From mboxrd@z Thu Jan 1 00:00:00 1970 From: Hannes Schmelzer Date: Fri, 29 Mar 2019 06:54:48 +0100 Subject: [U-Boot] [PATCH v1] colibri_vf: fix ethernet by adding explicit phy node In-Reply-To: <20190328222450.22308-1-marcel@ziswiler.com> References: <20190328222450.22308-1-marcel@ziswiler.com> Message-ID: List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de Marcel Ziswiler schrieb am 28.03.2019 23:24:50: > Von: Marcel Ziswiler > An: u-boot at lists.denx.de > Kopie: Hannes Schmelzer , Marcel Ziswiler > , Albert Aribaud , > Stefan Agner > Datum: 28.03.2019 23:25 > Betreff: [PATCH v1] colibri_vf: fix ethernet by adding explicit phy node > > From: Marcel Ziswiler > > The implicit fallback mechanism for searching the whole MDIO bus for at > least one PHY has been gone with the following commit b882005a18de > ("drivers/net/fec: phy_init: remove redundant logic"). This lead to the > Ethernet driver erroring out as follows: > > Net: Could not get PHY for FEC0: addr -19 > eth-1: fec at 400d1000 > Colibri VFxx # dhcp > Could not get PHY for FEC0: addr -19 > Could not get PHY for FEC0: addr -19 > Could not get PHY for FEC0: addr -19 > Could not get PHY for FEC0: addr -19 > No ethernet found. > Could not get PHY for FEC0: addr -19 > Colibri VFxx # > > This commit adds an explicit PHY node which makes it work again. > > Signed-off-by: Marcel Ziswiler > > --- Looks good to me, now the hardware is fully defined, which should be the normal case. I will provide today a patch which "brings back" the fallback mechanism within the "phy-subsystem". This should be the starting point for the disussion if we need such fallback or if we want describe our hardware. Reviewed-by: Hannes Schmelzer