# Re: \[PATCH net-next\] net: phy: micrel: Add support for lan9645x internal phy

- From: Jens Emil Schulz Ostergaard \<hidden\>
- Date: 2026-01-27 08:49:58 
- List: netdev
- Page: https://ratatoskr.run/netdev/2026/01/3388465
- Thread: https://ratatoskr.run/netdev/2026/01/3383752/t.md

```text
Hi Andrew,

On Fri, 2026-01-23 at 17:11 +0100, Andrew Lunn wrote:
> > +static int lan9645x_config_intr(struct phy_device *phydev)
> > +{
> > +     int err;
> > +
> > +     /* enable / disable interrupts */
> > +     if (phydev->interrupts == PHY_INTERRUPT_ENABLED) {
> > +             /* This is an internal PHY of lan9645x and is not possible to
> > +              * change the polarity of irq sources in the OIC (CPU_INTR)
> > +              * found in lan9645x. Therefore change the polarity of the
> > +              * interrupt in the PHY from being active low instead of active
> > +              * high.
> > +              */
> > +             phy_write(phydev, LAN8804_CONTROL,
> > +                       LAN8804_CONTROL_INTR_POLARITY);
> > +
> > +             /* By default interrupt buffer is open-drain in which case the
> > +              * interrupt can be active only low. Therefore change the
> > +              * interrupt buffer to be push-pull to be able to change
> > +              * interrupt polarity.
> > +              */
> 
> It is not stated here, but i assume that there are multiple of these
> PHYs in the switch, and each PHY has its own independent connection to
> the parent interrupt controller? There is no interrupt sharing going
> on, which is common with discreet PHYs.
> 
>         Andrew

Yes you are right, there are 5 of these PHYs in the chip, each with their
on connection to the interrupt controller. It is similar to lan966x,
which has two internal PHYs of an older generation.

Thanks,
Emil

```
