I don't have the nvidia for pass through to try this with; but I suggest that you try the following:
a) Start a windows vm running an older version unaffected by the bug and start a 2d test
b) run 'perf top' on the host while the test is running and capture the results
- make sure you have debug symbols for both qemu (and related libraries) and the kernel
c) now repeat a/b for the newer version of windows that's affected
add the results of the 'perf top' to this bug; hopefully we'll be able to see qemu or the kernel spending a lot more time in some particular routine in the new version.
I don't have the nvidia for pass through to try this with; but I suggest that you try the following:
a) Start a windows vm running an older version unaffected by the bug and start a 2d test
b) run 'perf top' on the host while the test is running and capture the results
- make sure you have debug symbols for both qemu (and related libraries) and the kernel
c) now repeat a/b for the newer version of windows that's affected
add the results of the 'perf top' to this bug; hopefully we'll be able to see qemu or the kernel spending a lot more time in some particular routine in the new version.