[U-Boot] [PATCH 2/2] ARM: omap3_logic: Enable SPL booting device tree

[Simon Goldschmidt]

Am Sa., 16. Feb. 2019, 22:34 hat Tom Rini <trini at konsulko.com> geschrieben:

> On Sat, Feb 16, 2019 at 03:30:08PM -0600, Adam Ford wrote:
> > On Sat, Feb 16, 2019 at 3:22 PM Tom Rini <trini at konsulko.com> wrote:
> > >
> > > On Sat, Feb 16, 2019 at 03:08:02PM -0600, Derald D. Woods wrote:
> > > > On Sat, Feb 16, 2019 at 03:49:01PM -0500, Tom Rini wrote:
> > > > > On Sat, Feb 16, 2019 at 01:31:24PM -0600, Adam Ford wrote:
> > > > >
> > > > > > With the generic omap3-u-boot.dtsi file available, this patch
> > > > > > increased the memory of the various incarnations of the
> omap3_logic
> > > > > > board, and points their respective u-boot.dtsi files to the newly
> > > > > > created generic one, and removes the PLATDATA from the board
> file.
> > > > > >
> > > > > > These are all done at once because the're all utilizing the same
> > > > > > omap3logic.c board file.
> > > > > >
> > > > > > Signed-off-by: Adam Ford <aford173 at gmail.com>
> > > > >
> > > > > Ah-haha!  Thank you!  I also needed to delete the platform data
> and then
> > > > > beagleboard is also booting.
> > > > >
> > > >
> > > > Can we get a quick summary of the conversion steps? My omap3-evm did
> not
> > > > get a serial console when applying patch[1] and using example
> > > > modifications from patch[2].
> > >
> > > I'm cleaning up Beagle now but:
> > > - Switch to using the new omap3-u-boot.dtsi
> > > - Remove all platdata from the board files
> > > - These are a must:
> > > CONFIG_SYS_MALLOC_F_LEN=0x4000
> > > CONFIG_SPL_SYS_MALLOC_F_LEN=0x400
> >
> > 0x400 or 0x4000?  I think mine is 0x4000
>
> It's 0x400, which is the default.  It may need to grow again as we set
> it to 0x1000 but only on AM33XX.  Maybe that's just a bit large, would
> have to check.
>

I just had a similar issue on socfpga (where this is set to 0x2000) and I
found most of the actually required ~1.2 KiB are used to instantiate (not
probe) uclasses and devices that are only probed after DDR is available.
I'm working on concepts to reduce this, but I haven't found a straight way
out, yet.

Regards,
Simon


> > > CONFIG_SPL_SYS_MALLOC_SIMPLE=y
> > > CONFIG_SPL_SEPARATE_BSS=y
> > > CONFIG_SPL_OF_CONTROL=y
> > > CONFIG_SPL_DM_SEQ_ALIAS=y
> > > CONFIG_SPL_OF_TRANSLATE=y
> > >
> > > And we tend to drop falcon mode due to size, but could find other
> things
> > > instead.
> >
> > I was able to drop EXT support in SPL.  At least the Logic PD
> > documentation states formatting the SD cards as FAT, so the stuff we
> > publish doesn't need EXT in SPL.  Falcon Mode was important for us to
> > keep, but I noticed it takes more space.
> > Luckily, I was able to retain it.
>
> I might be able to reduce the amount of removed parts on Beagle, but we
> don't have Falcon mode in use widely there I think.  And I need more
> space as I'm tempted to see if I can finally do things like support NAND
> env on Beagleboard and FAT env for xM and so forth.
>
> --
> Tom
> _______________________________________________
> U-Boot mailing list
> U-Boot at lists.denx.de
> https://lists.denx.de/listinfo/u-boot
>