The core version of EMAC on qcs615 has minor differences compared to that
on sm8150. During the bring-up routine, the loopback bit needs to be set,
and the Power-On Reset (POR) status of the registers isn't entirely
consistent with sm8150 either.
Therefore, it should be treated as a separate entity rather than a
fallback option.

Fixes: 32535b9410b8 ("dt-bindings: net: qcom,ethqos: add description for qcs615")
Signed-off-by: Yijie Yang <quic_yijiyang@quicinc.com>
---
 Documentation/devicetree/bindings/net/qcom,ethqos.yaml | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)

diff --git a/Documentation/devicetree/bindings/net/qcom,ethqos.yaml b/Documentation/devicetree/bindings/net/qcom,ethqos.yaml
index 0bcd593a7bd093d4475908d82585c36dd6b3a284..576a52742ff45d4984388bbc0fcc91fa91bab677 100644
--- a/Documentation/devicetree/bindings/net/qcom,ethqos.yaml
+++ b/Documentation/devicetree/bindings/net/qcom,ethqos.yaml
@@ -23,12 +23,9 @@ properties:
           - enum:
               - qcom,qcs8300-ethqos
           - const: qcom,sa8775p-ethqos
-      - items:
-          - enum:
-              - qcom,qcs615-ethqos
-          - const: qcom,sm8150-ethqos
       - enum:
           - qcom,qcs404-ethqos
+          - qcom,qcs615-ethqos
           - qcom,sa8775p-ethqos
           - qcom,sc8280xp-ethqos
           - qcom,sm8150-ethqos

-- 
2.34.1

On 24/12/2024 04:07, Yijie Yang wrote:
> The core version of EMAC on qcs615 has minor differences compared to that
> on sm8150. During the bring-up routine, the loopback bit needs to be set,
> and the Power-On Reset (POR) status of the registers isn't entirely
> consistent with sm8150 either.
> Therefore, it should be treated as a separate entity rather than a
> fallback option.

... and explanation of ABI impact? You were asked about this last time,
so this is supposed to end up here.

> 
> Fixes: 32535b9410b8 ("dt-bindings: net: qcom,ethqos: add description for qcs615")
> Signed-off-by: Yijie Yang <quic_yijiyang@quicinc.com>
> ---
>  Documentation/devicetree/bindings/net/qcom,ethqos.yaml | 5 +----
>  1 file changed, 1 insertion(+), 4 deletions(-)
Best regards,
Krzysztof

On 2024-12-24 17:02, Krzysztof Kozlowski wrote:
> On 24/12/2024 04:07, Yijie Yang wrote:
>> The core version of EMAC on qcs615 has minor differences compared to that
>> on sm8150. During the bring-up routine, the loopback bit needs to be set,
>> and the Power-On Reset (POR) status of the registers isn't entirely
>> consistent with sm8150 either.
>> Therefore, it should be treated as a separate entity rather than a
>> fallback option.
> 
> ... and explanation of ABI impact? You were asked about this last time,
> so this is supposed to end up here.

I actually replied to this query last time, but maybe it wasn't clear. 
Firstly, no one is using Ethernet on this platform yet. Secondly, the 
previous fallback to sm8150 is incorrect and causes packet loss. 
Instead, it should fall back to qcs404.

> 
>>
>> Fixes: 32535b9410b8 ("dt-bindings: net: qcom,ethqos: add description for qcs615")
>> Signed-off-by: Yijie Yang <quic_yijiyang@quicinc.com>
>> ---
>>   Documentation/devicetree/bindings/net/qcom,ethqos.yaml | 5 +----
>>   1 file changed, 1 insertion(+), 4 deletions(-)
> Best regards,
> Krzysztof

-- 
Best Regards,
Yijie

On 25/12/2024 09:58, Yijie Yang wrote:
> 
> 
> On 2024-12-24 17:02, Krzysztof Kozlowski wrote:
>> On 24/12/2024 04:07, Yijie Yang wrote:
>>> The core version of EMAC on qcs615 has minor differences compared to that
>>> on sm8150. During the bring-up routine, the loopback bit needs to be set,
>>> and the Power-On Reset (POR) status of the registers isn't entirely
>>> consistent with sm8150 either.
>>> Therefore, it should be treated as a separate entity rather than a
>>> fallback option.
>>
>> ... and explanation of ABI impact? You were asked about this last time,
>> so this is supposed to end up here.
> 
> I actually replied to this query last time, but maybe it wasn't clear. 
> Firstly, no one is using Ethernet on this platform yet. Secondly, the 
> previous fallback to sm8150 is incorrect and causes packet loss. 
> Instead, it should fall back to qcs404.
> 
And if you send v3 with same commit msg, I will ask the same. Your
commit msg must answer to this.

Best regards,
Krzysztof

On Wed, Dec 25, 2024 at 04:58:20PM +0800, Yijie Yang wrote:
> 
> 
> On 2024-12-24 17:02, Krzysztof Kozlowski wrote:
> > On 24/12/2024 04:07, Yijie Yang wrote:
> > > The core version of EMAC on qcs615 has minor differences compared to that
> > > on sm8150. During the bring-up routine, the loopback bit needs to be set,
> > > and the Power-On Reset (POR) status of the registers isn't entirely
> > > consistent with sm8150 either.
> > > Therefore, it should be treated as a separate entity rather than a
> > > fallback option.
> > 
> > ... and explanation of ABI impact? You were asked about this last time,
> > so this is supposed to end up here.
> 
> I actually replied to this query last time, but maybe it wasn't clear.
> Firstly, no one is using Ethernet on this platform yet. Secondly, the
> previous fallback to sm8150 is incorrect and causes packet loss. Instead, it
> should fall back to qcs404.

One of the purposes of the commit message is to answer questions
reviews might have. You were even asked this question, so that should
of been a clue to include the answer in the commit message.

	Andrew

On 2024-12-26 00:17, Andrew Lunn wrote:
> On Wed, Dec 25, 2024 at 04:58:20PM +0800, Yijie Yang wrote:
>>
>>
>> On 2024-12-24 17:02, Krzysztof Kozlowski wrote:
>>> On 24/12/2024 04:07, Yijie Yang wrote:
>>>> The core version of EMAC on qcs615 has minor differences compared to that
>>>> on sm8150. During the bring-up routine, the loopback bit needs to be set,
>>>> and the Power-On Reset (POR) status of the registers isn't entirely
>>>> consistent with sm8150 either.
>>>> Therefore, it should be treated as a separate entity rather than a
>>>> fallback option.
>>>
>>> ... and explanation of ABI impact? You were asked about this last time,
>>> so this is supposed to end up here.
>>
>> I actually replied to this query last time, but maybe it wasn't clear.
>> Firstly, no one is using Ethernet on this platform yet. Secondly, the
>> previous fallback to sm8150 is incorrect and causes packet loss. Instead, it
>> should fall back to qcs404.
> 
> One of the purposes of the commit message is to answer questions
> reviews might have. You were even asked this question, so that should
> of been a clue to include the answer in the commit message.
> 

I will include it in the commit message in the next version.

> 	Andrew

-- 
Best Regards,
Yijie