Graphics Card Fan Suddenly Stopped While Using Computer

[Dylan Taylor]

Binary package hint: xserver-xorg-video-ati

While I was using my computer, doing perfectly normal things, such as reading news articles in Firefox, my graphics card's fan suddenly completely stopped, causing the GPU to overheat. At first I thought that the fan was slowing down because there wasn't anything particularly GPU intensive going on, and system load was below normal, however, like 10 minutes after, the screen turned completely black, and the computer locked up. This almost certainly because the GPU overheated. The first thing I did was tried Alt-SysRq-K to reboot the X server. That did nothing, so I tried Alt-SysRq-REISUB, and the system did NOT reboot. I'm assuming there may have been a kernel panic. Therefore, I held in the power button on my computer until it turned off. Shortly after that, I hit the power button again to turn the computer back on, but immediately after pressing the power button, the system fans started spinning at maximum speed for a split second and the system immediately shut off. No BIOS POST information was displayed on the screen or anything. So, I assumed the system shut off because it overheated and it was trying to protect the hardware. I let it sit for a couple of minutes, and then I managed to get it to turn on. Something worth noting here is that the GPU fan worked perfectly fine after a full system reboot, making me believe this is some sort of driver problem. The first time I tried to boot it, Ubuntu froze before being able to load the splash screen, so I assumed it may either be because the GPU was still overheated and not working properly, or some file system corruption. I tried Alt-SysRq-B to reboot it, but surprisingly that did nothing, so I turned it off with the power button again. I then booted into recovery mode, and thankfully, I managed to get to the recovery options. I then proceeded to run the options "dpkg, clean, and resume", but when it finished loading services, I was in a text mode login instead of GDM, so I rebooted using Alt-SysRq-REISUB. I then booted Ubuntu with the normal boot method, and after running a quick fsck, I was able to login and post this bug report on the same computer.

ProblemType: Bug
DistroRelease: Ubuntu 10.04
Package: xserver-xorg-video-ati 1:6.12.192-2ubuntu2
ProcVersionSignature: Ubuntu 2.6.32-19.28-generic 2.6.32.10+drm33.1
Uname: Linux 2.6.32-19-generic x86_64
Architecture: amd64
Date: Sat Apr 3 15:40:04 2010
DkmsStatus:
 virtualbox-ose, 3.1.4, 2.6.32-17-generic, x86_64: installed
 virtualbox-ose, 3.1.4, 2.6.32-16-generic, x86_64: installed
 virtualbox-ose, 3.1.4, 2.6.32-19-generic, x86_64: installed
InstallationMedia: Ubuntu 9.10 "Karmic Koala" - Release amd64 (20091027)
MachineType: Shuttle Inc SG33
ProcCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.32-19-generic root=UUID=f75382cd-2436-4b42-96ef-0c1f545f6fa5 ro quiet splash
ProcEnviron:
 PATH=(custom, user)
 LANG=en_US.utf8
 SHELL=/bin/bash
SourcePackage: xserver-xorg-video-ati
dmi.bios.date: 04/25/2008
dmi.bios.vendor: Phoenix Technologies, LTD
dmi.bios.version: 6.00 PG
dmi.board.name: FG33
dmi.board.vendor: Shuttle Inc
dmi.board.version: V10
dmi.chassis.type: 3
dmi.chassis.vendor: Shuttle Inc
dmi.chassis.version: G5
dmi.modalias: dmi:bvnPhoenixTechnologies,LTD:bvr6.00PG:bd04/25/2008:svnShuttleInc:pnSG33:pvrV10:rvnShuttleInc:rnFG33:rvrV10:cvnShuttleInc:ct3:cvrG5:
dmi.product.name: SG33
dmi.product.version: V10
dmi.sys.vendor: Shuttle Inc
system:
 distro: Ubuntu
 codename: lucid
 architecture: x86_64
 kernel: 2.6.32-19-generic

[lspci]
01:00.0 VGA compatible controller [0300]: ATI Technologies Inc Radeon X1950 GT [1002:7288] (rev 9a)
     Subsystem: PC Partner Limited Device [174b:e190]

[Dylan Taylor (dylanmtaylor)]

I think this bug should be critical, because even though my hardware ended up being fine, if I or someone else was doing something GPU intensive like playing games or watching movies on their computer, this could _seriously_ mess up your hardware...

[Dylan Taylor (dylanmtaylor)]

I just checked -- _all_ of my packages are up to date, and they were at the time of this bug occurring. This might be a regression from an older version of one of the packages on my system.

[Bryce Harrington (bryce)]

Forwarding this bug from Ubuntu reporter Dylan Taylor:
http://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-ati/+bug/554783

[Problem]
Fan on card stopped during usage, leading to system overheat and lockups on reboot.

Is this a hardware error or software? How do we tell the difference?

[Original Description]
While I was using my computer, doing perfectly normal things, such as reading news articles in Firefox, my graphics card's fan suddenly completely stopped, causing the GPU to overheat. At first I thought that the fan was slowing down because there wasn't anything particularly GPU intensive going on, and system load was below normal, however, like 10 minutes after, the screen turned completely black, and the computer locked up. This almost certainly because the GPU overheated. The first thing I did was tried Alt-SysRq-K to reboot the X server. That did nothing, so I tried Alt-SysRq-REISUB, and the system did NOT reboot. I'm assuming there may have been a kernel panic. Therefore, I held in the power button on my computer until it turned off. Shortly after that, I hit the power button again to turn the computer back on, but immediately after pressing the power button, the system fans started spinning at maximum speed for a split second and the system immediately shut off. No BIOS POST information was displayed on the screen or anything. So, I assumed the system shut off because it overheated and it was trying to protect the hardware. I let it sit for a couple of minutes, and then I managed to get it to turn on. Something worth noting here is that the GPU fan worked perfectly fine after a full system reboot, making me believe this is some sort of driver problem. The first time I tried to boot it, Ubuntu froze before being able to load the splash screen, so I assumed it may either be because the GPU was still overheated and not working properly, or some file system corruption. I tried Alt-SysRq-B to reboot it, but surprisingly that did nothing, so I turned it off with the power button again. I then booted into recovery mode, and thankfully, I managed to get to the recovery options. I then proceeded to run the options "dpkg, clean, and resume", but when it finished loading services, I was in a text mode login instead of GDM, so I rebooted using Alt-SysRq-REISUB. I then booted Ubuntu with the normal boot method, and after running a quick fsck, I was able to login and post this bug report on the same computer.

DistroRelease: Ubuntu 10.04
Package: xserver-xorg-video-ati 1:6.12.192-2ubuntu2
ProcVersionSignature: Ubuntu 2.6.32-19.28-generic 2.6.32.10+drm33.1
Uname: Linux 2.6.32-19-generic x86_64
Architecture: amd64
Date: Sat Apr 3 15:40:04 2010
DkmsStatus:
 virtualbox-ose, 3.1.4, 2.6.32-17-generic, x86_64: installed
 virtualbox-ose, 3.1.4, 2.6.32-16-generic, x86_64: installed
 virtualbox-ose, 3.1.4, 2.6.32-19-generic, x86_64: installed
InstallationMedia: Ubuntu 9.10 "Karmic Koala" - Release amd64 (20091027)
MachineType: Shuttle Inc SG33
ProcCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.32-19-generic root=UUID=f75382cd-2436-4b42-96ef-0c1f545f6fa5 ro quiet splash
ProcEnviron:
 PATH=(custom, user)
 LANG=en_US.utf8SourcePackage: xserver-xorg-vi...

[Bryce Harrington (bryce)]

Created an attachment (id=34688)
XorgLog.txt

[Bryce Harrington (bryce)]

Created an attachment (id=34689)
CurrentDmesg.txt

[Bryce Harrington (bryce)]

Created an attachment (id=34690)
BootDmesg.txt

[agd5f (agd5f)]

The driver doesn't currently mess with the fan setup; it uses the default profile setup by the vbios at post. The vbios sets up a default card specific fan profile when the card is posted that is adequate for the default clocks. Did you flash the vbios on your card? Doing so could cause an incompatible fan profile to be loaded. If not, you may have failing hardware. Make sure the fan is clean and spins freely.

[Bryce Harrington (bryce)]

My initial guess would be a hardware problem (I had almost identical symptoms on a 5xx card a few months ago, that went away once I swapped it out with a new card), but just in case it isn't, I can send the bug report upstream first and see what they say.

[Bryce Harrington (bryce)]

Dylan Taylor - I've forwarded this bug upstream to http://bugs.freedesktop.org/show_bug.cgi?id=27475 - please subscribe yourself to this bug, in case they need further information or wish you to test something. Thanks ahead of time!

[Bryce Harrington (bryce)]

Fwiw, with my card I saw this set of symptoms happen intermittently for a few months, with increasing frequency over time. With several reboots, blowing canned air on it, etc. I could get the gpu fan to come on and work properly. But eventually it just stopped entirely and swapping out the card was the only option. I'd used the card pretty heavily for several years so it wasn't a surprise that the fan would die eventually. Pity though; aside from the fan that card had worked really well for me.

Anyway, again please follow up on the upstream bug report on this issue. I'm curious if your issue was indeed a hardware issue like mine or if there is something more to be concerned with in the software code.

[Bryce Harrington (bryce)]

Dylan, ok upstream's comments seem to confirm my own suspicions that this is most likely a hardware issue. Unfortunately if my experience with bad gpu fans is to go by, it's going to get worse over time... I'd recommend reporting it to your hardware vendor.

If you feel that it really is a software rather than hardware error, please work with upstream on the issue, and if it comes to a solution that looks relevant to include in Ubuntu please feel free to reopen this bug report.

[Bryce Harrington (bryce)]

I'm closing out the bug as likely hardware breakage. There's not been further comment from the reporter to confirm one way or the other, but he can reopen the bug if there's more to investigate here.