Ravi,
On 27/07/2023 08:36, Ravi Gunasekaran wrote:
On 7/26/23 6:22 PM, Nishanth Menon wrote:
On 14:44-20230726, Ravi Gunasekaran wrote: [...]
If it is the former, then would a duplicate mdio node help keep the changes to minimum?
That is worse hack. How does it make sense to have two mdio nodes to represent the same hardware block? we are trying to get closer to kernel dts support not farther away from it.
I know it's not a clean hack. In k3-am625-sk-u-boot.dtsi, as a hack, the mdio pinctrl gets added to cpsw node. So was wondering if a duplicate mdio pinctrl node could be added in the same file. Just wanted to know if we could skip the changes in CPSW driver posted by Maxime.
But Maxime's approach is much cleaner. We can just sync up kernel dts and not keeping adding the hack to every K3 SoC's DT.
Everything is sorted now as far as DT sync is concerned. Please see [1] and its dependencies.
Only thing missing is to have proper MDIO DM support for TI platforms in u-boot. Then we can get rid of the MDIO hack in am65-cpsw driver.
[1] - https://lore.kernel.org/all/20230726151027.2517151-1-nm@ti.com/