Bug#971343: Animated background/wallpaper (changing over time) causes system freeze on Nouveau
Em qua., 30 de set. de 2020 às 14:04, Simon McVittie escreveu: > On Wed, 30 Sep 2020 at 12:31:06 -0300, Leandro Cunha wrote: > > I did this test and there was no crash, but see the attached > > screenshot to see how the wallpaper. > > Hmm, interesting failure mode. I think there was a bug fixed recently > where the animated background draws one frame at the wrong brightness > before adjusting correctly to the time of day - this might mean your Shell > is actually showing half of one frame and half of another or something? > > This made me think that there could be a problem with opengl that may be occurring in more specific versions of the Kernel. After conducting studies in other distros and Debian, I realized that the same version of nouveau was working normally. I have also studied a number of bug reports associated with other problems that have to do with wallpapers in Gnome. However, I don't have many details to be able to state this until now, but I keep studying. https://www.omgubuntu.co.uk/2015/05/animated-wallpaper-adds-live-backgrounds-to-linux-distros > > In the bug attached see from 10:29 when I started testing with Kernel > > 4.19, until 10:44 that appears in the screen capture that shows both > > wallpapers side by side. Normally the interface stopped responding, > > but what I do is run su -c "cat /var/log/syslog > log" on tty to > > check. > > To confirm, are you saying that the log entries from Sep 30 10:29 to > 10:44 are with the 4.19 kernel from Debian 10, but with everything else > (including Mesa and GNOME) taken from unstable? > Yes. > > > What calls attention that the screen is black, but the interface does > > not freeze with this version of Kernel, being that I ran the tests on > > Xorg where I had the best performance so far. > > This looks interesting: > > > Sep 30 10:29:36 debian-pc gnome-shell[1707]: nouveau: kernel rejected > pushbuf: Não foi possível alocar memória > > Sep 30 10:29:36 debian-pc kernel: [ 1520.979894] nouveau :01:00.0: > Xorg[1558]: fail ttm_validate > > Sep 30 10:29:36 debian-pc kernel: [ 1520.979904] nouveau :01:00.0: > Xorg[1558]: validating bo list > > Sep 30 10:29:36 debian-pc kernel: [ 1520.979915] nouveau :01:00.0: > Xorg[1558]: validate: -12 > > Sep 30 10:29:36 debian-pc gnome-shell[1707]: nouveau: ch7: krec 0 pushes > 1 bufs 13 relocs 0 > > This also looks maybe bad: > > > Sep 30 10:44:10 debian-pc kernel: [ 2395.486040] nouveau :01:00.0: > fb: trapped read at 00473a4200 on channel 3 [1fa2 Xorg[1558]] engine 00 > [PGRAPH] client 0a [TEXTURE] subclient 00 [] reason 0002 > [PAGE_NOT_PRESENT] > > Sep 30 10:44:10 debian-pc kernel: [ 2395.493226] nouveau :01:00.0: > fb: trapped read at 0047913d00 on channel 3 [1fa2 Xorg[1558]] engine 00 > [PGRAPH] client 0a [TEXTURE] subclient 00 [] reason 0002 > [PAGE_NOT_PRESENT] > > which I think indicates a bug in the nouveau kernel-side driver? But I > might be wrong. > > > Sep 30 10:30:59 debian-pc gnome-shell[1707]: JS ERROR: TypeError: > this._dialog is null#012_onFocusChanged > @resource:///org/gnome/shell/ui/closeDialog.js:135:9#012reset > @resource:///org/gnome/shell/ui/viewSelector.js:452:26#012show > @resource:///org/gnome/shell/ui/viewSelector.js:275:14#012_animateVisible > @resource:///org/gnome/shell/ui/overview.js:580:27#012show > @resource:///org/gnome/shell/ui/overview.js:566:14#012_onShowAppsButtonToggled@ > /usr/share/gnome-shell/extensions/ > dash-to-d...@micxgx.gmail.com/docking.js:1908:31 > > It's probably best if you disable extensions like dash-to-dock while > testing this. > > smcv > After the release of version 5.9 of Linux, there was a deterioration in performance, causing frequent crashes with nouveau, which made me stay on the buster and not use the latest versions of gnome and also of kernel. I sent a message to a Gnome contributor to help study the bug and I am waiting for a response from him.
Bug#971343: Animated background/wallpaper (changing over time) causes system freeze on Nouveau
On Wed, 30 Sep 2020 at 12:31:06 -0300, Leandro Cunha wrote: > I did this test and there was no crash, but see the attached > screenshot to see how the wallpaper. Hmm, interesting failure mode. I think there was a bug fixed recently where the animated background draws one frame at the wrong brightness before adjusting correctly to the time of day - this might mean your Shell is actually showing half of one frame and half of another or something? > In the bug attached see from 10:29 when I started testing with Kernel > 4.19, until 10:44 that appears in the screen capture that shows both > wallpapers side by side. Normally the interface stopped responding, > but what I do is run su -c "cat /var/log/syslog > log" on tty to > check. To confirm, are you saying that the log entries from Sep 30 10:29 to 10:44 are with the 4.19 kernel from Debian 10, but with everything else (including Mesa and GNOME) taken from unstable? > What calls attention that the screen is black, but the interface does > not freeze with this version of Kernel, being that I ran the tests on > Xorg where I had the best performance so far. This looks interesting: > Sep 30 10:29:36 debian-pc gnome-shell[1707]: nouveau: kernel rejected > pushbuf: Não foi possível alocar memória > Sep 30 10:29:36 debian-pc kernel: [ 1520.979894] nouveau :01:00.0: > Xorg[1558]: fail ttm_validate > Sep 30 10:29:36 debian-pc kernel: [ 1520.979904] nouveau :01:00.0: > Xorg[1558]: validating bo list > Sep 30 10:29:36 debian-pc kernel: [ 1520.979915] nouveau :01:00.0: > Xorg[1558]: validate: -12 > Sep 30 10:29:36 debian-pc gnome-shell[1707]: nouveau: ch7: krec 0 pushes 1 > bufs 13 relocs 0 This also looks maybe bad: > Sep 30 10:44:10 debian-pc kernel: [ 2395.486040] nouveau :01:00.0: fb: > trapped read at 00473a4200 on channel 3 [1fa2 Xorg[1558]] engine 00 > [PGRAPH] client 0a [TEXTURE] subclient 00 [] reason 0002 > [PAGE_NOT_PRESENT] > Sep 30 10:44:10 debian-pc kernel: [ 2395.493226] nouveau :01:00.0: fb: > trapped read at 0047913d00 on channel 3 [1fa2 Xorg[1558]] engine 00 > [PGRAPH] client 0a [TEXTURE] subclient 00 [] reason 0002 > [PAGE_NOT_PRESENT] which I think indicates a bug in the nouveau kernel-side driver? But I might be wrong. > Sep 30 10:30:59 debian-pc gnome-shell[1707]: JS ERROR: TypeError: > this._dialog is > null#012_onFocusChanged@resource:///org/gnome/shell/ui/closeDialog.js:135:9#012reset@resource:///org/gnome/shell/ui/viewSelector.js:452:26#012show@resource:///org/gnome/shell/ui/viewSelector.js:275:14#012_animateVisible@resource:///org/gnome/shell/ui/overview.js:580:27#012show@resource:///org/gnome/shell/ui/overview.js:566:14#012_onShowAppsButtonToggled@/usr/share/gnome-shell/extensions/dash-to-d...@micxgx.gmail.com/docking.js:1908:31 It's probably best if you disable extensions like dash-to-dock while testing this. smcv
Bug#971343: Animated background/wallpaper (changing over time) causes system freeze on Nouveau
On Wed, 30 Sep 2020 at 11:11:04 -0300, Leandro Cunha wrote: > It has a greater effect on use. I think it is more appropriate > to lower the severity to important. Thanks, I agree. We usually use 'important' for crash/freeze bugs that seem hardware-dependent or otherwise difficult to reproduce. smcv
Bug#971343: Animated background/wallpaper (changing over time) causes system freeze on Nouveau
Analyzing at what point the problem was solved without the whole system stopping, which occurred specifically with Wayland and not with Xorg. It has a greater effect on use. I think it is more appropriate to lower the severity to important. https://www.debian.org/Bugs/Developer#severities
Bug#971343: Animated background/wallpaper (changing over time) causes system freeze on Nouveau
On Tue, 29 Sep 2020 at 22:29:29 -0300, Leandro Cunha wrote: > > * Is this a new installation, or have you been using this hardware with > > Linux for a while? If you've been using it previously, have you had > > other graphics- or freeze-related issues with it? > > Only in version 3.36.3 onwards, if I'm not mistaken. But it is the first time. Presumably you've also been upgrading other packages, like the kernel and graphics drivers. Please try installing the kernel from Debian 10 (linux-image-4.19.0-11-amd64) and booting into that - you can have multiple kernels available on the same system, so this is a lot easier than switching between versions of user-space packages, which is why I'm suggesting to try it first. Then see whether you still get this same bug with the older kernel. If you still see the bug, another thing to try would be to confirm which user-space packages cause the regression. The best way to do this is probably to set up a temporary parallel installation of Debian 10 on the same hardware (maybe on an external USB disk if you can get a suitable one), and gradually upgrade it towards current testing: * reduce the size of a partition enough to fit a new partition for Debian 10, or use a USB disk (you'll need at least 10G for an installation with just gnome-session and its dependencies, 20G is safer) * install Debian 10 with at least the gnome-session package (if you have enough space, using the GNOME desktop task in the installer will be easiest; if you don't have much space, just install gnome-session) * upgrade to 3.34.4-1 from: deb [check-valid-until=no] https://snapshot.debian.org/archive/debian/20200226T030551Z sid main * try to reproduce the bug * upgrade to 3.36.2-1 from: deb [check-valid-until=no] https://snapshot.debian.org/archive/debian/20200501T154658Z sid main * try to reproduce the bug * upgrade to 3.36.3-1 from: deb [check-valid-until=no] https://snapshot.debian.org/archive/debian/20200604T025809Z sid main * try to reproduce the bug * if none of those show the bug, try upgrading to the latest version from testing If your guess about 3.36.3 is right, then this must have been triggered by something that changed between 3.36.2 and 3.36.3, and that's a much smaller change to examine. When doing that, it would be helpful if you can upgrade only gnome-shell and the packages that it requires - try to avoid upgrading the kernel or the Mesa graphics drivers, to reduce the number of things you're varying. If none of that makes the bug appear, try the same process with the Mesa graphics drivers: libgles2, libgl1, libglx-mesa0, libgl1-mesa-dri, libegl1 and libegl-mesa0. The apt logs in /var/log/apt will tell us exactly what was upgraded in the apt transaction that triggered the regression. > > What do you mean by "restored the original system settings"? > > I delete the configuration file. Which one? > > What do you mean by "it would only start if [...]"? GNOME would only > > start if you did that? The laptop would only boot up if you did that? > > Something else? > > The answer is yes. Somehow you can't log in to a section. OK, this might give us more useful information: if you can't log in, but the computer doesn't freeze and you are able to work around it without rebooting, then we might be able to get a log with more information. Please could you try this: * make a note of the time: time before reproducing bug * reproduce the bug * recover by rebooting * make a note of the time again: time after rebooting * try to log in (which you say will fail) * delete the configuration file * note the time again: time before login * log in successfully * note the time again when login has finished: time after login * send the log from that whole period, and the times that you noted That will help to correlate log messages with what you observe. > > I don't think we can necessarily treat GNOME freezing on > > more-than-10-year-old hardware as release-critical > > Gnome works well on this hardware that was on the market for years, > devices from 2011 and 2012 were still sold with this hardware is the > case with mine. Obviously we'd like it to continue to work, but hardware-specific failures are difficult (especially for people who don't have the hardware in front of them). If we treated every hardware-specific failure as a release-critical bug that prevented the next Debian release from happening, then we'd never be able to make a new release, because there's always at least one device that regresses. I've configured my (Intel-based) laptop to have the animated background to see whether I can reproduce anything similar, but it's had this configuration for 24 hours now without problems. smcv
Bug#971343: Animated background/wallpaper (changing over time) causes system freeze on Nouveau
> > Control: found -1 3.36.6-1 > Control: found -1 3.38.0-2 > > This is on a GT218M GPU, which is a Tesla 2.0 device from 2007. > https://www.notebookcheck.net/NVIDIA-GeForce-310M.22439.0.html suggests > that this is comparable in performance to Intel integrated graphics from > the Ivybridge (2012) generation, but probably with a much higher power > consumption. > > I don't think we can necessarily treat GNOME freezing on > more-than-10-year-old hardware as release-critical, particularly since > there's a workaround (not using the animated background). Gnome works well on this hardware that was on the market for years, devices from 2011 and 2012 were still sold with this hardware is the case with mine. Who still likes it and who knows even the rest of the laptop. The problem may be occurring for several reasons that may not be related to the age of the hardware. This problem only occurs in the latest versions of Gnome. > Some other questions I should have asked: > > * Are you using any GNOME Shell extensions? (If yes, please try > disabling them all and see whether the problem persists.) The problem still persists. > * Is this a new installation, or have you been using this hardware with > Linux for a while? If you've been using it previously, have you had > other graphics- or freeze-related issues with it? Only in version 3.36.3 onwards, if I'm not mistaken. But it is the first time. > * Has the animated/changing background worked in earlier versions of > Debian and/or GNOME, or is 3.36 the oldest version you've tried? All previous versions work normally. > * Are you able to do an installation of the Debian 10 'buster' stable > release on this hardware, if you haven't already tried that? That would > give us a baseline for whether this is a situation that has been there > for a while. In the Buster (stable current) version it works normally. > On Tue, 29 Sep 2020 at 16:48:21 -0300, Leandro Cunha wrote: > > Graphics: > > Device-1: NVIDIA GT218M [GeForce 310M] driver: nouveau v: kernel > > Device-2: Suyin type: USB driver: uvcvideo > > Is this Suyin USB device an input (camera), or an output (display)? > Is it part of the laptop, or a removable device? I have no idea what this is. > If it's a removable output device, please check whether this issue still > occurs with the USB device disconnected and just the NVIDIA graphics > device, to keep things as simple as possible. There is only one mouse and one USB network adapter. > Also, if you have external screens attached, please check whether it > still occurs with just the laptop's built-in screen, again to try to > keep things as simple as possible. > > > Sep 29 10:50:17 debian-pc kernel: [ 60.247471] nouveau :01:00.0: > > firmware: failed to load nouveau/nva8_fuc084 (-2) > > Sep 29 10:50:17 debian-pc kernel: [ 60.247475] firmware_class: See > > https://wiki.debian.org/Firmware for information about missing firmware > > Sep 29 10:50:17 debian-pc kernel: [ 60.247478] nouveau :01:00.0: > > Direct firmware load for nouveau/nva8_fuc084 failed with error -2 > > Sep 29 10:50:17 debian-pc kernel: [ 60.247492] nouveau :01:00.0: > > firmware: failed to load nouveau/nva8_fuc084d (-2) > > Sep 29 10:50:17 debian-pc kernel: [ 60.247494] nouveau :01:00.0: > > Direct firmware load for nouveau/nva8_fuc084d failed with error -2 > > Sep 29 10:50:17 debian-pc kernel: [ 60.247497] nouveau :01:00.0: > > msvld: unable to load firmware data > > Sep 29 10:50:17 debian-pc kernel: [ 60.247499] nouveau :01:00.0: > > msvld: init failed, -19 > > This firmware blob is not available in Debian (it seems we cannot legally > distribute it, even in non-free, although it can be extracted from > proprietary NVIDIA drivers) but apparently it's for 2D video encoder/decoder > acceleration (VDPAU) rather than anything GNOME Shell would need, so this > warning is probably harmless? Even without it it still occurs. > > Sep 29 10:54:53 debian-pc gnome-shell[1633]: 0xa000f6: frame_complete > > callback never occurred for frame 3729 > > Sep 29 10:54:54 debian-pc gnome-shell[1633]: 0xa000f6: frame_complete > > callback never occurred for frame 3735 > > Sep 29 10:55:01 debian-pc gnome-shell[1633]: 0xa000f6: frame_complete > > callback never occurred for frame 3766 > > (and lots more) > > This is maybe interesting. If the Shell isn't reliably getting frame > completion notifications back from the driver or hardaware, that might be > related to the display freezing. > > > With Wayland it completely crashed and got no response to force the > > boot and had to use the on and off button. In Xorg I was able to > > restart with keyboard via tty, but after locking it would only start > > if I restored the original system settings and this is necessary for > > both. Now I am with wayland. > > What do you mean by "restored the original system settings"? I delete the configuration file. > What do you mean by "it would only start if [
Bug#971343: Animated background/wallpaper (changing over time) causes system freeze on Nouveau
Control: found -1 3.36.6-1 Control: found -1 3.38.0-2 This is on a GT218M GPU, which is a Tesla 2.0 device from 2007. https://www.notebookcheck.net/NVIDIA-GeForce-310M.22439.0.html suggests that this is comparable in performance to Intel integrated graphics from the Ivybridge (2012) generation, but probably with a much higher power consumption. I don't think we can necessarily treat GNOME freezing on more-than-10-year-old hardware as release-critical, particularly since there's a workaround (not using the animated background). Some other questions I should have asked: * Are you using any GNOME Shell extensions? (If yes, please try disabling them all and see whether the problem persists.) * Is this a new installation, or have you been using this hardware with Linux for a while? If you've been using it previously, have you had other graphics- or freeze-related issues with it? * Has the animated/changing background worked in earlier versions of Debian and/or GNOME, or is 3.36 the oldest version you've tried? * Are you able to do an installation of the Debian 10 'buster' stable release on this hardware, if you haven't already tried that? That would give us a baseline for whether this is a situation that has been there for a while. On Tue, 29 Sep 2020 at 16:48:21 -0300, Leandro Cunha wrote: > Graphics: > Device-1: NVIDIA GT218M [GeForce 310M] driver: nouveau v: kernel > Device-2: Suyin type: USB driver: uvcvideo Is this Suyin USB device an input (camera), or an output (display)? Is it part of the laptop, or a removable device? If it's a removable output device, please check whether this issue still occurs with the USB device disconnected and just the NVIDIA graphics device, to keep things as simple as possible. Also, if you have external screens attached, please check whether it still occurs with just the laptop's built-in screen, again to try to keep things as simple as possible. > Sep 29 10:50:17 debian-pc kernel: [ 60.247471] nouveau :01:00.0: > firmware: failed to load nouveau/nva8_fuc084 (-2) > Sep 29 10:50:17 debian-pc kernel: [ 60.247475] firmware_class: See > https://wiki.debian.org/Firmware for information about missing firmware > Sep 29 10:50:17 debian-pc kernel: [ 60.247478] nouveau :01:00.0: Direct > firmware load for nouveau/nva8_fuc084 failed with error -2 > Sep 29 10:50:17 debian-pc kernel: [ 60.247492] nouveau :01:00.0: > firmware: failed to load nouveau/nva8_fuc084d (-2) > Sep 29 10:50:17 debian-pc kernel: [ 60.247494] nouveau :01:00.0: Direct > firmware load for nouveau/nva8_fuc084d failed with error -2 > Sep 29 10:50:17 debian-pc kernel: [ 60.247497] nouveau :01:00.0: msvld: > unable to load firmware data > Sep 29 10:50:17 debian-pc kernel: [ 60.247499] nouveau :01:00.0: msvld: > init failed, -19 This firmware blob is not available in Debian (it seems we cannot legally distribute it, even in non-free, although it can be extracted from proprietary NVIDIA drivers) but apparently it's for 2D video encoder/decoder acceleration (VDPAU) rather than anything GNOME Shell would need, so this warning is probably harmless? > Sep 29 10:54:53 debian-pc gnome-shell[1633]: 0xa000f6: frame_complete > callback never occurred for frame 3729 > Sep 29 10:54:54 debian-pc gnome-shell[1633]: 0xa000f6: frame_complete > callback never occurred for frame 3735 > Sep 29 10:55:01 debian-pc gnome-shell[1633]: 0xa000f6: frame_complete > callback never occurred for frame 3766 (and lots more) This is maybe interesting. If the Shell isn't reliably getting frame completion notifications back from the driver or hardaware, that might be related to the display freezing. > With Wayland it completely crashed and got no response to force the > boot and had to use the on and off button. In Xorg I was able to > restart with keyboard via tty, but after locking it would only start > if I restored the original system settings and this is necessary for > both. Now I am with wayland. What do you mean by "restored the original system settings"? What do you mean by "it would only start if [...]"? GNOME would only start if you did that? The laptop would only boot up if you did that? Something else? > > If Ctrl+Alt+Delete doesn't work, does the system respond to > > the "magic sysrq key" sequences, in particular AltGr+SysRq+o > > (immediate power off) and AltGr+SysRq+b (immediate reboot)? (See > > https://en.wikipedia.org/wiki/Magic_SysRq_key for more details) > > I manage to force the shutdown that way. OK, so the kernel is still working to at least some extent, otherwise those key sequences wouldn't work. The problem could be in (from highest to lowest level) GNOME Shell; libmutter or some other library it uses; the Mesa user-space graphics driver; or the Nouveau kernel-side graphics driver. > Log attached, after the problem occurs this log. You opened a gnome-terminal at 12:44 and the Nautilus file manager at 12:46, a block of zero bytes was written at
Bug#971343: Animated background/wallpaper (changing over time) causes system freeze on Nouveau
Control: reassign -1 gnome-shell Control: tags -1 + moreinfo bullseye sid Control: retitle -1 Animated background/wallpaper (changing over time) causes system freeze on Nouveau Retitling to make this show up in more searches. gnome-control-center just changes settings, and it is GNOME Shell that is actually responsible for setting the wallpaper. Please run "reportbug --template gnome-shell" and send the result to this bug address so that we have details of the version you are using, including its dependencies. On Tue, 29 Sep 2020 at 00:46:25 -0300, Leandro Cunha wrote: > When changing the wallpaper to the one that changes throughout the day in both > sections xorg and wayland through the gnome control center gnome-control-center does not have separate sections for Xorg and Wayland settings. Do you mean this affects both the Xorg and Wayland session types? > the system freezes > using nouveau and it is necessary to force the system shutdown How quickly does this happen? (Within seconds/minutes/hours/days?) If your keyboard has a Caps Lock or Num Lock light, can you still toggle it by pressing the appropriate key after the freeze has occurred? (I ask because it gives an indication of how many layers of the system have frozen) Similarly, after the freeze occurs, does the system respond to repeated presses of Ctrl+Alt+Delete? (Pressing it more than 7 times in 2 seconds might cause systemd to do a controlled reboot, even if no change is visible on the screen.) If Ctrl+Alt+Delete doesn't work, does the system respond to the "magic sysrq key" sequences, in particular AltGr+SysRq+o (immediate power off) and AltGr+SysRq+b (immediate reboot)? (See https://en.wikipedia.org/wiki/Magic_SysRq_key for more details) After forcing the system off and rebooting, are there errors or warnings logged in /var/log/syslog or the systemd journal? To make journal messages persist across reboots, you would have to enable the persistent mode of the journal by creating the directory /var/log/journal (I can't remember whether this is done by default yet). What hardware are you running on? (In particular, is it a laptop or desktop, and which NVIDIA graphics device does it have?) > As the problem > occurs in other distributions and I found out by doing tests I put the > upstream > tag. Do you have any references for bug reports in other distributions about this? Or, if you tested other distributions yourself, which ones? > I haven't tested the current version 3.38 Please provide the information above first, and then upgrade GNOME packages (particularly gnome-shell and gnome-control-center) to their versions from unstable and try again. Thanks, smcv
