Description of problem: upgrade mesa-dri-drivers from 20.2.0-2 to 20.3.3-3 causes desktop icons smearing. Moving tears of icons along bottom of screen also occurring. Version-Release number of selected component (if applicable): mesa-dri-drivers-20.3.3.33-fc33 x86_64 How reproducible: Steps to Reproduce: 1. upgrade mesa-dri-drivers from 20.2.0-2 to 20.3.3-3, reboot. 2. login and move mouse around. Open windows and drag close to bottom of screen. Actual results: 1. observe desktop icons ghosting/smearing as mouse moved and clicked. 2. drag windows close to the bottom of screen and observe tearing that moves over time. Expected results: no altering of the icons and other items on screen. Additional info: doing a dnf downgrade mesa-dri-drivers to previous version and reboot doesn't permanently fix problem. UPDATE: problems lessened considerably but still occurs occasionally. Booting to previous kernel 5.10.9-201.fc33.x86_64 lessens remaining issues. Operating System: Fedora 33 KDE Plasma Version: 5.20.5 KDE Frameworks Version: 5.78.0 Qt Version: 5.15.2 Kernel Version: 5.10.10-200.fc33.x86_64 OS Type: 64-bit Processors: 4 × Intel® Core™ i5-3230M CPU @ 2.60GHz Memory: 5.7 GiB of RAM Graphics Processor: Mesa DRI Intel® HD Graphics 4000 Manufacturer: SAMSUNG ELECTRONICS CO., LTD. Product Name: 3570R/370R/470R/450R/510R/4450RV NP510R5E-A01UB
I have reported something similar in askfedora "Unstable Graphics on Fedora 33 with Kernels 5.10.9 and 5.10.10" Partial verbatim report as follows: "I’ve been having some odd things happen since upgrading to Kernels 5.10.9. and 5.10.10. Note that I am using a HPProDesk 600 G1 SFF with an i3-4160 CPU, 8GB RAM and integrated graphics. Desktop is Plasma. On 5.10.9, I found that my lower panel would flicker sometimes as I moved my mouse various places. Later on, my screen completely blacked out, so I went back to using 5.10.8. I upgraded to 5.10.10 when it became available. I still use it, but I have found (occasional) poor rendering of the numbers on my calendar/clock in the system tray, streaking on icons for my adblocker and “Heartbleed” display on my Google Chrome browser, streaking of widgets on my desktop and blackout of the screen background that goes away as I move my mouse over my widgets, and comes back when I move it away from the widgets (maybe vice-versa. I get a bit excited when it happens and don’t remember exactly how it happens)." A few screenshots appear on the askfedora thread. I also have a .mkv file I'll send if you like, but askfedora doesn't allow their upload and I don't know how to do it here. Give us a shout if you like; also if I should send another bug report.Thank you!
Created attachment 1752663 [details] Screen recording of one example of my (mister_t_stauss) anomaly I earlier noted I had no way to send you this .mkv file, but here it is. Note the other screen shot examples at askfedora "Unstable Graphics on Fedora 33 with Kernels 5.10.9 and 5.10.10"
Created attachment 1752664 [details] The rest of the screen shots Here are the rest of the screen shots. Sorry for all the posts.
Having the same issue with kernel versions 5.10.10 & 5.10.11 (didn't test with 5.10.9 because didn't update when released). I'm not sure if it's because of the mesa-dri-drivers package but I also have it installed (mesa-dri-drivers-20.3.4, haven't tested downgrading the package). My temporal solution is to keep using 5.10.8 to stop getting graphics issues. Systems specs are: HP ProDesk 600 G1 DM Intel i5-4570T (4) @ 3.600GHz This is the info dump given by the command `inxi -Gxx`: Graphics: Device-1: Intel Xeon E3-1200 v3/4th Gen Core Processor Integrated Graphics vendor: Hewlett-Packard driver: i915 v: kernel bus ID: 00:02.0 chip ID: 8086:0412 Display: x11 server: Fedora Project X.org 1.20.10 driver: modesetting unloaded: fbdev,vesa resolution: 1600x900~60Hz s-dpi: 96 OpenGL: renderer: Mesa DRI Intel HD Graphics 4600 (HSW GT2) v: 4.5 Mesa 20.3.4 compat-v: 3.0 direct render: Yes The info dump by `lspci -nnv | grep -iA10 vga`: 00:02.0 VGA compatible controller [0300]: Intel Corporation Xeon E3-1200 v3/4th Gen Core Processor Integrated Graphics Controller [8086:0412] (rev 06) (prog-if 00 [VGA controller]) DeviceName: Onboard IGD Subsystem: Hewlett-Packard Company Device [103c:21d0] Flags: bus master, fast devsel, latency 0, IRQ 29 Memory at f7800000 (64-bit, non-prefetchable) [size=4M] Memory at e0000000 (64-bit, prefetchable) [size=256M] I/O ports at f000 [size=64] Expansion ROM at 000c0000 [virtual] [disabled] [size=128K] Capabilities: <access denied> Kernel driver in use: i915 Kernel modules: i915
After a lot more tests and searching for others with similar problems, have following conclusions: 1. Is not a problem with mesa-dri-drivers based on trying different versions. Removed mesa subsystem from bug report and replaced with kernel. 2. Based on internet searches, seems to be tied to Intel i915 integrated display hardware and across multiple distros (fedora, arch, ...). Also some users report display freezes in some cases but the user can ssh into system. 3. Showed up after loading kernel 5.10.9 thru 5.10.12 (as of today). 4. Easiest way to trigger display flaws is to put system under some load greater than a load of 2 or 2.5 and then move mouse around the screen. Once load is lessened, the problem goes away after screen refreshed. On my system, running clamscan creates load of 2 or more reliably for a long period of time. I am not saying clamscan is the cause, just easy way to load system while poking at system. 5. So far, have not found VGA HANG messages while this display flaw is occurring on my system.
Have found an apparent workaround mentioned in https://gitlab.freedesktop.org/drm/intel/-/issues/3099 by applying i915.mitigations=off in grub. Works on my laptop and a workstation so far with 5.10.16 on each system. Time will tell if this continues to work. Will post update in a few days.
(In reply to Bruce Petrie from comment #6) > Have found an apparent workaround mentioned in > https://gitlab.freedesktop.org/drm/intel/-/issues/3099 by applying > i915.mitigations=off in grub. Works on my laptop and a workstation so far > with 5.10.16 on each system. Time will tell if this continues to work. Will > post update in a few days. Ok, so this is the same issue which we've been working on fixing / trying to pin down in bug 1925346. There are 3 fixes upstream for this which help fix this for almost all users (there is 1 report/case of the fixes not fully fixing things). It would be great if you (and other people seeing this) can try a kernel which I build with the 3 fixes added: https://koji.fedoraproject.org/koji/taskinfo?taskID=61751956 Here are some generic instructions on directly installing a kernel from koji (the Fedora buildsystem): https://fedorapeople.org/~jwrdegoede/kernel-test-instructions.txt Note this kernel is a test-build and as such is not signed, so to run this one you do need to disable secure-boot. *** This bug has been marked as a duplicate of bug 1925346 ***