[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <87pmb9vdti.fsf@miraculix.mork.no>
Date: Fri, 20 Jan 2023 08:56:57 +0100
From: Bjørn Mork <bjorn@...k.no>
To: "Russell King (Oracle)" <linux@...linux.org.uk>
Cc: netdev@...r.kernel.org, Felix Fietkau <nbd@....name>,
John Crispin <john@...ozen.org>,
Sean Wang <sean.wang@...iatek.com>,
Mark Lee <Mark-MC.Lee@...iatek.com>,
Lorenzo Bianconi <lorenzo@...nel.org>,
Daniel Golle <daniel@...rotopia.org>,
Alexander Couzens <lynxis@...0.eu>
Subject: Re: [PATCH net 2/3] net: mediatek: sgmii: autonegotiation is required
"Russell King (Oracle)" <linux@...linux.org.uk> writes:
> If you have 'managed = "in-band-status";' in your DT, that will set
> "mode" to be MLO_AN_INBAND, and phylink_autoneg_inband(mode) will be
> true - which should result in the link being programmed for in-band
> mode. You should also find that mtk_pcs_get_state() gets called.
>
> Hmm, it looks like setting ss->pcs[i].pcs.poll to true was missed
> when support for inband was properly added, so that might be the
> issue there - as the mtk ethernet driver doesn't make use of
> phylink_mac_change().
OK, this doesn't just work as-is either. But I guess that's something
else. Will try to debug some more.
But I wonder: Why would I want to use "in-band-status" here? Is that
the preferred mode? Probably stupid question. But shouldn't we also
make this link work without it, whatever that takes?
Note that I don't have to care about unknown phys. No SFP cage on this
board.
Bjørn
Powered by blists - more mailing lists