Comment 60 for bug 1011792

Unfortunately, it looks like we're still freezing up with noautogroup enabled (set at boot using Grub).

Booted with:

Oct 17 00:07:45 localhost kernel: [ 0.000000] Command line: root=UUID=3ad27d04-4ecf-493d-bb19-4710c3caf924 ro console=hvc0 noautogroup

> uname -a
Linux moonshine21-readslave-ssd-pv 3.2.0-31-virtual #50-Ubuntu SMP Fri Sep 7 16:36:36 UTC 2012 x86_64 x86_64 x86_64 GNU/Linux

> cat /proc/sys/kernel/sched_autogroup_enabled
0

But within a few hours (which, granted, is longer than Precise lasted without the autogrouping disabled):

Oct 17 02:38:00 localhost kernel: [1653329.620312] INFO: rcu_sched detected stalls on CPUs/tasks: { 1 4 6 8} (detected by 0, t=15002 jiffies)
Oct 17 02:38:00 localhost kernel: [1653329.620330] sending NMI to all CPUs:
Oct 17 02:41:00 localhost kernel: [1653509.748321] INFO: rcu_sched detected stalls on CPUs/tasks: { 1 4 6 8} (detected by 0, t=60034 jiffies)
Oct 17 02:41:00 localhost kernel: [1653509.748340] sending NMI to all CPUs:
Oct 17 02:42:24 localhost kernel: [1653594.056348] INFO: rcu_bh detected stalls on CPUs/tasks: { 1 3 4 6 8} (detected by 0, t=15002 jiffies)
Oct 17 02:42:24 localhost kernel: [1653594.056362] sending NMI to all CPUs:
Oct 17 02:44:00 localhost kernel: [1653689.876311] INFO: rcu_sched detected stalls on CPUs/tasks: { 1 4 6 8} (detected by 0, t=105066 jiffies)
Oct 17 02:44:00 localhost kernel: [1653689.876326] sending NMI to all CPUs:
Oct 17 02:45:24 localhost kernel: [1653774.184313] INFO: rcu_bh detected stalls on CPUs/tasks: { 1 3 4 6 8} (detected by 0, t=60034 jiffies)
Oct 17 02:45:24 localhost kernel: [1653774.184331] sending NMI to all CPUs:
Oct 17 02:47:00 localhost kernel: [1653870.004311] INFO: rcu_sched detected stalls on CPUs/tasks: { 1 4 6 8} (detected by 0, t=150098 jiffies)
Oct 17 02:47:00 localhost kernel: [1653870.004326] sending NMI to all CPUs:
Oct 17 02:48:24 localhost kernel: [1653954.312311] INFO: rcu_bh detected stalls on CPUs/tasks: { 1 3 4 6 8} (detected by 0, t=105066 jiffies)

Any other variations worth trying, or other data that would be interesting to produce?