Hi Simon,
2016-10-06 1:09 GMT+09:00 Simon Glass <s...@chromium.org>: > Hi Masahiro, > > On 4 October 2016 at 21:27, Masahiro Yamada > <yamada.masah...@socionext.com> wrote: >> Hi Simon, >> >> 2016-10-05 0:37 GMT+09:00 Simon Glass <s...@chromium.org>: >> >>>> diff --git a/common/image-fdt.c b/common/image-fdt.c >>>> index 3d23608..91970d4 100644 >>>> --- a/common/image-fdt.c >>>> +++ b/common/image-fdt.c >>>> @@ -458,6 +458,11 @@ __weak int ft_verify_fdt(void *fdt) >>>> return 1; >>>> } >>>> >>>> +__weak int arch_fixup_fdt(void *blob) >>>> +{ >>>> + return 0; >>>> +} >>> >>> Do we have to have a weak function? I was hoping we could avoid these >>> since they make it hard to figure out at build time what code is >>> executed. >>> >> >> >> This hunk is just reverting Michal's commit e2f88dfd2d9671. >> >> Is it better to add an empty stub to every architecture that may call it? > > IMO all the FDT fixups need work. Perhaps we need a linker list > approach so we can declare these fixups more easily? Or perhaps that > will just make things harder to figure out? This is up to you. My interest is to not touch memory node, but need other DT fixups. I sent v2 with no-op stubs instead of the weak function. -- Best Regards Masahiro Yamada _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot