Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
Yes, this is a known issue.Checked the past posts, let me know if this is known.
Items in the drop-down notification window sometimes freeze if I intentionally (or unintentionally) swipe them to the side. It is a useful feature if you want to get rid of a single notification, but not if it freezes the whole notification window.
Only fix I found was a restart. Anyone seen this?
WhyKernel-02-08.zip
Uploaded a new kernel (02-08). The occasional "event" wakelock should be fixed. I looked at some other 2.6.37 kernels and found out 2 input events running together caused it (likely accelerometer, magnetometer, compass, microphone or some combination), it was only showing up as 'event2' and 'event4'. Now /proc/wakelocks confirms they're not active anymore.
Also I experimented with removing some Big Kernel Locks (lock_kernel) calls, by modeling after 2.6.37 kernels which have this old method phased out (in favor of smaller scale mutex-locks). It seems to have improved touchscreen reliability. But I've said this before & have been wrong, maybe these are all small incremental improvements, only really noticeable when switching back to a stock kernel.
I'd like to hear if anyone still encounters any of these issues:
1) "event" kernel wakelocks
2) freezing capacitive buttons
3) any other odd behavior that did NOT occur before this kernel (assuming you've ruled out any setting changes or new app installs)
WhyKernel-02-08.zip
Uploaded a new kernel (02-08). The occasional "event" wakelock should be fixed. I looked at some other 2.6.37 kernels and found out 2 input events running together caused it (likely accelerometer, magnetometer, compass, microphone or some combination), it was only showing up as 'event2' and 'event4'. Now /proc/wakelocks confirms they're not active anymore.
Also I experimented with removing some Big Kernel Locks (lock_kernel) calls, by modeling after 2.6.37 kernels which have this old method phased out (in favor of smaller scale mutex-locks). It seems to have improved touchscreen reliability. But I've said this before & have been wrong, maybe these are all small incremental improvements, only really noticeable when switching back to a stock kernel.
I'd like to hear if anyone still encounters any of these issues:
1) "event" kernel wakelocks
2) freezing capacitive buttons
3) any other odd behavior that did NOT occur before this kernel (assuming you've ruled out any setting changes or new app installs)
I seem to have more idle drain than you have. maybe 10-15% over 6 hours or so with nothing but radio on. For some reason 2 hours before my alarm goes off in the morning it shows constantly awake. Maybe it needs that?? I'll try uninstalling and reinstalling it.
This kernel does fix the capacative buttons and is a God send. I am having a new problem though. My music does skip a lot now like a scratched CD. I listen to music all the time as I'm walking to and from classes and this hasn't happened ever before. I've had no installs or added any new music or any other data to my phone since trying to use this kernel.
And I have confirmed that it's every song, not just a few. And the longer I am listening to my music the more it's occuring.
10%-15% over 6 hours? WHAT THE HELL I'd be lucky to have over 30% in that amount of time.... CM9 seems to do it better, or maybe the app that was causing it is gone.
Probably a background app or service. I would start with Settings - Applications - Running Services to start the investigation. And the phone should not need to be awake for the alarm to go off. Rewipe alarm app cache, recovery mode cache & dalvik-cache, or a clean reinstall.
I just tested playing about 5 MP3 songs with screen off without problems. Sounds like you're using a governor that locks to min frequency when screen is off. Change your CM Settings - Performance to 245-1024 Interactive governor to see if that fixes it. Or try wiping the music app's cache or wipe cache & dalvik-cache again in recovery mode. Also titanium restores can cause problems, so a clean install may be needed if all else fails.
Mine's about the same. If i push it down to 61, music starts skipping.Thanks Whyzor. Coincidentally I was just listening to music today (I still have the 2-02 TG-R installed) and noticed the scratching while listening to mp3s. Was running 61-1024 clock and pushed it up to 122-1024. Confirming the mp3's scratching is minimized. 245 eliminates the scratch.
10%-15% over 6 hours? WHAT THE HELL I'd be lucky to have over 30% in that amount of time.... CM9 seems to do it better, or maybe the app that was causing it is gone.
Thanks Whyzor. Coincidentally I was just listening to music today (I still have the 2-02 TG-R installed) and noticed the scratching while listening to mp3s. Was running 61-1024 clock and pushed it up to 122-1024. Confirming the mp3's scratching is minimized. 245 eliminates the scratch.
Mine's about the same. If i push it down to 61, music starts skipping.
whyzor, this kernal... too good. we dont deserve it.
its like cocain infused.
edit: holy moses typing accuracy has shot over 9000!
Hm, that's interesting, because I also use interactive (per your suggestion) and it still does it. Even if I'm just playing music and doing nothing else, it does it. I have to keep my min at 122 to prevent it.You two probably have a min freq locking governor, which I don't recommend. I use 61-1024 interactive and no need to mess with the lowest value, it AUTOMATICALLY scales up to enough speed for whatever load is required. For example if 61 is not enough with no background apps running, it'll automatically bump it up to 122 if I'm downloading something in the background. Locking min frequency sounds good and gives you control, but really unnecessary. Phones with screen off should be in deep sleep. Locking it to a low awake frequency just prevents any apps/service that needs to run from running properly. The correct solution is to disable that app/service from running if that's the intention.
hey so i got a wake locj all night long after the kernal upgrade. how do i check current kernal installed? i can only see current version.
Hm, that's interesting, because I also use interactive (per your suggestion) and it still does it. Even if I'm just playing music and doing nothing else, it does it. I have to keep my min at 122 to prevent it.
WhyKernel-02-08.zip
Uploaded a new kernel (02-08). The occasional "event" wakelock should be fixed. I looked at some other 2.6.37 kernels and found out 2 input events running together caused it (likely accelerometer, magnetometer, compass, microphone or some combination), it was only showing up as 'event2' and 'event4'. Now /proc/wakelocks confirms they're not active anymore.
Also I experimented with removing some Big Kernel Locks (lock_kernel) calls, by modeling after 2.6.37 kernels which have this old method phased out (in favor of smaller scale mutex-locks). It seems to have improved touchscreen reliability. But I've said this before & have been wrong, maybe these are all small incremental improvements, only really noticeable when switching back to a stock kernel.
I'd like to hear if anyone still encounters any of these issues:
1) "event" kernel wakelocks
2) freezing capacitive buttons
3) any other odd behavior that did NOT occur before this kernel (assuming you've ruled out any setting changes or new app installs)
I tried to theme with another theme. That theme turned all of my icons in the icon bar (the bottom of the screen) white, in the shape of the icons. I went back to the default theme. Still white. Any way to fix this?
CPU Spy app will show the build date of the kernel. Or in terminal emulator 'uname -a'.
Yeah that's a bit odd. What kinds of music files and which music player? I just use the old default music app (not the new Google Music) and regular constant bit-rate mp3s. Maybe other apps or codecs require higher CPU.
I have a question for the WhyKernel-02-08.zip. I am trying to install this like I do with every build and that is by going to clock work mod recovery. I tried to install it twice and both times it stated that install was successful but when I go to terminal emulator and type 'uname -a' it states that the kernel is still 2.6. Am I missing something here? Sorry for the noob question. I guess i will download the zip again and try once more.