On 23/07/11 15:12, lee wrote:
Hm, I can imagine that intel_gpu_dump tries to load (i. e. mmap() )some file which isn't there because the X server didn't crash and thus gives you the error message that it cannot mmap() a file that doesn't exist.
Thank you for helping.
Obviously I only disabled the nvidia GPU, left the intel operating, which can be checked by unplugging the AC adapter and checking around 23 Watt battery draining with nvidia ON, and about 13 Watt turning it off with acpi_call module.You're saying in [1] that you have disabled GPUs to save power. Is the problem still there when don't disable the GPUs?
Ok, so I remove the module, even renamed the file just to be sure, reboot, checked battery drain just to be sure and triggered the freeze, which happened again. Connected by ssh and checked Xorg.0.log which had a backtrace, which doesn't always happen. I was then preparing to do the gpu_dump when I noticed that I had forgot to mount debugfs.
So, cycled the whole thing again, mounted debugfs and this time the log didn't have a backtrace and doesn't do gpu_dump.
Apparently, acpi_call doesn't contribute to the freeze, do you agree? Hope you have more ideas. rgds jss