Bug 103753

Summary: Scrolling artifacts in Firefox on GTX 1060 6GB
Product: xorg Reporter: Artem S. Tashkinov <aros>
Component: Driver/nouveauAssignee: Nouveau Project <nouveau>
Status: RESOLVED DUPLICATE QA Contact: Xorg Project Team <xorg-team>
Severity: major    
Priority: medium    
Version: git   
Hardware: x86-64 (AMD64)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
XFCE panel screenshot
none
Xorg.0.log
none
Firefox none

Description Artem S. Tashkinov 2017-11-15 10:04:01 UTC
Created attachment 135483 [details]
XFCE panel screenshot

This is clearly a regression because kernel 4.12.x worked just fine here and all 4.13.x releases exhibit this problem.

I'm running XFCE under Fedora 27 (recently updated from 26 which also exhibited this bug) without a compositing manager.

Upon logon XFCE panel icons are always broken in one way or another. If you move any window above the panel then everything gets back to normal. Yet when you run `xfce4-panel -r` glitches return.

Also this bug manifests itself in Firefox when you're scrolling. Text often gets strange artifacts.

Here are relevant kernel messages:

nouveau 0000:01:00.0: NVIDIA GP106 (136000a1)
nouveau 0000:01:00.0: bios: version 86.06.0e.00.28
nouveau 0000:01:00.0: fb: 6144 MiB GDDR5
nouveau 0000:01:00.0: DRM: VRAM: 6144 MiB
nouveau 0000:01:00.0: DRM: GART: 1048576 MiB
nouveau 0000:01:00.0: DRM: BIT table 'A' not found
nouveau 0000:01:00.0: DRM: BIT table 'L' not found
nouveau 0000:01:00.0: DRM: TMDS table version 2.0
nouveau 0000:01:00.0: DRM: DCB version 4.1
nouveau 0000:01:00.0: DRM: DCB outp 00: 01000f42 04620030
nouveau 0000:01:00.0: DRM: DCB outp 01: 04811f96 04600020
nouveau 0000:01:00.0: DRM: DCB outp 02: 04011f92 04620020
nouveau 0000:01:00.0: DRM: DCB outp 03: 04822f86 04600010
nouveau 0000:01:00.0: DRM: DCB outp 04: 04022f82 04620010
nouveau 0000:01:00.0: DRM: DCB outp 06: 02033f62 04620010
nouveau 0000:01:00.0: DRM: DCB outp 07: 02844f76 04600020
nouveau 0000:01:00.0: DRM: DCB outp 08: 02044f72 00020020
nouveau 0000:01:00.0: DRM: DCB conn 00: 00001031
nouveau 0000:01:00.0: DRM: DCB conn 01: 02000146
nouveau 0000:01:00.0: DRM: DCB conn 02: 01000246
nouveau 0000:01:00.0: DRM: DCB conn 03: 00010361
nouveau 0000:01:00.0: DRM: DCB conn 04: 00020446
nouveau 0000:01:00.0: DRM: Pointer to flat panel table invalid
[drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
[drm] Driver supports precise vblank timestamp query.
nouveau 0000:01:00.0: DRM: MM: using COPY for buffer copies
nouveau 0000:01:00.0: DRM: allocated 1920x1080 fb: 0x60000, bo ffff908246f26800
fbcon: nouveaufb (fb0) is primary device
nouveau 0000:01:00.0: fb0: nouveaufb frame buffer device
[drm] Initialized nouveau 1.3.1 20120801 for 0000:01:00.0 on minor 0
Comment 1 Ilia Mirkin 2017-11-15 14:14:51 UTC
Please confirm that you're using the nouveau ddx and not the modesetting ddx? [Include your xorg log.]

Was acceleration working for you with the 4.12.x kernel? I don't really see what would have changed in the kernel to affect this one way or the other.

We're now using the full BAR2 amount (bar/gf100: fix access to upper half of BAR2)... Could be somehow different on GP10x?
Comment 2 Artem S. Tashkinov 2017-11-16 05:23:39 UTC
Created attachment 135509 [details]
Xorg.0.log

(In reply to Ilia Mirkin from comment #1)
> Please confirm that you're using the nouveau ddx and not the modesetting
> ddx? [Include your xorg log.]

I've no idea what I'm using. I have a default Fedora 27 installation without Xorg.conf file or any configuration for kernel modules.

> 
> Was acceleration working for you with the 4.12.x kernel? I don't really see
> what would have changed in the kernel to affect this one way or the other.

I'm not sure what you're talking about. I don't use any applications which use 3D acceleration and I've no idea if 2D acceleration is in use.

> 
> We're now using the full BAR2 amount (bar/gf100: fix access to upper half of
> BAR2)... Could be somehow different on GP10x?

Er, no clue ;-)
Comment 3 Ilia Mirkin 2017-11-16 05:36:30 UTC
[    42.818] (II) modeset(0): using drv /dev/dri/card0

You're using xf86-video-modesetting. Please switch to xf86-video-nouveau. This would normally be the default, but I believe Fedora does something to break that (on purpose). You should reach out to their support forums if you need help unbreaking it. I believe an xorg.conf that forces nouveau to load will do the trick, but I'm not sure.

BTW - 2d, 3d - it's all acceleration. And it either works or doesn't. If it doesn't, then you're fully cpu-rendered for everything.
Comment 4 Artem S. Tashkinov 2017-11-16 06:04:17 UTC
(In reply to Ilia Mirkin from comment #3)
> [    42.818] (II) modeset(0): using drv /dev/dri/card0
> 

/etc/X11/xorg.conf.d/01-nouveau.conf                                                                                                                                Section "Device"
        Identifier      "Videocard0"
        Driver          "nouveau"
EndSection

has solved the issue.
Comment 5 Artem S. Tashkinov 2017-11-16 06:13:00 UTC
I've file a bug report at Fedora's bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1513848
Comment 6 Artem S. Tashkinov 2017-11-17 09:30:29 UTC
Font artifacts while scrolling in Firefox are still present even with correct settings.
Comment 7 Artem S. Tashkinov 2017-11-17 09:38:18 UTC
Created attachment 135542 [details]
Firefox

I've noticed that it happens only with italic web fonts.

If you open the same web page for the second time the problem disappears.

If you cover the browser window and put it back into foreground the problem disappears.
Comment 8 caguduzexi 2018-01-29 14:08:19 UTC
I wont recommend using/keeping the GP106 (GTX 1060). It cant ever run with free software: https://www.theregister.co.uk/2015/04/15/nvidia_gtx_900_linux_driver_roadbloack/
https://www.phoronix.com/scan.php?page=news_item&px=Nouveau-XDC2017
https://www.phoronix.com/scan.php?page=news_item&px=Nouveau-XDC2016-NVIDIA

Sell this crappy GP102 card away and go away from nvidia. Nvidia died with the 780ti card. Its the last end-user card that can be used normaly. Everything else is in some countries even a legal problem. Because the manufacturer (nvidia) blocks the users from beeing able to boot the software they want on THEIR hardware - happyly illegal in some countries. Hopefully some layer would sue the heck out of nvidia so that they would have to release the private signing key or close their doors.
Blocking the freedom of the users on such way should not be accepted by anyone.
Comment 9 Martin Peres 2018-01-29 17:33:27 UTC
(In reply to caguduzexi from comment #8)
> I wont recommend using/keeping the GP106 (GTX 1060). It cant ever run with
> free software:
> https://www.theregister.co.uk/2015/04/15/
> nvidia_gtx_900_linux_driver_roadbloack/
> https://www.phoronix.com/scan.php?page=news_item&px=Nouveau-XDC2017
> https://www.phoronix.com/scan.php?page=news_item&px=Nouveau-XDC2016-NVIDIA
> 
> Sell this crappy GP102 card away and go away from nvidia. Nvidia died with
> the 780ti card. Its the last end-user card that can be used normaly.
> Everything else is in some countries even a legal problem. Because the
> manufacturer (nvidia) blocks the users from beeing able to boot the software
> they want on THEIR hardware - happyly illegal in some countries. Hopefully
> some layer would sue the heck out of nvidia so that they would have to
> release the private signing key or close their doors.
> Blocking the freedom of the users on such way should not be accepted by
> anyone.

User banned
Comment 10 Artem S. Tashkinov 2018-01-30 12:04:27 UTC
This is reproducible under NVIDIA binary drivers as well.

In fact it's an X.org server bug.

*** This bug has been marked as a duplicate of bug 25497 ***

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.