Description of problem:Dell XPS13 9310 computer. Mageia 8 updated. 0000:00:02.0 VGA compatible controller: Intel Corporation TigerLake-LP GT2 [Iris Xe Graphics] (rev 01) At times the text has altrnative lines displaced by a few (EG 5) PIXELS which of course make the text unreadable. Usually this clears up almost immediately, but occasionally it does not, unless I move around the window. I will include a picture I took of the screen of Google Chrome. Sometimes it is adjacent ines that are displaced, sometimes it is groups of 3 or 4 lines that are displaced. Brand new machine. Version-Release number of selected component (if applicable): How reproducible: It happens sporadically but at least once a day. Steps to Reproduce: 1.No idea. It seems to just happen when new text is written to a window, or the window is opened. 2. 3.
I cannot send the jpeg of the screen as it is too large, so here is a link to the picture. http://www.theory.physics.ubc.ca/xps13-shattered-lines.jpg Note that the Moire pattern is not part of the problem-- just the shattered lines below the middle and right of the screen
> I cannot send the jpeg of the screen as it is too large [Max size is 1Mb] This is surprising. Perhaps it is due to a huge screen; mine is 1366x768. I have just made 3 different (Gnome, Spectacle, Flameshot) PNG (usually bigger than JPG) screenshots of the browser window of this bug, and they were all <110Kb. I found the link you posted above rather poor to evaluate. 1. Please post the console output of: $ inxi -MSGxx 2. You cite Google Chrome; is the problem tied to that? When you see it, can you simultaneously open a different browser to show exactly the same URL/screen? And see what shows. 3a. Take a screenshot of the bad display window, which might well fill the screen anyway, but many screenshotters offer an 'active window' option. If you have & can use Flameshot, this has the powerful facility to directly annotate the screenshot before saving it; for example, draw a circle around the offending area. 3b. Or take a screenshot selecting the 'bad' area, and some of the surround to put it into context.
Status: NEW => NEEDINFOCC: (none) => lewyssmith
Created attachment 12962 [details] reduced png file for the screen picture without moire artifact Here is a reduced size picture which fits the limits and without the moire artifacts of the jpg file on the web. See the test in the bottom right quarter of the screen which is shadowed by pixel displacement of some of the rows of pixels on the screen.
Do you have an external speaker or anything else that would be generating a magnetic field near the affected part of the screen?
CC: (none) => davidwhodgins
Actually an LCD or OLED screen is not influenced by magnetic fields. Also the effect would be strongest near the magnet, and here it is constant strength in a whole area and then zero strength in the rest of the screen. But no there is nothing in the vicinity of the laptop which could be doing this.
This sounds like an issue with the TGL Xe GPU I'd suggest you try the 5.14.14-1 kernel from backports that have a lot of gpu fixes that upstream wont backport to 5.10 series
I just installed 5.14.14-1 and things are looking better. The slight weird behaviours of text where half lines disappear before the rest do not seem to be there anymore. I will test a few days but it looks like this is a vast improvement. Thanks:
Oops spoke too soon. Moving a termianl window over the chrome window I got the messed up stuff in the picture in attachment 2 [details]
Created attachment 12963 [details] new shattered text after kernel 14.14.0-1 The text below Summary-- the third line to sixth line-- again has the shadowing but by more than one character width.
Are all the screenshots taken with a camera photographing the screen? If you use a screenshot program, can you catch the problem? (and when you view the shot do it show the same problem?) Also, can you attach an external screen, and do the same problem show there? Can you boot another OS? What I am after is it may be a hardware problem in the transmission to the screen, so lines occasionally get shifted (i.e missed clock pulses) This may be temperature dependant, if you want to stress it, so also dependant on system load, etc, also on internal electrical noise from various sources...
CC: (none) => fri
(In reply to w unruh from comment #8) > Oops spoke too soon. Moving a termianl window over the chrome window I got > the messed up stuff in the picture in attachment 2 [details] I cant see whats wrong with it. Or I am just tired? Is this a software (not camera) screenshot?
Summary: Some text gets lines displaced giving a mess for text. => Random display lines get shifted right. Dell XPS13 9310
Yes, all are taken by camera. I have no idea whatsoever what the attachment pointer is comment 8 is. It is certainly not mine, and is not anything I have ever seen. The attachment in comment 9 however is what I posted. Notice under Summary, the first two lines are fine, but the 3,4,5 lines are a mess. 6 is again fine.
Your photo 'screenshots' are unclear & unnecessary (+ complicated for you) in the light of you having various proper screenshot programs available to do the job. See comment 2 Q3. > I have no idea whatsoever what the attachment pointer [2] is comment 8 is. > It is certainly not mine, and is not anything I have ever seen. But it seems that you posted it. In addition to which, you have not answered my simple questions comment 2: 1. Please post the console output of: $ inxi -MSGxx 2. You cite Google Chrome; is the problem tied to that? When you see it, can you simultaneously open a different browser to show exactly the same URL/screen? And see what shows. > Usually this clears up almost immediately, > but occasionally it does not, unless I move around the window If it is this easy to get rid of, it is a small annoyance. P.S. Thank you tmb for your intervention; and w unruh for putting it to the test.
Sorry, inxi sliped past my attention System: Host: crawler Kernel: 5.14.14-desktop-1.mga8 x86_64 bits: 64 compiler: gcc v: 10.3.0 Console: tty 2 wm: kwin_x11 DM: LightDM, SDDM Distro: Mageia 8 mga8 Machine: Type: Laptop System: Dell product: XPS 13 9310 v: N/A serial: 70JX2J3 Chassis: type: 10 serial: 70JX2J3 Mobo: Dell model: 07P9Y7 v: A00 serial: /70JX2J3/CNCMK0019O02B3/ UEFI: Dell v: 3.1.0 date: 07/22/2021 Graphics: Device-1: Intel TigerLake-LP GT2 [Iris Xe Graphics] vendor: Dell driver: i915 v: kernel bus ID: 0000:00:02.0 chip ID: 8086:9a49 Device-2: Realtek Integrated_Webcam_HD type: USB driver: uvcvideo bus ID: 3-6:3 chip ID: 0bda:5517 most Display: server: Mageia X.org 1.20.12 compositor: kwin_x11 driver: intel,v4l resolution: 1920x1200 s-dpi: 96 OpenGL: renderer: Mesa Intel Xe Graphics (TGL GT2) v: 4.6 Mesa 21.2.4 direct render: Yes ----------------------------------------------- Yes, it is a relatively small annoyance. It does seem to almost always disappear almost immediately Very occasionally it lasts longer but again, if I change focus, it clears up. But it should not do that. The new kernel did seem to help, not solve but make is better. Re the photos rather than the screenshots, as I said, it is too evanescent to get a screenshot of. Re Chrome-- that is where I have seen it most blatantly but it also affects Konsole text sometimes, (which is what I mainly use). Re the Devaid Hodgins screenshot from 2011, I have never looked at that or had any reason to download it to my machine, so again, I have no idea where it came from. Ie, I cannot see how I could have posted it. It is like if you suddenly discovered that you posted my water bill. You don't know me. How would you have gotten my water bill. Ie, there is also the possibiity that Mageia's bugzilla has a bug in it. But yes, there is also the possibility that I did something weird to post it. Anyway, no great harm except in that it confused some readers.
Bugzilla converts the word attachment followed by a number into a hyper link automatically, just like it does for the word bug followed by a number. By referencing the second attachment as attachment 2 [details], it becomes a hyper link to the second attachment ever posted to Mageia's bugzilla.
Ah, that explains where that attachment came from. I did call it attachment two (but using 2) meaning the second attachment I had made for this problem. I would call that a bug in bugzilla, but might well have trouble convincing others of that.
It is documented feature. :-) https://bugzilla.readthedocs.io/en/5.0/using/tips.html?highlight=hyperlink The triggers are: bug # bugs #,# comment # bug # comment# attachment # as well as any properly formatted url or email address
Thanks Dave for sorting the attachement link misunderstanding :) (In reply to w unruh from comment #14) > Re the photos rather than the screenshots, as I said, it is too evanescent > to get a screenshot of. Cant you just press printscreen button when you see the problem, and save directly? (crop the image etc if you want afterwards) At least Spectacle on Plasma capture whole screen instantly before it shows up.
I have tried that, but the prtscr button does not seem to do anything. Is spectacle supposed to be running constantly? Is it supposed to be tied to the prtscr button somehow? And there seems to be no documentation of spectacle and no starting of the service.
The spectacle workaround is described in Mageia 8 errata. https://wiki.mageia.org/en/Mageia_8_Errata#Plasma
They were all already marked as Default. The problem seems to be that spectacle is not running at all. There seems to be an /lib/systemd/user/app-org.kde.spectacle.service service but as far as I can tell is never run. Hitting fn-Shift-PrtSc produces nothing in Pictures. Ie, the setup of spectacle in Mageia 8 seems to leave a lot to be desired. But that would be the subject of another bug report. Just mentioning it here as an excuse for not being able to supply screenshots.
If the remedies described in Errata (for Bug 27183) does not work, it would be good to have them enhanced. First, are you sure you need Fn-Shift for the printscreen button? And can you launch Spectacle directly? Quickest is Alt+F2, start typing "spect" and hit enter You could also try to define another key combo for Spectacle.
The correct key combo for this keyboard layout from this laptop is either PrtSc or Fn+PrtSc
But doesn't spectacle need to be running for this to work? I have tried Prtsc or fn+PrtSc and neither did anything except type a ~ if a konsole window was active. Ie, the key was not being intercepted by the non-running spectacle
(In reply to w unruh from comment #24) > But doesn't spectacle need to be running for this to work? No. > I have tried Prtsc or fn+PrtSc and neither did anything except type a ~ if a > konsole window was active. Ie, the key was not being intercepted by the > non-running spectacle Have you tried per the errata, to delete/rename the mentioned file? Or launch Spectacle, and set other keys to trig it (in Spectacle the second left button at bottom.) and see if that works? Another test for Spectacle is to create another ruser login to see if it works with completely fresh Plasma settings that way.
I still have not gotten PrtScr to work, depite trying the various suggestions. The problem exists most reliably on google chrome windows, but I have also had it occur on a Konsole window.so it is not solely a Chrome problem. It can be "gotten around" by forcing a rewrite on the window, but it is sure still an annoyance.
It seems to happen if another window is opened covering part of the window in question. There is (sometimes) then a mess (displaced lines) on the visible parts of the window in question.
This problem is continuing to occur, now on two different machines. One is a laptop, the other a desktop system. lspci Desktop 00:02.0 VGA compatible controller: Intel Corporation CoffeeLake-S GT2 [UHD Graphics 630] Dell xps 9310 laptop 0000:00:02.0 VGA compatible controller: Intel Corporation TigerLake-LP GT2 [Iris Xe Graphics] (rev 01) It is still sporadic. I primarily see it on Google chrome, but also on Konsole terminal at times. Mageia 8 on both systems, uname -a Desktop Linux tunnel.physics.ubc.ca 5.15.55-desktop-2.mga8 #1 SMP Fri Jul 15 20:28:24 UTC 2022 x86_64 x86_64 x86_64 GNU/Linux Dell Laptop Linux crawler.physics.ubc.ca 5.15.35-desktop-2.mga8 #1 SMP Fri Apr 22 20:40:06 UTC 2022 x86_64 x86_64 x86_64 GNU/Linux
Created attachment 13382 [details] screen grab indicating the problem Here is a screengrab I managed to get -- the text and pictures on Googel Chrome are a mess. I seems to be a function of the Audion Volume window open to the right. Note that sometimes the mess is in a Konsole window rather than in a Chrome window, so it is not limited to Chrome.
Bill, is this still an issue? Have you found any work-around, like changing to another desktop, then back? We may have to put it in ERRATA for M9.
No I have not found a work around. It happens on a number of machines, including my laptop (a Dell xps13) all of which use intel graphics.
Sorry to have left you. > all of which use intel graphics and the same driver (i91) ? Note that Mageia 9 is newly out, and it will be necessary to know whether your problem still happens. I say 'your' problem, because Intel graphics are widespread, and we do not seem to have similar complaints from other users. > It happens on a number of machines Are all these machines under your care, or independant? I wonder whether there is some special configuration that you invoke unwittingly. Of all the example attachments, it is the last (comment 29) which best/clearly shows the complaint.
Closing moribund; can be reopened.
Status: NEEDINFO => RESOLVEDResolution: (none) => OLD
It is still a problem. Yesterday half the screen looked like bad Chinese charaters (it was the display of an English file) on my laptop. One problem is that it is not stable. Almost anything I do and the screen redraws and the artifact disappears. I have also not been alble to get prtscr to work. No idea why, but I really cannot spare the time to hnt for how to fix that. I can do a screen crab with xv, but whenever I try the redraw of the XV screens also redraws the rest and thus I cannot grab a screenshot of the problem. It seems to alwas occur when there is another window open on top of part of the window with the problem-- often something to do with chrome (eg the horribly ugly 1/4 inch black bordered messages chrome has suddenly taken to displaying). Dell XPS-13 9310 0000:00:02.0 VGA compatible controller: Intel Corporation TigerLake-LP GT2 [Iris Xe Graphics] (rev 01) Uptodate MGA8 (Since Mga9 came out I have been travelling and cannot afford to have problems with my laptop, or spend the time needed to update everything.)
Status: RESOLVED => REOPENEDResolution: OLD => (none)
Created attachment 14214 [details] picture of mangles screen Here is another photo (ScrPrt does not work and besides would probably destry the mangling)of a screen with mangled text in lowest widnow in display. Toward bottom, scan lines displaced about 10 pixels Mageia8 updated,
The attachment just has this URL: htp://www.theory.physics.ubc.ca/tmp/screen-mangled-letters.jpg which even if corrected does not lead to the image. Try attaching the image directly - unless it does not show anything more than the attachments from comment 1, comment 3, which do display. If kernel updates do not help with this, we have no choice to put it in ERRATA. Just 8 for now, possibly 9 if that changes nothing.
Keywords: (none) => FOR_ERRATA8Summary: Random display lines get shifted right. Dell XPS13 9310 => Random display lines get shifted right. Intel Coffee|TigerLake .... GT2 graphics
Oops. Sorry. It should have read http://www.theory.physics.ubc.ca/screen-mangled-letters.jpg ), and told me to enter the url which I did. Unfortunately not without a couple of typos. I have tested the above to make sure it works.
Created attachment 14216 [details] Correction to URL for "Picture of mangled screen"
To me this seem to rare to be mentioned in Errata, but i could be convinced. There are other artifacts seen on other systems in some cases. I see some occasionally in firefox when resuming from suspend. Nvidia driver. Another laptop i had a while ago had other artifacts problem resuming from suspend. Also, Mageia 8 is EOL about now... It would be interesting to see if your problem is seen using mga9.
As I have said I have seen it as well on other machines I have )all intel graphics based). The XPS13 is the only one I have access to (and use regualarly) right now.. So it is not as rare as it might seem (mind you they all have the same machine in front of the keyboard but it is hard to imagine how bugs in that machine would manage to display themselves on the computer -:) Yes, after new year I hope to update to Mga9. Note that this has been there since Mga 7 with all of the changes in kernels since then.
Well, entered in Errata. https://wiki.mageia.org/en/Mageia_8_Errata#Intel I wish you (and Intel) a happy new year then :) - and please feedback regarding results on mga9.
Keywords: FOR_ERRATA8 => IN_ERRATA8Status comment: (none) => User to test in mga9