Question about framebuffer

First of all, thank you for your great work, I learned a lot from it. The CPU of my device is also i7-8550U (with UHD620), so I try to use the framebuffer patch of this repo. Usually, I use the laptop’s Type-C and HDMI port to connect to two monitors. In general, it works well but has some little problems:

1. When two monitors are just connected (the macOS power on or monitor plugged in), they display solid green and get well after one or two seconds.
2. Sometimes a horizontal line of the opposite color to the appearing position will flicker on the screen (it appears as white if the background is black and appears as black if the background is white)
3. Sometimes the monitor which connected by Type-C will go off suddenly and get well after a second or two (the monitor did not lose the connection because the other two monitors have not changed, if they lose the connection, they will also go off first)

I have been learning WhateverGreen related knowledge recently to solve these problems, and I have some questions about the framebuffer patch of this repo:

1. Why should the pipe of con1 be set to 10

I saw that WhateverGreen’s explanation of pipe is just one sentence:

Appears to be used for grouping ports just like Piker says, but I cannot find the usage.

I can't search through Google about the role of pipe and the reason for modifying it.

After removing the framebuffer-con1-pipe, anything seems to work as same.

2. Why modify framebuffer-flags and framebuffer-conX-flags

The relevant comments in the https://github.com/acidanthera/WhateverGreen/blob/master/Manual/IntelFramebuffer.bt file seem hard to comprehend, and I noticed that some patches even modified unknown flags (such as framebuffer-con1-flags and framebuffer-con2-flags added flag CNUnknownFlag_200)

I also tried another AAPL,ig-platform-id = 0000C087 according to https://dortania.github.io/OpenCore-Install-Guide/config-laptop.plist/kaby-lake.html#deviceproperties but nothing seems different.

Moreover, is it possible that framebuffer-unifiedmem can help solve this problem?