On Thu, Mar 26, 2026 at 03:13:19PM +0000, Russell King (Oracle) wrote:

On Thu, Mar 26, 2026 at 01:04:15PM +0100, Louis-Alexis Eyraud wrote:

quoted

+static int an8801r_set_wol(struct phy_device *phydev,
+			   struct ethtool_wolinfo *wol)
+{
+	struct net_device *attach_dev = phydev->attached_dev;
+	const unsigned char *macaddr = attach_dev->dev_addr;

This isn't a criticism for this patch, but a discussion point for
phylib itself.

It occurs to me that there's a weakness in the phylib interface for WoL,
specifically when WoL is enabled at the PHY, but someone then changes
the interface's MAC address - there doesn't seem to be a way for the
address programmed into the PHY to be updated. Should there be?

Do we instead expect users to disable WoL before changing the MAC for
a network interface?

Program the MAC address during suspend? I assume userspace is no
longer active at this point, so the address should be stable.

       Andrew