On Fri, Jul 21, 2017 at 6:18 PM, Tom Rini <tr...@konsulko.com> wrote: > On Fri, Jul 21, 2017 at 03:10:10PM -0400, Rob Clark wrote: > >> Snapdragon is a bit of a funny case, where we want to build the fdt to >> pass to lk, which loads us, but otherwise want to treat it as OF_BOARD, >> since lk will patch the fdt (for example, to insert simple-framebuffer >> node). >> >> Signed-off-by: Rob Clark <robdcl...@gmail.com> >> --- >> Makefile | 6 ++++++ >> 1 file changed, 6 insertions(+) >> >> diff --git a/Makefile b/Makefile >> index 452596485d..0749cdfc5d 100644 >> --- a/Makefile >> +++ b/Makefile >> @@ -780,6 +780,12 @@ ALL-$(CONFIG_SPL_FRAMEWORK) += u-boot.img >> endif >> ALL-$(CONFIG_TPL) += tpl/u-boot-tpl.bin >> ALL-$(CONFIG_OF_SEPARATE) += u-boot.dtb >> +# Snapdragon devices, where lk/aboot loads u-boot is kind of a special >> +# case, because lk loads the fdt which is embedded (along with u-boot) >> +# in a boot.img. But it also does some fdt fixups. So generally we >> +# want to treat it like CONFIG_OF_BOARD=y, except that we also want >> +# to build the dtb's >> +ALL-$(CONFIG_ARCH_SNAPDRAGON) += u-boot.dtb >> ifeq ($(CONFIG_SPL_FRAMEWORK),y) >> ALL-$(CONFIG_OF_SEPARATE) += u-boot-dtb.img >> endif > > We really can't just select OF_SEPARATE here? Because the code in > lib/fdtdec.c is wrong, as it stands, yes? If so, could we shuffle that > code right to allow for the case of "we get our DT modified by firmware > that loads us" ? That doesn't soune like a super uncommon case moving > forward.
from a quick look, I think the only reason we need OF_BOARD is the call to board_fdt_blob_setup() in fdtdec.. I guess we could make that a weak sym, and call it unconditionally.. in that case I think OF_SEPARATE would work. If you think that is a cleaner solution, I can give that a try. BR, -R _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de https://lists.denx.de/listinfo/u-boot
