Re: [RFC PATCH net-next v3 4/5] net: dsa: qca8k: flag QCA8337 internal CPU PHYs for SmartSpeed

[Yongzhao Chen <yongzhao.derek@gmail.com>]

Hi Andrew,

> How are the two ports coupled?
>
> Are there magnetics? Capacitor coupling?

I don't have a schematic and have not inspected the board closely
myself. What I have are third-party photos of the Redmi AX5400 board
(board marking M74). A teardown with the shields removed:

https://www.acwifi.net/19451.html

Full board: https://www.acwifi.net/wp-content/uploads/2022/03/DSC_4718.jpg
Close-up:   https://www.acwifi.net/wp-content/uploads/2022/03/DSC_4725.jpg

And a repair forum photo of the same board with the shields on:

https://remont-aud.net/forum/103-124057-1
Photo: https://remont-aud.net/_fr/1240/2027121.jpg

In the full board photo the QCA8337 is under the small shield at the top
left and the IPQ5018 is under the diagonal shield in the middle. They
are a few centimetres apart, judging from the photo, and I cannot see
magnetics or common-mode chokes between them.

Just outside the chamfered corner of the QCA8337 shield, between the two
chips, the silkscreen shows eight capacitor positions in pairs, C548 to
C555. In the forum photo, parallel traces appear to run from this group
towards the IPQ5018 shield. That looks like AC coupling on the four
pairs, but from the photos I cannot tell whether they are populated,
what their values are, or whether they are capacitors at all rather than
0 ohm links. I also cannot identify any bias or termination parts, so I
cannot say whether it matches the arrangement in the Micrel note.

The OpenWrt DTS for this board sets qcom,dac-preset-short-cable on the
IPQ5018 GE PHY. The qca,ar803x.yaml binding describes it as adjusting
amplitude, bias current and error correction for a PHY connected to
another PHY over a short distance. There is no equivalent setting on the
QCA8337 side.

Thanks,
Yongzhao Chen