Sorry, I don't have any ideas. I don't think we've seen anything like this either; we'd probably need a machine that uses the same touchpad or touchpad driver.
If the Nouveau driver also supports the RandR 1.4 provider source capability, it might be an interesting experiment to configure PRIME with Nouveau -> modesetting to see if it's a generic interaction between the touchpad driver and RandR 1.4. Alternatively, it might be useful to attempt configuring an X screen on the Intel GPU using the modesetting driver and no offloading, to see if it's a general interaction between the touchpad driver and the modesetting X driver. That should be a lot easier to set up.
Or... another possibility would be to try configuring the offloading with the Intel driver as the provider output instead of the modesetting driver. I can't remember the specific problems we ended up having with the Intel driver that led us to recommend the modesetting driver in the README instead, but if the Intel driver works as an output for you, and the problem is due to an interaction between the modesetting and touchpad drivers, then this may be a way to avoid the problem.
These are all just wild guesses though; I don't have any good ideas as to why this might be happening.
Sorry, I don't have any ideas. I don't think we've seen anything like this either; we'd probably need a machine that uses the same touchpad or touchpad driver.
If the Nouveau driver also supports the RandR 1.4 provider source capability, it might be an interesting experiment to configure PRIME with Nouveau -> modesetting to see if it's a generic interaction between the touchpad driver and RandR 1.4. Alternatively, it might be useful to attempt configuring an X screen on the Intel GPU using the modesetting driver and no offloading, to see if it's a general interaction between the touchpad driver and the modesetting X driver. That should be a lot easier to set up.
Or... another possibility would be to try configuring the offloading with the Intel driver as the provider output instead of the modesetting driver. I can't remember the specific problems we ended up having with the Intel driver that led us to recommend the modesetting driver in the README instead, but if the Intel driver works as an output for you, and the problem is due to an interaction between the modesetting and touchpad drivers, then this may be a way to avoid the problem.
These are all just wild guesses though; I don't have any good ideas as to why this might be happening.