On Sat, Oct 10, 2026 at 12:51:49PM +0200, Łukasz Majewski wrote: > Hi Tom, Bradley, > > > On Fri, Oct 09, 2026 at 07:36:39PM +0000, Bradley Morgan wrote: > > > > > MPC8548 silicon is no longer manufactured, and the MPC8548CDS board > > > is obsolete. Remove the board and its SoC support. > > > > > > Signed-off-by: Bradley Morgan <[email protected]> > > > > So, globally. As has been discussed on IRC already, please take a more > > neutral removal message in v2. The problem isn't that the boards > > aren't manufactured, it's that (likely) they aren't being actively > > tested and maintained in the community. We're fine with "obsolete" > > hardware if someone is maintaining it. Also, please rework the CC > > list (outside the cover letter) to include the listed board > > maintainers, and then only people that have touched the relevant code > > recently, perhaps with a manual audit or exclude some global or > > common files, when looking for who else to CC. Thanks. > > > > Maybe I would add my 2 cents... > > There are a lot of MPC85xx systems out there - and probably (my guess) > is that at some point some customers would be "forced" to refurbish > them. > > Some companies need very long time to realize that their crucial legacy > products may be removed from u-boot. I would opt for longer and more > "broadcast" obsolence process. > > > I had similar situation in the past when some Samsung SoCs were removed > from u-boot - nobody was willing to pay for re-adding them upstrem - > this means that they were kept out of tree.
Thanks for the feedback here. Let me make this a message to the future, when someone is needing to bring back an old platform we've removed. As a project we are happy to bring back removed platforms. You can see https://git.u-boot-project.org/u-boot/u-boot/-/commit/93d204f7175c as an example where we've done that, even, for the exact reasons of "someone has a long term support need". There's not a demand to fix all of the code problems of the SoC support, etc, itself. There is a need to bring things up to modern standards. This is something that even if you keep the board(s) out of tree are likely needing to be done anyways as we'll have dropped other legacy APIs too. The time and effort needed to bring something back *and* upstream it is not going to be much more than to bring it back to start with. And to anyone who finds this helpful now, to pass along to people who control budgets, a little bit of time now to keep a reference platform supported and tested will save money down the road when you need to bring support back and fix things that had already broken before they were removed. There's a cost the project has to always pay for a given platform. We currently support building for 1541 different platforms (and even more physical devices). We're always happy to add more. The cost is simply someone saying "I will test this device sometimes to make sure it still runs". We already make sure it still builds. The project makes 4 releases a year, on a regular schedule. 4 days a year should not be a hard amount of time for a company to dedicate. And with that, thank you so very much to all of the people doing this as a hobby or otherwise just for fun in their free time. And thank you too, to the people doing this as part of their job (and their company for supporting them). -- Tom
signature.asc
Description: PGP signature
