Comment 609 for bug 359392

I have a regression in Karmic: after living with the "green" drivers for a couple of week and everything was fine, I upgraded to Karmic on Thursday and ever since I had multiple freezes. I don't know if this is related to this bug, but I'll assume that it is unless someone tell me otherwise:

I got only one freeze while working, and it looks exactly the same as this problem: the screen freezes but the mouse still responds and I can ssh into the machine. But most of the freezes where while the laptop was on power with the lid closed, which means the screen is actually off and when I open it several hours later nothing responds - the computer is still on but the screen is disabled and nothing I do brings it back. I haven't yet tried to ssh into the laptop when its like that as usually I need to do something right away and I have to power cycle it.

Anyway, going through syslog I found this which might be related:
Jun 20 15:17:57 sepiroth kernel: [ 9960.560081] INFO: task events/1:10 blocked for more than 120 seconds.
Jun 20 15:17:57 sepiroth kernel: [ 9960.560088] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
Jun 20 15:17:57 sepiroth kernel: [ 9960.560095] events/1 D 0000000100213a35 0 10 2
Jun 20 15:17:57 sepiroth kernel: [ 9960.560108] ffff88007d3c9db0 0000000000000046 0000000000000000 0000000000013000
Jun 20 15:17:57 sepiroth kernel: [ 9960.560119] ffff88007d3c03a8 0000000000013000 0000000000013000 0000000000013000
Jun 20 15:17:57 sepiroth kernel: [ 9960.560128] 0000000000013000 0000000000013000 ffff88007d3c03a8 0000000000013000
Jun 20 15:17:57 sepiroth kernel: [ 9960.560138] Call Trace:
Jun 20 15:17:57 sepiroth kernel: [ 9960.560157] [<ffffffff806d9467>] __mutex_lock_slowpath+0xd7/0x160
Jun 20 15:17:57 sepiroth kernel: [ 9960.560170] [<ffffffff802436b1>] ? finish_task_switch+0x51/0x110
Jun 20 15:17:57 sepiroth kernel: [ 9960.560179] [<ffffffff806d9186>] mutex_lock+0x26/0x50
Jun 20 15:17:57 sepiroth kernel: [ 9960.560215] [<ffffffffa0132ec8>] i915_gem_retire_work_handler+0x38/0x90 [i915]
Jun 20 15:17:57 sepiroth kernel: [ 9960.560238] [<ffffffffa0132e90>] ? i915_gem_retire_work_handler+0x0/0x90 [i915]
Jun 20 15:17:57 sepiroth kernel: [ 9960.560248] [<ffffffff802643d5>] run_workqueue+0x95/0x170
Jun 20 15:17:57 sepiroth kernel: [ 9960.560256] [<ffffffff80264554>] worker_thread+0xa4/0x120
Jun 20 15:17:57 sepiroth kernel: [ 9960.560266] [<ffffffff80268e90>] ? autoremove_wake_function+0x0/0x40
Jun 20 15:17:57 sepiroth kernel: [ 9960.560274] [<ffffffff802644b0>] ? worker_thread+0x0/0x120
Jun 20 15:17:57 sepiroth kernel: [ 9960.560282] [<ffffffff80268a35>] kthread+0x55/0xa0
Jun 20 15:17:57 sepiroth kernel: [ 9960.560290] [<ffffffff802130ca>] child_rip+0xa/0x20
Jun 20 15:17:57 sepiroth kernel: [ 9960.560299] [<ffffffff802689e0>] ? kthread+0x0/0xa0
Jun 20 15:17:57 sepiroth kernel: [ 9960.560306] [<ffffffff802130c0>] ? child_rip+0x0/0x20

There are several copies of these over a period of about 20 minutes, each almost exactly 2 minutes after another (as can probably be deduced from the error message) before it falls silent again - I can't tell for sure whether its related to the freeze, except that these messages have appeared after I closed the computer when it was working and before I opened it again when it was not.

Going back through the logs, I found another set of events like this at another time that I remember the computer freezes, but it was yesterday and I may be mistaken about this.