On 7/3/26 14:40, Patrice CHOTARD wrote: > > > On 7/2/26 15:00, Dario Binacchi wrote: >> Hi Patrice >> >> On Mon, Jun 29, 2026 at 9:29 AM Patrice CHOTARD >> <[email protected]> wrote: >>> >>> >>> >>> On 5/29/26 17:31, Dario Binacchi wrote: >>>> The STM32MP157F-DK2 board is a direct derivative of the STM32MP157C-DK2. >>>> As explained by ST [1], the "STM32MP157F-DK2 is assembled with latest >>>> STM32MP15 RevZ cut which is the only one available for new product now. >>>> The major differences in this 'F' variant SoC are the increased max >>>> frequency of the Cortex-A7 cores and the correction of some HW bugs >>>> (Device Rev.Z, see errata ES0438)". >>>> >>>> On the software/architecture side, this variant utilizes SCMI (System >>>> Control and Management Interface) for secure world resource management. >>>> As a consequence, the underlying stm32mp15-scmi.dtsi explicitly deletes >>>> several fixed clock nodes (such as clk_hse, clk_hsi, clk_lse, clk_lsi, >>>> and clk_csi), making it impossible to include stm32mp157c-dk2-u-boot.dtsi >>>> without triggering compilation errors due to these missing clock labels. >>>> >>>> So let's introduce a dedicated stm32mp157f-dk2-u-boot.dtsi that directly >>>> defines the fwu-mdata node. This provides the necessary FWU metadata >>>> configuration required to allow automatic A/B bank selection, supported >>>> for STM32MP15 boards by commit 560d8f32703f ("board: st: stm32mp15: >>>> support dynamic A/B bank bootup"). >>>> >>>> [1] >>>> https://community.st.com/t5/stm32-mpus-products-and-hardware/i-m-wondering-what-s-the-difference-point-between-stm32mp157c/td-p/213460 >>>> Signed-off-by: Dario Binacchi <[email protected]> >>>> >>>> --- >>>> >>>> Changes in v2: >>>> - Drop the inclusion of stm32mp157c-dk2-u-boot.dtsi to avoid compilation >>>> errors caused by deleted clock nodes in stm32mp15-scmi.dtsi. >>>> - Directly define the fwu-mdata node in stm32mp157f-dk2-u-boot.dtsi to >>>> properly support automatic A/B bank selection. >>>> >>>> arch/arm/dts/stm32mp157f-dk2-u-boot.dtsi | 11 +++++++++++ >>>> 1 file changed, 11 insertions(+) >>>> create mode 100644 arch/arm/dts/stm32mp157f-dk2-u-boot.dtsi >>>> >>>> diff --git a/arch/arm/dts/stm32mp157f-dk2-u-boot.dtsi >>>> b/arch/arm/dts/stm32mp157f-dk2-u-boot.dtsi >>>> new file mode 100644 >>>> index 000000000000..77805417bfa9 >>>> --- /dev/null >>>> +++ b/arch/arm/dts/stm32mp157f-dk2-u-boot.dtsi >>>> @@ -0,0 +1,11 @@ >>>> +// SPDX-License-Identifier: GPL-2.0+ >>>> +/* >>>> + * Copyright (C) 2026 Amarula Solutions, Dario Binacchi >>>> <[email protected]> >>>> + */ >>>> + >>>> +/ { >>>> + fwu-mdata { >>>> + compatible = "u-boot,fwu-mdata-gpt"; >>>> + fwu-mdata-store = <&sdmmc1>; >>>> + }; >>>> +}; >>> >>> Hi Dario >>> >>> What about including stm32mp157c-dk2-u-boot.dtsi instead of defining >>> fwu-mdata node ? >> >> This was actually the approach used in v1. However, as you pointed out in the >> v1 thread [1], it does not build because stm32mp15-scmi.dtsi removes the >> clk_hse, clk_hsi, clk_lse, clk_lsi and clk_csi labels referenced by >> stm32mp157c-dk2-u-boot.dtsi. >> >> That's why I dropped the include in v2 and defined only the fwu-mdata node >> locally. >> >> Or am I perhaps missing something? > > Hi Dario > > Sorry, you are right, it's just following an internal discussion with Patrick > and i totally forgot your > v1 implementation already included stm32mp157c-dk2-u-boot.dtsi and has > compilatioj issue. > > Patrice > > >> >> [1] >> https://patchwork.ozlabs.org/project/uboot/patch/[email protected]/ >> >> Thanks and regards, >> Dario >> >>> >>> Thanks >>> Patrice >> >> >> > > _______________________________________________ > Uboot-stm32 mailing list > [email protected] > https://st-md-mailman.stormreply.com/mailman/listinfo/uboot-stm32 Applied to u-boot-stm32/next Thanks Patrice
