[GIT PULL 2/5] soc: dt changes for 6.18

Linus Torvalds torvalds at linux-foundation.org
Wed Oct 1 21:43:49 PDT 2025


On Wed, 1 Oct 2025 at 15:27, Arnd Bergmann <arnd at arndb.de> wrote:
>
>   https://git.kernel.org/pub/scm/linux/kernel/git/soc/soc.git tags/soc-dt-6.18

Bah. I've merged this, but this causes build failures on arm64.

I only noticed after already having pushed it out.

> SeonGu Kang (1):
>       arm64: dts: axis: Add ARTPEC-8 Grizzly dts support

Doing a basic allmodconfig build, I get a

  In file included from arch/arm64/boot/dts/exynos/axis/artpec8-grizzly.dts:12
  arch/arm64/boot/dts/exynos/axis/artpec8.dtsi:12:10: fatal error:
dt-bindings/clock/axis,artpec8-clk.h: No such file or directory
     12 | #include <dt-bindings/clock/axis,artpec8-clk.h>
        |          ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

and sure enough, no such file exists. It looks like it was never added
by commit 639f8e36baf1 ("arm64: dts: exynos: axis: Add initial
ARTPEC-8 SoC support")

That commit came in through the Soc tree and in that tree it came in
through the 'samsung-dt64-6.18' branch.

Pls fix.  My arm64 tree builds otherwise, but I have to disable that
whole ARTPEC mess just to even check.

It appears that there's some serious lack of sanity checking in the
arm64 dts land. I _think_ that file ends up coming in through a
different branch, but this damn well should not have been merged AT
ALL in a branch that didn't have all the prerequisites.

Yes, this is annoying.

         Linus



More information about the linux-riscv mailing list