Fedora 12 freezes - Page 6

AlexDudko · #76 28th February 2010, 12:28 PM

Can confirm that my laptop faces the same problem of random freezes in Ubuntu Lucid alpha-3. And it doesn't if nomodeset is a kernel option.

eliashickman · #77 28th February 2010, 04:01 PM

That's very promising puntarenas. I can't wait for that to be added to the yum repos.

puntarenas · #78 1st March 2010, 09:02 AM

Quote:

Originally Posted by AlexDudko

Can confirm that my laptop faces the same problem of random freezes in Ubuntu Lucid alpha-3. And it doesn't if nomodeset is a kernel option.

In Fedora you can either add that to the boot options or recreate the initrd manually. Otherwise GDM wouldn't come up at all, but that was not what I meant with "random freezes" on my system.

Quote:

Originally Posted by eliashickman

That's very promising puntarenas. I can't wait for that to be added to the yum repos.

You may want to try this excellent Guide to build nvidia-195.36.08 rpm's (rpmfusion based) .

Edit: Okay, this drives me nuts. I'm sorry for all the confusion, but while Ubuntu Lucid Alpha3 seemed to be stable with Nvidia's beta driver (I still believe it is, have to try again), Fedora 12 with 195.36.08 still freezes now and then.

lzlu123 · #79 4th March 2010, 03:28 AM

I don't have nvidia video card but do experience random freeze, average once a day, even after I had several update. I run in the kernel 2.6.31.12-174.2.22.fc12.x86-64.

I never had any problem when I used FC10 and FC11. don' know why there are so much trouble with FC12

mawkus · #80 8th March 2010, 10:00 PM

Howdy,

Same problem here - using nouveau or nvidia drivers (also 195.36.08) the system consistently freezes after ~5mins to 1,5h of use. Usually SSH unavailable and all.

I'm using a GeForce 8500 GT and the problem has persisted since i upgraded from F10 to F12 a couple of weeks ago. F10 worked without any problems using proprietary nvidia drivers.

The only working solution so far has been using "nv" drivers, which have only crashed on me once after +1 week of uptime.

Cheers,
-Markus

puntarenas · #81 13th March 2010, 10:47 AM

Power Management crash with IRQ nobody cares

I did a fresh install of F13 Alpha (+updates) with 195.36.08 again. Guess what I found in /var/log/messages:

Code:

Mar 13 00:25:11 client01 pulseaudio[2992]: pid.c: Daemon already running.
Mar 13 00:25:12 client01 setroubleshoot: SELinux not enabled, setroubleshootd exiting...
Mar 13 00:25:12 client01 kernel: NVRM: Xid (0001:00): 6, PE0001
Mar 13 00:25:12 client01 kernel: NVRM: Xid (0001:00): 6, PE0001
Mar 13 00:25:15 client01 kernel: irq 16: nobody cared (try booting with the "irqpoll" option)
Mar 13 00:25:15 client01 kernel: Pid: 0, comm: swapper Tainted: P           2.6.33-1.fc13.i686 #1
Mar 13 00:25:15 client01 kernel: Call Trace:
Mar 13 00:25:15 client01 kernel: [<c0489e39>] __report_bad_irq+0x33/0x74
Mar 13 00:25:15 client01 kernel: [<c0489f74>] note_interrupt+0xfa/0x152
Mar 13 00:25:15 client01 kernel: [<c048a556>] handle_fasteoi_irq+0x8f/0xb2
Mar 13 00:25:15 client01 kernel: [<c0404ef0>] handle_irq+0x40/0x4c
Mar 13 00:25:15 client01 kernel: [<c040475f>] do_IRQ+0x46/0x9f
Mar 13 00:25:15 client01 kernel: [<c0403975>] common_interrupt+0x35/0x3c
Mar 13 00:25:15 client01 kernel: [<c0409745>] ? mwait_idle+0x68/0x78
Mar 13 00:25:15 client01 kernel: [<c04025be>] cpu_idle+0x9b/0xb5
Mar 13 00:25:15 client01 kernel: [<c07adcaa>] start_secondary+0x204/0x242
Mar 13 00:25:15 client01 kernel: handlers:
Mar 13 00:25:15 client01 kernel: [<c06adaca>] (usb_hcd_irq+0x0/0x8d)
Mar 13 00:25:15 client01 kernel: [<f8bb2801>] (nv_kern_isr+0x0/0x59 [nvidia])
Mar 13 00:25:15 client01 kernel: Disabling IRQ #16
Mar 13 00:25:20 client01 kernel: NVRM: Xid (0001:00): 16, Head 00000001 Count 00000000
Mar 13 00:25:21 client01 kernel: NVRM: Xid (0001:00): 16, Head 00000000 Count 00000003
Mar 13 00:25:27 client01 kernel: NVRM: Xid (0001:00): 8, Channel 0000007f

micca · #82 13th March 2010, 12:16 PM

Leaving only one prefLevel availible to powermizer (stopping it from changing powermode) keeps my card responding. If its card specific or otherwise - no clue.

My Gfx card

Code:

lspci | grep nV
02:00.0 VGA compatible controller: nVidia Corporation GT200 [GeForce GTX 260] (rev a1)

Module option

Code:

cat /etc/modprobe.d/powerstate-nvidia.conf 
options nvidia NVreg_RegistryDwords="PerfLevelSrc=0x2222"

puntarenas · #83 13th March 2010, 12:47 PM

Well, I was too optimistic in this thread before, but for now the freezes seem to be under control again. However the only thing that kept me away from Nouveau is it's lack of power management and putting Nvidia's binary blob to "Maximum Performance" seemed a bit *@?!!"%&d to me.

Thanks to this excellent blog post I have a solution now, as a workaround one can force Power Mizer to maximum power saving:

nVidia PowerMizer powersaving/cooling in Linux (updated)

From my xorg.conf:

Code:

Section "Device"
	Option  "Coolbits" "1"
	Option  "RegistryDwords" "PowerMizerEnable=0x1; PerfLevelSrc=0x2222; PowerMizerLevel=0x3; PowerMizerDefault=0x3; PowerMizerDefaultAC=0x3"
	Identifier  "Videocard0"
	Driver      "nvidia"
	Option	    "AddARGBGLXVisuals" "True"
EndSection

Works fine, temperatures and power consumption as I want it to be and hopefully the freezes finally vanished.

Edit: ...and still freezes:

Code:

Mar 13 16:48:22 client01 kernel: irq 16: nobody cared (try booting with the "irqpoll" option)
Mar 13 16:48:22 client01 kernel: Pid: 0, comm: swapper Tainted: P           2.6.33-1.fc13.i686 #1
Mar 13 16:48:22 client01 kernel: Call Trace:
Mar 13 16:48:22 client01 kernel: [<c0489e39>] __report_bad_irq+0x33/0x74
Mar 13 16:48:22 client01 kernel: [<c0489f74>] note_interrupt+0xfa/0x152
Mar 13 16:48:22 client01 kernel: [<c048a556>] handle_fasteoi_irq+0x8f/0xb2
Mar 13 16:48:22 client01 kernel: [<c0404ef0>] handle_irq+0x40/0x4c
Mar 13 16:48:22 client01 kernel: [<c040475f>] do_IRQ+0x46/0x9f
Mar 13 16:48:22 client01 kernel: [<c0403975>] common_interrupt+0x35/0x3c
Mar 13 16:48:22 client01 kernel: [<c0409745>] ? mwait_idle+0x68/0x78
Mar 13 16:48:22 client01 kernel: [<c04025be>] cpu_idle+0x9b/0xb5
Mar 13 16:48:22 client01 kernel: [<c07adcaa>] start_secondary+0x204/0x242
Mar 13 16:48:22 client01 kernel: handlers:
Mar 13 16:48:22 client01 kernel: [<c06adaca>] (usb_hcd_irq+0x0/0x8d)
Mar 13 16:48:22 client01 kernel: [<fa16f801>] (nv_kern_isr+0x0/0x59 [nvidia])
Mar 13 16:48:22 client01 kernel: Disabling IRQ #16
Mar 13 16:48:50 client01 kernel: SysRq : Keyboard mode set to system default
Mar 13 16:48:51 client01 kernel: SysRq : Terminate All Tasks
Mar 13 16:48:51 client01 kernel: Kernel logging (proc) stopped.
Mar 13 16:48:51 client01 rsyslogd: [origin software="rsyslogd" swVersion="4.4.2" x-pid="977" x-info="http://www.rsyslog.com"] exiting on signal 15.

spacecadet · #84 16th March 2010, 01:40 AM

I also have been experiencing freezes.
Gigabyte 8I845GVM-RZ MB, 2GB RAM (using the on-board Intel 82845GV video, which steal 8MB from system RAM)

Running: 2.6.32.9-70.fc12.i686.PAE

Lately, screen goes black (even while typing away). It used to just freeze although mouse pointer movement was still possible. I notice the num-lock, caps-lock & sroll-lock keys don't function either.

However I was able to ssh to the box from another one. However, I've been away from linux so long I'm not sure what to look at while ssh'd to it.

This is from the message log; these snipped messages are the same every time the box freezes:

~~~~~
Mar 15 20:18:28 localhost kernel: [drm:i915_hangcheck_elapsed] *ERROR* Hangcheck timer elapsed... GPU hung
Mar 15 20:18:28 localhost kernel: render error detected, EIR: 0x00000000
Mar 15 20:18:28 localhost kernel: [drm:i915_do_wait_request] *ERROR* i915_do_wait_request returns -5 (awaiting 82953 at 82949)
Mar 15 20:18:29 localhost abrt[2092]: saved core dump of pid 1292 (/usr/bin/Xorg) to /var/cache/abrt/ccpp-1268698708-1292.new/coredump (13500416 bytes)
Mar 15 20:18:29 localhost abrtd: Directory 'ccpp-1268698708-1292' creation detected
Mar 15 20:18:29 localhost gnome-keyring-daemon[1438]: dbus failure unregistering from session: Connection is closed
Mar 15 20:18:29 localhost gnome-keyring-daemon[1438]: dbus failure unregistering from session: Connection is closed
Mar 15 20:18:29 localhost abrtd: Crash is in database already (dup of /var/cache/abrt/ccpp-1268573315-1322)
Mar 15 20:18:29 localhost abrtd: Deleting crash ccpp-1268698708-1292 (dup of ccpp-1268573315-1322), sending dbus signal
Mar 15 20:18:30 localhost abrt[2104]: not dumping repeating crash in '/usr/bin/Xorg'
Mar 15 20:18:30 localhost gdm-binary[1266]: WARNING: GdmDisplay: display lasted 0.642989 seconds
Mar 15 20:18:30 localhost abrt[2109]: not dumping repeating crash in '/usr/bin/Xorg'
Mar 15 20:18:30 localhost gdm-binary[1266]: WARNING: GdmDisplay: display lasted 0.396725 seconds
Mar 15 20:18:30 localhost abrt[2114]: not dumping repeating crash in '/usr/bin/Xorg'
Mar 15 20:18:30 localhost gdm-binary[1266]: WARNING: GdmDisplay: display lasted 0.400225 seconds
Mar 15 20:18:31 localhost abrt[2119]: not dumping repeating crash in '/usr/bin/Xorg'
Mar 15 20:18:31 localhost gdm-binary[1266]: WARNING: GdmDisplay: display lasted 0.386829 seconds
Mar 15 20:18:31 localhost abrt[2124]: not dumping repeating crash in '/usr/bin/Xorg'
Mar 15 20:18:31 localhost gdm-binary[1266]: WARNING: GdmDisplay: display lasted 0.387543 seconds
Mar 15 20:18:31 localhost abrt[2129]: not dumping repeating crash in '/usr/bin/Xorg'
Mar 15 20:18:31 localhost gdm-binary[1266]: WARNING: GdmDisplay: display lasted 0.385009 seconds
Mar 15 20:18:31 localhost gdm-binary[1266]: WARNING: GdmLocalDisplayFactory: maximum number of X display failures reached: check X server log for errors
Mar 15 20:18:32 localhost init: prefdm main process (1266) terminated with status 1
Mar 15 20:18:32 localhost init: prefdm main process ended, respawning
~~~~~~~

I had been blaming the hardware since it's a bit older. I did remove memory and clean the edge connector and re-seated it. But no joy.

Unfortunately I cannot determine when this all started (which kernel).

spacecadet

mako999 · #85 19th March 2010, 08:25 AM

I also was doing hang with FC12.

Code:

$ lspci | grep nV
01:00.0 VGA compatible controller: nVidia Corporation GeForce 8800 GTS 512 (rev a2)

FC13 alpha is tried now.　Got nouveau going on my 8800gts and not freezes. And Compiz moves usually, too.

puntarenas · #86 23rd March 2010, 11:25 AM

Nvidia is at least aware of the freezes now, maybe some more user feedback would help them to track the issue down and start fixing it:

Freezes after upgrade from Fedora 10 to 12

I think they would love to receive some more nvidia-bug-report.log.gz created by nvidia-bug-report.sh attached to "me too postings", allthough it is hard to register on their forums with common mail addresses.

lensman3 · #87 23rd March 2010, 02:53 PM

I disagree. Should I as a user be debugging their hardware/software? This is the same problem that M$ had with them when Vista was introduced. Nvidia's own policies are hurting them in the open source arena. I'm using the nouveau graphics driver which is being reverse-engineered. Maybe this reverse-engineering project will show the world that the hardware is the problem. The microcode in the GPU is faulty and needs to be fixed.

I can make F12 lock just using Firefox. The current 64-bit kernel will lock a machine that never had any problems. (The boot virtual disk, the one that the booting kernel uses temporily to load needed drivers) doubled in size from the previous kernel release. My machine usually runs 24/7 and the last kernel locks up under conditions that previous kernel didn't!

<FLAME ON>I'm tired of debugging lazy developers code!<FLAME OFF>