Hi Paolo,

On 6/1/26 10:59 AM, Paolo Abeni wrote:

On 5/26/26 11:47 AM, Bastien Curutchet (Schneider Electric) wrote:

quoted

@@ -3798,9 +3795,6 @@ int ksz_set_wol(struct dsa_switch *ds, int port,
  	if (wol->wolopts & ~(WAKE_PHY | WAKE_MAGIC))
  		return -EINVAL;
  
-	if (!is_ksz9477(dev) && !ksz_is_ksz87xx(dev))
-		return -EOPNOTSUPP;

Sashiko noted there is still a ksz_is_ksz87xx(dev) check in
ksz_wol_pre_shutdown() which would probably keep WoL broken for the
newly enabled switches.

Good catch from Sashiko here. I'm going to replace the is_ksz() checks 
in ksz_wol_pre_shutdown() with a if (ds->ops->get_wol && 
ds->ops->set_wol) (as done here in ksz_port_set_mac_address()) to ensure 
consistency in the WoL handling.


Best regards,
Bastien