On 12/12/2021 19:39, Krzysztof Kozlowski wrote:

On 06/12/2021 16:31, David Virag wrote:

quoted

Just like on Exynos850, the clock controller driver is designed to have
separate instances for each particular CMU, so clock IDs start from 1
for each CMU in this bindings header too.

Reviewed-by: Krzysztof Kozlowski <redacted>
Signed-off-by: David Virag <redacted>
---
Changes in v2:
  - Added R-b tag by Krzysztof Kozlowski

Changes in v3:
  - Nothing

Changes in v4:
  - Nothing

 include/dt-bindings/clock/exynos7885.h | 115 +++++++++++++++++++++++++
 1 file changed, 115 insertions(+)
 create mode 100644 include/dt-bindings/clock/exynos7885.h

Hi Sylwester,

The DTS/DTSI patch (7/7) depends on this one, just like the clock driver.

Since some time Arnd and Olof prefer not to have external trees going
into the arm-soc, even if this is only the header change. They recommend
one of:
1. to hard-code the numbers in DTS and replace numbers->macros later,
2. merge headers to arm-soc tree with DTS and provide the header to an
external (e.g. clk) tree,
3. wait with merging DTSI till headers reach mainline.

I propose that I take the clock headers, put on separate branch and
provide them to you as stable tag. You can base clk driver changes on
top of it.

Are you okay with this?

Hi Sylwester,

I see you applied the patches, so I understand we are not going with
this proposal.

David,

Your DTSI and DTS will have to either wait for next cycle or please
resend with clock macros replaced with numbers + TODO note.

Best regards,
Krzysztof