[RFC PATCH 0/1] riscv: dts: Allow BUILTIN_DTB for all socs
Conor Dooley
conor at kernel.org
Fri Feb 23 03:39:10 PST 2024
On Fri, Feb 23, 2024 at 03:42:01PM +0800, Yangyu Chen wrote:
> To sum up, I agree with the reasons to refuse it for debugging purposes. I
> am wondering what to do next. After reviewing the code carefully, I found
> this feature not only for K210 but also for other SoCs. But for other SoCs,
> it is broken as it will link multiple dtbs to the kernel, but the kernel
> will always pick up the first dtb to use as discussed on this thread [2].
> That is because SOC_BUILTIN_DTB_DECLARE is removed from this patch [3] then
> no code to select for multiple dtbs now. And makefiles on other soc which is
> on the mainline kernel currently will build every dtb object file from this
> patch [4]. So this feature for other SoCs is broken now.
>
> Choices might be one of the following:
>
> 1. Remove BUILTIN_DTB feature if K210 support get removed
> 2. Continue to add this feature to get this work for other socs
>
> I prefer to continue to get this feature to work. Not only for my debugging
> purposes but also for fixes.
My email might have been a bit unclear because I responded to things out
of order. My final thoughts were that we should apply a v2 (or 3 etc) of
this patch if for no other reason than fixing BUILTIN_DTB. I could not
really justify refusing your changes when a patch fixing BUILTIN_DTB for
the non-k210 platforms that support it would look almost identical to
this patch. Apologies if you misunderstood.
I left some comments on the patch itself with a v2 in mind.
Thanks,
Conor.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 228 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-riscv/attachments/20240223/523e3e8a/attachment.sig>
More information about the linux-riscv
mailing list