nouveau "GPU lockup" and gnome-shell/X crash [HAVE WORKAROUND]

mitch_feaster · 2014-11-18 18:23:26

For the past week or so I've been seeing crashes in nouveau which lead to GPU lockup and eventually a hung system. It seems like out of nowhere my entire system will freeze and after 15 or 20 seconds I get booted out to a VT login. Then, about 1 minute after that the VT console becomes unresponsive and I have to do a hard reset.

I've pasted the tail end of my system logs for the past three crashes below.

The problem seems to have started after an update on November 12th which touched the following packages:

kmod,hwids,libsystemd,systemd,colord,gtk3,aisleriot,boost-libs,libpng,lib32-gcc-libs,gcc-libs-multilib,elfutils,mesa-dri,mesa,mesa-libgl,qt4,akonadi,at-spi2-core,avahi,miniupnpc,protobuf,bitcoin-qt,bluez,bluez-libs,bluez-utils,boost,btrfs-progs,ca-certificates-mozilla,gstreamer,gst-plugins-base-libs,curl,gnutls,gst-plugins-bad,gst-plugins-base,gst-plugins-good,cheese,libjpeg-turbo,poppler,cups-filters,dhcpcd,eog,nss,epiphany,libgdata,evolution-data-server,libpst,evolution,evolution-ews,fakeroot,file-roller,firefox,four-in-a-row,gcc-multilib,gdb,gtksourceview3,gedit,gettext,gnome-chess,vte-common,vte3,gnome-color-manager,gnome-contacts,gnome-devel-docs,gnome-dictionary,gnome-disk-utility,gnome-klotski,libffi,libtracker-sparql,poppler-glib,tracker,gnome-photos,gnome-robots,gnome-sudoku,gnome-terminal,go,grub,gst-libav,gvfs,hdparm,iagno,ipython,ipython2,jansson,ldns,lib32-elfutils,libltdl,lib32-libltdl,lib32-libxml2,lib32-mesa-dri,lib32-mesa,lib32-mesa-libgl,libaio,libedit,libgit2-glib,libreoffice-still-en-US,libreoffice-still-common,libreoffice-still-base,libreoffice-still-calc,libreoffice-still-draw,libreoffice-still-gnome,libreoffice-still-impress,libreoffice-still-kde4,libreoffice-still-math,libreoffice-still-postgresql-connector,libreoffice-still-writer,libtool,libwbclient,logrotate,man-db,mercurial,nodejs,ocaml-findlib,opencv,orca,pacman-mirrorlist,procps-ng,parallel,perl-uri,phonon-qt4,pinentry,poppler-qt4,python-pygments,python2-logilab-common,python2-pygments,python2-pylint,rygel,tdb,tevent,smbclient,samba,screen,source-highlight,systemd-sysvcompat,tcl,tk,totem,usbutils,vinagre,vino,wine,wine_gecko,xorg-server-common,xorg-server,android-sdk-build-tools,go-mtpfs-git,google-chrome,google-talkplugin,openchange

Looks like there a few packages related to graphics so that's a little suspicious... Are there any known regressions with nouveau or any of these other packages at the moment?

Thanks!

Nov 13 13:47:19 mitchelh-linux kernel: nouveau E[  PGRAPH][0000:01:00.0] TRAP ch 5 [0x001fbd4000 gnome-shell[983]]
Nov 13 13:47:19 mitchelh-linux kernel: nouveau E[  PGRAPH][0000:01:00.0] GPC0/PROP trap: RT_LINEAR_MISMATCH
Nov 13 13:47:19 mitchelh-linux kernel: nouveau E[  PGRAPH][0000:01:00.0] x = 0, y = 0, format = 3b, storage type = fe
Nov 13 13:47:19 mitchelh-linux kernel: nouveau E[  PGRAPH][0000:01:00.0] GPC0/TPC0/TEX: 0x80000009
Nov 13 13:47:19 mitchelh-linux kernel: nouveau E[  PGRAPH][0000:01:00.0] GPC0/TPC1/TEX: 0x80000009
Nov 13 13:48:18 mitchelh-linux kernel: nouveau E[   PFIFO][0000:01:00.0] read fault at 0x0014508000 [PAGE_NOT_PRESENT] from PGRAPH/GPC0/TEX on channel 0x001fbd4000 [gnome-shell[983]]
Nov 13 13:48:18 mitchelh-linux kernel: nouveau E[   PFIFO][0000:01:00.0] PGRAPH engine fault on channel 5, recovering...
Nov 13 13:49:35 mitchelh-linux kernel: nouveau E[Xorg.bin[664]] failed to idle channel 0xcccc0000 [Xorg.bin[664]]
Nov 13 13:49:50 mitchelh-linux kernel: nouveau E[Xorg.bin[664]] failed to idle channel 0xcccc0000 [Xorg.bin[664]]
Nov 13 13:49:50 mitchelh-linux kernel: nouveau E[   PFIFO][0000:01:00.0] read fault at 0x000001b000 [PAGE_NOT_PRESENT] from PFIFO/BAR_READ on channel 0x001fcc0000 [unknown]
Nov 13 13:50:06 mitchelh-linux kernel: nouveau E[gnome-shell[983]] failed to idle channel 0xcccc0000 [gnome-shell[983]]
Nov 13 13:50:18 mitchelh-linux kernel: nouveau E[     DRM] GPU lockup - switching to software fbcon
Nov 13 13:50:21 mitchelh-linux kernel: nouveau E[gnome-shell[983]] failed to idle channel 0xcccc0000 [gnome-shell[983]]
Nov 13 13:50:36 mitchelh-linux kernel: nouveau E[chrome[1423]] failed to idle channel 0xcccc0000 [chrome[1423]]
Nov 13 13:50:50 mitchelh-linux kernel: usb 1-1.6.3: USB disconnect, device number 7
Nov 13 13:50:51 mitchelh-linux kernel: nouveau E[chrome[1423]] failed to idle channel 0xcccc0000 [chrome[1423]]
Nov 13 13:50:52 mitchelh-linux kernel: usb 1-1.6.3: new low-speed USB device number 10 using ehci-pci
Nov 13 13:50:52 mitchelh-linux kernel: input: PixArt HP USB Optical Mouse as /devices/pci0000:00/0000:00:1a.0/usb1/1-1/1-1.6/1-1.6.3/1-1.6.3:1.0/0003:03F0:134A.0005/input/input21
Nov 13 13:50:52 mitchelh-linux kernel: hid-generic 0003:03F0:134A.0005: input,hidraw3: USB HID v1.11 Mouse [PixArt HP USB Optical Mouse] on usb-0000:00:1a.0-1.6.3/input0
Nov 13 13:51:08 mitchelh-linux kernel: nfsd: last server has exited, flushing export cache

Nov 15 11:25:48 mitchelh-linux kernel: nouveau E[   PFIFO][0000:01:00.0] read fault at 0x000ceea000 [PAGE_NOT_PRESENT] from PGRAPH/GPC0/UNK04 on channel 0x001fbd4000 [gnome-shell[24711]]
Nov 15 11:25:48 mitchelh-linux kernel: nouveau E[   PFIFO][0000:01:00.0] PGRAPH engine fault on channel 5, recovering...
Nov 15 11:25:48 mitchelh-linux kernel: nouveau E[  PGRAPH][0000:01:00.0] TRAP ch 5 [0x001fbd4000 gnome-shell[24711]]
Nov 15 11:25:48 mitchelh-linux kernel: nouveau E[  PGRAPH][0000:01:00.0] GPC0/TPC0/TEX: 0x80000049
Nov 15 11:25:48 mitchelh-linux kernel: nouveau E[  PGRAPH][0000:01:00.0] GPC0/TPC1/TEX: 0x80000049
Nov 15 11:27:05 mitchelh-linux kernel: nouveau E[Xorg.bin[24030]] failed to idle channel 0xcccc0000 [Xorg.bin[24030]]
Nov 15 11:27:20 mitchelh-linux kernel: nouveau E[Xorg.bin[24030]] failed to idle channel 0xcccc0000 [Xorg.bin[24030]]
Nov 15 11:27:20 mitchelh-linux kernel: nouveau E[   PFIFO][0000:01:00.0] read fault at 0x000001b000 [PAGE_NOT_PRESENT] from PFIFO/BAR_READ on channel 0x001fcc0000 [unknown]
Nov 15 11:27:21 mitchelh-linux kernel: dleyna-server-s[30665]: segfault at 8 ip 00007f01c34956b9 sp 00007fff2018b028 error 4 in libdleyna-server-1.0.so.1.0.3[7f01c3485000+22000]
Nov 15 11:27:21 mitchelh-linux kernel: traps: dleyna-renderer[30656] general protection ip:7fc80ca14158 sp:7fffa1f9a2e0 error:0 in libdleyna-renderer-1.0.so.1.0.3[7fc80ca0a000+17000]
Nov 15 11:27:35 mitchelh-linux kernel: nouveau E[gnome-shell[24711]] failed to idle channel 0xcccc0000 [gnome-shell[24711]]
Nov 15 11:27:35 mitchelh-linux kernel: nouveau E[     DRM] GPU lockup - switching to software fbcon
Nov 15 11:27:50 mitchelh-linux kernel: nouveau E[gnome-shell[24711]] failed to idle channel 0xcccc0000 [gnome-shell[24711]]
Nov 15 11:28:05 mitchelh-linux kernel: nouveau E[chrome[24962]] failed to idle channel 0xcccc0000 [chrome[24962]]
Nov 15 11:28:20 mitchelh-linux kernel: usb 1-1.6.3: USB disconnect, device number 7
Nov 15 11:28:20 mitchelh-linux kernel: nouveau E[chrome[24962]] failed to idle channel 0xcccc0000 [chrome[24962]]
Nov 15 11:28:21 mitchelh-linux kernel: usb 1-1.6.3: new low-speed USB device number 10 using ehci-pci
Nov 15 11:28:22 mitchelh-linux kernel: input: PixArt HP USB Optical Mouse as /devices/pci0000:00/0000:00:1a.0/usb1/1-1/1-1.6/1-1.6.3/1-1.6.3:1.0/0003:03F0:134A.0005/input/input21
Nov 15 11:28:22 mitchelh-linux kernel: hid-generic 0003:03F0:134A.0005: input,hidraw3: USB HID v1.11 Mouse [PixArt HP USB Optical Mouse] on usb-0000:00:1a.0-1.6.3/input0
Nov 15 11:29:38 mitchelh-linux kernel: nouveau E[gnome-session-c[30516]] failed to idle channel 0xcccc0000 [gnome-session-c[30516]]
Nov 15 11:29:53 mitchelh-linux kernel: nouveau E[gnome-session-c[30516]] failed to idle channel 0xcccc0000 [gnome-session-c[30516]]
Nov 15 11:30:08 mitchelh-linux kernel: nouveau E[gnome-session-c[30516]] failed to idle channel 0xcccc0000 [gnome-session-c[30516]]
Nov 15 11:30:23 mitchelh-linux kernel: nouveau E[gnome-session-c[30516]] failed to idle channel 0xcccc0000 [gnome-session-c[30516]]
Nov 15 11:30:38 mitchelh-linux kernel: nouveau E[gnome-session-c[30516]] failed to idle channel 0xcccc0000 [gnome-session-c[30516]]
Nov 15 11:30:53 mitchelh-linux kernel: nouveau E[gnome-session-c[30516]] failed to idle channel 0xcccc0000 [gnome-session-c[30516]]

Nov 18 09:45:05 mitchelh-linux kernel: nouveau E[  PGRAPH][0000:01:00.0] TRAP ch 5 [0x001fbd4000 gnome-shell[980]]
Nov 18 09:45:05 mitchelh-linux kernel: nouveau E[  PGRAPH][0000:01:00.0] GPC0/PROP trap: RT_LINEAR_MISMATCH
Nov 18 09:45:05 mitchelh-linux kernel: nouveau E[  PGRAPH][0000:01:00.0] x = 0, y = 0, format = 3b, storage type = fe
Nov 18 09:45:05 mitchelh-linux kernel: nouveau E[  PGRAPH][0000:01:00.0] GPC0/TPC0/TEX: 0x80000009
Nov 18 09:45:05 mitchelh-linux kernel: nouveau E[  PGRAPH][0000:01:00.0] GPC0/TPC1/TEX: 0x80000009
Nov 18 09:47:25 mitchelh-linux kernel: nouveau E[   PFIFO][0000:01:00.0] read fault at 0x0012c6a000 [PAGE_NOT_PRESENT] from PGRAPH/GPC0/TEX on channel 0x001fbd4000 [gnome-shell[980]]
Nov 18 09:47:25 mitchelh-linux kernel: nouveau E[   PFIFO][0000:01:00.0] PGRAPH engine fault on channel 5, recovering...
Nov 18 09:47:27 mitchelh-linux kernel: nouveau E[Xorg.bin[647]] nv50cal_space: -16
Nov 18 09:47:28 mitchelh-linux kernel: nouveau E[Xorg.bin[647]] nv50cal_space: -16
Nov 18 09:47:28 mitchelh-linux kernel: nouveau E[Xorg.bin[647]] nv50cal_space: -16
[...] (repeated about 50 times)
Nov 18 09:47:56 mitchelh-linux kernel: nouveau E[Xorg.bin[647]] nv50cal_space: -16
Nov 18 09:47:56 mitchelh-linux kernel: nouveau E[Xorg.bin[647]] nv50cal_space: -16
Nov 18 09:48:12 mitchelh-linux kernel: nouveau E[Xorg.bin[647]] failed to idle channel 0xcccc0000 [Xorg.bin[647]]
Nov 18 09:48:12 mitchelh-linux kernel: nouveau E[Xorg.bin[647]] failed to idle channel 0xcccc0000 [Xorg.bin[647]]
Nov 18 09:48:12 mitchelh-linux kernel: nouveau E[   PFIFO][0000:01:00.0] read fault at 0x000001b000 [PAGE_NOT_PRESENT] from PFIFO/BAR_READ on channel 0x001fcc0000 [unknown]
Nov 18 09:48:16 mitchelh-linux kernel: dconf worker[918]: segfault at 1 ip 00007f5f80e3bc9a sp 00007f5f766e64b8 error 4 in libc-2.20.so[7f5f80dba000+199000]
Nov 18 09:48:16 mitchelh-linux kernel: dleyna-server-s[21242]: segfault at 8 ip 00007f4d23fd56b9 sp 00007fffc64f18e8 error 4 in libdleyna-server-1.0.so.1.0.3[7f4d23fc5000+22000]
Nov 18 09:48:23 mitchelh-linux kernel: nouveau E[     DRM] GPU lockup - switching to software fbcon
Nov 18 09:48:28 mitchelh-linux kernel: nouveau E[chrome[1383]] failed to idle channel 0xcccc0000 [chrome[1383]]
Nov 18 09:48:43 mitchelh-linux kernel: nouveau E[chrome[1383]] failed to idle channel 0xcccc0000 [chrome[1383]]
Nov 18 09:48:58 mitchelh-linux kernel: nouveau E[gnome-shell[980]] failed to idle channel 0xcccc0000 [gnome-shell[980]]
Nov 18 09:49:13 mitchelh-linux kernel: usb 1-1.6.3: USB disconnect, device number 7
Nov 18 09:49:13 mitchelh-linux kernel: nouveau E[gnome-shell[980]] failed to idle channel 0xcccc0000 [gnome-shell[980]]
Nov 18 09:49:14 mitchelh-linux kernel: usb 1-1.6.3: new low-speed USB device number 9 using ehci-pci
Nov 18 09:49:14 mitchelh-linux kernel: input: PixArt HP USB Optical Mouse as /devices/pci0000:00/0000:00:1a.0/usb1/1-1/1-1.6/1-1.6.3/1-1.6.3:1.0/0003:03F0:134A.0005/input/input21
Nov 18 09:49:14 mitchelh-linux kernel: hid-generic 0003:03F0:134A.0005: input,hidraw3: USB HID v1.11 Mouse [PixArt HP USB Optical Mouse] on usb-0000:00:1a.0-1.6.3/input0
Nov 18 09:49:32 mitchelh-linux kernel: nfsd: last server has exited, flushing export cache
Nov 18 09:49:32 mitchelh-linux systemd[1]: var-lib-nfs-rpc_pipefs.mount mount process exited, code=exited status=32
Nov 18 09:49:32 mitchelh-linux systemd[1]: Unit rpcbind.service entered failed state.
Nov 18 09:49:32 mitchelh-linux systemd[1]: rpcbind.service failed.

Last edited by mitch_feaster (2014-11-26 21:21:19)

jakub · 2014-11-19 17:08:00

Recently I had two crashes on fresh arch installation where for the first time I used nouveau driver instead of nVidia proprietary one. I couldn't track down anything in logs, but the GPU driver is the only thing I can blame as it's the only thing that changed since I reinstalled arch. I you want to look at it here's my thread: https://bbs.archlinux.org/viewtopic.php?id=189948.

* first crash happend when I clicked thunerbird on bar to maximize it (then I got system freezed for a while and restart)
* second crash happend just after running starting my laptop and trying to move firefox window

I'm not sure how much this was a GPU-related job but for sure it was a little bit - both crashes happend while GPU had to render / move windows on screen.

Last edited by jakub (2014-11-19 17:11:15)

davidm · 2014-11-19 17:24:37

I'm using nouveau on a Quadro FX 570 and it is working okay here. I even enabled VDPAU which works most of the time.

Which GPU do you have?

Does it still happen if you try the linux-lts kernel? It's pretty easy to install that for testing. Basically just install 'linux-lts' and then rerun your Grub configure and then reboot and select it. You might even try 'linux-mainline' in the AUR as well to test if it is kernel related or more the fault of something like 'mesa-dri'.

Last edited by davidm (2014-11-19 17:25:19)

mitch_feaster · 2014-11-19 19:29:01

davidm wrote:

I'm using nouveau on a Quadro FX 570 and it is working okay here. I even enabled VDPAU which works most of the time.
Which GPU do you have?

I have:

01:00.0 VGA compatible controller: NVIDIA Corporation GF108GL [Quadro 600] (rev a1)

davidm wrote:

Does it still happen if you try the linux-lts kernel? It's pretty easy to install that for testing. Basically just install 'linux-lts' and then rerun your Grub configure and then reboot and select it. You might even try 'linux-mainline' in the AUR as well to test if it is kernel related or more the fault of something like 'mesa-dri'.

I'm a tad nervous switching around kernel versions since I use btrfs, which is still fairly volatile at the moment... If I'm feeling adventurous I'll give it a shot.

davidm · 2014-11-19 19:42:13

mitch_feaster wrote:

davidm wrote:
I'm using nouveau on a Quadro FX 570 and it is working okay here. I even enabled VDPAU which works most of the time.
Which GPU do you have?
I have:
01:00.0 VGA compatible controller: NVIDIA Corporation GF108GL [Quadro 600] (rev a1)
davidm wrote:
Does it still happen if you try the linux-lts kernel? It's pretty easy to install that for testing. Basically just install 'linux-lts' and then rerun your Grub configure and then reboot and select it. You might even try 'linux-mainline' in the AUR as well to test if it is kernel related or more the fault of something like 'mesa-dri'.
I'm a tad nervous switching around kernel versions since I use btrfs, which is still fairly volatile at the moment... If I'm feeling adventurous I'll give it a shot.

If it helps I use a multi-device (4 disk) btrfs raid1 filesystem too. linux-lts was pretty stable and it's on 3.16.x which does not have the nasty bugs 3.17.1 had (until fixed). I last tried it a week ago and it was fine then. I had that problem with the 'btrfs' hook and having to put it in as a module. I haven't tried linux-lts since doing that but it's probably safe.

On the current core kernel 3.17.3-1-ARCH (and previous ones) I do get a lot of errors like this:

Nov 18 22:27:30  kernel: nouveau E[  PGRAPH][0000:01:00.0] TRAP_PROP - TP 0 - RT_FAULT - Address 00203ed080
Nov 18 22:27:30  kernel: nouveau E[  PGRAPH][0000:01:00.0] TRAP_PROP - TP 0 - e0c: 00000000, e18: 00000000, e1c: 00340020, e20: 00001100, e24: 00030000
Nov 18 22:27:30  kernel: nouveau E[  PGRAPH][0000:01:00.0] ch 8 [0x000f355000 mplayer[27662]] subc 3 class 0x8297 mthd 0x1b0c data 0x1000f010
Nov 18 22:27:30  kernel: nouveau E[     PFB][0000:01:00.0] trapped write at 0x00203ed080 on channel 0x0000f355 [mplayer[27662]] PGRAPH/PROP/RT0 reason: PAGE_NOT_PRESENT
Nov 18 22:27:30  kernel: nouveau E[  PGRAPH][0000:01:00.0] TRAP_PROP - TP 0 - RT_FAULT - Address 00203f1080
Nov 18 22:27:30  kernel: nouveau E[  PGRAPH][0000:01:00.0] TRAP_PROP - TP 0 - e0c: 00000000, e18: 00000000, e1c: 00340020, e20: 00001100, e24: 00030000
Nov 18 22:27:30  kernel: nouveau E[  PGRAPH][0000:01:00.0] ch 8 [0x000f355000 mplayer[27662]] subc 3 class 0x8297 mthd 0x1b0c data 0x1000f010
Nov 18 22:27:30  kernel: nouveau E[     PFB][0000:01:00.0] trapped write at 0x0020482000 on channel 0x0000f355 [mplayer[27662]] PGRAPH/PROP/RT0 reason: PAGE_NOT_PRESENT
Nov 18 22:27:30  kernel: nouveau E[  PGRAPH][0000:01:00.0] TRAP_PROP - TP 0 - RT_FAULT - Address 0040f80000
Nov 18 22:27:30  kernel: nouveau E[  PGRAPH][0000:01:00.0] TRAP_PROP - TP 0 - e0c: 00000000, e18: 00000000, e1c: 09c80000, e20: 00001100, e24: 00030000
Nov 18 22:27:30  kernel: nouveau E[  PGRAPH][0000:01:00.0] ch 8 [0x000f355000 mplayer[27662]] subc 3 class 0x8297 mthd 0x1b0c data 0x1000f010
Nov 18 22:27:30  kernel: nouveau E[     PFB][0000:01:00.0] trapped write at 0x0020461080 on channel 0x0000f355 [mplayer[27662]] PGRAPH/PROP/RT0 reason: PAGE_NOT_PRESENT

but no lockups and the only thing I ever notice wrong are occasional artifacts mostly when typing into forum text boxes but I am very satisfied with nouveau on my GPU (Quadro FX 570) which seems to be somewhat close to yours.

Last edited by davidm (2014-11-19 19:44:43)

mitch_feaster · 2014-11-19 21:19:14

davidm wrote:

mitch_feaster wrote:
davidm wrote:
I'm using nouveau on a Quadro FX 570 and it is working okay here. I even enabled VDPAU which works most of the time.
Which GPU do you have?
I have:
01:00.0 VGA compatible controller: NVIDIA Corporation GF108GL [Quadro 600] (rev a1)
davidm wrote:
Does it still happen if you try the linux-lts kernel? It's pretty easy to install that for testing. Basically just install 'linux-lts' and then rerun your Grub configure and then reboot and select it. You might even try 'linux-mainline' in the AUR as well to test if it is kernel related or more the fault of something like 'mesa-dri'.
I'm a tad nervous switching around kernel versions since I use btrfs, which is still fairly volatile at the moment... If I'm feeling adventurous I'll give it a shot.
If it helps I use a multi-device (4 disk) btrfs raid1 filesystem too. linux-lts was pretty stable and it's on 3.16.x which does not have the nasty bugs 3.17.1 had (until fixed). I last tried it a week ago and it was fine then. I had that problem with the 'btrfs' hook and having to put it in as a module. I haven't tried linux-lts since doing that but it's probably safe.

All right, I'm convinced . I'm now running linux-lts. Since this was showing up every 2 or 3 days I guess I'll be able to report back after a week with results.

davidm wrote:

On the current core kernel 3.17.3-1-ARCH (and previous ones) I do get a lot of errors like this:

Nov 18 22:27:30  kernel: nouveau E[  PGRAPH][0000:01:00.0] TRAP_PROP - TP 0 - RT_FAULT - Address 00203ed080
Nov 18 22:27:30  kernel: nouveau E[  PGRAPH][0000:01:00.0] TRAP_PROP - TP 0 - e0c: 00000000, e18: 00000000, e1c: 00340020, e20: 00001100, e24: 00030000
Nov 18 22:27:30  kernel: nouveau E[  PGRAPH][0000:01:00.0] ch 8 [0x000f355000 mplayer[27662]] subc 3 class 0x8297 mthd 0x1b0c data 0x1000f010
Nov 18 22:27:30  kernel: nouveau E[     PFB][0000:01:00.0] trapped write at 0x00203ed080 on channel 0x0000f355 [mplayer[27662]] PGRAPH/PROP/RT0 reason: PAGE_NOT_PRESENT
Nov 18 22:27:30  kernel: nouveau E[  PGRAPH][0000:01:00.0] TRAP_PROP - TP 0 - RT_FAULT - Address 00203f1080
Nov 18 22:27:30  kernel: nouveau E[  PGRAPH][0000:01:00.0] TRAP_PROP - TP 0 - e0c: 00000000, e18: 00000000, e1c: 00340020, e20: 00001100, e24: 00030000
Nov 18 22:27:30  kernel: nouveau E[  PGRAPH][0000:01:00.0] ch 8 [0x000f355000 mplayer[27662]] subc 3 class 0x8297 mthd 0x1b0c data 0x1000f010
Nov 18 22:27:30  kernel: nouveau E[     PFB][0000:01:00.0] trapped write at 0x0020482000 on channel 0x0000f355 [mplayer[27662]] PGRAPH/PROP/RT0 reason: PAGE_NOT_PRESENT
Nov 18 22:27:30  kernel: nouveau E[  PGRAPH][0000:01:00.0] TRAP_PROP - TP 0 - RT_FAULT - Address 0040f80000
Nov 18 22:27:30  kernel: nouveau E[  PGRAPH][0000:01:00.0] TRAP_PROP - TP 0 - e0c: 00000000, e18: 00000000, e1c: 09c80000, e20: 00001100, e24: 00030000
Nov 18 22:27:30  kernel: nouveau E[  PGRAPH][0000:01:00.0] ch 8 [0x000f355000 mplayer[27662]] subc 3 class 0x8297 mthd 0x1b0c data 0x1000f010
Nov 18 22:27:30  kernel: nouveau E[     PFB][0000:01:00.0] trapped write at 0x0020461080 on channel 0x0000f355 [mplayer[27662]] PGRAPH/PROP/RT0 reason: PAGE_NOT_PRESENT

but no lockups and the only thing I ever notice wrong are occasional artifacts mostly when typing into forum text boxes but I am very satisfied with nouveau on my GPU (Quadro FX 570) which seems to be somewhat close to yours.

Interesting. Yeah those look similar to some that I see near my crashes. Anyways, I'll post an update after a few days of linux-lts usage.

simark · 2014-11-22 15:09:18

I get something similar the original post, although for me it generally freezes after a couple of days. I have never seen it happening, I just come back to the computer, and the monitors don't show anything. I can ssh to the machine and use non-graphical programs without problems though.

Here is my dmesg output:
http://pastebin.com/eRhXNnFg

My GPU is:
01:00.0 VGA compatible controller: NVIDIA Corporation GF104 [GeForce GTX 460] (rev a1)

I am going to try the linux-lts kernel for a few days and report if it helps (although it's hard to confirm that something doesn't crash when the crashes are a few days apart).

mitch_feaster · 2014-11-26 21:19:24

So I moved back off of linux-lts due to issues with my disks coming up out-of-order. I think I've solved that problem now so I could technically try linux-lts again, but in the meantime I switched to the nvidia proprietary driver and am no longer seeing this issue. So:

   Known broken: [linux 3.17.3-1] + nouveau
  Known working: [linux 3.17.3-1] + nvidia
        Unknown: [linux-lts 3.14.24-1 (or similar)] + nouveau
Assumed working: [linux-lts 3.14.24-1 (or similar)] + nvidia

If someone could get that data point of whether [linux-lts + nouveau] is working that'd be great.

Last edited by mitch_feaster (2014-11-26 21:20:18)

Arch Linux

#1 2014-11-18 18:23:26

nouveau "GPU lockup" and gnome-shell/X crash [HAVE WORKAROUND]

#2 2014-11-19 17:08:00

Re: nouveau "GPU lockup" and gnome-shell/X crash [HAVE WORKAROUND]

#3 2014-11-19 17:24:37

Re: nouveau "GPU lockup" and gnome-shell/X crash [HAVE WORKAROUND]

#4 2014-11-19 19:29:01

Re: nouveau "GPU lockup" and gnome-shell/X crash [HAVE WORKAROUND]

#5 2014-11-19 19:42:13

Re: nouveau "GPU lockup" and gnome-shell/X crash [HAVE WORKAROUND]

#6 2014-11-19 21:19:14

Re: nouveau "GPU lockup" and gnome-shell/X crash [HAVE WORKAROUND]

#7 2014-11-22 15:09:18

Re: nouveau "GPU lockup" and gnome-shell/X crash [HAVE WORKAROUND]

#8 2014-11-26 21:19:24

Re: nouveau "GPU lockup" and gnome-shell/X crash [HAVE WORKAROUND]

Board footer