Re: [PATCH net-next v8 2/3] dt-bindings: net: wiznet,w5100: add link status interrupt

[Arthur Crepin Leblond <arthur@marmottus.net>]

On Thu, Sep 17, 2026 at 08:08:50PM +0000, netdev-bot+sashiko@kernel.org wrote:
>[Severity: Low]
>Is this else branch describing the hardware, or the current Linux driver?
>
>The LINKLED output exists on the w5100 and w5200 packages as well, so a
>board that routes w5100/w5200 LINKLED to a host GPIO cannot be described
>here:
>
>  compatible = "wiznet,w5200";
>  interrupts = <25 IRQ_TYPE_LEVEL_LOW>, <24 IRQ_TYPE_EDGE_BOTH>;
>
>The second entry is rejected by maxItems: 1 during dt_binding_check.
>
>The only in-tree justification found for the w5500 restriction is the
>driver gate added later in this series, in w5100_probe():
>
>  priv->link_irq = ops->chip_id == W5500 ?
>                   fwnode_irq_get(dev_fwnode(dev), 1) : -EINVAL;
>
>which is a software limitation (register readback of link state via w5500
>PHYCFGR) rather than absence of the pin.  Would it be preferable to allow
>two interrupts for all three compatibles and note in the interrupts
>description which parts can actually report link state, so this constraint
>does not have to be relaxed later?

The LINKLED pin exists on w5100/w5200 but cannot be reliably used as a
trigger to check the carrier state as it also blinks on RX/TX activity.

   Link LED
     Low: Link (10/100M)
     High: Un-Link
     link: TX or RX state on Link

Arthur