[331.20 & earlier] Geforce 9600M GS causes black screen/hang when powermizer reaches level 0

I have a Acer Aspire 7720G laptop with a Geforce 9600M GS MXM-II card inserted into it.
Whenever I have the Nvidia driver running, my laptop hangs with a black screen when powermizer level 0 is reached. Levels 1-3 are functioning ok. I can’t ssh into the laptop anymore when this happens. The only thing I can do is hardreset.

I am using the latest Nvidia Driver (331.32) on Ubuntu 13.10 64-bit, but have also tested this with the 172 driver version and several newer drivers, on ubuntu 12.04 and 13.04, 32 and 64 bit. All had the same issue.

Unfortunately there is no entry of these crashes in the kern.log, dmesg, syslog or Xorg log. (I did not enable extra logging function, don’t know what I should do for that).

I have attached a nvidia bug-report from when powermizer is set to max performance (lvl 3) and a xorg log with verbosity 6.

Let me know if I can debug more.

I have tested this card on Windows 7 and it had no problems there like this.
nvidia-bug-report.log.gz (73.1 KB)
Xorg.0.log (179 KB)

I installed kernel 3.12 and now the laptop hangs, but the screen doesn’t go black anymore, and I can get dmesg log and nvidia-bug-report.log.gz while the device has crashed using ssh.

Relevant dmesg:

[  120.990424] NVRM: os_pci_init_handle: invalid context!
[  120.990435] NVRM: os_pci_init_handle: invalid context!
[  120.990440] NVRM: GPU at 0000:01:00.0 has fallen off the bus.
[  120.990442] NVRM: os_pci_init_handle: invalid context!

relevant Xorg.0.log

[   126.473] (WW) NVIDIA(0): WAIT (2, 6, 0x8000, 0x000023d4, 0x00002ec0)
(EE) [mi] EQ overflowing.  Additional events will be discarded until existing events are processed.
(EE) Backtrace:
(EE) 0: /usr/bin/X (xorg_backtrace+0x3d) [0x7f8d8a0f902d]
(EE) 1: /usr/bin/X (mieqEnqueue+0x22b) [0x7f8d8a0daf6b]
(EE) 2: /usr/bin/X (QueuePointerEvents+0x52) [0x7f8d89fc26d2]
(EE) 3: /usr/lib/xorg/modules/input/evdev_drv.so (0x7f8d8013c000+0x566f) [0x7f8d8014166f]
(EE) 4: /usr/bin/X (0x7f8d89f57000+0x91af8) [0x7f8d89fe8af8]
(EE) 5: /usr/bin/X (0x7f8d89f57000+0xba520) [0x7f8d8a011520]
(EE) 6: /lib/x86_64-linux-gnu/libpthread.so.0 (0x7f8d89057000+0xfbb0) [0x7f8d89066bb0]
(EE) 7: /usr/lib/x86_64-linux-gnu/xorg/extra-modules/nvidia_drv.so (0x7f8d80b4a000+0x6149b) [0x7f8d80bab49b]
(EE) 8: /usr/lib/x86_64-linux-gnu/xorg/extra-modules/nvidia_drv.so (0x7f8d80b4a000+0x61ec1) [0x7f8d80babec1]
(EE) 9: /usr/lib/x86_64-linux-gnu/xorg/extra-modules/nvidia_drv.so (0x7f8d80b4a000+0xd225b) [0x7f8d80c1c25b]
(EE) 10: /usr/lib/x86_64-linux-gnu/xorg/extra-modules/nvidia_drv.so (0x7f8d80b4a000+0x90cf2) [0x7f8d80bdacf2]
(EE) 11: /usr/lib/x86_64-linux-gnu/xorg/extra-modules/nvidia_drv.so (0x7f8d80b4a000+0x51397c) [0x7f8d8105d97c]
(EE) 12: /usr/lib/x86_64-linux-gnu/xorg/extra-modules/nvidia_drv.so (0x7f8d80b4a000+0x4f56d9) [0x7f8d8103f6d9]
(EE) 13: /usr/bin/X (BlockHandler+0x44) [0x7f8d89fb0054]
(EE) 14: /usr/bin/X (WaitForSomething+0x124) [0x7f8d8a0f6574]
(EE) 15: /usr/bin/X (0x7f8d89f57000+0x54d61) [0x7f8d89fabd61]
(EE) 16: /usr/bin/X (0x7f8d89f57000+0x4456a) [0x7f8d89f9b56a]
(EE) 17: /lib/x86_64-linux-gnu/libc.so.6 (__libc_start_main+0xf5) [0x7f8d87c92de5]
(EE) 18: /usr/bin/X (0x7f8d89f57000+0x448af) [0x7f8d89f9b8af]
(EE) [mi] These backtraces from mieqEnqueue may point to a culprit higher up the stack.
(EE) [mi] mieq is *NOT* the cause.  It is a victim.
[   133.473] (WW) NVIDIA(0): WAIT (1, 6, 0x8000, 0x000023d4, 0x00002ec0)
[   133.473] [mi] Increasing EQ size to 512 to prevent dropped events.
[   133.474] [mi] EQ processing has resumed after 23 dropped events.
[   133.474] [mi] This may be caused my a misbehaving driver monopolizing the server's resources.
[   136.486] (WW) NVIDIA(0): WAIT (2, 6, 0x8000, 0x000023d4, 0x0000f910)
[   143.486] (WW) NVIDIA(0): WAIT (1, 6, 0x8000, 0x000023d4, 0x0000f910)

Hmm, sometimes it fully hangs with black screen, sometimes it just stops responding but available on ssh.

nvidia-bug-report.ondevice.log.gz (58.6 KB)
Xorg.0.crashed.log (182 KB)

–post removed-- not relevant.