Re: [PATCH v2 2/5] dt-bindings: arm: stm32: Add clock-controller sub-node to STM32MP2 SYSCFG

From: Gabriel FERNANDEZ

Date: Wed Nov 18 2026 - 07:15:03 EST


Hi Krzysztof,

On 11/13/26 09:24, Krzysztof Kozlowski wrote:
On Tue, Nov 10, 2026 at 02:10:22PM +0100, Gabriel Fernandez wrote:
The SYSCFG of the STM32MP23x and STM32MP25x SoCs is described as the
provider of the LTDC pixel clock, but the inputs of this mux (LTDC
kernel clock, LVDS PHY pixel clock and DSI PHY byte lane clock) can't
be described by the "clocks" property of the SYSCFG node: the SYSCFG is
Not true, you can.

a syscon, whose first clock is taken as the clock of its registers and
enabled around each register access.
That's a driver problem, fix drivers.

Add a clock-controller sub-node to describe the mux with its own clock
inputs. It is required on STM32MP23x and STM32MP25x and not allowed on
the other SoCs.

Deprecate the '#clock-cells' property of the SYSCFG node, which is no
longer required. No driver ever registered a clock on the SYSCFG node,
so this doesn't change anything for existing devicetrees.

Signed-off-by: Gabriel Fernandez <gabriel.fernandez@xxxxxxxxxxx>
---
.../devicetree/bindings/arm/stm32/st,stm32-syscon.yaml | 11 ++++++++++-
1 file changed, 10 insertions(+), 1 deletion(-)

diff --git a/Documentation/devicetree/bindings/arm/stm32/st,stm32-syscon.yaml b/Documentation/devicetree/bindings/arm/stm32/st,stm32-syscon.yaml
index a559f8b03b84..13c02844a91f 100644
--- a/Documentation/devicetree/bindings/arm/stm32/st,stm32-syscon.yaml
+++ b/Documentation/devicetree/bindings/arm/stm32/st,stm32-syscon.yaml
@@ -38,6 +38,10 @@ properties:
"#clock-cells":
const: 0
+ deprecated: true
+ description:
+ The LTDC pixel clock multiplexer is described by the clock-controller
+ sub-node.
"#address-cells":
const: 1
@@ -45,6 +49,10 @@ properties:
"#size-cells":
const: 1
+ clock-controller:
+ type: object
+ $ref: /schemas/clock/st,stm32mp25-syscfg-clk.yaml#
+
patternProperties:
"^phy@2[48]00$":
As you can easily see, each child needs addressing. This is one more
argument why this is a fake device split, done only to match your driver
limitation. Please fix your drivers.
Agreed. Sorry, I sent v3 [1] before answering here.

v3 drops the clock-controller sub-node and its binding: the mux inputs
are described by the clocks/clock-names properties of the SYSCFG node,
which remains the clock provider. It also keeps '#clock-cells' in the
SoC devicetrees, as you requested on patches 4/5 and 5/5.

The drivers are fixed instead: the STM32MP2 SYSCFG driver creates the
SYSCFG regmap and registers it with of_syscon_register_regmap(), so the
syscon core no longer attaches the first clock of the node to it. The
clock driver is an auxiliary driver, whose device is created by the
SYSCFG driver.

[1] https://lore.kernel.org/r/20261118-ups_mp25_clk_syscfg-v3-0-935de791a4ce@xxxxxxxxxxx

Best regards,

Gabriel


Best regards,
Krzysztof