You’re not really supposed to do anything with it - it’s aimed more towards the kernel devs.
It’s basically saying that the code to redraw the damaged parts of the screen (i.e. parts that have changed in the frame buffer) is taking more time than expected or desired; more than 10 ms, and did so four times.
Unless it’s an issue for you, or it’s flooding your logs, you can ignore it.
They run the same debug kernel that you’re running and see the same messages which suggest that there’s either a code path with needs optimisation, a parameter that needs tuning or that there is a problem with the interaction with a driver (the video driver in this case - whichever one is handling the framebuffer in your specific situation)