View Issue Details

IDProjectCategoryView StatusLast Update
0005774CentOS-6kernelpublic2012-06-15 13:37
Reporterjweage 
PrioritynormalSeveritycrashReproducibilityalways
Status newResolutionopen 
Product Version6.2 
Target VersionFixed in Version 
Summary0005774: Anaconda kernel crashes due to nouveau driver on GTX 680
DescriptionWhen installing on a workstation with a GeForce GTX 680 graphics card, the installer crashes with the following error:

Kernel panic - not syncing: Fatal exception
Pid: 348, comm: modprobe Tainted: G D -------- 2.6.32.220.e16.x86_64 #1
Call Trace:
[<fffffff814ec321>] ? panic +0x78/0x143
[<fffffff814f04b4>] ? oops_end +0xe4/0/100
[<fffffff8104230b>] ? no_context+0xfb/0x260
[<fffffff81042595>] ? __bad_area_nosemaphore+0x125/0x1e0
[<fffffff810426be>] ? bad_area+0x4e/0x60
[<fffffff81042dc3>] ? __do_page_fault+0x3c3/0x480
[<fffffff8109694f>] ? up+0x2f/0x50
[<fffffff814f246e>] ? release_console_sem+0x1cf/0x220
[<fffffffa0233af0>] ? do_page_fault+0x3e/0xa0
? probe_monitoring_device+0x0/0x80 [nouveau]

Adding the following to the boot kernel command line avoids the crash:

xdriver=vesa rdblacklist=nouveau
Steps To ReproduceAttempt to install on a system with NVIDIA GeForce GTX 680 graphics.
Additional InformationSame/similar bug is reported here: https://bugzilla.redhat.com/show_bug.cgi?id=768888

but RH seems to think this should happen with their kernel.
TagsNo tags attached.

Activities

jweage

jweage

2012-06-13 20:16

reporter   ~0015256

last line should read: "but RH seems to think this shouldn't happen with their kernel."
toracat

toracat

2012-06-13 20:23

manager   ~0015257

The CentOS kernel must be binary compatible with the upstream kernel. Did you install something that may alter the kernel? I see "Tainted" in your output.
jweage

jweage

2012-06-14 00:46

reporter   ~0015258

This is the boot kernel from the CentOS 6.2 PXE boot image, so no, nothing has been installed.
toracat

toracat

2012-06-14 21:28

manager   ~0015265

I'm still puzzled by the "Pid: 348, comm: modprobe Tainted: G D" line. Can you check the output from:

cat /proc/sys/kernel/tainted

If it is other than zero, you have some module that taints the kernel. Then look in /proc/modules and look for (G).
jweage

jweage

2012-06-15 13:25

reporter   ~0015272

I can't do that as the install crashes before anything starts. If I use the VESA driver, then the kernel isn't tainted.

For the normal install, I'm using the following kernel arguments:
initrd=centos-6.2-x86_64/initrd.img ramdisk_size=15489 method=http://192.168.2.18/centos/6/os/x86_64 ip=dhcp

For the VESA driver I use these:
initrd=centos-6.2-x86_64/initrd.img ramdisk_size=15489 method=http://192.168.2.18/centos/6/os/x86_64 ip=dhcp xdriver=vesa rdblacklist=nouveau

Maybe nouveau is considered to be tainted?
jweage

jweage

2012-06-15 13:37

reporter   ~0015273

This problem is reported on the forums here:

https://www.centos.org/modules/newbb/viewtopic.php?topic_id=37072&forum=55

I copy/pasted the call trace from that message. I see nearly the same thing. I see the "Detecting Hardware" message right before it crashes.

The full message on my screen is as follows:

[<ffffffff8100b0f2>] system_call_fastpath+0x16/0x1b
Code: 50 48 c7 [rest deleted]
RIP [<ffffffffa0182636>] i2c_transfer+0x16/0x140 [i2c_core]
 RSP <ffff88062bee7828>
CR2: 0000000000000010
---[ end trace e55b503012e36e3e ]---
Kernel panic - not syncing: Fatal exception
Pid: 526, comm: modprobe Tainted: G D -------- 2.6.32-220.e16.x86_64 #1
Call Trace:
[<ffffffff814ec321>] ? panic +0x78/0x143
[<ffffffff814f04b4>] ? oops_end +0xe4/0/100
[<ffffffff8104230b>] ? no_context+0xfb/0x260
[<ffffffff81042595>] ? __bad_area_nosemaphore+0x125/0x1e0
[<ffffffff810426be>] ? bad_area+0x4e/0x60
[<ffffffff81042dc3>] ? __do_page_fault+0x3c3/0x480
[<ffffffff8109694f>] ? up+0x2f/0x50
[<ffffffff814f246e>] ? release_console_sem+0x1cf/0x220
[<ffffffffa0233af0>] ? do_page_fault+0x3e/0xa0
[<ffffffffa02699f0>] ? probe_monitoring_device+0x0/0x80 [nouveau]
[<ffffffff814ef825>] ? page_fault+0x25/0x30
[<ffffffffa02699f0>] ? probe_monitoring_device+0x0/0x80 [nouveau]

Issue History

Date Modified Username Field Change
2012-06-13 20:11 jweage New Issue
2012-06-13 20:16 jweage Note Added: 0015256
2012-06-13 20:23 toracat Note Added: 0015257
2012-06-14 00:46 jweage Note Added: 0015258
2012-06-14 21:28 toracat Note Added: 0015265
2012-06-15 13:25 jweage Note Added: 0015272
2012-06-15 13:37 jweage Note Added: 0015273