On 10/9/26 09:23, Simon Glass wrote:
On 2026-09-23T11:49:31, Ludwig Nussel <[email protected]> wrote:
- Note: commit 68e65953400 introduced forcing OF_OMIT_DTB which is
incompatible with the intent of the qemu DT merging feature, ie merge
built-in signature nodes into qemu's DT. To be sorted out at some
point.
One thing I'd like to check before this series goes in.. with this
series applied, qemu_arm64_defconfig still ends up with
CONFIG_OF_OMIT_DTB=y, since ARCH_QEMU implies OF_HAS_PRIOR_STAGE and
BINMAN is not enabled. OF_OMIT_DTB no longer has a prompt, so the 'is
not set' line added to the defconfig in patch 1 has no effect and
savedefconfig drops it. As a result qemu_arm64 never has a built-in
DT, the merge code in patch 1 always takes the early exit, and the
problem in the cover letter (QEMU not seeing the public keys) remains
by default. The patch 1 commit message also says the defconfig change
enables the feature, which is no longer true.
Please can you either fix the Kconfig (e.g. stop defaulting
OF_OMIT_DTB to y for QEMU when FIT_SIGNATURE is enabled, or let
ARCH_QEMU opt out some other way), or drop the defconfig change and
say clearly that the merge only happens on builds that include a DT? A
test for the merge itself would be good, perhaps as a follow-up, since
none of the vboot tests run on QEMU. This may also be worth checking
against the CI failure Tom reported.
When I developed the patch OF_OMIT_DTB was "default". 68e65953400
changed it to "def_bool". Would something like this make sense as a fix?
--- a/dts/Kconfig
+++ b/dts/Kconfig
@@ -213,7 +213,8 @@ config OF_HAS_PRIOR_STAGE
disabled in the board's defconfig file.
config OF_OMIT_DTB
- def_bool y if OF_HAS_PRIOR_STAGE && !BINMAN
+ def_bool y if OF_HAS_PRIOR_STAGE && !BINMAN && !ARCH_QEMU
+ def_bool n if ARCH_QEMU
help
As a special case, avoid writing a device tree file
u-boot.dtb when
building. Also don't include that file in u-boot.bin
cu
Ludwig
--
Ludwig Nussel
Siemens AG
www.siemens.com
