[PATCH 4/8] Revert "fdt: Allow the devicetree to come from a bloblist"

[Simon Glass]

Hi Tom,

On Tue, 3 Dec 2024 at 12:45, Simon Glass <sjg at chromium.org> wrote:
>
> Hi Tom,
>
> On Tue, 3 Dec 2024 at 09:51, Tom Rini <trini at konsulko.com> wrote:
> >
> > On Tue, Dec 03, 2024 at 09:37:34AM -0700, Simon Glass wrote:
> >
> > > Hi Tom,
> > >
> > > If you like, you can apply the reverts first, since they break boards.
> > > That should not be controversial, particularly as I explained at the
> > > time that this was likely to cause problems.
> > >
> > > Then you can leave the discussion on OF_BLOBLIST open.
> > >
> > > Just a thought.
> >
> > Absolutely not. Again, 70fe23859437ffe4efe0793423937d8b78ebf9d6 says:
> > "Standard passage provides for a bloblist to be passed from one firmware
> > phase to the next. That can be used to pass the devicetree along as well.
> > Add an option to support this.
> >
> > Tests for this will be added as part of the Universal Payload work."
> >
> > And your new patch says exactly the same thing, but then:
> > "Note: This is the correct way to deal with bloblist, since it allows
> > boards to choose whether they want to use the devicetree from there, or
> > not." and a large number of patchwork links. And after squashing the
> > three commits together to see what's being changed, yes, you're
> > rewriting a whole bunch of the logic without any explanation of why. It
> > seems to fix 3 Chromebooks and isn't tested on anything else. This is
> > why I'm saying it's unreviewable.
> >
> > So no, at the moment I'm inclined to wait for further reports of
> > problems, and we're coming up on a year since 70fe23859437 was merged,
> > before making any sort of last-minute change here.
>
> OK, thanks for confirming that.

We could of course imply OF_BLOBLIST which would still allow my boards
to work, by disabling it.

The opposition to this patch is mystifying to me, particularly as it
seems like a clear mess-up.

I'll apply this to my tree for now. It is certainly correct.

Regards,
Simon

Applied to ci/master, thanks!