Re: [PATCH RFC net-next] net: phylink: add quirk for disabling in-band-status for mediatek pcs at 2500base-x

[Eric Woudstra]

Sorry,  I'm on Android for now and have trouble find a suitable cliënt. This should have no http.


Anyway, I see now, I had another version which I should have send:


+       if (phylink_major_no_inband(pl, state->interface) && (!!pl->phydev)) {
+               if (pl->cur_link_an_mode == MLO_AN_INBAND)
+                       pl->cur_link_an_mode = MLO_AN_PHY;
+       }
+       else
+              /* restore mode if it was changed before */
+              pl->cur_link_an_mode = pl->cfg_link_an_mode;



To prevent it from toggling all the time. So I do need to spend some more
attention to this part, cause this also may not be 100% ok, if changing 
phylink core would be done.




The reason I do it in phylink, because of changing to MLO_AN_PHY when
there is a PHY attached. mtk_eth_soc would not. Be aware of PHY attached.

So that's why I've opened the rfc.

See which way to go with this in an acceptable way, preferably using MLO_AN_PHY.