Buildroot Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Niklas Cassel <Niklas.Cassel@wdc.com>
To: Fiona Klute <fiona.klute@gmx.de>
Cc: "Romain Naour" <romain.naour@smile.fr>,
	"buildroot@buildroot.org" <buildroot@buildroot.org>,
	"Gaël PORTAY" <gael.portay+rtone@gmail.com>,
	"Julien Grossholtz" <julien.grossholtz@openest.io>,
	"Martin Bark" <martin@barkynet.com>,
	"Yann E . MORIN" <yann.morin.1998@free.fr>,
	"José Mendes" <jp.baltazar.mendes@gmail.com>
Subject: Re: [Buildroot] [PATCH v3 2/3] board/raspberrypi: allow non-matching globs for DTBs
Date: Mon, 17 Feb 2025 13:36:25 +0000	[thread overview]
Message-ID: <Z7M7WcyqKWCycgI4@ryzen> (raw)
In-Reply-To: <10299ed2-643e-4ced-9ddd-579c8ef33bfb@gmx.de>

On Mon, Feb 17, 2025 at 01:33:52PM +0100, Fiona Klute wrote:
> Hi Niklas!
> 
> Am 10.02.25 um 13:47 schrieb Niklas Cassel:
> > On Fri, Feb 07, 2025 at 06:14:57PM +0100, Fiona Klute via buildroot wrote:
> > > From: "Fiona Klute (WIWA)" <fiona.klute@gmx.de>
> > > 
> > > When installing DTBs from the rpi-firmware package
> > > (BR2_PACKAGE_RPI_FIRMWARE_INSTALL_DTBS=y), there's no DTB directly in
> > > the images directory because the kernel build doesn't provide one.
> > 
> > Your new defconfig does not seem to have BR2_PACKAGE_RPI_FIRMWARE_INSTALL_DTBS
> > set. What am I missing?
> 
> This change is independent of the defconfig, which is aimed at a board
> that's well supported by in-tree DTS (both mainline and RPi downstream).
> The only connection is that both are about support for RPi boards.
> 
> > Regardless, perhaps clarify/rephrase your commit message.
> 
> What needs clarifying in your opinion?

This series adds a new defconfig:
configs/raspberrypicm4io_64_uboot_defconfig

Since these 3 patches are in a series, it is logical that patches 1-2 are
actually needed for patch 3.

The commit message in patch 2 references
BR2_PACKAGE_RPI_FIRMWARE_INSTALL_DTBS=y
However, the defconfig in patch 3 does not select this config.

So, if patch 2 is completely unrelated to patch 3, why submit it in the
same series?


After looking at package/rpi-firmware/Config.in it appears that
BR2_PACKAGE_RPI_FIRMWARE_INSTALL_DTBS is "default y"
(and you new defconfig does select BR2_PACKAGE_RPI_FIRMWARE).

So I guess the answer is: because package/rpi-firmware/Config.in is silly.

A config option to install something is usually never "default y".
To me it is very counter-intuitive that defconfigs that want firmware
installed, but do not want DTBs installed would have to do:
BR2_PACKAGE_RPI_FIRMWARE=y
# BR2_PACKAGE_RPI_FIRMWARE_INSTALL_DTBS is not set

Rather than configs that want both installed would do:
BR2_PACKAGE_RPI_FIRMWARE=y
BR2_PACKAGE_RPI_FIRMWARE_INSTALL_DTBS=y

But this is unrelated to you series, so feel free to ignore my comment.


> 
> > > Using pre-compiled DTBs is discouraged, but useful for using mainline
> > > kernels until suitable DTS can be merged, or to provide U-Boot with a
> > > DTB (via the firmware bootloader) independent of whatever kernel DTB
> > > is in use.
> > 
> > I don't really understand which pre-compiled DTB you are using, since
> > your series is not adding any DTB, and your defconfig does not have
> > BR2_PACKAGE_RPI_FIRMWARE_INSTALL_DTBS set.
> > 
> > However, wouldn't it be much better to rebase this series on either:
> > 
> > 1) Romain's series:
> > https://lore.kernel.org/buildroot/20250204163622.192932-3-romain.naour@smile.fr/
> > which introduces BR2_LINUX_KERNEL_CUSTOM_DTS_PATH
> > 
> > or
> > 
> > 2) Your own series:
> > https://lore.kernel.org/buildroot/20250207183455.1468220-1-fiona.klute@gmx.de/
> > 
> > 
> > Then this series could provide a DT source file (.dts) instead of a binary
> > blob (.dtb). Having a DTS file would make it easier for other buildroot
> > contributors to make changes, and once upstream has a DTS merged, you can
> > remove your DTS file provided by buildroot.
> 
> This patch has nothing to do with either series. Both of those series
> aim to support building out-of-tree DTS and installing the resulting DTB.
> 
> This patch allows the post-image script to work if the user does not
> have the kernel (or tools added by one of those series) build a DTB, and
> instead sets BR2_PACKAGE_RPI_FIRMWARE_INSTALL_DTBS=y. If they have a
> suitable DTS to build, there's no point in using that option. The
> rpi-firmware package installs its pre-compiled DTBs into
> "${BINARIES_DIR}"/rpi-firmware/, so the "${BINARIES_DIR}"/*.dtb glob
> won't match anything, and including the literal glob in the FILES list
> will lead to "file not found" errors.
> 
> In short: The purpose of this patch is only to make
> BR2_PACKAGE_RPI_FIRMWARE_INSTALL_DTBS=y work without forcing the user to
> adjust the post-image script.

Ok, thank you for clarifying.


Kind regards,
Niklas
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

  reply	other threads:[~2025-02-17 13:36 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-02-07 17:14 [Buildroot] [PATCH v3 0/3] U-Boot on RPi boards Fiona Klute via buildroot
2025-02-07 17:14 ` [Buildroot] [PATCH v3 1/3] board/raspberrypi: Support boot.scr for U-Boot Fiona Klute via buildroot
2025-02-07 17:14 ` [Buildroot] [PATCH v3 2/3] board/raspberrypi: allow non-matching globs for DTBs Fiona Klute via buildroot
2025-02-10 12:47   ` Niklas Cassel via buildroot
2025-02-17 12:33     ` Fiona Klute via buildroot
2025-02-17 13:36       ` Niklas Cassel [this message]
2025-02-07 17:14 ` [Buildroot] [PATCH v3 3/3] configs/raspberrypicm4io_64_uboot_defconfig: new defconfig Fiona Klute via buildroot
2025-02-07 21:36   ` Arnout Vandecappelle via buildroot
2025-02-07 22:36     ` bryce.schober
2025-02-10  7:44       ` Arnout Vandecappelle via buildroot
2025-02-13 19:25         ` Fiona Klute via buildroot
2025-02-14  8:12           ` Arnout Vandecappelle via buildroot

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=Z7M7WcyqKWCycgI4@ryzen \
    --to=niklas.cassel@wdc.com \
    --cc=buildroot@buildroot.org \
    --cc=fiona.klute@gmx.de \
    --cc=gael.portay+rtone@gmail.com \
    --cc=jp.baltazar.mendes@gmail.com \
    --cc=julien.grossholtz@openest.io \
    --cc=martin@barkynet.com \
    --cc=romain.naour@smile.fr \
    --cc=yann.morin.1998@free.fr \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox