> This last two lines doesn't make sense to me. Isn't the grab > toggling entirely in control of the QEMU process, regardless > of what state the guest is at ?
Actually, you're right, they do not make sense. This issue of having the guest taking a while to start and the toggle keys not working, only seem to appear when running the VM under libvirt. I was not able to reproduce this issue when running qemu directly from the command line. So either this is a libvirt issue or something related to my setup (VFIO with GPU passthrough, so a lot can go wrong). Should I send a new version of the patch with an updated commit message that does not mention this issue? Regards, Justinien