device: 0.170 semantic: 0.097 other: 0.096 socket: 0.080 graphic: 0.079 PID: 0.073 files: 0.065 vnc: 0.062 network: 0.058 debug: 0.055 performance: 0.052 permissions: 0.046 boot: 0.037 KVM: 0.029 debug: 0.251 files: 0.154 PID: 0.098 other: 0.080 device: 0.076 network: 0.067 graphic: 0.057 socket: 0.044 semantic: 0.038 performance: 0.035 KVM: 0.035 permissions: 0.022 boot: 0.022 vnc: 0.019 GTK display and mouse input area scaling fails when using vfio-pci device Version qemu 3.0.0-1 running on Arch. Found on Windows 8.1 and Windows 10 VM's when using Intel gvt-g device. While in fullscreen the GTK display is scaled larger than the x11 screen or virtual machine resolution. Without choosing zoom-to-fit portions of the VM display are not shown on x11 screen regardless of the VM resolution. When zoom-to-fit is done the mouse that's shown on screen and actual input are off sync. The mouse can wander off screen when going left and down. This message is shown when changing from gxl-vga to vfio-pci in view menu. (qemu-system-x86_64:6472): Gtk-WARNING **: 09:50:06.663: drawing failure for widget 'GtkDrawingArea': NULL pointer (qemu-system-x86_64:6472): Gtk-WARNING **: 09:50:06.664: drawing failure for widget 'GtkNotebook': NULL pointer (qemu-system-x86_64:6472): Gtk-WARNING **: 09:50:06.664: drawing failure for widget 'GtkBox': NULL pointer (qemu-system-x86_64:6472): Gtk-WARNING **: 09:50:06.664: drawing failure for widget 'GtkWindow': NULL pointer Hi, I have a hack for this issue. In general, the gvt-g is buggy in that the kernel driver i915 failed to provide the actual resolution for the vgpu and always report the largest available one. The following patch modified the gtk-egl implementation for DMA Buf mode, to convert mouse movement/tablet position to the scaled window coordinates. See https://pastebin.com/96iBLdDx Hi there, Would anyone be so kind to review my patch? Thanks. https://lists.nongnu.org/archive/html/qemu-devel/2018-10/msg06806.html Chen, have you seen the following bug report? https://bugs.launchpad.net/qemu/+bug/1592351 See my comment there, with recent git build of 3.1.0 I don't have a pointer offset issue. Perhaps this is due to this issue https://github.com/intel/gvt-linux/issues/71 https://yadi.sk/i/Ejg_6JJEBIc-Zw The QEMU project is currently considering to move its bug tracking to another system. For this we need to know which bugs are still valid and which could be closed already. Thus we are setting older bugs to "Incomplete" now. If you still think this bug report here is valid, then please switch the state back to "New" within the next 60 days, otherwise this report will be marked as "Expired". Or please mark it as "Fix Released" if the problem has been solved with a newer version of QEMU already. Thank you and sorry for the inconvenience. [Expired for QEMU because there has been no activity for 60 days.]