Hi Hans, On 29 October 2014 03:32, Hans de Goede <hdego...@redhat.com> wrote: > Hi, > > On 10/29/2014 04:02 AM, Simon Glass wrote: >> Hi Hans, >> >> On 28 October 2014 03:13, Hans de Goede <hdego...@redhat.com> wrote: >>> Hi, >>> >>> On 10/28/2014 01:04 AM, Simon Glass wrote: >>>> Hi Hans, >>>> >>>> On 24 October 2014 02:38, Hans de Goede <hdego...@redhat.com> wrote: >>>>> Hi, >>>>> >>>>> On 10/23/2014 06:02 AM, Simon Glass wrote: >>>>>> For now we won't want to mess with the existing configurations. Create a >>>>>> new one which will enable device tree and driver model. Note that this >>>>>> brings the device tree binary into u-boot-sunxi-with-spl.bin. >>>>>> >>>>>> Signed-off-by: Simon Glass <s...@chromium.org> >>>>>> --- >>>>>> >>>>>> Changes in v2: None >>>>>> >>>>>> Makefile | 3 ++- >>>>>> arch/arm/dts/Makefile | 1 + >>>>>> board/sunxi/MAINTAINERS | 1 + >>>>>> configs/Linksprite_pcDuino3_fdt_defconfig | 8 ++++++++ >>>>>> include/configs/sun7i.h | 4 ++++ >>>>>> 5 files changed, 16 insertions(+), 1 deletion(-) >>>>>> create mode 100644 configs/Linksprite_pcDuino3_fdt_defconfig >>>>>> >>>>>> diff --git a/Makefile b/Makefile >>>>>> index 99097e1..816f859 100644 >>>>>> --- a/Makefile >>>>>> +++ b/Makefile >>>>>> @@ -941,7 +941,8 @@ u-boot-nand.gph: u-boot.bin FORCE >>>>>> ifneq ($(CONFIG_SUNXI),) >>>>>> OBJCOPYFLAGS_u-boot-sunxi-with-spl.bin = -I binary -O binary \ >>>>>> --pad-to=$(CONFIG_SPL_PAD_TO) >>>>>> --gap-fill=0xff >>>>>> -u-boot-sunxi-with-spl.bin: spl/sunxi-spl.bin u-boot.img FORCE >>>>>> +u-boot-sunxi-with-spl.bin: spl/sunxi-spl.bin \ >>>>>> + u-boot$(if $(CONFIG_OF_CONTROL),-dtb,).img FORCE >>>>>> $(call if_changed,pad_cat) >>>>>> endif >>>>>> >>>>>> diff --git a/arch/arm/dts/Makefile b/arch/arm/dts/Makefile >>>>>> index 3299817..564eb76 100644 >>>>>> --- a/arch/arm/dts/Makefile >>>>>> +++ b/arch/arm/dts/Makefile >>>>>> @@ -1,3 +1,4 @@ >>>>>> +dtb-$(CONFIG_SUN7I) += sun7i-a20-pcduino3.dtb >>>>> >>>>> Please make this >>>>> >>>>> dtb-$(CONFIG_PCDUINO3) += sun7i-a20-pcduino3.dtb >>>> >>>> Well I can do that, but this should have no effect if we don't define >>>> CONFIG_OF_CONTROL. The way I have it we will be able to add more >>>> device tree files for sun7x without anyone having to change it. The >>>> target should build all sun7i device trees eventually. >>> >>> Ah, I see, ok lets keep this as is then. >>> >>>> >>>>> >>>>>> dtb-$(CONFIG_S5PC100) += s5pc1xx-smdkc100.dtb >>>>>> dtb-$(CONFIG_S5PC110) += s5pc1xx-goni.dtb >>>>>> dtb-$(CONFIG_EXYNOS4) += exynos4210-origen.dtb \ >>>>>> diff --git a/board/sunxi/MAINTAINERS b/board/sunxi/MAINTAINERS >>>>>> index 4f32195..3dbfe15 100644 >>>>>> --- a/board/sunxi/MAINTAINERS >>>>>> +++ b/board/sunxi/MAINTAINERS >>>>>> @@ -21,6 +21,7 @@ F: configs/A20-OLinuXino_MICRO_defconfig >>>>>> F: configs/Bananapi_defconfig >>>>>> F: configs/i12-tvbox_defconfig >>>>>> F: configs/Linksprite_pcDuino3_defconfig >>>>>> +F: configs/Linksprite_pcDuino3_fdt_defconfig >>>>>> F: configs/qt840a_defconfig >>>>>> >>>>>> CUBIEBOARD2 BOARD >>>>>> diff --git a/configs/Linksprite_pcDuino3_fdt_defconfig >>>>>> b/configs/Linksprite_pcDuino3_fdt_defconfig >>>>>> new file mode 100644 >>>>>> index 0000000..e5f6236 >>>>>> --- /dev/null >>>>>> +++ b/configs/Linksprite_pcDuino3_fdt_defconfig >>>>>> @@ -0,0 +1,8 @@ >>>>>> +CONFIG_SPL=y >>>>>> +CONFIG_SYS_EXTRA_OPTIONS="PCDUINO3,AXP209_POWER,SUNXI_GMAC,AHCI,SATAPWR=SUNXI_GPH(2),USB_EHCI,DM" >>>>>> +CONFIG_FDTFILE="sun7i-a20-pcduino3.dtb" >>>>>> +CONFIG_DEFAULT_DEVICE_TREE="sun7i-a20-pcduino3" >>>>> >>>>> Hmm, I don't like this doubling of info. CONFIG_FDTFILE is sort of a >>>>> sunxi specific thing >>>>> here (some other boards have local usage of its too, but it is not used >>>>> globably. >>>>> >>>>> Can you do a patch (and prepend it to this one in the next posting of >>>>> this series), >>>>> which replaces sunxi's CONFIG_FDTFILE usage with >>>>> CONFIG_DEFAULT_DEVICE_TREE, dropping >>>>> the .dtb appending in the defconfigs and instead append the .dtb in >>>>> include/configs/sunxi-commom.h >>>>> >>>>> Note sunxi uses CONFIG_FDTFILE in 3 places: >>>>> >>>>> 1) board/sunxi/Kconfig >>>>> 2) various board/*_defconfig files >>>>> 3) include/configs/sunxi-common.h >>>> >>>> But isn't that used for passing to the kernel? Do you want to drop the >>>> ability to have a separate file for each? >>> >>> A good question, there are 2 things here: >>> >>> 1) This is just the filename, it gets combined with a "basedir" set through >>> extlinux.conf >>> when booting the real kernel, so they can still be different >>> >>> 2) TBH even if this would make it mandatory for them to be 1 and the same >>> file, I would >>> not consider that a problem, I would very much like to avoid the need to >>> have separate >>> u-boot and kernel devicetree flavors for sunxi. >> >> OK fair enough, but actually I feel that this should be unrelated to >> my change. The CONFIG_DEFAULT_DEVICE_TREE setting is used by all >> boards that use device tree. It seems that CONFIG_FDTFILE is used in a >> few places but is not very widespread. >> >> So I suppose what I am saying is, how about you do the patch you are >> talking about? I need to be careful to focus on drivel model changes >> and not boil the ocearn, or I will never get done! > > Ok, I lets leave this bit of the patch as is then, and I'll put fixing this > on my todo list. > > So...
Thank you, you are very kind :-) I still have a sunxi board and expect to be doing more here in future, but right now am a bit tight for time. > >> >>> >>> >>>> >>>>> >>>>> >>>>>> +CONFIG_OF_CONTROL=y >>>>>> +CONFIG_OF_SEPARATE=y >>>>>> ++S:CONFIG_ARM=y >>>>>> ++S:CONFIG_TARGET_SUN7I=y >>>>>> diff --git a/include/configs/sun7i.h b/include/configs/sun7i.h >>>>>> index a902b84..500d0e3 100644 >>>>>> --- a/include/configs/sun7i.h >>>>>> +++ b/include/configs/sun7i.h >>>>>> @@ -36,6 +36,10 @@ >>>>>> #define CONFIG_ARMV7_SECURE_BASE SUNXI_SRAM_B_BASE >>>>>> #define CONFIG_SYS_CLK_FREQ 24000000 >>>>>> >>>>>> +#if !defined(CONFIG_SPL_BUILD) && defined(CONFIG_DM) >>>>>> +# define CONFIG_CMD_DM >>>>>> +#endif >>>>>> + >>>>> >>>>> Can you please put this in include/configs/sunxi-common.h instead ? >>>> >>>> OK will do. > > If you can do a respin based on tip of the latest master and with this one > small item fixed, then I'll queue up your patches in u-boot-sunxi/next I will respin with the changes and retest. I was planning to bring it through -dm with suitable acks, but it is all sunxi stuff now and there are no DM-dependent patches outstanding, so going through sunxi sounds good. But would like to make sure it merges to mainline next week. Regards, Simon _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
