Re: [PATCH RESEND v9 5/5] arm64: dts: apm: Harmonize DWC USB3 DT nodes name

[Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>]

On 24/06/2022 22:59, Serge Semin wrote:
> On Fri, Jun 24, 2022 at 07:17:53PM +0200, Krzysztof Kozlowski wrote:
>> On 24/06/2022 16:16, Serge Semin wrote:
>>> In accordance with the DWC USB3 bindings the corresponding node
>>> name is suppose to comply with the Generic USB HCD DT schema, which
>>> requires the USB nodes to have the name acceptable by the regexp:
>>> "^usb(@.*)?" . Make sure the "snps,dwc3"-compatible nodes are correctly
>>> named despite of the warning comment about possible backward
>>> compatibility issues.
>>
> 
>> Sometimes node name is exposed to user-space which depends on it. How
>> did you check there is no issue here?
> 
> I well remember the Qcom problem caused by one of my patch:
> https://lore.kernel.org/lkml/CALAqxLX_FNvFndEDWtGbFPjSzuAbfqxQE07diBJFZtftwEJX5A@mail.gmail.com/
> 
> The next patch caused the same problem, but hasn't been reverted.
> https://lore.kernel.org/lkml/CALAqxLWGujgR7p8Vb5S_RimRVYxwm5XF-c4NkKgMH-43wEBaWg@mail.gmail.com/
> 
> As before I am more inclined to thinking that the problem was mainly caused
> by the improper node-name utilization. Anyway John later noted that the
> problem was fixed in the user-space. 

Yes, I remember. The node names are not considered ABI, therefore any
reliance on them is not correct.

I wonder however what was the reasoning for this comment in APM DTS.

> That why afterwards you were able
> to provide the commit b77a1c4d6b05 ("arm64: dts: qcom: correct DWC3
> node names and unit addresses").
> 
> Anyway I am not able to track the way the node-name is used on the
> affected platform and can't make sure that the dts would be still
> working well on that devices. But seeing nobody responded/commented on
> this patch for more than a year we can at least try to merge this in
> and see whether it causes any problem should the denoted platform is
> still in use. If it does we can revert the update back and forget
> about it.

The APM is kind of abandoned, so indeed we might never get a reply.

I'll take it.

Best regards,
Krzysztof