On Tue, Oct 25, 2022 at 02:07:55PM +0200, Stefano Babic wrote: > On 25.10.22 11:28, Stefano Babic wrote: > > On 25.10.22 10:50, Oliver Graute wrote: > > > On 24/10/22, Stefano Babic wrote: > > > > On 24.10.22 13:59, Oliver Graute wrote: > > > > > On 24/10/22, Stefano Babic wrote: > > > > > > On 24.10.22 13:37, Fabio Estevam wrote: > > > > > > > Hi Stefano and Oliver, > > > > > > > > > > > > > > On Tue, May 3, 2022 at 7:49 AM Oliver Graute > > > > > > > <oliver.gra...@kococonnector.com> wrote: > > > > > > > > > > > > > > > > Switch to use binman to pack images > > > > > > > > > > > > > > > > Signed-off-by: Oliver Graute <oliver.gra...@kococonnector.com> > > > > > > > > Reviewed-by: Fabio Estevam <feste...@denx.de> > > > > > > > > Reviewed-by: Simon Glass <s...@chromium.org> > > > > > > > > --- > > > > > > > > Changes for v3 > > > > > > > > -added Reviewed-by > > > > > > > > -rebased on master > > > > > > > > > > > > > > > > arch/arm/dts/imx8qm-cgtqmx8.dts | 1 + > > > > > > > > arch/arm/mach-imx/imx8/Kconfig | 1 + > > > > > > > > board/congatec/cgtqmx8/imximage.cfg | 4 ++-- > > > > > > > > configs/cgtqmx8_defconfig | 2 +- > > > > > > > > doc/board/congatec/cgtqmx8.rst | 3 +-- > > > > > > > > 5 files changed, 6 insertions(+), 5 deletions(-) > > > > > > > > > > > > > > Just realized that v3 does not include > > > > > > > imx8qm-u-boot.dtsi and this breaks CI. > > > > > > > > > > > > > > > > > > > Right - as well as V4, this is what I tried to merge and > > > > > > breaks. I have drop > > > > > > it. > > > > > > > > > > > > > The previous v2 did include it: > > > > > > > https://patchwork.ozlabs.org/project/uboot/patch/20220221090139.10066-1-oliver.gra...@kococonnector.com/ > > > > > > > > > > > > Thanks for noting this - Oliver, can you check and repost a V5, > > > > > > then ? > > > > > > > > > > The arch/arm/dts/imx8qm-u-boot.dtsi is applied on the binman Patch > > > > > for > > > > > "PATCH v3 imx: imx8qm-rom7720: switch to binman" > > > > > > > > > > So if you first apply the patches for the imx8qm-rom7720 > > > > > Board it should work. > > > > > > > > > > > > Ok, thanks for tipp - I apply both and will test again > > > > > > I saw the CI Pipeline failing. Perhaps the reason is a > > > missing CONFIG_FIT_EXTERNAL_OFFSET? > > > > > > > I see this, but just applying the two patches, multiple boards are broken: > > imx8qm_rom7720_a1_4G cgtqmx8 imx8qm_mek imx8qxp_mek deneb giedi > > > > > > > +++ b/configs/cgtqmx8_defconfig > > > @@ -20,6 +20,7 @@ CONFIG_HAS_CUSTOM_SYS_INIT_SP_ADDR=y > > > CONFIG_CUSTOM_SYS_INIT_SP_ADDR=0x80200000 > > > CONFIG_REMAKE_ELF=y > > > CONFIG_FIT=y > > > +CONFIG_FIT_EXTERNAL_OFFSET=0x3000 > > > CONFIG_SPL_LOAD_FIT=y > > > # CONFIG_USE_SPL_FIT_GENERATOR is not set > > > > > > Can you confirm? > > > > > > > I let CI run just to see if cgtqmx8 is fixed. > > So this seems to fix the Congatec board. Rather, all DTS for i.MX8 are built > together, and other boards have not switched to binman. I feel this > dependency too hard. Should we maybe split the targets in > arch/arm/dts/Makefile ? All DTS are build according to CONFIG_ARCH_IMX8, so > congatec DTS is compiled even in case of deneb and gied. > > See https://source.denx.de/u-boot/custodians/u-boot-imx/-/pipelines/13928.
IMHO at this point we need to just untested-convert those last outstanding board or boards, or remove them. This migration has been dragging on and on and I'm sure I've mailed the board maintainers a few times. -- Tom
signature.asc
Description: PGP signature