[U-Boot] [PATCH v3 09/14] tegra20: add u-boot.t2 target

[Simon Glass]

Hi Allen,

On Mon, Jun 11, 2012 at 4:09 PM, Allen Martin <amartin at nvidia.com> wrote:

> On Mon, Jun 11, 2012 at 01:21:13PM -0700, Simon Glass wrote:
> > Hi Allen,
> >
> > On Mon, Jun 11, 2012 at 12:22 PM, Allen Martin <amartin at nvidia.com
> <mailto:amartin at nvidia.com>> wrote:
> > On Sat, Jun 09, 2012 at 12:12:09PM -0700, Simon Glass wrote:
> > > Hi Allen,
> > >
> > > On Fri, Jun 8, 2012 at 2:16 PM, Allen Martin <amartin at nvidia.com
> <mailto:amartin at nvidia.com><mailto:amartin at nvidia.com<mailto:
> amartin at nvidia.com>>> wrote:
> > > Add target for tegra20 u-boot image.  This is a concatenation of tegra
> > > spl and normal u-boot binaries.
> > >
> > > Signed-off-by: Allen Martin <amartin at nvidia.com<mailto:
> amartin at nvidia.com><mailto:amartin at nvidia.com<mailto:amartin at nvidia.com>>>
> > > ---
> > >  .gitignore                      |    1 +
> > >  Makefile                        |   11 +++++++++++
> > >  board/nvidia/seaboard/config.mk<http://config.mk><http://config.mk>
> |    1 +
> > >  3 files changed, 13 insertions(+)
> > >  create mode 100644 board/nvidia/seaboard/config.mk<http://config.mk><
> http://config.mk>
> > >
> > > diff --git a/.gitignore b/.gitignore
> > > index 0f32fd8..b9192bf 100644
> > > --- a/.gitignore
> > > +++ b/.gitignore
> > > @@ -39,6 +39,7 @@
> > >  /u-boot.ais
> > >  /u-boot.dtb
> > >  /u-boot.sb<http://u-boot.sb><http://u-boot.sb>
> > > +/u-boot.t2
> > >
> > > What does t2 mean? If it is a binary file of some sort  perhaps
> > u-boot-t2.bin would be better?
> >
> > It's just means "tegra2".  I was following the convention that other
> > SPL builds use.  I don't have a strong opinion on the name though.
> >
> > OK, still would prefer a .bin on the end, but up to you.
>
> I'm ok with that, it keeps in line with the u-boot-dtb.bin
>
> >
> >
> > > +ifeq ($(CONFIG_OF_SEPARATE),y)
> > > +T2_UBOOT=$(obj)u-boot-dtb.bin
> > > +else
> > > +T2_UBOOT=$(obj)u-boot.bin
> > > +endif
> > >
> > > What is this logic for? The dtb file is separate but that doesn't
> > > necessarily mean that it must be immediately after the U-Boot
> > > image. We provide other options for packaging it, like
> > > getenv(). Maybe if you want to create this composite binary you
> > > should change its name (u-boot-dtb-t2.bin or u-boot-t2.bin) to
> > > indicate what it contains?
> >
> > This just picks up the name of the u-boot binary so it glues the SPL
> > to the right thing depending if devicetree is in use or not.
> >
> > OK, to avoid confusion I think you should change the name to include
> > the -dtb or not, rather than having a filename without the -dtb part
> > which does in fact include a dtb.
>
> Ok.
>
> > > I echo Stephen's comments. But also SPL is supposed to load U-Boot,
> > > so shouldn't this t2 binary do that?
> >
> > The t2 binary is the SPL and u-boot concatenated together into one
> > binary.  The whole thing will get loaded into memory by the tegra
> > BootROM.  The SPL knows the address of the real u-boot at compile time
> > and uses that as the address for the Cortex A9 to jump to when it
> > comes out of reset.
> >
> > OK I see, makes sense. I am interested in your comments as to whether we
> might move to a 'true SPL' later, where U-Boot is actually loaded by SPL.
>
> To keep this already complicated patch series under control I want to
> just assume the SPL and regular u-boot are glued together for now, but
> I'm definately thinking about the case where they are not.
>

Yes, no suggestion that we do this now, I was just interested in your plans.


>
> In particular part of my motivation for this work is the ability to
> have a version of the SPL that you can run in recovery mode when you
> can't trust or don't have a BCT to initialize RAM.  In that case the
> SPL would run out of IRAM and could take a BCT and u-boot from USB
> DFU.  There's no reason it couldn't be extended to take those from
> somewhere else the BootROM wouldn't normally be able to boot from like
> SATA or network.  And it wouldn't have to be just in recovery mode
> either I suppose.
>

OK I see. Yes SPL from IRAM would be more 'traditional'.


>
> -Allen
> --
> nvpublic
>

Regards,
Simon