Welcome to XDA

Search to go directly to your device's forum

Register an account

Unlock full posting privileges

Ask a question

No registration required
Post Reply

[KERNEL][UV][CWM] Entropy's Daily Driver-GB, 03/07/2012 (Small Fixes))

OP Entropy512

28th February 2012, 08:47 PM   |  #1991  
OP Senior Recognized Developer
Flag Owego, NY
Thanks Meter: 24,855
 
13,541 posts
Join Date:Joined: Aug 2007
Donate to Me
More
Quote:
Originally Posted by shinji257

Just to let you know I was running the 2/18 builds (the one actually dated 2/18) and had the weird shutdown issue. Upon turning on the button leds flash real quick then a normal boot. An attempt to get a /proc/last_kmsg failed as it didn't exist. Ideas?

Was this when connected to a charger?

Which 2/18 build? A or B? Were you on 15C before?

crash-to-LPM is pissing me off - it's damn near impossible to get a log.

I'm going to look into maybe putting something in the initramfs to dump last_kmsg in LPM.
28th February 2012, 08:53 PM   |  #1992  
shoman94's Avatar
Recognized Contributor
Flag Maine
Thanks Meter: 12,255
 
8,478 posts
Join Date:Joined: Mar 2009
Donate to Me
More
I want to say that 2/15c is awesome on my rom so far. 2/18b was a step backwards. I know the changes aren't great and I don't understand why the big difference but it is. Finally something better then 12/8/11.
Take it for what it is.....

Sent from my SAMSUNG-SGH-I777 using Tapatalk
The Following User Says Thank You to shoman94 For This Useful Post: [ View ]
Red5
28th February 2012, 08:55 PM   |  #1993  
Guest
Thanks Meter: 0
 
n/a posts
Another thumbs up for 2/15-C, awesome!
28th February 2012, 10:03 PM   |  #1994  
OP Senior Recognized Developer
Flag Owego, NY
Thanks Meter: 24,855
 
13,541 posts
Join Date:Joined: Aug 2007
Donate to Me
More
Quote:
Originally Posted by shoman94

I want to say that 2/15c is awesome on my rom so far. 2/18b was a step backwards. I know the changes aren't great and I don't understand why the big difference but it is. Finally something better then 12/8/11.
Take it for what it is.....

Sent from my SAMSUNG-SGH-I777 using Tapatalk

PLEASE compare 18A to 18B!!!!!! The ability to do so is the whole point of releasing kernels with one patch per release. Also provide more details as to the circumstances where it performs more poorly power-wise.

I'm much more inclined to investigate/rework/revert the patch in 18A than 18B. The 18B patch is from Linux mainline, which means - it's gonna be in ICS whether you like it or not.

Red: Can you confirm again that 26 was the one problematic for you? I've already decided to revert that patch for now and investigate other solutions for Nizda1's issue (if he even has it again, ashmem may have been a red herring) because unlike shoman94, you narrowed it down exactly to a specific patch and you were able to see a night-and-day significant difference. (I don't think I've seen any feedback on 18A from shoman.)
Last edited by Entropy512; 28th February 2012 at 10:08 PM.
28th February 2012, 10:41 PM   |  #1995  
rav4kar's Avatar
Senior Member
Flag Edison, NJ
Thanks Meter: 539
 
1,499 posts
Join Date:Joined: Oct 2011
More
Quote:
Originally Posted by Entropy512

Was Android OS percentage very high along with lots of wlan_rx wakelocks in the kernel wakelock list? If so - this is one of two widely documented bugs with the UCKK6 base.

Entropy,
We can add even XWKL1 base to bug list; in idle state, XWKL1 base drains 10%/ hour on my work wifi with AOS topping >67% with lots of wlan_rx wake locks.
28th February 2012, 10:45 PM   |  #1996  
Senior Member
Thanks Meter: 363
 
1,056 posts
Join Date:Joined: Oct 2011
More
Quote:
Originally Posted by jazzboyrules

Hi Entropy,

These are the logs I got after I kept my phone running on battery for about an hour at home.
logcat, dmesg log (after setting debug_mask = 35 ): http://db.tt/zZVL7NNn
BBS log: http://db.tt/8HJk20yv
About 28k wakelocks in an hour!
I looked for wake_lock_destroyed in dmesg but didn't find it heavily populated with those. What else should I look for in it?

I did not get that many wakelocks at work, but still heavier than usual battery drain.
logcat, dmesg log (with debug_mask = 35): http://db.tt/LwMkZVEb
BBS log: http://db.tt/w9iS8XRX
1698 wakelocks in 38 mins.

At home, I have D-Link DIR-655 router and Microcell.

I will be testing 02/26 overnight. Tomorrow, I will switch back to 02/15-C and post similar logs.

Bump!
Probably this post was missed. It shows that 02/26 was a problem for me.

BBS log of 02/26 overnight with Wi-Fi on.
Staying on 02/26 overnight was a disaster too. Weird.. if I turn off Wi-Fi, I get deleted wake locks and if I keep Wi-Fi on during night, I get wlan_rx wake locks (because of the bug in 2.3.6). Then what's the point of turning off Wi-Fi? Above logs are for 2/26 on SHOstock 1.9.2.

Now I am testing a different ROM on 2/15-C
Last edited by jazzboyrules; 28th February 2012 at 10:52 PM.
28th February 2012, 10:46 PM   |  #1997  
Senior Member
Thanks Meter: 33
 
429 posts
Join Date:Joined: Feb 2010
More
Quote:
Originally Posted by Entropy512

Was this when connected to a charger?

Which 2/18 build? A or B? Were you on 15C before?

crash-to-LPM is pissing me off - it's damn near impossible to get a log.

I'm going to look into maybe putting something in the initramfs to dump last_kmsg in LPM.

I believe it was 2/18b (confirm definitely 2/18b) since the kernel was actually dated 2/18. It was charging via PC USB rather than the standalone adapter. I had skipped 2/15c coming from build 1/30/12.

Sent from my SAMSUNG-SGH-I777 using xda premium
Last edited by shinji257; 28th February 2012 at 10:49 PM.
Red5
28th February 2012, 10:54 PM   |  #1998  
Guest
Thanks Meter: 0
 
n/a posts
Quote:
Originally Posted by Entropy512


Red: Can you confirm again that 26 was the one problematic for you? I've already decided to revert that patch for now and investigate other solutions for Nizda1's issue (if he even has it again, ashmem may have been a red herring) because unlike shoman94, you narrowed it down exactly to a specific patch and you were able to see a night-and-day significant difference. (I don't think I've seen any feedback on 18A from shoman.)

Ya let me go back through both 2/18's and then 26. Give me some time though as I want my phone to use then for a bit to be double sure.
28th February 2012, 11:36 PM   |  #1999  
shoman94's Avatar
Recognized Contributor
Flag Maine
Thanks Meter: 12,255
 
8,478 posts
Join Date:Joined: Mar 2009
Donate to Me
More
Quote:
Originally Posted by Entropy512

PLEASE compare 18A to 18B!!!!!! The ability to do so is the whole point of releasing kernels with one patch per release. Also provide more details as to the circumstances where it performs more poorly power-wise.

I'm much more inclined to investigate/rework/revert the patch in 18A than 18B. The 18B patch is from Linux mainline, which means - it's gonna be in ICS whether you like it or not.

Red: Can you confirm again that 26 was the one problematic for you? I've already decided to revert that patch for now and investigate other solutions for Nizda1's issue (if he even has it again, ashmem may have been a red herring) because unlike shoman94, you narrowed it down exactly to a specific patch and you were able to see a night-and-day significant difference. (I don't think I've seen any feedback on 18A from shoman.)

You haven't seen any feedback on 18a yet because I haven't used it. I can only run one kernel at a time and I need a full day so I can test it. I was planning on flashing 18a tonight.
So again I will be flashing 18a tonight and report at the end of the day tomorrow.
Between 15c and 18b I haven't had any problems but 15c just performed a touch better and was quite a bit better on battery life.

Sent from my SAMSUNG-SGH-I777 using Tapatalk
29th February 2012, 01:48 AM   |  #2000  
Alucardis666's Avatar
Senior Member
Miramar
Thanks Meter: 229
 
1,951 posts
Join Date:Joined: Jul 2009
More
Running the latest release and things seem super stable and battery is much improved. Thanks again.

Post Reply Subscribe to Thread

Tags
cwm, kernel
Previous Thread Next Thread
Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes