RE: [PATCH] ARM: dts: imx6qdl-sabresd: change phy-mode to use rgmii-id

[Steve Twiss <stwiss.opensource@diasemi.com>]

Hi Fabio,

On 21 March 2019 11:17, Fabio Estevam,

> Subject: Re: [PATCH] ARM: dts: imx6qdl-sabresd: change phy-mode to use rgmii-id
> 
> Hi Abel,
> 
> On Thu, Mar 21, 2019 at 5:42 AM Abel Vesa <abel.vesa@nxp.com> wrote:
> 
> > > It seems we have other boards that need to be fixed and we can not
> > > have an old dtb with functional Ethernet with a new kernel.
> > >
> > > Does anyone know if this issue is AR8031 specific?
> >
> > I can confirm the same fix is works on imx6sx too.
> 
> imx6sx-sabresd also uses an AR8031 Ethernet PHY.
> 
> I also tested that changing the phy-mode to "rgmii-id" fixes Ethernet
> on pico-imx7d with AR8035.
> 
> So yes, we currently have lots of broken dtb's in mainline and I am
> wondering what is the proper fix here.

Agreed!
The DT should be an ABI. 

> Does anyone know what was the kernel commit that introduced such
> regressions?

I bisected to the original breakage (for the NFS rootfs) back to this commit:
commit 13d0ab6750b20957ac1466da4e44dc0af746ff28

Reverting this commit fixed the problem, but only for the kernel up to that
point: it was a long time ago, and several other fixes were added after that
which meant that by the time the kernel got to v5.0 it was working for me
again.

I bisected the breakage between v5.0 to v5.1-rc1 as this:
commit 3acca1dd17060332cfab15693733cdaf9fba1c90
... which doesn't make sense to me.

Regards,
Steve

---

commit 13d0ab6750b20957ac1466da4e44dc0af746ff28
Author: Heiner Kallweit <hkallweit1@gmail.com>
Date:   Wed Jan 16 08:07:38 2019 +0100

    net: phy: check return code when requesting PHY driver module
    
    When requesting the PHY driver module fails we'll bind the genphy
    driver later. This isn't obvious to the user and may cause, depending
    on the PHY, different types of issues. Therefore check the return code
    of request_module(). Note that we only check for failures in loading
    the module, not whether a module exists for the respective PHY ID.
    
    v2:
    - add comment explaining what is checked and what is not
    - return error from phy_device_create() if loading module fails
    
    Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
    Reviewed-by: Andrew Lunn <andrew@lunn.ch>
    Signed-off-by: David S. Miller <davem@davemloft.net>

---

commit 3acca1dd17060332cfab15693733cdaf9fba1c90
Author: Heiner Kallweit <hkallweit1@gmail.com>
Date:   Mon Mar 4 19:39:03 2019 +0100

    net: dsa: mv88e6xxx: add call to mv88e6xxx_ports_cmode_init to probe for new DSA framework
    
    In the original patch I missed to add mv88e6xxx_ports_cmode_init()
    to the second probe function, the one for the new DSA framework.
    
    Fixes: ed8fe20205ac ("net: dsa: mv88e6xxx: prevent interrupt storm caused by mv88e6390x_port_set_cmode")
    Reported-by: Shaokun Zhang <zhangshaokun@hisilicon.com>
    Suggested-by: Andrew Lunn <andrew@lunn.ch>
    Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
    Reviewed-by: Andrew Lunn <andrew@lunn.ch>
    Signed-off-by: David S. Miller <davem@davemloft.net>