Created attachment 130203 [details] screenshot of corruption Bug description: When dragging the window width handle back and forth, the text in the window can become corrupted as in the screenshot or can also become blacked out. System environment: -- chipset: IVB -- system architecture: 64-bit -- xf86-video-intel: 2.99.917+git20161206-1 -- xserver: X.Org X Server 1.19.2 -- mesa: OpenGL core profile version string: 3.3 (Core Profile) Mesa 13.0.5 -- libdrm: Version: 2.4.74 -- kernel: 4.9.0-2-amd64 -- Linux distribution: Debian Stretch -- Machine or mobo model: -- Display connector: DVI Reproducing steps: Grab the handle on the left or right side and pull the window size in either direction Additional info:
Please attach Xorg.0.log. Are you using a compositor and dri2 or dri3?
Created attachment 130206 [details] Xorg.0.log
I'm using Gnome 3.22, so the compositor is Mutter 3.22. There's nothing in the changelog about switching to dri3, so I assume it's dri2.
Hmm, all the versions in that Xorg.0.log don't match your listed values -- I guess your system has moved away from storing Xorg.0.log. journalctl? My first thought is that it is a badly synchronized buffer exchange, though if you can quickly check kernels 4.4 or 4.10 (or anything else you might have precompiled) just to rule out v4.9.
I can check an older kernel. I don't know about a newer one. I have had similar corruption in tooltips for a long time, possibly going back a year or two. Maybe it's the same issue?
The tooltip corruption is the same type of buffer exchange issue as I suspect is happening here. I'm trying not to jump to conclusions! If you are happy to compile your own mesa, we can try those patches. First though, it is easy to jump between a few major versions, one package at a time, to see if we can narrow down the culprit.
Which Mesa patches?
I tested the 4.9 kernel, and the tooltips and Gedit bugs remain. However, I tested the 4.4 kernel, and while I can still reproduce the tooltips bug, which is years old, I can't reproduce the Gedit bug, which seems to be much more recent. There may be a regressions somewhere in between 4.4 and 4.9, and maybe the tooltips issue would be fixed by the Mesa patches you mentioned.
Created attachment 130213 [details] [review] dri2: Serialize with all X rendering on glXWaitX() This should fix the tooltip corruption (dri2)
I think the tooltips bug I was seeing is chromium bug 442111. because I can't reproduce it in Firefox. Also, that bug is 2 years old, which seems about right. So, it's just down to the kernel issue to fix the Gedit regression.
(In reply to Brandon from comment #10) > I think the tooltips bug I was seeing is chromium bug 442111. because I > can't reproduce it in Firefox. Also, that bug is 2 years old, which seems > about right. So, it's just down to the kernel issue to fix the Gedit > regression. Hello Brandon, Is this bug still valid with the last kernel? Is there any new information in this case? Thank you.
(In reply to Elizabeth from comment #11) > (In reply to Brandon from comment #10) > > I think the tooltips bug I was seeing is chromium bug 442111. because I > > can't reproduce it in Firefox. Also, that bug is 2 years old, which seems > > about right. So, it's just down to the kernel issue to fix the Gedit > > regression. > > Hello Brandon, > Is this bug still valid with the last kernel? Is there any new information > in this case? Thank you. My software versions have not significantly changed since I posted this bug. After Debian Testing switches up to Gnome 3.24 and the kernels are upgraded past 4.9, I will be able to add information. FWIW, the bug remains.
(In reply to Brandon from comment #12) > My software versions have not significantly changed since I posted this bug. > After Debian Testing switches up to Gnome 3.24 and the kernels are upgraded > past 4.9, I will be able to add information. FWIW, the bug remains. Hello Brandon, Could you please attach dmesg with parameter drm.debug=0xe on grub with the latest kernel? Once you provide the information, please mark as REOPEN. Thanks.
After upgrading to the 4.11 kernel, this issue seems to be fixed. I am therefore closing this bug.
(In reply to Brandon from comment #14) > After upgrading to the 4.11 kernel, this issue seems to be fixed. I am > therefore closing this bug. Thanks for the information Brandon. Status: CLOSED.
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.