Re: [PATCH v2 1/2] dt-bindings: iio: tyhx,hx9023s: Add performance tuning configuration

From: Yasin Lee
Date: Thu Sep 26 2024 - 03:58:45 EST



On 9/26/24 14:53, Krzysztof Kozlowski wrote:
On Thu, Sep 26, 2024 at 12:40:17AM +0800, Yasin Lee wrote:
When hardware design introduces significant sensor data noise,
performance can be improved by adjusting register settings.

Signed-off-by: Yasin Lee <yasin.lee.x@xxxxxxxxx>
---
.../bindings/iio/proximity/tyhx,hx9023s.yaml | 19 +++++++++++++++++++
1 file changed, 19 insertions(+)

diff --git a/Documentation/devicetree/bindings/iio/proximity/tyhx,hx9023s.yaml b/Documentation/devicetree/bindings/iio/proximity/tyhx,hx9023s.yaml
index 64ce8bc8bd36..0673c40472bd 100644
--- a/Documentation/devicetree/bindings/iio/proximity/tyhx,hx9023s.yaml
+++ b/Documentation/devicetree/bindings/iio/proximity/tyhx,hx9023s.yaml
@@ -28,6 +28,18 @@ properties:
vdd-supply: true
+ tyhx,performance-tuning:
+ description:
+ Optional, When hardware design introduces significant sensor data noise,
+ performance can be improved by adjusting register settings, including
+ but not limited to sample integration time and average sample count.
+ Performance tuning parameters represented as register address-value pairs.
+ The array consists of 8-bit values, where each pair represents a register
+ address followed by the value to be written to that register.
+ $ref: /schemas/types.yaml#/definitions/uint8-array
I think you wanted here in such case uint8-matrix with one dimension set
to two and other 256 or whatever.

I also wonder why do allow any register to be tuned. To me "tuning"
sounds like improving or tweaking the configuration, so for example I
would not expect to change some control-like registers, e.g. interrupts,
enable/disable, reset etc.


+ minItems: 2
+ maxItems: 512
+
"#address-cells":
const: 1
@@ -65,6 +77,13 @@ examples:
interrupt-parent = <&pio>;
interrupts = <16 IRQ_TYPE_EDGE_FALLING>;
vdd-supply = <&pp1800_prox>;
+ tyhx,performance-tuning = [00 00
This above is clearing entirely the GLOBAL_CTRL0 register, which is not
tuning.

You are basically avoiding hardware description as properties in DT and
just dump a blob.


Best regards,
Krzysztof
Hi Krzysztof,

Thank you for your feedback.

You're absolutely right about the operation on the 00 register. That was indeed a mistake on my part, and it was not my intention to clear the GLOBAL_CTRL0 register. It was only meant to serve as an example.

I also now understand that importing a data block in this manner doesn't align with the design philosophy of device trees. Moving forward, I will abstract each possible tuning configuration into separate properties. Since there are quite a few properties to handle, this will result in a larger codebase.

Given the additional work required, I will take a few more days to revise the patch and resubmit it.

Thank you again for your valuable input.

Best regards,
Yasin Lee