Unfortunately, no. I don't have any real understanding of the problem, or of how virtual COM ports work at a low level in general.
I can't reproduce the problem because I don't own a board with the sketchy CH340 chip so everything I know about the problem comes from the user reports.
I am interested in learning more about the subject since this is impacting a lot of users and increasing the burden on the forum helpers (some of the less distinctive mysterious upload errors reported by ESP32/ESP8266 board users are also surely caused by this problem), but it is also not something I can put a significant amount of effort towards since it is a self-inflicted problem caused by people buying low quality hardware from unreliable sources.
It is very interesting. There was a previous report of something similar here:
I also remember another report that the Arduino IDE 1.x Serial Monitor can connect to these boards even though the IDE 2.x Serial Monitor can not:
So it seems that there is something about the way the Arduino IDE 1.x Serial Monitor connects to the port that puts it in a state where AVRDUDE can work with the port even when the latest driver is in use. This points to the possibility of fixing the problem through a patch in the AVRDUDE codebase as well as the Arduino Serial Monitor codebase.
Congratulations on finding a solution! I'm glad it is working now. Thank you for taking the time to post an update.
Yeah, it is. I was previously using that SparkFun tutorial as the driver source in my instructions but later found a way to get the newer version directly from the chip manufacturer's website so I switched to that one since it is a more official driver source and because I figure it is generally better to use the newest possible version of a driver. But I never heard any reports of problems with the version of the driver from the SparkFun tutorial so you will probably be fine with that one.