Hi,
On Mon, Oct 19, 2015 at 11:05:58AM +0200, Nicolas Dechesne wrote:
> On Mon, Oct 19, 2015 at 11:03 AM, Stefan Christ <[email protected]> wrote:
> > Hmm, the meaning of the code seems to be non obvious. I just looked at sed
> > replacement and concluded that it's about the file name ending.
> >
> > Using the full path to the dtb file is not supported by the code anyway,
> > e.g.
> >
> > DTB_PATH="${B}/arch/${ARCH}/boot/dts/${DTB}"
> >
> > would be wrong.
> >
> > So the correct solution would check both cases
> >
> > - full path or only filename and
> > - ending is dts or dtb
> >
> > Correct?
>
>
> If you look at the commit that introduced this code, e.g. 72980d5b,
> you can see that this code was added to support the 'legacy' case
> which was the absolute path of the .dts file, or the new case which is
> the .dtb file name. e.g.:
>
> ==
> ,----[ Original definition ]
> | KERNEL_DEVICETREE = "${S}/arch/arm/boot/dts/imx6q-sabresd.dts"
> `----
>
> Becomes:
>
> ,----[ New definition ]
> | KERNEL_DEVICETREE = "imx6q-sabresd.dtb"
> `----
> ==
>
> so the '/dts/' was used to detect the legacy case, and print a warning
> in that condition.
>
Yeah. Your right. The code is correct. The line
DTB=`basename ${DTB} | sed 's,\.dts$,.dtb,g'`
will strip off the directories and replaces the ending. I overlooked that
somehow. So drop my 'fixes', please.
What about my rework patches? I will resend them if they are ok.
Mit freundlichen Grüßen / Kind regards,
Stefan Christ
--
_______________________________________________
Openembedded-core mailing list
[email protected]
http://lists.openembedded.org/mailman/listinfo/openembedded-core
