[U-Boot] [PATCH] fdt: Fix handling of paths with options in them

[Hans de Goede]

Hi,

On 24-04-15 14:42, Simon Glass wrote:
> Hi Hans,
>
> On 23 April 2015 at 10:15, Simon Glass <sjg at chromium.org> wrote:
>> Hi Hans,
>>
>> On 23 April 2015 at 00:55, Hans de Goede <hdegoede at redhat.com> wrote:
>>> Hi,
>>>
>>>
>>> On 22-04-15 19:20, Simon Glass wrote:
>>>>
>>>> Hi Hans,
>>>>
>>>> On 20 April 2015 at 12:10, Hans de Goede <hdegoede at redhat.com> wrote:
>>>>>
>>>>> Hi,
>>>>>
>>>>> On 20-04-15 17:39, Simon Glass wrote:
>>>>>>
>>>>>>
>>>>>> Hi Hans,
>>>>>>
>>>>>> On 20 April 2015 at 03:13, Hans de Goede <hdegoede at redhat.com> wrote:
>>>>>>>
>>>>>>>
>>>>>>> After syncing the sunxi dts files with the upstream kernel dm/fdt sunxi
>>>>>>> builds would no longer boot.
>>>>>>>
>>>>>>> The problem is that stdout-path is now set like this in the upstream
>>>>>>> dts
>>>>>>> files: stdout-path = "serial0:115200n8". The use of options in
>>>>>>> of-paths,
>>>>>>> either after an alias name, or after a full path, e.g. stdout-path =
>>>>>>> "/soc at 01c00000/serial at 01c28000:115200", is standard of usage, but
>>>>>>> something
>>>>>>> which the u-boot dts code so far did not handle.
>>>>>>>
>>>>>>> This commit fixes this, adding support for both path formats.
>>>>>>>
>>>>>>> Signed-off-by: Hans de Goede <hdegoede at redhat.com>
>>>>>>> ---
>>>>>>>     arch/arm/dts/sun7i-a20-pcduino3.dts |  2 +-
>>>>>>>     lib/libfdt/fdt_ro.c                 | 25 ++++++++++++++++++++++---
>>>>>>
>>>>>>
>>>>>>
>>>>>> I haven't looked. but is this change in dtc upstream or just in the
>>>>>> kernel?
>>>>>
>>>>>
>>>>>
>>>>> This is just a change in the dts files shipped with the kernel not in
>>>>> dtc,
>>>>> the dts files for sunxi used to not set stdout-path, and you patched in
>>>>> a stdout-path setting for u-boot:
>>>>
>>>>
>>>> In that case, can we change this in the fdt support /fdtdec code,
>>>> instead of making a change to libfdt that will never go upstream?
>>>>
>>>> If that doesn't work or is too painful, then we should take this patch.
>>>
>>>
>>> Actually I started with fixing this the fdtdev level, but then I noticed
>>> that the kernel does this at the of_find_node_by_path level, so if we
>>> fix this for stdout-path only (which a fdtdec patch would do) then we may
>>> get bit by this again later.
>>>
>>> Also fixing it at the fdtdec level means adding a strdup + error checking
>>> since then we need to pass a truncated (options removed) copy of the path
>>> to fdt_path_offset(), while with the current patch we can keep using
>>> read only access to the fdt.
>>>
>>> So I've a slight preference for going this way. Why would libfdt upstream
>>> not
>>> take this patch ?
>>
>> I'm not sure, but please go ahead and send it upstream. Thanks for the
>> explanation, I'll pick this up.
>>
>> Acked-by: Simon Glass <sjg at chromium.org>
>>
>
> Can I just check if you are OK to send this upstream? We don't need to
> wait for it to be accepted, just want to know it will be sent.

Yes I will submit this upstream.

Regards,

Hans