On 2025/1/14 18:11, Krzysztof Kozlowski wrote:
On 14/01/2025 11:00, Dmitry Baryshkov wrote:
On Tue, 14 Jan 2025 at 09:57, Krzysztof Kozlowski <krzk@xxxxxxxxxx> wrote:
On Mon, Jan 13, 2025 at 04:03:10PM +0800, Yongxing Mou wrote:
+patternProperties:
+ "^display-controller@[0-9a-f]+$":
+ type: object
+ additionalProperties: true
+
+ properties:
+ compatible:
+ items:
+ - const: qcom,qcs8300-dpu
+ - const: qcom,sa8775p-dpu
+
+ "^displayport-controller@[0-9a-f]+$":
+ type: object
+ additionalProperties: true
+
+ properties:
+ compatible:
+ items:
+ - const: qcom,qcs8300-dp
+ - const: qcom,sm8650-dp
Parts of qcs8300 display are compatible with sa8775p, other parts with
sm8650. That's odd or even not correct. Assuming it is actually correct,
it deserves explanation in commit msg.
It seems to be correct. These are two different IP blocks with
different modifications. QCS8300's DP configuration matches the SM8650
([1]), though the DPU is the same as the one on the SA8775P platform.
[1] https://lore.kernel.org/dri-devel/411626da-7563-48fb-ac7c-94f06e73e4b8@xxxxxxxxxxx/
That's the driver, so you claim that qcs8300, which is a sa8775p, is not
compatible with sa8775p because of current driver code? You see the
contradiction? sa8775p is not compatible with sa8775p because of current
driver patch?
I don't think it is correct, but let's repeat: if you think otherwise,
this should be explain in commit msg.
Best regards,
Krzysztof
Hi,let me explain this: qcs8300 uses the same DPU as sa8775p, which is
DPU_8_4. Therefore, for the DPU driver, qcs8300 reuses the driver of
sa8775p. However, for the DisplayPort controller of qcs8300, it's
different with sa8775p. qcs8300 only supports one DisplayPort output
port, while sa8775p has two DPUs and supports four DisplayPort outputs.
Therefore, the DisplayPort controller driver of sa8775p cannot be reused
for qcs8300. Additionally, the base offset of qcs8300's DisplayPort
controller is the same as that of sm8650, so the DisplayPort controller
reuses "qcom,sm8650-dp". I explained this in the commit messages of the
previous two bindings, but it might not have been explained in the
commit message of this patch.i will update the commit msg in next patchset.
