Hi Francesco,
On 28.07.22 09:53, Francesco Dolcini wrote:
Hello Tom, and all
On Thu, Jul 28, 2022 at 06:23:05AM +0000, Josua Mayer wrote:
This approach does work in Linux, but the maintainers there have rejected it for producing unnecessary errors in the boot log. For U-Boot imo this approach would be acceptable.
My current understanding is that U-Boot device tree are supposed to be the same as the Linux one. They are somehow expected to get out-of-sync, but it should be always possible to just take the Linux DTS and merge those back in U-Boot.
Right - else there should not be the effort to take the DTS in sync with Linux, as much as possible.
This is what Marcel did, and it was really triggered by me sending a one line patch to fix a specific DTS issue [1]. Unfortunately once you go down the rabbit hole because of the way dts/dtsi include tree is designed he had to update lot of boards and not just the Toradex one we could test.
I understand the reason from Linux Maintainer. I had really preferred a way where all phys are disabled instead of "okay", and U-Boot can detect which is really available (maybe in board code) and fix the fdt for Linux. Is this possible ? The FEC should be on, and yes, there could be something in board code (what the DTS are assume to remove), but DTS can remain the same.
This is going into the opposite direction. What should we do?
I see the same, we asked the maintainers to sync DTs with Linux and avoid U-Boot's versions.
Regards, Stefano
Francesco
[1] https://lore.kernel.org/all/20210923115759.GD31748@bill-the-cat/