Re: [PATCH net] net: freescale: ucc_geth: Return early when TBI found can't be found

[Maxime Chevallier <maxime.chevallier@bootlin.com>]

Hi Russell,

On 13/01/2026 19:44, Russell King (Oracle) wrote:
> On Tue, Jan 13, 2026 at 09:16:29AM +0100, Maxime Chevallier wrote:
>> Hi,
>>
>> On 13/01/2026 08:43, Maxime Chevallier wrote:
>>> In ucc_geth's .mac_config(), we configure the TBI block represented by a
>>> struct phy_device that we get from firmware.
>>>
>>> While porting to phylink, a check was missed to make sure we don't try
>>> to access the TBI PHY if we can't get it. Let's add it and return early
>>> in case of error
>>>
>>> Reported-by: kernel test robot <lkp@intel.com>
>>> Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
>>> Closes: https://lore.kernel.org/r/202601130843.rFGNXA5a-lkp@intel.com/
>>> Fixes: 53036aa8d031 ("net: freescale: ucc_geth: phylink conversion")
>>> Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
>>
>> Heh that's what I get from sending patches while having mild fever, the
>> patch title is all wrong and should be :
>>
>> net: freescale: ucc_geth: Return early when TBI PHY can't be found
>>
>> I'll wait for the 24h cooldown, grab some honey + milk and resend after :)
> 
> A question - based on dwmac:
> 
> When implementing dwmac to support 1000base-X, the dwmac doesn't
> implement the _full_ 1000base-X, but only up to the PCS. The PCS
> provides a TBI interface to the SerDes PHY provided by the SoC
> designer which acts as the PMA layer.
> 
> The talk here of TBI makes me wonder whether the same thing is going
> on with ucc_geth. Is the "TBI PHY" in fact the SerDes ?

Yeah I think it is indeed.

> Traditionally, we've represented the SerDes using drivers/phy rather
> than the drivers/net/phy infrastructure, mainly because implementations
> hvaen't provided anything like an 802.3 PHY register set, but moreover
> because the SerDes tends to be generic across ethernet, PCIe, USB, SATA
> etc (basically, anything that is a high speed balanced pair serial
> communication) and thus the "struct phy" from drivers/phy can be used
> by any of these subsystems.
> 

True, and I completely agree with that. The reason I didn't touch that
when porting to phylink is that the device I'm using, that has a
Motorola/Freescale/NXP MPC832x, doesn't have that TBI/RTBI block, so I
can't test that at all should we move to a more modern SerDes driver
(modern w.r.t when this driver was written) :(

Maxime