off-by: Greg Kurz <gr...@kaod.org>
> > > > Signed-off-by: Cédric Le Goater <c...@kaod.org>
> > > > ---
> > > > hw/ppc/pnv_psi.c | 19 ++++++++++++-------
> > > > 1 file changed, 12 insertions(+), 7 deletions(-)
> > > >
> > > > diff --git a/hw/ppc/pnv_psi.c b/hw/ppc/pnv_psi.c
> > > > index 6c94781e377d..546232106756 100644
> > > > --- a/hw/ppc/pnv_psi.c
> > > > +++ b/hw/ppc/pnv_psi.c
> > > > @@ -469,6 +469,16 @@ static void pnv_psi_reset_handler(void *dev)
> > > > device_reset(DEVICE(dev));
> > > > }
> > > > +static void pnv_psi_realize(DeviceState *dev, Error **errp)
> > > > +{
> > > > + PnvPsi *psi = PNV_PSI(dev);
> > > > +
> > > > + /* Default BAR for MMIO region */
> > > > + pnv_psi_set_bar(psi, psi->bar | PSIHB_BAR_EN);
> > > > +
> > > > + qemu_register_reset(pnv_psi_reset_handler, dev);
> > > > +}
> > > > +
> > > > static void pnv_psi_power8_instance_init(Object *obj)
> > > > {
> > > > Pnv8Psi *psi8 = PNV8_PSI(obj);
> > > > @@ -528,9 +538,6 @@ static void pnv_psi_power8_realize(DeviceState
> > > > *dev, Error **errp)
> > > > memory_region_init_io(&psi->regs_mr, OBJECT(dev), &psi_mmio_ops,
> > > > psi,
> > > > "psihb", PNV_PSIHB_SIZE);
> > > > - /* Default BAR for MMIO region */
> > > > - pnv_psi_set_bar(psi, psi->bar | PSIHB_BAR_EN);
> > > > -
> > > > /* Default sources in XIVR */
> > > > for (i = 0; i < PSI_NUM_INTERRUPTS; i++) {
> > > > uint8_t xivr = irq_to_xivr[i];
> > > > @@ -538,7 +545,7 @@ static void pnv_psi_power8_realize(DeviceState
> > > > *dev, Error **errp)
> > > > ((uint64_t) i << PSIHB_XIVR_SRC_SH);
> > > > }
> > > > - qemu_register_reset(pnv_psi_reset_handler, dev);
> > > > + pnv_psi_realize(dev, errp);
> > > > }
> > > > static int pnv_psi_dt_xscom(PnvXScomInterface *dev, void *fdt, int
> > > > xscom_offset)
> > > > @@ -873,9 +880,7 @@ static void pnv_psi_power9_realize(DeviceState
> > > > *dev, Error **errp)
> > > > memory_region_init_io(&psi->regs_mr, OBJECT(dev),
> > > > &pnv_psi_p9_mmio_ops, psi,
> > > > "psihb", PNV9_PSIHB_SIZE);
> > > > - pnv_psi_set_bar(psi, psi->bar | PSIHB_BAR_EN);
> > > > -
> > > > - qemu_register_reset(pnv_psi_reset_handler, dev);
> > > > + pnv_psi_realize(dev, errp);
> > > > }
> > > > static void pnv_psi_power9_class_init(ObjectClass *klass, void *data)
> > > >
> > > >
> > >
> >
>
--
David Gibson | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_
| _way_ _around_!
http://www.ozlabs.org/~dgibson
signature.asc
Description: PGP signature
