"No valid device tree blob found" error

[Luiz Otavio O Souza]

On 17 February 2014 16:11, Ian Lepore wrote:
> On Mon, 2014-02-17 at 16:03 -0300, Luiz Otavio O Souza wrote:
>> On 16 February 2014 18:31, Glen Barber wrote:
>> > On Sun, Feb 16, 2014 at 04:30:01PM -0500, Glen Barber wrote:
>> >> Images for RPI-B and BEAGLEBONE (and I suspect PANDABOARD) are failing
>> >> to boot this week.
>> >>
>> >> The images are built against r261948.  Console messages during boot:
>> >>
>> >>   ## Starting application at 0x88000054 ...
>> >>   Consoles: U-Boot console
>> >>   Compatible API signature found @9f242240
>> >>   MMC Device 2 not found
>> >>   MMC Device 3 not found
>> >>   Number of U-Boot devices: 2
>> >>
>> >>   FreeBSD/armv6 U-Boot loader, Revision 1.2
>> >>   (root at grind.freebsd.org, Sun Feb 16 18:10:43 UTC 2014)
>> >>   DRAM:    512MB
>> >>
>> >>   Device: disk
>> >>   Loading /boot/defaults/loader.conf
>> >>   /boot/kernel/kernel data=0x460bc8+0x2c7438
>> >>   syms=[0x4+0x85a60+0x4+0x50c89]
>> >>
>> >>   Hit [Enter] to boot immediately, or any other key for command prompt.
>> >>   Booting [/boot/kernel/kernel]...
>> >>   Using DTB provided by U-Boot.
>> >>   No valid device tree blob found!WARNING! Trying to fire up the kernel,
>> >>   but no device tree blob found!
>> >>
>> >> Any ideas if this is error on my part, or a problem in head/ ?  The
>> >> stable/10/ images boot fine, so I do not suspect any code changes in the
>> >> build process.
>> >>
>> >
>> > Correction:  RPI-B fails to boot.  BEAGLEBONE boots after pressing 'q'
>> > when this message is displayed.
>> >
>> > Glen
>> >
>>
>>
>> Yeah, i had noted this difference already (and forgot to ask about it...).
>>
>> It works on BEAGLEBONE because the BEAGLEBONE kernel still has the
>> FDT_DTB_STATIC option.
>>
>> I have booted mine without the static dtb blob included in kernel
>> without any issue (using crochet images - other images which doesn't
>> use ubldr may be broken by this change).
>>
>> If you guys think it is appropriate i can ask to commit the attached patch.
>>
>> Luiz
>
> I think it's a good idea to leave the static dtb compiled in on
> platforms where it'll work.  The code in initarm() tries to use the dtb
> passed in by ubldr or by u-boot using the linux boot abi, and only falls
> back to the static one if those aren't available.
>
> -- Ian

Yeah. I agree.

Luiz