Re: [PATCH v5 1/2] dt-bindings: usb: dwc3: Clean up hs_phy_irq in binding

From: Krishna Kurapati PSSNV
Date: Tue Dec 26 2023 - 10:04:23 EST




On 12/26/2023 5:52 PM, Krzysztof Kozlowski wrote:


This does not answer why, you sc8280xp and x1e80100 not get one optional
interrupt. I asked "why" you are doing this change. Why do you need it?
What is the rationale?

Then I grunted about unmanageable commit, because all my troubles to
review it are the effect of it: it is very difficult to read. It is also
difficult for you, because you keep making here mistakes. So if you
cannot write this commit properly and I cannot review it, then it is way
over-complicated, don't you think? But this is still second problem
here, don't ignore the fist - "why?"

HI Krzysztof,

Thanks for the review.
To answer the question,

"why ?" : The interrupts have been mis-interpreted on many platforms or
many interrupts are missing.

I asked about these two specific platforms. Please explain these
changes. Above is so generic that tells me nothing.


Is the question, "Why do x1e80100 and sc8280 don't have hs_phy_irq ?"
If so, I checked the SC8280 HW specifics and I see one small error. The name was printed wrong. I got it from another source. Will move sc8280 to list having 5 interrupts. As per x1e80100, I wasn't able to get my hands on the hw specifics and I followed the following link by Abel Vesa:

https://lore.kernel.org/r/20231214-x1e80100-usb-v1-1-c22be5c0109e@xxxxxxxxxx

As per the above patch, x1e80100 had only 4 interrupts.
For ipq5332, it has no hs_phy_irq and so I kept it under this section.


Now, if I am adding the missing interrupts, I need to segregate targets
also into respective buckets in the same patch and that is what making
this patch a little complicated. Is it possible / acceptable to split
this into two patches if this is the case. Can you help with suggestions
from your end ? Or may be I am understanding your question wrong ? 😅

Split the patch into manageable chunks.


I will try to split it up, but not sure if it is a good idea. I say so because all permutations should be added in single patch and I can't split that.

Regards,
Krishna,