On 26/11/2021 10:16, conor wrote:

On 26/11/2021 09:48, Krzysztof Kozlowski wrote:

quoted

EXTERNAL EMAIL: Do not click links or open attachments unless you know 
the content is safe

On 25/11/2021 16:31, Geert Uytterhoeven wrote:

quoted

"make dtbs_check" reports:

     arch/riscv/boot/dts/microchip/microchip-mpfs-icicle-kit.dt.yaml: 
soc: refclk: {'compatible': ['fixed-clock'], '#clock-cells': [[0]], 
'clock-frequency': [[600000000]], 'clock-output-names': 
['msspllclk'], 'phandle': [[7]]} should not be valid under {'type': 
'object'}
       From schema: dtschema/schemas/simple-bus.yaml

Fix this by moving the node out of the "soc" subnode.
While at it, rename it to "msspllclk", and drop the now superfluous
"clock-output-names" property.

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
---
  arch/riscv/boot/dts/microchip/microchip-mpfs.dtsi | 13 ++++++-------
  1 file changed, 6 insertions(+), 7 deletions(-)

It is also logical because refclk usually is not a property of the SoC.
It actually might be a property of board...

This is one of the fun FPGAisms like the GPIO interrupt configuration. 
This clock setting is determined by what design has been loaded onto the 
FPGA - the msspll outputs are configurable, I could redo my FPGA design 
and change this to 500 MHz etc. In turn the msspll clock is set by 
another clock source that is actually on the board of either 100 or 125 
MHz.

Since it's not set until bitstream programming time, I would agree that 
that property should be moved to out of mpfs.dtsi. (typo fixed)

Geert/Krzysztof,
Would the following make sense:
- Since the refclk hardware is a part of the chip, move the refclk out 
of the soc node but leave it in mfps.dtsi
- The clk freq itself is set by the fpga bitstream, so move the 
clock-frequency property to mpfs-icicle-kit.dts?

quoted

Acked-by: Krzysztof Kozlowski <redacted>


Best regards,
Krzysztof