[EOL][KERNEL] [OnePlusX] [OOS 3.x.x/2.x.x] Boeffla-Kernel (12-06-2017)

[Naman Bhalla]

Thanks a lot Andi for your Hard work for the OnePlus X. I finally received my OPX today. Already started working on my separate profiles for all 3 kinds of usage.

BTW Boeffla Kernel Profiles Thread is UP :-
http://forum.xda-developers.com/showthread.php?t=3330187

Rain of profiles starting soon

Sent from my ONE A2003 using Tapatalk

 

[nasif69]

Which color?

 

[andip71]

Result of recent poll about displaying actual charge current in Boeffla-Config.

https://twitter.com/andip71/status/706758503717847040

In short - feature will be back in next updates, however slightly modified to indicate the information is potentially very inaccurate.

Andi

 

[taaeem]

Okay the kernel is really good and I get really good battery life, but for some reason it seems like if I leave my phone at the charger overnight that the battery percentage hangs on 100% for a while (1-2 hours) and then drops into the 80s.
Never had these issues with Stock kernel or blu_spark.
I'm running OOS 2.2.0-X with Xposed, greenify, underclocked, undervolted.

 

[andip71]

Okay the kernel is really good and I get really good battery life, but for some reason it seems like if I leave my phone at the charger overnight that the battery percentage hangs on 100% for a while (1-2 hours) and then drops into the 80s.
Never had these issues with Stock kernel or blu_spark.
I'm running OOS 2.2.0-X with Xposed, greenify, underclocked, undervolted.

Not an "issue" really. It will sort itself after 2-3 charge cycles. This is sometimes the case after flashing something (and destroying the battery calibration).

At least I do not have that here anymore (had it once, as said).

Also - there is zero changed that could impact battery capacity control in my kernel.

And just by the way - xposed + undervolted = you are out of my support, see OP.

Andi

 

[andip71]

Update

Hi all,

new kernel update for you:

Version 4.0-beta7
07.03.2016

• Note: This kernel only runs on Oxygen OS 2.2.0-x for the OnePlus X - nothing else !!!
  
• Note: You need at least version 2.2.34 of Boeffla-Config V2 to allow all functionalities to be available
• Change: double tap 2 wake without haptic feedback does properly use proximity sensor now (avoid unwanted wake ups)
• Change: swipe 2 wake gestures are now following the haptic feedback setting in Boeffla-Config as well
• Change: Actual charge current is displayed again in Boeffla-Config, however prefixed with a tilde (~) sign to indicate it is very inaccurate
• Fix: xbox module not loading correctly is fixed
• New: Kexec hard boot implemented (untested as I have no use for it, maybe it works, maybe not)
• Update: Internal kernel superuser updated to su 2.65

Enjoy
Andi

 

[hellcat50]

Hi all,

new kernel update for you:

Version 4.0-beta7
07.03.2016

• Note: This kernel only runs on Oxygen OS 2.2.0-x for the OnePlus X - nothing else !!!
  
• Note: You need at least version 2.2.34 of Boeffla-Config V2 to allow all functionalities to be available
• Change: double tap 2 wake without haptic feedback does properly use proximity sensor now (avoid unwanted wake ups)
• Change: swipe 2 wake gestures are now following the haptic feedback setting in Boeffla-Config as well
• Change: Actual charge current is displayed again in Boeffla-Config, however prefixed with a tilde (~) sign to indicate it is very inaccurate
• Fix: xbox module not loading correctly is fixed
• New: Kexec hard boot implemented (untested as I have no use for it, maybe it works, maybe not)
• Update: Internal kernel superuser updated to su 2.65

Enjoy
Andi

This kernel works on 2.2.0 too (already said that)

2.2.0 and 2.2.0-x

 

[taaeem]

Mmh is it governor and governor settings depended how much I can undervolt? On zzmove battery extreme yank I can go like - 100 without any impact on performance or hot reboots. On blu_spark and blu_active at - 100 it just reboots and with - 50 it's unstable. So is it kernel or governor depended?

 

[ZaneZam]

Mmh is it governor and governor settings depended how much I can undervolt? On zzmove battery extreme yank I can go like - 100 without any impact on performance or hot reboots. On blu_spark and blu_active at - 100 it just reboots and with - 50 it's unstable. So is it kernel or governor depended?

it is uv dependent! the moment u change the stock voltages things r getting to be a lottery game. so everything is possible. mostly "everthing" means "nothing good"
do u know that u r a lucky one regarding zzmoove? as normally it's vice versa u can't do much uv with zzmoove (due to it's fast freq switching) and more with other govs.
but this "less and more" is anyway and always in the "red zone". in short uv = expect everything

 

[MarocoKong]

it is uv dependent! the moment u change the stock voltages things r getting to be a lottery game. so everything is possible. mostly "everthing" means "nothing good"
do u know that u r a lucky one regarding zzmoove? as normally it's vice versa u can't do much uv with zzmoove (due to it's fast freq switching) and more with other govs.
but this "less and more" is anyway and always in the "red zone". in short uv = expect everything

Don't make yourself smaller than you are
From what I can say after only 3 days of testing this seems to be a very very great governor! Here to with zzmoove game-profile plus native zzmoove-hotplug and -75 mV superstable and and fluid phone and almost obscenely battery friendly too. Got to -75 with Blu_Spark as well though - but it ate more battery...
So huge thanks to you for the governor and The Lord for putting it into this wonderfull Kernel!
Cheers

 

[CertifiedBlyndGuy]

Don't make yourself smaller than you are
From what I can say after only 3 days of testing this seems to be a very very great governor! Here to with zzmoove game-profile plus native zzmoove-hotplug and -75 mV superstable and and fluid phone and almost obscenely battery friendly too. Got to -75 with Blu_Spark as well though - but it ate more battery...
So huge thanks to you for the governor and The Lord for putting it into this wonderfull Kernel!
Cheers

I think it needs to be pointed out that an undervolt won't have anything to do with other kernel settings but rather the limitation of hardware and how far that hardware is being pushed. Zzmoove seems to be better about switching frequencies than most other governors which is why you and @taaeem are hitting walls much earlier. If a UV is stable with one kernel and unstable with another it's not the kernel, it's your undervolt. For example, on Blu Spark I was able to maintain a 925mv undervolt for the maximum frequency for a while however Blue Active seemed to stick with lower frequencies most of the time and spike up only when needed which was a rarity. However due to Zzmoove and it's great scaling it hits those higher frequencies for much longer. Now that same frequency required 975mv to operate safely. Because of this, I now know what settings truly work. When resting frequencies, try to test them individually with the performance governor. Lock your maximum clockspeed to the frequency on question and your good to go.

 

[MarocoKong]

I think it needs to be pointed out that an undervolt won't have anything to do with other kernel settings but rather the limitation of hardware and how far that hardware is being pushed. Zzmoove seems to be better about switching frequencies than most other governors which is why you and @taaeem are hitting walls much earlier. If a UV is stable with one kernel and unstable with another it's not the kernel, it's your undervolt. For example, on Blu Spark I was able to maintain a 925mv undervolt for the maximum frequency for a while however Blue Active seemed to stick with lower frequencies most of the time and spike up only when needed which was a rarity. However due to Zzmoove and it's great scaling it hits those higher frequencies for much longer. Now that same frequency required 975mv to operate safely. Because of this, I now know what settings truly work. When resting frequencies, try to test them individually with the performance governor. Lock your maximum clockspeed to the frequency on question and your good to go.

Thanks for your reply. Just to clarify, I got -75 mv undervolt stable in both boeffla and blu_spark. Used Kernel Adiutor for blu_spark which allows 5 mV steps instead of 25 mV in Boeffla Config so there I could get to - 85 stable (Actually -95 gave some seldom forced restarts when waking up from deep sleep and directly doing something demanding, so I gave it a little buffer ).
I didn't even try -100 on boeffla. With zzmoove game and hotplug it's so battery friendly anyways I didn't see the need to push to the absolute limit.
Makes sense though, that it's not too much Kernel related apart from "how far the hardware is pushed" by governors.
Anyways, wasn't in any way intended as a complaint rather to praise you two for the great work!
Cheers

 

[taaeem]

Thanks for all the responses, can somebody tell me what the stock voltages without undervolting are? I need to check something

 

[andip71]

Thanks for all the responses, can somebody tell me what the stock voltages without undervolting are? I need to check something

No, there are no single stock voltages, as this is dependent on your specific SoC quality.

Simplest way is: In boeffla config select the "no undervolting" pattern, then you get the stock values back and you can see them.

Andi

 

[Naman Bhalla]

Ok. So some talks regarding undervolting were going on... So I thing I should share my experience with it on OPX too. While working for my profile, I was finding the best possible undervolting combo on OPX... To my surprise, I was able to achieve voltage of even 600mv (-250mv) for Frequencies upto 768 or so (Not using OPX as I type). After that the stable Frequency for the next was 627mv and it went till 1011mv (Only some 20mv undervolt) for the highest Frequency . This is the maximum possible undervolting on an OPX. You can't ever go beyond it...
Now the thing is how it can be used effectively. In the first impressions using Perfmon, I found that the kernel with stock settings touches 4 Frequencies in its cycle (mainly). These are the lowest, the highest, 7xx and 14xx. So till 14xx part, it works perfectly. When it jumps from that to the highest (I tried 19xx as highest), a reboot happens for sure. So, in case anyone is working on a heavy undervolting profile, just to make stability, tune your profile in such a way such that it touches atleast 8 Frequencies in its cycle (if you are taking 300 to 25xx).Then you will achieve a great stability with very awesome battery life too.

 

[CertifiedBlyndGuy]

Ok. So some talks regarding undervolting were going on... So I thing I should share my experience with it on OPX too. While working for my profile, I was finding the best possible undervolting combo on OPX... To my surprise, I was able to achieve voltage of even 600mv (-250mv) for Frequencies upto 768 or so (Not using OPX as I type). After that the stable Frequency for the next was 627mv and it went till 1011mv (Only some 20mv undervolt) for the highest Frequency . This is the maximum possible undervolting on an OPX. You can't ever go beyond it...
Now the thing is how it can be used effectively. In the first impressions using Perfmon, I found that the kernel with stock settings touches 4 Frequencies in its cycle (mainly). These are the lowest, the highest, 7xx and 14xx. So till 14xx part, it works perfectly. When it jumps from that to the highest (I tried 19xx as highest), a reboot happens for sure. So, in case anyone is working on a heavy undervolting profile, just to make stability, tune your profile in such a way such that it touches atleast 8 Frequencies in its cycle (if you are taking 300 to 25xx).Then you will achieve a great stability with very awesome battery life too.

Every chip is binned differently. For instance mind won't do below 660mv however the maximum frequency is only using 1035mv. Your undervolts may not work with others devices.
@Lord Boeffla When using a schedular such as deadline or ZEN for external storage, internal storage switches over to ROW. I have no idea why. I've tried both fiops and SIO for internal but it's the same story. I'm sure that there is a simple explanation for this. It's probably a stupid thing to bring up in the first place

---------- Post added at 01:02 AM ---------- Previous post was at 12:47 AM ----------

Investigated the issue further and it's with every governer. No matter what internal storage defaults to ROW...

 

[Lazy Rich]

Do I have a super chip? I've set - 100mV on all and it runs totally stable, tried tons of apps and games. Max frequency runs with only 970mV, min with 725, maybe I can go a bit deeper with that.

I really love this chipset

 

[andip71]

@Lord Boeffla When using a schedular such as deadline or ZEN for external storage, internal storage switches over to ROW. I have no idea why. I've tried both fiops and SIO for internal but it's the same story. I'm sure that there is a simple explanation for this. It's probably a stupid thing to bring up in the first place

---------- Post added at 01:02 AM ---------- Previous post was at 12:47 AM ----------

Investigated the issue further and it's with every governer. No matter what internal storage defaults to ROW...

Hi,

no, there is definitely no simple explanation for this. It should not happen.

I am running on noop for both int and ext memory. And it sticks.

So next questions are:
- when does internal switch to row? after a reboot, or after some time running etc?
- are you sure you have not allowed scheduler overwrite in Boeffla-Config settings?
- anyone else seeing schedulers not sticking? (which should never happen as I lock them down in Boeffla-Config so nobody else can change them)

Andi

 

[CertifiedBlyndGuy]

Hi,

no, there is definitely no simple explanation for this. It should not happen.

I am running on noop for both int and ext memory. And it sticks.

So next questions are:
- when does internal switch to row? after a reboot, or after some time running etc?
- are you sure you have not allowed scheduler overwrite in Boeffla-Config settings?
- anyone else seeing schedulers not sticking? (which should never happen as I lock them down in Boeffla-Config so nobody else can change them)

Andi

1. After reboot.
2. I didn't have this enabled, but after enabling it settings stuck... weird (should have had the opposite affect)

I have no idea what it could've been.

BTW, when compiling from source for NetHunter support I only had to apply one patch and modify the defconfig... Your build script handled everything else. Thanks for making such a great kernel and providing clean code

 

[andip71]

1. After reboot.
2. I didn't have this enabled, but after enabling it settings stuck... weird (should have had the opposite affect)

I have no idea what it could've been.

BTW, when compiling from source for NetHunter support I only had to apply one patch and modify the defconfig... Your build script handled everything else. Thanks for making such a great kernel and providing clean code

Yep, this does not make any sense, so let's investigate it.

Can you please create a debug info file in Boeffla-Config (it is last poin in Action menu) and send it to [email protected]

I want to see what happens, please untick the overwrite option again and reboot before you create the file.

Thanks
Andi