On Thu, Jun 8, 2023 at 11:28 AM Rob Herring <robh@xxxxxxxxxx> wrote:
>
> +Saravana
>
> On Mon, May 08, 2023 at 07:58:33PM +0530, Anup Patel wrote:
> > The RISC-V APLIC interrupt controller driver is a regular platform
> > driver so we need to ensure that it is probed as soon as possible.
> > To achieve this, we mark the interrupt controller device nodes with
> > FWNODE_FLAG_BEST_EFFORT (just like console DT node).
> >
> > Fixes: 8f486cab263c ("driver core: fw_devlink: Allow firmware to mark devices as best effort")
>
> It is good practice to Cc the commit author of what you are fixing.
>
> > Signed-off-by: Anup Patel <apatel@xxxxxxxxxxxxxxxx>
> > ---
> > drivers/of/irq.c | 10 ++++++++++
> > 1 file changed, 10 insertions(+)
> >
> > diff --git a/drivers/of/irq.c b/drivers/of/irq.c
> > index 174900072c18..94e1d9245cff 100644
> > --- a/drivers/of/irq.c
> > +++ b/drivers/of/irq.c
> > @@ -535,6 +535,16 @@ void __init of_irq_init(const struct of_device_id *matches)
> > INIT_LIST_HEAD(&intc_desc_list);
> > INIT_LIST_HEAD(&intc_parent_list);
> >
> > + /*
> > + * We need interrupt controller platform drivers to work as soon
>
> If it's platform drivers/devices you care about, then perhaps this
> should be done when platform devices are created.
>
> > + * as possible so mark the interrupt controller device nodes with
> > + * FWNODE_FLAG_BEST_EFFORT so that fw_delink knows not to delay
> > + * the probe of the interrupt controller device for suppliers
> > + * without drivers.
> > + */
> > + for_each_node_with_property(np, "interrupt-controller")
>
> This function only operates on nodes matching 'matches', but this
> operates on everything.
>
> It does make sense that if we init an interrupt controller here, then we
> will surely want to probe its driver later on. So maybe just move
> setting FWNODE_FLAG_BEST_EFFORT within
> for_each_matching_node_and_match() below.
>
> > + np->fwnode.flags |= FWNODE_FLAG_BEST_EFFORT;
> > +
Definite Nack. You are pretty much disabling fw_devlink for all
interrupt controllers. There are plenty of examples of the IRQ drivers
being needed very early on and still probing properly without the need
for this flag. Please fix your driver/DT.
-Saravana
> > for_each_matching_node_and_match(np, matches, &match) {
> > if (!of_property_read_bool(np, "interrupt-controller") ||
> > !of_device_is_available(np))
> > --
> > 2.34.1
> >
