On Thursday 28 October 2021 10:42:27 Robert Marko wrote: > On Wed, Oct 27, 2021 at 10:34 PM Tom Rini <tr...@konsulko.com> wrote: > > > > On Wed, Oct 27, 2021 at 10:11:36PM +0200, Robert Marko wrote: > > > On Wed, Oct 27, 2021 at 10:07 PM Tom Rini <tr...@konsulko.com> wrote: > > > > > > > > On Wed, Oct 27, 2021 at 09:53:24PM +0200, Robert Marko wrote: > > > > > On Wed, Oct 27, 2021 at 9:48 PM Marek Behún <marek.be...@nic.cz> > > > > > wrote: > > > > > > > > > > > > On Wed, 27 Oct 2021 21:40:11 +0200 > > > > > > Robert Marko <robert.ma...@sartura.hr> wrote: > > > > > > > > > > > > > To me, it also doesn't make sense to do it like that as what's the > > > > > > > purpose of having stuff that is completely unsupported in U-boot > > > > > > > in > > > > > > > DTS? > > > > > > > > > > > > It makes sense because the device tree should compeltely describe > > > > > > the > > > > > > device. It doesn't matter if some of it's components don't have > > > > > > U-Boot > > > > > > drivers. Devicetree is independent on sw, and should be the same for > > > > > > Linux, BSD, and U-Boot. > > > > > > > > > > Trust me, I wish this was true, but I don't see it happening anytime > > > > > soon > > > > > as bindings will always change but it is something to aspire to. > > > > > > > > > > > > > Also, take a look at other Armada 7k/8k devices, they are all > > > > > > > doing it > > > > > > > the same way as I am. > > > > > > > > > > > > Some of them (most?) are, but that's just because we haven't synced > > > > > > them yet. > > > > > > > > > > > > But you don't need to do it now, IMO, if it would take too much > > > > > > work. > > > > > > > > > > The thing is that you cant just copy/paste the Linux DTS, it won't > > > > > compile > > > > > as the SoC DTSI as well as CP DTSI are different. > > > > > Otherwise having a -u-boot.dtsi would work and I honestly don't have > > > > > time > > > > > to sync those and then fix all of the regressions. > > > > > > > > The intention is that you copy/paste (well, cp) the DTS files in to > > > > U-Boot and then use -u-boot.dtsi to do whatever further is needed. > > > > > > I understand that, but SoC and CP DTSI-s are really out of sync with > > > Linux, > > > it would require to pretty much override/delete half the DTS for > > > -u-boot.dtsi > > > for it to work. > > > Otherwise, other 7k/8device would use -u-boot.dtsi as well. > > > > OK, so it needs to be done then and fixed up. The files weren't > > supposed to get out of sync to start with. > > I wish it were that simple, the biggest issue is dealing with the > regressions after > as U-boot drivers have also diverged in binding to some upstream ones. > And I seriously don't have time to do this all on my own and also > refactor all existing > 7k/8k boards.
Now I have done this for Armada 3720 code and all its boards: https://patchwork.ozlabs.org/project/uboot/patch/20211126135715.26485-10-ka...@kernel.org/ https://patchwork.ozlabs.org/project/uboot/list/?series=285999&state=* So I hope and expect that somebody else who develops Armada 7040 would do it for Armada 7040 code and board files...
