Re: [PATCH] of: property: Add fw_devlink support for interrupt-map property

[Rob Herring]

On Fri, Apr 12, 2024 at 6:29 AM Anup Patel <apatel@xxxxxxxxxxxxxxxx> wrote:
>
> Some of the PCI controllers (such as generic PCI host controller)
> use "interrupt-map" DT property to describe the mapping between
> PCI endpoints and PCI interrupt pins.

I would go as far as saying that's the only case as that's the only
case where the interrupts are not described in DT.

> Currently, there is no fw_devlink created based on "interrupt-map"
> DT property so interrupt controller is not guaranteed to be probed
> before PCI host controller. This mainly affects RISC-V platforms
> where both PCI host controller and interrupt controllers are probed
> as regular platform devices.

That's *every* system with PCI really.

> This creates fw_devlink between consumers (PCI host controller) and
> supplier (interrupt controller) based on "interrupt-map" DT property.
>
> Signed-off-by: Anup Patel <apatel@xxxxxxxxxxxxxxxx>
> ---
>  drivers/of/property.c | 53 +++++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 53 insertions(+)
>
> diff --git a/drivers/of/property.c b/drivers/of/property.c
> index a6358ee99b74..ccbbb651a89a 100644
> --- a/drivers/of/property.c
> +++ b/drivers/of/property.c
> @@ -1311,6 +1311,58 @@ static struct device_node *parse_interrupts(struct device_node *np,
>         return of_irq_parse_one(np, index, &sup_args) ? NULL : sup_args.np;
>  }
>
> +static struct device_node *parse_interrupt_map(struct device_node *np,
> +                                              const char *prop_name, int index)
> +{
> +       struct device_node *tn, *ipar, *supnp = NULL;
> +       u32 addrcells, intcells, cells;
> +       const __be32 *imap, *imap_end;
> +       int i, imaplen;
> +
> +       if (!IS_ENABLED(CONFIG_OF_IRQ))
> +               return NULL;
> +
> +       if (strcmp(prop_name, "interrupt-map"))
> +               return NULL;
> +
> +       ipar = of_node_get(np);
> +       do {
> +               if (!of_property_read_u32(ipar, "#interrupt-cells", &intcells))
> +                       break;
> +               tn = ipar;
> +               ipar = of_irq_find_parent(ipar);
> +               of_node_put(tn);
> +       } while (ipar);
> +       if (!ipar)
> +               return NULL;
> +       addrcells = of_bus_n_addr_cells(ipar);
> +       of_node_put(ipar);
> +
> +       imap = of_get_property(np, "interrupt-map", &imaplen);
> +       if (!imap || imaplen <= (addrcells + intcells))
> +               return NULL;
> +       imap_end = imap + imaplen;
> +
> +       for (i = 0; i <= index && imap < imap_end; i++) {
> +               if (supnp)
> +                       of_node_put(supnp);
> +
> +               imap += addrcells;
> +               imap += intcells;
> +
> +               supnp = of_find_node_by_phandle(be32_to_cpu(imap[0]));
> +               if (!supnp)
> +                       return NULL;
> +               imap += 1;
> +
> +               if (of_property_read_u32(supnp, "#interrupt-cells", &cells))
> +                       return NULL;
> +               imap += cells;

This is wrong. Technically, you can have #address-cells too.

The bigger problem I have is this creates 2 sets of 'interrupt-map'
parsing code. Your version skips a lot of things like whether the
interrupt controller is available and there's the list of
'interrupt-map' abusers to think about.

Rob