Hi

Just thought I'd post this to see if there are any thoughts.

pre-47 - n900 worked flawlessly, long battery life in idle of several days, autodisconnect worked perfectly.

kp47 - n900 constant low background cpu chatter, battery life lasts around 6hrs in idle, autodisconnect fails to disconnect reliably.

kp48 - n900 worked flawlessly, long battery life in idle of several days, autodisconnect worked perfectly.

kp49 - n900 constant low background cpu chatter, battery life lasts around 6hrs in idle, autodisconnect fails to disconnect reliably.

cpu is slightly overclocked to 750/720(49).
same software throughout as near as makes no difference.
all background tasks relating to network (wifi) are off so no, skype or other IM's etc.

rgds
uTMY

As far as I can tell, KP49 is behaving at least as good KP48 battery-wise here. Since this is (also) quite critical to me, I'm sure I'll notice if something has changed

Still sleepy?
http://talk.maemo.org/showpost.php?p...postcount=2100

It's the background cpu chatter thats killing it for me, I never did find out what was causing it in kp47, earlier versions were perfect almost a flat line on the battery meter, kp48 was also perfect.

kp49 has reintroduced the constant background chatter.

:-(

rgds

I upgraded today to kp49 and see also a higher CPU usage while the phone is idle, so far my batter life is acceptable though. I will keep an eye on it.

True, I edited that post 30 seconds after writing, then left it on edit state without submitting My bad.

Anyway, thanks freemangordon and knobby - decreasing dsp for certain frequencies fixed issues. It wasn't necessary to decrease it so much (as in knobby post), anyway it's working rock solid now.

---

On the other hand, I managed to catch "full" framebuffer output of booting device after reboot, (aka mysterious reboot-shutdown bug), so here it is - another episode of "Ancient mysteries of the N900":

(be sure, to view it in original size).

I've marked things that seems important to me, (with "!" - some fatal errors about fcam and entering SHUTDOWN state + "?" on things that doesn't seem important, but are still weird) but I have no idea if it's real cause. Knowledgeable ones, any ideas?

/Estel

// Edit

Of course it only happens on reboot, powering-up from shutdown doesn't result in problems. Also, as others said it, it doesn't happen on every reboot (but, in my case, on 99% of it) - I had 2 reboots (in last few weeks) without this issue.

@Estel - seems like incorrect boot state after reboot, I dont think "SHUTDOWN" is OK. Though not sure who is responsible for that.

I have error messages about iommu too and I did not see any problems...

The next problem is with fcam kernel module. That error message means that you do not have compiled that module against actual kernel. Try to uninstall fcam (or recompile it!)

@Estel - is it possible to edit /etc/event.replace.d/dsme and change:

script
export BOOTSTATE=`getbootstate 2>/dev/null`

to

script
cat /proc/bootreason
export BOOTSTATE=`getbootstate 2>/bootstate.log`
cat /bootstate.log

so an additional info for startup conditions to be available.

Thanks, will do it and report back. By the way, totally noobish question - is there a way to dump boottime framebuffer output into a file? Taking pictures isn't most convenient,had to do 5 sessions to get material for mixing continuous output.

On the other hand, while using vdd1=1 and recommended changes (limited dsp overclock on lower frequencies), sometimes during video playback dmesg is trashed with tens of errors like this:
... while playback is unaffected. Does it mean I should limit DSP overclock further, or is it harmless? Just for the record, it doesn't happen with vdd1=0

/Estel

//Edit
Pali, I'll try to compile fCam against actual kp headers - in meantime, should't it refuse to work at all, if module fail? After normal boot, fCam (program) works flawlessly. What exactly this module is responsible for?