Develop Bugs

[Q] Odin Problem, CPU Scaling Problem

I'm running SyndicateROM Frozen 1.2, fresh install with twilightzone 1.1.1. For some reason, in setCPU, the cpu is always running at the max setting, eg 1200, using conservative governor. It won't throttle down. According to the SystemPanel app, the cpu is not actually doing anything, it's just stuck at max clock setting.
So I decided to goto stock with odin and start over, only that the phone in download mode will not be recognized by two computers running win7 32/64 bit. I can turn the phone on, use CWM. It behaves normally otherwise. The phone is recognized when turned on, and the appropriate drivers install. USB debugging is turned on. Any ideas how to get my phone recognized in download mode?
-Greg

I think this is an issue on several kernels. I noticed this months ago but haven't figured out a fix myself. Anything I set over 1000 on conservative doesn't stick. I always use "On Demand." CPU Spy is a good free app to monitor this behavior.
Sent from my SPH-D700 using xda premium

[Q] Eugene's sleeper script

I just wanted to confirm my understanding of eugene's sleeper script since I never fully got a handle on it. As I understand it, prior to this our phones second core never truly went offline (even though system tuner read it as offline). Now, our second core always goes offline when the screen is off (even though system tuner reads it as running at 192mhz). Is this right? Also, does it go offline at low load time periods when the screen is off?

yoft1 said:
I just wanted to confirm my understanding of eugene's sleeper script since I never fully got a handle on it. As I understand it, prior to this our phones second core never truly went offline (even though system tuner read it as offline). Now, our second core always goes offline when the screen is off (even though system tuner reads it as running at 192mhz). Is this right? Also, does it go offline at low load time periods when the screen is off?
Click to expand...
Click to collapse
On stock rooted ICS (though de-bloted) second core goes into deep sleep no problem when task scheduler/governor commands so and regardless if screen is On or Off.
What Eugene's script is doing: constantly looping every 2/4 sec (or whatever) and checking if screen is On or Off and then force CPU1 to go offline regardless what Linux Governor think is appropriate for current CPU load.
It might very much happened that next moment (in between loop cycles) governor would turn CPU1 back online if load on CPU0 is above threshold.
If script loop would be too short then script itself would burn battery juice... so, I think best thing to improve battery life is just keep phone clean from unwanted background processes.
Also, it might happened in reverse, when screen is on and there is no load - governor would turn CPU1 off, but script will continue to force CPU1 online.

Vlad_z said:
On stock rooted ICS (though de-bloted) second core goes into deep sleep no problem when task scheduler/governor commands so and regardless if screen is On or Off.
What Eugene's script is doing: constantly looping every 2/4 sec (or whatever) and checking if screen is On or Off and then force CPU1 to go offline regardless what Linux Governor think is appropriate for current CPU load.
It might very much happened that next moment (in between loop cycles) governor would turn CPU1 back online if load on CPU0 is above threshold.
If script loop would be too short then script itself would burn battery juice... so, I think best thing to improve battery life is just keep phone clean from unwanted background processes.
Also, it might happened in reverse, when screen is on and there is no load - governor would turn CPU1 off, but script will continue to force CPU1 online.
Click to expand...
Click to collapse
So if I understand you right, the script does nothing radically different, all it does is override the original cpu handling to force cpu1 offline when screen is off? It doesn't put cpu1 offline a different way than before?

yoft1 said:
So if I understand you right, the script does nothing radically different, all it does is override the original cpu handling to force cpu1 offline when screen is off? It doesn't put cpu1 offline a different way than before?
Click to expand...
Click to collapse
What do you mean by "different way"? (Script writes 1/0 into Linux CPU driver)

Vlad_z said:
What do you mean by "different way"? (Script writes 1/0 into Linux CPU driver)
Click to expand...
Click to collapse
I'm not sure, I just remember Eugene saying that cpu1 wasn't truly going offline before his script?

I don't pretend to understand any of it, but I will tell you that when the screen is off, it eats much less battery then before using it.

Vibrantnew-b said:
I don't pretend to understand any of it, but I will tell you that when the screen is off, it eats much less battery then before using it.
Click to expand...
Click to collapse
that's very possible. If you have some background processes, using this script, you basically trying to make from dual core CPU single.
or this is just placebo effect... how did you test it? My phone goes to deep sleep no problem, overnight 4% for 8 hours.

My Nexus not waking up?

basically just recently i have noticed my Nexus not waking up.
will simply go sleep perfectly fine but sometimes will not wake and sort of reboots in away.
tried all sort of ROMS, kernels, radios with no effect. even switched back to ICS but still get the same results.
Galaxy nexus
Android 4.1.1
XXLF1
Rom: Latest CM10 Nightly
Kenel: Stock (even tried Francos)
Logcats:
http://pastebin.com/tpeuQYqx
Any ideas guys?

Did you try wiping everything on the phone and starting fresh. Go in the Android Revolution thread and run the superwipe script and wipe again then reinstall everything see if that helps.
I didn't look at the logcats, too tedious.

See here. http://forum.xda-developers.com/showthread.php?p=29839968
good day.

clarkey1269 said:
basically just recently i have noticed my Nexus not waking up.
will simply go sleep perfectly fine but sometimes will not wake and sort of reboots in away.
tried all sort of ROMS, kernels, radios with no effect. even switched back to ICS but still get the same results.
Galaxy nexus
Android 4.1.1
XXLF1
Rom: Latest CM10 Nightly
Kenel: Stock (even tried Francos)
Logcats:
http://pastebin.com/tpeuQYqx
Any ideas guys?
Click to expand...
Click to collapse
Does this happen on stock. This logcat is using franco's kernel, are you UV'ing? OC'ing? Smartreflex? I don't know if actually Francisco is using official stock voltages today, or is already uv'ing a little from source.
And I just finished going through your log, there's nothing useful there; you should be getting it while it happens, or better yet would be /proc/last_kmsg immediately after the reboot; Grab it like this, for example:
Code:
adb shell cat /proc/last_kmsg > last_kmsg_on_my_pc
Also, what you're experiencing is called Sleep of Death.

chopper the dog said:
See here. http://forum.xda-developers.com/showthread.php?p=29839968
good day.
Click to expand...
Click to collapse
also happens on ICS not just JB
full wipe is taken out with every flash or nightly update without fail.
even did a fresh start with locking and relocking the bootloader with no effect i still get the annoying wake

Are you using any app to control your min CPU voltage? the problem might be that your phone doesn't have enough power to come back from sleep or just takes really long if the CPU voltage is too low...

cayte2 said:
Are you using any app to control your min CPU frequency? the problem might be that your phone doesn't have enough power to come back from sleep or just takes really long if the CPU frequency is too low...
Click to expand...
Click to collapse
Why? It's more likely that it's the voltages that are causing it, they're probably too low for his chip. What you're saying is that "it doesnt have enough power", so you're kind of contradicting yourself. And CPU ramps you know? So, if it meets a intensive task (if you call waking up the screen and bringing up UI an intensive task), it will ramp UP to the next step, until it reaches the defined workload percentage for max cpu speed defined in the governor.

bk201doesntexist said:
Why? It's more likely that it's the voltages that are causing it, they're probably too low for his chip. What you're saying is that "it doesnt have enough power", so you're kind of contradicting yourself. And CPU ramps you know? So, if it meets a intensive task (if you call waking up the screen and bringing up UI an intensive task), it will ramp UP to the next step, until it reaches the defined workload percentage for max cpu speed defined in the governor.
Click to expand...
Click to collapse
Sorry I meant to say voltage

cayte2 said:
Are you using any app to control your min CPU frequency? the problem might be that your phone doesn't have enough power to come back from sleep or just takes really long if the CPU frequency is too low...
Click to expand...
Click to collapse
cayte2 said:
Sorry I meant to say voltage
Click to expand...
Click to collapse
It seems you didn't understand where you're contradicting yourself.
min cpu frequency is not equal to min cpu frequency voltage. If you set min cpu freq to 700mhz step, but lower the voltage to a value where that same cpu step it's not stable, it will still have issues.
Another thing I recalled: OP, are you limiting your MAX cpu freq, WHEN screen is off? Like with cpu profilers and such apps?

bk201doesntexist said:
It seems you didn't understand where you're contradicting yourself.
min cpu frequency is not equal to min cpu frequency voltage. If you set min cpu freq to 700mhz step, but lower the voltage to a value where that same cpu step it's not stable, it will still have issues.
Another thing I recalled: OP, are you limiting your MAX cpu freq, WHEN screen is off? Like with cpu profilers and such apps?
Click to expand...
Click to collapse
when using Francos kernel i have being using his app.
haven't messed with any frequency's and is set to what ever the kernel is standard. with no effect.
no fancy profiles set in his app either.
really starting to bug me now

clarkey1269 said:
when using Francos kernel i have being using his app.
haven't messed with any frequency's and is set to what ever the kernel is standard. with no effect.
no fancy profiles set in his app either.
really starting to bug me now
Click to expand...
Click to collapse
Does it happen when using google's factory image?

Do you have Eugene cpu sleeper installed? I was having that happen to mine and I just uninstalled, wiped caches, fixed permissions and I haven't had a problem since.
Sent from my Galaxy Nexus using xda premium

bk201doesntexist said:
Does it happen when using google's factory image?
Click to expand...
Click to collapse
fresh instal of googles stock image and il report back

Note 2 charging CPU max clock

I just did a fresh install of a stock rom 4.1.2 and did full factory reset , wiped dalvik and noticed that my Note 2 is getting hot while charging 38-39 Battery temp and around 50c+ CPU temp , i used CPU spy and the phone is running on 1.6ghz while charging , its always 1.6ghz. Is this a normal behavior because i am pretty worried?
Edit: its running 1.6ghz only when its charging and the screen is off , the moment i turn the screen on and unlock the phone it starts to run normally 200mhz-800mhz .

Fygaroo said:
I just did a fresh install of a stock rom 4.1.2 and did full factory reset , wiped dalvik and noticed that my Note 2 is getting hot while charging 38-39 Battery temp and around 50c+ CPU temp , i used CPU spy and the phone is running on 1.6ghz while charging , its always 1.6ghz. Is this a normal behavior because i am pretty worried?
Edit: its running 1.6ghz only when its charging and the screen is off , the moment i turn the screen on and unlock the phone it starts to run normally 200mhz-800mhz .
Click to expand...
Click to collapse
Since you did flash your phone recently, it is quite possible that your phone is running some background maintenance activities (like indexing) during screen-off. If it is not too hot, you could try and see whether it settles down after a 24 hours. However, if it is getting really hot, why don't you try and put the phone into recovery and complete a full charge. Then do a full reset after unplugging from charger and check whether CPU over heating still occurs.
One other option is that you could try to use 'System Tuner' to manually control the CPU Governor. Check out the following thread http://forum.xda-developers.com/showthread.php?t=2509233 for some details on how to control CPU governor settings during 'Screen-off'.

degakc said:
Since you did flash your phone recently, it is quite possible that your phone is running some background maintenance activities (like indexing) during screen-off. If it is not too hot, you could try and see whether it settles down after a 24 hours. However, if it is getting really hot, why don't you try and put the phone into recovery and complete a full charge. Then do a full reset after unplugging from charger and check whether CPU over heating still occurs.
One other option is that you could try to use 'System Tuner' to manually control the CPU Governor. Check out the following thread http://forum.xda-developers.com/showthread.php?t=2509233 for some details on how to control CPU governor settings during 'Screen-off'.
Click to expand...
Click to collapse
I wrote this post to fast and didn't let my phone settle on the new rom , it was running on max clock only on the first charge cycle to 100% its all working normally now. This post can be closed now sorry for making it for nothing .

CPU Cores 2 and 3 never turn off

I'm on the US NG1 firmware on the Tab Pro 8.4 (all stock) and I seem to have this situation where CPU Cores #2 and #3 are always on, and always running at least 1.5GHz. In fact they seem locked at 1.5GHz (they don't seem to go faster or slower) which is very strange. Both CPU-Z and the "device info" screen of AnTuTu show evidence of this.
Any ideas?
I've already tried killing all programs/tasks/recent apps ... no change.
( A few days ago I had the issue where even when the tablet was idle (screen turned off) for several hours it was basically almost fully awake, and the main culprit was "Google Play Services" which used way more battery than even the SCREEN. Very weird. But after 2 days of this weird behaviour, it stopped and the tablet was able to stay in deep sleep most of the time in Idle )

Strange behaviour indeed...I have troubles with deep sleep to, but not related to the cores locked to max freq.
Try hard reset, and if it doesn't work, go for replacement. ..

I cleared Cache from Settings->Storage and did a cold boot (turn off, wait, then turn on)
The issue seems to be gone. I quickly opened each of my most-freq-used apps at least once ... then checked CPU-Z again, and it's still fine (CPU cores are sleeping).
So it's not one of my frequently used apps.
I'll probably check my Battery specs on a daily basis for the next little while to ensure it doesn't come back.
BTW are there any apps for unrooted phones which can detect wakelocks? All the ones I saw on the Play Store require Root to detect wakelocks