I submitted this Fedora bug a while back: https://bugzilla.redhat.com/show_bug.cgi?id=1352325 The latest discoveries I have made tend to indicate it is a problem with the intel video drivers when running under a 4.6 kernel. The strange behavior disappeared when I plugged in an nvidia card and stopped using the onboard intel video on my motherboard. The most obvious problem happens with the cursor in a Windows 10 virtual machine using QXL/Spice for the display. A video attached to the fedora bugzilla above shows the horrible flickering plus the mouse images it leaves behind when I move the mouse. I also have an ancient Motif app displaying some xpm icons which start randomly flickering on and off (they are not animated icons :-). There is an attached video for that as well. On the 4.5 kernel, none of these problems happen. On the 4.6 kernel they happen all the time.
Can you add information about the system and also include logs if possible to better root cause the problem
Created attachment 125321 [details] A brand new run of dmesg This is the current dmesg output on my system (which currently has the new nvidia card installed, so it won't say much useful about intel graphics).
Created attachment 125322 [details] An old dmesg file I had around showing the intel video This is a few months old, but the hardware info hasn't changed any, so anything it says about the hardware should still be relevant.
Created attachment 125323 [details] A new dmidecode run showing motherboard info
Created attachment 125324 [details] /var/log/messages The date range in here shows booting with the old and new video, and almost certainly spans the time range where I had problems with the virtual machine cursor, but I don't think any kind of relevant errors showed up in the log.
Created attachment 125325 [details] libvirt qemu logs from running windows10 virtual machine This also has a range of dates in it. Some of these runs will have had cursor problems, but I don't see any messages that are different from one run to the next.
Created attachment 125326 [details] The xml definition of the virtual machine with cursor problems I don't know why this cursor madness only happens in the Windows 10 virtual machine, but if there is a clue, it might be here.
Created attachment 125327 [details] A new xorg log from booting with nvidia card This is my current Xorg.0.log file which was generated while the nvidia card is in the system.
Created attachment 125328 [details] An older xorg log showing intel drivers in use Like the old dmesg this is a few months old, but the hardware was the same when it was generated.
Can't think of anything else I could add, and I'm not sure any of this is actually useful, but there might be a hint buried in there somewhere :-).
Tom, can you try to reproduce the issue with drm.debug=0xe in the boot command line in order to collect more debug info in the dmesg? Then attach this new log. thanks
I'm afraid it is too late :-). It was a lot of trouble to move the box to a place I could work on it to put the nvidia card in, and now that it is in there, I can't run intel graphics (unless you know some magical kernel parameter I can use to disable the nvidia card without actually removing it from the system).
I am not sure if only black listing the nvidia driver, and may be setting in BIOS use of onboard gpu, should be sufficient.
I don't think that works. The onboard graphics completely disappears in the BIOS when "real" hardware is using the PCI bus.
Can you try to reproduce this bug with latest drm-tip kernel from (https://cgit.freedesktop.org/drm-tip branch drm-tip). Thanks
Fraid not. Like I said in comment 12, the hardware has changed. I can't get it to use the intel hardware now that an nvidia card is in the system (attempting to blacklist nvidia and nouveau just gives me no video at all).
not able to reproduced and hardware has change... moving bug to closure
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.