So I think I figured out the problem but I am not sure what the correct fix is.
It seems to be a change in the toolchain. It comes down to the define for CONFIG_SYS_BOARD. For some reason strcat does not like it.
If I undefine CONFIG_SYS_BOARD and then define it as "panda\0" just before the board_name env variable is set then it works fine.
I had been compiling with linaro 4.9-2015.05. I compiled it with an older toolchain I had laying around. I am not even sure where I got it but version prints out arm-linux-gnu-gcc (GCC) 4.9.2 20150212 (Red Hat Cross 4.9.2-5). Compiled with that toolchain it works fine without modification.
Right now I am leaning toward the fix being a patch that has the config generate a define for CONFIG_SYS_BOARD as "panda\0" but I am not sure if that is the best way to handle it.
By the way, on the non working u-boot, if you drop to a u-boot prompt and setenv board_name to panda-es it works.
-------- Original Message --------
Subject: Re: Source, patches, config for u-boot
From: Peter Robinson <pbrobinson@xxxxxxxxx>
Date: Fri, August 21, 2015 5:29 am
To: dbatzle@xxxxxxxxxxxx
Cc: arm@xxxxxxxxxxxxxxxxxxxxxxx
On Fri, Aug 21, 2015 at 1:21 PM, <dbatzle@xxxxxxxxxxxx> wrote:
> Thank you for the head start.
NP
> Yes, I suspected I would find the problem in panda.c. I was planning to do a
> diff of that file between the versions and see if anything jumped out at me.
I looked at that and didn't see any obvious change, the setting of the
variants is wrapped in ifdef statements so it might be somewhere else.
Would love help fixing it, I've just not had the time needed.
> I should be able to get it sorted this weekend. I'll let you know what I
> find.
Excellent, also happy to help herd fixes upstream too
> Thanks again.
>
>
> -------- Original Message --------
> Subject: Re: Source, patches, config for u-boot
> From: Peter Robinson <pbrobinson@xxxxxxxxx>
> Date: Fri, August 21, 2015 1:58 am
> To: dbatzle@xxxxxxxxxxxx
> Cc: arm@xxxxxxxxxxxxxxxxxxxxxxx
>
> On Fri, Aug 21, 2015 at 3:26 AM, <dbatzle@xxxxxxxxxxxx> wrote:
>> I was wondering if I could get the source and configuration files that are
>> used to build u-boot for Fedora on ARM.
>
> It's all in the Fedora package repository for uboot-tools.
>
> "fedpkg clone uboot-tools; cd uboot-tools; fedpkg prep" will get you a
> fully patched tree.
>
>> I have a Pandaboard ES. Fedora 21 runs fine on it. Fedora 22 will not boot
>> out of the box. U-boot loads omap4-panda.dtb instead of
>> omap4-panda-es.dtb.
>> If I rename the es dtb file the kernel still hangs, it just runs longer
>> before it hangs.
>
> Yes, it's a problem even with vanilla upstream. I see it on my
> Panda-ES too but just ran out of time to get to the bottom of the
> problem with it. I did report it upstream [1] but never got a response
> but I suspect there's some issue in the board/ti/panda/panda.c which
> might not be setting up the board specifics
>
> You can interrupt the boot and run the command "setenv fdtfile
> omap4-panda-es.dtb" and then just run "boot" and it'll have the
> correct .dtb set and it might get you further and might help you
> debug.
>
> [1] http://lists.denx.de/pipermail/u-boot/2015-July/219376.html
>
>> I have successfully booted 22 by using u-boot 2014.10 from Fedora 21.
>>
>> I have u-boot source from git.denx.de and have applied patches from rcn-ee
>> but the resulting u-boot does not work on the Fedora /boot image.
>>
>> I suppose I could wait for Fedora 23 but I would like to dig into this a
>> more.
>
> You could test the 2015.04 rc1 in F-24 on F-22 and see if it's fixed
> it, I've been traveling and not had time to look at it closer.
>
> Peter
>
>
> _______________________________________________
> arm mailing list
> arm@xxxxxxxxxxxxxxxxxxxxxxx
> https://admin.fedoraproject.org/mailman/listinfo/arm
_______________________________________________ arm mailing list arm@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/arm
