View Single Post
Posts: 1,680 | Thanked: 3,685 times | Joined on Jan 2011
#21
Ok, things are not as bad as they seem. The second powertop you pasted shows what I would expect the phone to be like from a fresh boot. That implies that there is not a rogue service on the loose. Now re-enable things 1 by 1 and check with powertop for each one.

With the phone in online mode (without a data connection) I would expect around 800-900 wake ups per second.

Connected to wifi I would expect that to go up to around 1400 wake ups.

i.e. Connect to wifi, check for wakeups. Success? Enable IM accounts, check. Enable modes, check.

I.e. slowly build your system back to what it was and check each piece as you add it. To make things easier you can run powertop with the parameter -t 15 however bear in mind that this will halve the number of wake ups measured (as it is measuring for half the amount of time!).

If you have enabled EVERYTHING and the system is still running clear then there is obviously some other far more subtle interaction.

I have found that if I have kernel-power-settings installed for some reason, under some circumstance that I have not yet realised (my notion is towards PC suite mode with wifi on) I get an extra 7000 hardware wakeups per second and DMA is through the roof. I have solved this by copying the kernel-power-settings script to /bin, removing the kernel power settings package and triggering my overclock profiles with dbus when the phone goes inactive (super low power underclocked mode) and normal when I unlock it again.
__________________
N900: One of God's own prototypes. A high-powered mutant of some kind never even considered for mass production. Too weird to live, and too rare to die.