On 6.4.2016 03:28, Masahiro Yamada wrote: > Hi. > > > 2016-04-06 4:09 GMT+09:00 Simon Glass <s...@chromium.org>: >> Hi Michal, >> >> On 5 April 2016 at 04:15, Michal Simek <michal.si...@xilinx.com> wrote: >>> Hi Simon, >>> >>> On 5.4.2016 02:03, Simon Glass wrote: >>>> Hi Michal, >>>> >>>> On 4 April 2016 at 11:50, Michal Simek <michal.si...@xilinx.com> wrote: >>>>> Create CMD_FDT Kconfig entry to have an option to disable fdt command >>>>> which is not required for small configuration which requires libfdt >>>>> only. >>>>> Enable it by default for all targets which enables OF_LIBFDT. >>>>> >>>>> Signed-off-by: Michal Simek <michal.si...@xilinx.com> >>>>> --- >>>>> >>>>> cmd/Kconfig | 7 +++++++ >>>>> cmd/Makefile | 2 +- >>>>> 2 files changed, 8 insertions(+), 1 deletion(-) >>>>> >>>>> diff --git a/cmd/Kconfig b/cmd/Kconfig >>>>> index fe8b4f0510da..8703cdb4a9be 100644 >>>>> --- a/cmd/Kconfig >>>>> +++ b/cmd/Kconfig >>>>> @@ -173,6 +173,13 @@ config CMD_ELF >>>>> help >>>>> Boot an ELF/vxWorks image from the memory. >>>>> >>>>> +config CMD_FDT >>>>> + bool "Flattened Device Tree utility commands" >>>>> + default y >>>> >>>> Should that be: >>>> >>>> default y if OF_LIBFDT >>>> >>>> ? >>>> >>>>> + depends on OF_LIBFDT >>>>> + help >>>>> + Do FDT related setup before booting into the Operating System. >>>>> + >>> >>> >>> In recent commits to this file both formats are used. >>> >>> +config CMD_BLOCK_CACHE >>> + bool "blkcache - control and stats for block cache" >>> + depends on BLOCK_CACHE >>> + default y if BLOCK_CACHE >>> >>> even looks non standard. >>> >>> +config CMD_BOOTEFI >>> + bool "bootefi" >>> + depends on EFI_LOADER >>> + default y >>> >>> I am happy to change whatever style you prefer but I think it should be >>> synchronized. The efi one was Reviewed by you. :-) >> >> I think Masahiro knows most about this. If it works it's fine with me. >> The way you have it is more intuitive and I prefer it. But he did >> point at a problem at some point. > > > I think "depends on OF_LIBFDT" > is correct in this case. > > > do_fdt() calls fdt_fixup_memory(), which is defined in common/fdt_support.c, > which is enabled by CONFIG_OF_LIBFDT. > > So, CMD_FDT should depend on OF_LIBFDT. > Otherwise, "make menuconfig" would allow users > to enable CMD_FDT without OF_LIBFDT, > which would cause link error. > > >> One other question - won't this disable the 'fdt' command for many boards? > > > config CMD_FDT > bool "Flattened Device Tree utility commands" > default y > depends on OF_LIBFDT > > > "default y" cares about it. > So, if CONFIG_OF_LIBFDT is enabled in the defconfig, > CONFIG_CMD_FDT should be enabled as well. > > > > But the following 6 boards opt out of Kconfig. > They define CONFIG_OF_LIBFDT in their config headers, > so this patch would disable "fdt" command for them. > > include/configs/legoev3.h > include/configs/ma5d4evk.h > include/configs/pic32mzdask.h > include/configs/stm32f746-disco.h > include/configs/xilinx-ppc.h > include/configs/zipitz2.h > > > > Could you move them to defconfigs? >
I have sent v2 to address these. I have used buildman and there is up to 2k difference when symbol is in Kconfig. Not sure why but it shouldn't be big deal. Thanks, Michal _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot