mailarchive of the ptxdist mailing list
 help / color / mirror / Atom feed
From: Michael Olbrich <m.olbrich@pengutronix.de>
To: Bruno Thomsen <bruno.thomsen@gmail.com>
Cc: ptxdist@pengutronix.de
Subject: Re: [ptxdist] [RFC PATCH 2/2] kernel: validate device trees against the DT schema
Date: Fri, 14 Aug 2026 10:06:24 +0200	[thread overview]
Message-ID: <an7MgCfriVe89gvD@pengutronix.de> (raw)
In-Reply-To: <20260730130740.10851-3-bruno.thomsen@gmail.com>

On Thu, Jul 30, 2026 at 03:07:40PM +0200, Bruno Thomsen wrote:
> Make it possible to do device tree validation on
> dts/dtsi during development and preparation for
> upstream.
> 
> Signed-off-by: Bruno Thomsen <bruno.thomsen@gmail.com>
> ---
>  platforms/kernel.in | 25 +++++++++++++++++++++++++
>  rules/kernel.make   | 36 ++++++++++++++++++++++++++++++++++++
>  2 files changed, 61 insertions(+)
> 
> diff --git a/platforms/kernel.in b/platforms/kernel.in
> index 2c6d906e8..79a414fd8 100644
> --- a/platforms/kernel.in
> +++ b/platforms/kernel.in
> @@ -11,6 +11,8 @@ menuconfig KERNEL
>  	select HOST_LZ4			if KERNEL_LZ4
>  	select HOST_LIBKMOD		if KERNEL_MODULES
>  	select HOST_SYSTEM_BC
> +	select HOST_SYSTEM_PYTHON3	if KERNEL_DTB_CHECK
> +	select HOST_SYSTEM_PYTHON3_DTSCHEMA	if KERNEL_DTB_CHECK
>  	select HOST_OPENSSL		if KERNEL_OPENSSL
>  	select HOST_LIBELF		if KERNEL_LIBELF
>  	select HOST_BINDGEN_CLI		if KERNEL_RUST
> @@ -232,6 +234,29 @@ config KERNEL_DTSO
>  
>  endif
>  
> +config KERNEL_DTB_CHECK
> +	bool
> +	depends on KERNEL_DTB
> +	prompt "validate device trees against the DT schema"
> +	help
> +	  Additionally build the device trees with the kernel build system
> +	  ('make CHECK_DTBS=y <board>.dtb') to validate them against the
> +	  device tree bindings in Documentation/devicetree/bindings.
> +	  Problems are reported as warnings, they do not break the build.
> +
> +	  This requires the 'dtschema' python package on the build host and
> +	  it noticeably increases the build time of the kernel, because the
> +	  schema of all bindings must be processed first.
> +
> +	  Only device trees that are part of the kernel tree can be checked.
> +	  Each file in KERNEL_DTS is looked up in KERNEL_DTS_PATH and the
> +	  first match below a 'boot/dts' directory is checked, even if the
> +	  device tree itself is built from an earlier entry in the search
> +	  path. Device trees without such a match are skipped with a
> +	  warning. So for device trees in a vendor directory, that directory
> +	  must be part of KERNEL_DTS_PATH, e.g.
> +	  "${KERNEL_DIR}/arch/${GENERIC_KERNEL_ARCH}/boot/dts/ti/omap".
> +
>  config KERNEL_CODE_SIGNING
>  	prompt "use code signing infrastructure"
>  	select KERNEL_OPENSSL
> diff --git a/rules/kernel.make b/rules/kernel.make
> index af0dd3913..eaed71023 100644
> --- a/rules/kernel.make
> +++ b/rules/kernel.make
> @@ -215,6 +215,36 @@ KERNEL_MAKE_OPT		= \
>  	$(KERNEL_IMAGE) \
>  	$(call ptx/ifdef, PTXCONF_KERNEL_MODULES,modules)
>  
> +#
> +# device tree validation
> +#
> +# The kernel build system can only validate device trees from the kernel tree.
> +# Look up the device trees in KERNEL_DTS_PATH and use the first match below a
> +# 'boot/dts' directory.
> +#
> +kernel/dtb-check-file	= \
> +	$(firstword $(foreach dts, \
> +		$(call ptx/in-path-all,KERNEL_DTS_PATH,$(1)), \
> +		$(if $(findstring /boot/dts/,$(dts)),$(dts))))
> +
> +# the dtb make targets are relative to arch/<arch>/boot/dts
> +kernel/dtb-check-target	= \
> +	$(patsubst %.dts,%.dtb,$(patsubst /%,%,$(word 2, \
> +		$(subst /boot/dts,$(ptx/def/space), \
> +			$(call kernel/dtb-check-file,$(1))))))
> +
> +KERNEL_DTB_CHECK_FILES	= $(strip \
> +	$(foreach dts,$(KERNEL_DTS),$(call kernel/dtb-check-target,$(dts))))
> +
> +KERNEL_DTB_CHECK_SKIPPED = $(strip \
> +	$(foreach dts,$(KERNEL_DTS), \
> +		$(if $(call kernel/dtb-check-file,$(dts)),,$(dts))))
> +

I think this is too much complexity for the kernel.make. Instead add this
to ptxd_make_world_dtb(). That really only works with the Linux kernel
anyways.

> +KERNEL_DTB_CHECK_OPT	= \
> +	$(KERNEL_SHARED_OPT) \
> +	CHECK_DTBS=y \
> +	$(KERNEL_DTB_CHECK_FILES)

Make this conditional based on the kconfig option.

Add
	pkg_dtb_check_opt="$($(1)_DTB_CHECK_OPT)"

to world/dtb/env

and use that in ptxd_make_world_dtb() to run the check if set. It will be
much cleaner to find the in-tree files there as well.

Michael

> +
>  KERNEL_TOOL_PERF_OPTS	:= \
>  	-C $(KERNEL_DIR)/tools/perf \
>  	O=$(KERNEL_BUILD_DIR)/tools/perf \
> @@ -263,6 +293,12 @@ $(STATEDIR)/kernel.compile:
>  		$(KERNEL_BUILD_DIR)/usr/initramfs_data.cpio.* \
>  		$(KERNEL_BUILD_DIR)/usr/.initramfs_data.cpio.*
>  	@$(call world/compile, KERNEL)
> +ifdef PTXCONF_KERNEL_DTB_CHECK
> +	@$(foreach dts, $(KERNEL_DTB_CHECK_SKIPPED), \
> +		echo "dtb-check: skipping '$(dts)': no match below 'boot/dts' in KERNEL_DTS_PATH";)
> +	@$(if $(KERNEL_DTB_CHECK_FILES), \
> +		$(call world/execute, KERNEL, $(MAKE) $(KERNEL_DTB_CHECK_OPT)))
> +endif
>  ifdef PTXCONF_KERNEL_TOOL_PERF
>  	@mkdir -p $(KERNEL_BUILD_DIR)/tools/perf
>  	@$(call compile, KERNEL, $(KERNEL_TOOL_PERF_OPTS))
> -- 
> 2.55.0
> 
> 
> 

-- 
Pengutronix e.K.                           |                             |
Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |



      reply	other threads:[~2026-08-14  8:06 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-07-30 13:07 [ptxdist] [RFC PATCH 0/2] Run DT schema check on selected device trees Bruno Thomsen
2026-07-30 13:07 ` [ptxdist] [RFC PATCH 1/2] host-system-python: add dtschema Bruno Thomsen
2026-07-30 13:07 ` [ptxdist] [RFC PATCH 2/2] kernel: validate device trees against the DT schema Bruno Thomsen
2026-08-14  8:06   ` Michael Olbrich [this message]

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=an7MgCfriVe89gvD@pengutronix.de \
    --to=m.olbrich@pengutronix.de \
    --cc=bruno.thomsen@gmail.com \
    --cc=ptxdist@pengutronix.de \
    /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