Hi Tom,
On Sun, 9 Feb 2025 at 09:49, Tom Rini <tr...@konsulko.com> wrote:
>
> On Sun, Feb 09, 2025 at 07:27:37AM -0700, Simon Glass wrote:
> > Hi Jonas,
> >
> > On Fri, 20 Dec 2024 at 09:50, Jonas Karlman <jo...@kwiboo.se> wrote:
> > >
> > > Hi Quentin and Simon,
> > >
> > > On 2024-12-13 15:30, Quentin Schulz wrote:
> > > > Hi Jonas,
> > > >
> > > > On 12/13/24 12:57 AM, Jonas Karlman wrote:
> > > >> The u-boot.rom image contain u-boot.img FIT instead of the FIT
> > > >> generated
> > > >> by binman for the u-boot-rockchip.bin image.
> > > >>
> > > >> Change to include the binman generated FIT for the u-boot.rom image.
> > > >>
> > > >> This change result in TF-A being included and the use sha256 instead of
> > > >> crc32 checksum in the u-boot.rom FIT.
> > > >>
> > > >> Signed-off-by: Jonas Karlman <jo...@kwiboo.se>
> > > >
> > > > IIRC Simon was fine with removing u-boot.rom support for RK3399 when I
> > > > asked months (probably years?) ago, if that hasn't changed, maybe we
> > > > should now.
> > >
> > > Good to know and fully agree, we should try to remove it now.
> > >
> > > I did a compare between u-boot-rockchip-spi.bin and u-boot.rom and after
> > > this series they are now identical for the full size of
> > > u-boot-rockchip-spi.bin.
> > >
> > > However, the u-boot.rom also has a copy of ~u-boot.bin and a fdtmap
> > > starting at 0x300000. I have no idea if they are used for anything,
> > > if they are it is probably not for bare metal booting (TPL+SPL).
> >
> > I'm not sure about the extra u-boot.bin, but the fdtmap is so that
> > 'binman ls -i xxx' works.
> >
> > >
> > > Regards,
> > > Jonas
> > >
> > > >
> > > > Especially since the only RK3399 with CONGIG_HAS_ROM set are Bob and
> > > > Kevin.
> > > >
> > > > @Simon, you said you tested the patch series on them, with which image
> > > > did you do that? Are you still fine with removing the u-boot.rom for
> > > > RK3399?
> >
> > Yes that's fine.
> >
> > Basically all my testing is in my lab now, which you should be able to
> > use with gitlab.
>
> I really wish you would use the community gitlab instance and not
> encourage people to use your personal instance instead.
I wish I didn't need one and anyway, that wasn't what I said. Jonas
has u-boot-rockchip, I believe, so should be able to push things and
run on my lab.
Perhaps the problem is that it's too complicated? If you could find a
way to allow this change in, people could just click the button and
may be more likely to use it?
.lab_template: &lab_dfn
stage: sjg-lab
rules:
- if: $SJG_LAB == "1"
when: always
- if: $SJG_LAB != "1"
when: manual
allow_failure: true
tags: [ 'lab' ]
Regards,
Simon
