On Fri, Jul 17, 2020 at 12:14 PM Helen Koike [off-list ref] wrote:

Hi Rob,

Thanks for your review.

On 7/17/20 2:49 PM, Rob Herring wrote:

quoted

On Thu, Jul 2, 2020 at 1:13 PM Helen Koike [off-list ref] wrote:

quoted

aclk_isp_wrap is a child of aclk_isp, and hclk_isp_wrap is a child of
hclk_isp, thus we can remove parents from the list.

But it looks like it is the wrap clocks you are removing.

From this binding yes, but the idea is to add in the dt wherever clock
responsible for the full ACLK path for instance.
In the example below, clock aclk_isp is ACLK_ISP0_WRAPPER.
Does this make sense?

Just perhaps clarify the renaming.

quoted

quoted

Also, for the isp0, we only need the ISP clock, ACLK and HCLK.
In the future we'll need a pixel clock for RK3288 and RK3399, and a JPEG
clock for RK3288.

So with the goal to cleanup the dt-bindings and remove it from staging,
simplify clock names to isp, aclk and hclk.

For reference, this is the isp clock topology on RK3399:

 xin24m
    pll_npll
       npll
          clk_isp1
          clk_isp0
    pll_cpll
       cpll
          aclk_isp1
             aclk_isp1_noc
             hclk_isp1
                aclk_isp1_wrapper
                hclk_isp1_noc
          aclk_isp0
             hclk_isp1_wrapper
             aclk_isp0_wrapper
             aclk_isp0_noc
             hclk_isp0
                hclk_isp0_wrapper
                hclk_isp0_noc
 pclkin_isp1_wrapper

Signed-off-by: Helen Koike <redacted>

---

Changes in V4:
- update binding according to suggestion by Robin Murphy
on https://patchwork.kernel.org/patch/11475007/

Changes in V3:
- this is a new patch in the series
---
 .../bindings/media/rockchip-isp1.yaml         | 30 +++++++++----------
 drivers/staging/media/rkisp1/rkisp1-dev.c     |  8 ++---
 2 files changed, 17 insertions(+), 21 deletions(-)

diff --git a/drivers/staging/media/rkisp1/Documentation/devicetree/bindings/media/rockchip-isp1.yaml b/drivers/staging/media/rkisp1/Documentation/devicetree/bindings/media/rockchip-isp1.yaml
index 4d111ef2e89c7..f10c53d008748 100644
--- a/drivers/staging/media/rkisp1/Documentation/devicetree/bindings/media/rockchip-isp1.yaml
+++ b/drivers/staging/media/rkisp1/Documentation/devicetree/bindings/media/rockchip-isp1.yaml

@@ -24,20 +24,20 @@ properties:
     maxItems: 1

   clocks:
-    items:
-      - description: ISP clock
-      - description: ISP AXI clock clock
-      - description: ISP AXI clock  wrapper clock
-      - description: ISP AHB clock clock
-      - description: ISP AHB wrapper clock

This is the correct way to describe multiple clocks.

The idea was to prepare for rk3288 and rk3399 isp1, as suggested here https://patchwork.kernel.org/patch/11475007/#23462085

Or should we do:

clocks:
  oneOf:
    # rk3288 clocks
    - items:
      - description: ISP clock
      - description: ISP AXI clock
      - description: ISP AHB clock
      - description: ISP Pixel clock
      - description: ISP JPEG source clock

The main section should have this and 'minItems: 3'. IOW, it's a
superset of what's valid. Then you can restrict specific compatibles
further with an if/then schema. For rk3288, you need one with
'minItems: 5'.

    # rk3399 isp0 clocks
    - items:
      - description: ISP clock
      - description: ISP AXI clock
      - description: ISP AHB clock

And this would be an if/then schema based on the compatible string and
defining 'maxItems: 3'.

    # rk3399 isp1 clocks
    - items:
      - description: ISP clock
      - description: ISP AXI clock
      - description: ISP AHB clock
      - description: ISP Pixel clock

And an if/then with { minItems: 4, maxItems: 4 }. Or really since
these are just different instances, just combine them into 1
conditional allowing 3 or 4 clocks.

There are lots of examples to follow in the tree.

Rob