[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <0c4dc285-f218-4c88-87f3-b7c7c786cdba@monstr.eu>
Date: Fri, 8 Nov 2024 12:53:53 +0100
From: Michal Simek <monstr@...str.eu>
To: Masahiro Yamada <masahiroy@...nel.org>, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] microblaze: use the common infrastructure to support
built-in DTB
On 9/18/24 06:52, Masahiro Yamada wrote:
> MicroBlaze is the only architecture that supports a built-in DTB in
> its own way.
>
> Other architectures (e.g., ARC, NIOS2, RISC-V, etc.) use the common
> infrastructure introduced by commit aab94339cd85 ("of: Add support for
> linking device tree blobs into vmlinux").
>
> This commit migrates MicroBlaze to this common infrastructure.
>
> Signed-off-by: Masahiro Yamada <masahiroy@...nel.org>
> ---
>
> I do not know why MicroBlaze still adopts its own way.
> Perhaps, because MicroBlaze supports the built-in DTB
> before aab94339cd85 and nobody attempted migration.
> Anyway, I only compile-tested this patch.
> I hope the maintainer can do boot-testing.
I took a look at it and it is changing current behavior.
If you look at linux.bin and there is no DT inside. But when you patch is
applied linux.bin contains system.dtb inside which is regression.
Or is it intention of this patch?
I think there was any documentation about it's usage in past but never really
described in upstream kernel.
But idea was that linux.bin requires DT to be passed from bootloader via R7 reg
but simpleImage.X is linux.bin+DTB inside and can be used without bootloader.
Thanks,
Michal
Powered by blists - more mailing lists