Hi Patrick,
On 5/17/22 4:12 AM, Patrick DELAUNAY wrote:
Hi,
On 5/11/22 18:44, Sean Anderson wrote:
Hi Patrick,
On 5/10/22 3:51 AM, Patrick Delaunay wrote:
Add a minimal support for STM32MP13 RCC, the reset and clock controller
- update of the RCC MISC driver to bind the correct clock and reset driver
- reset driver, same than STM32MP15x = drivers/reset/stm32-reset.c
- clock driver, add a empty driver for STM32MP13x =
drivers/clk/stm32/clk-stm32mp13.c
- Add RCC node in SOC device tree with u-boot,dm-pre-reloc property
This serie is only a preliminary step for STM32MP13 clock and reset support in U-Boot, based on Linux kernel binding introduced by [1] and it prepares the next device tree alignment with Linux kernel.
The functional STMP13 clock driver based on CCF and on SCMI clocks provided by OP-TEE and the clock and reset references in SOC device tree will be pushed when the associated patches in [1] will be accepted.
[1] Introduction of STM32MP13 RCC driver (Reset Clock Controller) https://lore.kernel.org/linux-arm-kernel/20220316131000.9874-1-gabriel.ferna...
I'm not really sure what the purpose of this series is. Can you elaborate a bit on why we need a dummy clock driver? Why don't you just add the binding to the device tree without the associated driver?
After this serie, the RCC reset part is functional on STM32MP13 (probe and ops)
even if the associated binding is not present in device tree.
tested with:
------------------------- arch/arm/dts/stm32mp131.dtsi ------------------------- index fcb0af09b5..d9c6185bcf 100644 @@ -197,6 +197,7 @@ interrupt-names = "cmd_irq"; clocks = <&clk_pll4_p>; clock-names = "apb_pclk"; + resets = <&rcc 14224>; cap-sd-highspeed; cap-mmc-highspeed; max-frequency = <130000000>;
A dummy STM32MP13 clock driver is requested to allow RCC MISC and RCC RESET
binding and probe without issue.
Shouldn't the solution be to make the clock optional in the user?
This reset support was requested by SDMCC driver and SD-Card boot, before the patch:
http://patchwork.ozlabs.org/project/uboot/patch/20220506160540.13.I39b69e8dc...
As we solve the SDMCC dependency issue (reset become optionnal), so this serie is no more mandatory.
This serie is a just a cleanup / preliminary step, but I can drop this dummy RCC driver if it is disturbing.
I would like that, since this is not mandatory.
--Sean