[U-Boot] [PATCH v2 3/5] sunxi: DT: A64: update device tree file for Allwinner A64 SoC

[Andre Przywara]

On Wed, 17 Oct 2018 10:09:55 -0700
Vasily Khoruzhick <anarsoul at gmail.com> wrote:

> On Wed, Oct 17, 2018 at 9:54 AM Andre Przywara
> <andre.przywara at arm.com> wrote:
> >
> > On Wed, 17 Oct 2018 17:29:58 +0200
> > Maxime Ripard <maxime.ripard at bootlin.com> wrote:
> >
> > Hi,
> >  
> > > On Wed, Oct 17, 2018 at 04:18:41PM +0100, Andre Przywara wrote:  
> > > > On Tue, 16 Oct 2018 22:09:30 -0700
> > > > Vasily Khoruzhick <anarsoul at gmail.com> wrote:
> > > >
> > > > Hi,
> > > >  
> > > > > Updates the device tree file from the the Linux tree as of
> > > > > v4.19-rc4, exactly Linux commit:  
> > > >
> > > > Does this work easily without syncing the .dts files as well?
> > > >  
> > > > > commit 7876320f8880 (tag: v4.19-rc4)
> > > > > Author: Linus Torvalds <torvalds at linux-foundation.org>
> > > > > Date:   Sun Sep 16 11:52:37 2018 -0700
> > > > >
> > > > >     Linux 4.19-rc4  
> > > >
> > > > So this sounds like the right thing to do, but in this
> > > > particular case breaks Ethernet with UEFI booting in all
> > > > distribution installers or kernels out there (except >= Linux
> > > > 4.19-rc1). I consider this a major use case of U-Boot's DTB, so
> > > > what do we do about this?
> > > > The reason is that we dropped the "syscon" compatible string at
> > > > the end of the system-controller node, which older kernels rely
> > > > on to find the syscon node.
> > > > I suggested to re-add this[1], but didn't have much success,
> > > > unfortunately.
> > > > The easiest would be to re-add (or not remove) "syscon" for
> > > > U-Boot's copy, but this would mean a deviation from the Linux
> > > > DT's. I am fine with this, but would like to hear more
> > > > opinions.  
> > >
> > > tl; dr: You want to build something robust on top of assumptions
> > > that have never been guaranteed.  
> >
> > I just find it sad that "we do not guarantee" translates into "we
> > actually don't even try". My understanding of your concern is that
> > you can't rule this breakage out forever, which I can understand.
> > But if it is possible to avoid breaking compatibility, I believe we
> > should try this, on a case-by-case base.
> > We have been pretty good so far (leaving alone the MMC regulator
> > breakage in v4.15, but A64 support was quite basic before that
> > anyway). 
> > > I'm fine with having it as a U-Boot
> > > addition, if it's what it takes.  
> >
> > Sounds good to me! Thanks!
> >
> > I will check what's the easiest path to get the 4.20 DTs into
> > U-Boot, considering a stopover at the Pinebook DTs, and send
> > patches ASAP.  
> 
> So do you want me to wait for you to submit a patch that syncs
> sun50i-a64.dtsi?

I will review the rest of your series later tonight and try to suggest
something that's the easiest for you then.

Cheers,
Andre