Re: [Ltib] Device tree location patch

[Mike Goins]

On Mon, Sep 2, 2013 at 2:41 PM, Stuart Hughes <address@hidden> wrote:
> Hi Mike,
>
> The patch looks good, please go ahead and check-in if you wish.
>
> Thanks, Stuart

I'm actually having a couple issues with it [the patch].

Since the dtc compile is now inside of a doubly nested loop, I an not
sure how to cleanly break completely out once the dtc is discovered
without a state flag.  The current result is if the same name is exits
in the platform dts dir and in the source tree, the source tree wins.
This can be avoided by using a unique dtc name and/or by re-ordering
the nested search.  May have unexpected results and not sure what the
convention should be.

The other, which I just ran into, is that editing the dtc is not
picked up by "checking to see of sources have been updated" in a
typical directory build.  This is obviously outside the
kernel-common.tmpl and looks to require modifications to core ltib to
handle just as if the config was updated.   Hadn't dug into a solution
as I am buried in getting usb otg working on a particular platform.


> On 22/08/13 21:51, Mike Goins wrote:
>> Attached is a patch against kernel-common.tmpl that first checks for
>> the dts in the platform folder and then in powerpc/boot/dts.  I figure
>> since the platform config, kernel config, and busybox config reside
>> there, so could the dts.
>>
>> It seems that kernel work (particularly going through various kernel
>> versions) is much easier if I keep as much as I can out of the kernel
>> tree.
>>
>> Feedback welcomed.
>>
>>
>>
>> _______________________________________________
>> LTIB home page: http://ltib.org
>>
>> Ltib mailing list
>> address@hidden
>> https://lists.nongnu.org/mailman/listinfo/ltib
>>