Development Goals:
- stability
- energy savings due to more efficient algorithms (whether theoretical or not is unimportant)
- strictly no overclocking unless approved by the manufacturer or my source base integrates it (also, even if my source base integrates it, expect no support for it)
- no undervolting as well unless the manufacturer approves it since it's relatively pointless IMHO...
- all improvements should require MINIMAL user interaction (e.g. you don't need to do anything except flash the kernel or at the very least use SetCPU or the like to set fixed options)
- stability
*I've not uploaded versions for 4.2.1 stock.
**This section contains the STABLE kernels (i.e. tested to be stable for at least a day or two with no problems) for TEST kernels, proceed to the 3rd post.
Latest Kernels Here
20130227_11XX
- reverted zram and zcache backports to hopefully fix error
- BFQ v6
- more optimizations
20130224_17XX:
- more optimization
20130217_08XX:
- includes updates made in 4.2.2
*please note that zcache/zram does not work properly and may cause a restart
20130210_17XX:
- new optimization
20130204_17XX:
- rebased everything
- has more improvements specifically to ARM and some upstream commits
20130111_14XX:
- updated compiler to December version
- updated "random" stuff
20130101_15XX:
- added optimized swahb32 helper
- added ticket spinlock implementation for ARMv6+
- Proportional Rate TCP algorithm
20121231_12XX:
- new compiler flag
- USB OTG crash fix
20121228_17XX:
- with ROW IO scheduler
- allow use of highmem for binder buffer
- reverted backport latest interactive cpu governor from my N4 test kernel
- integrated kexec-hardboot patch
- stock JB support
20121222_11XX:
- dynamic sync control from faux123
- integrated ARM optimized RWSEM algorithm
- added the 1MB readahead optimization
- integrated BFQ v5.1
- removed alignment trap disabling since it seems to cause some kernel faults in certain situations
20121213_08XX:
- disabled cleancache since it eventually consumes all memory which causes a kernel fault as well
- replace lib/string and lib/memcpy with GNU implementation
20121211_20XX:
- enabled cleancache
- uses Linaro compiler 2012.11 release
20121204_22XX:
- updated LZO compression to upstream version
- replace do_fsync call in fs/sync.c to fget_light
- added ARM optimized AES and SHA1 routines
- tweak to the automated scheduler grouping algorithm
20121202_09XX:
- asynchronous compaction in lowmemorykiller driver
20121112_08XX-22XX:
- initial release
- uses Linaro compiler 2012.10 release
- compiler flags for tuning
- software alignment trap disabled (hardware supports unaligned accesses which is faster)
- added ARM CPU topology
- updated zcache, zram, and binder to 3.3 version
- updated lowmemorykiller to 3.4 version
You can find my kernels at:
intersectRaven's Kernels
GitHub is at:
intersectRaven's GitHub
Special Thanks To:
Steve Kondik - CyanogenMod
faux123 - some patches I integrated are from his repo
franciscofranco - some patches I integrated are from his repo
TEST Kernel Section:
- THESE AREN'T GUARANTEED AS STABLE! USE AT YOUR OWN RISK!
- please report problems when using in a comment in this thread with attached kmsg/last_kmsg if possible
Latest TEST Kernels Here
20130310_16XX:
- workqueue fix due to hashtable optimization
20130305_18XX:
- added lzo unaligned access optimization from linux-next plus
20130305_10XX:
- workqueue and ksm now use the new hashtable implementation
- new scatter library optimization
20130304_21XX:
- additional optimization
20130302_14XX:
- additional optimization flags
Great to see your skills applied to the N7 IR.
Many Thanks!
Sent from my Nexus 7 using xda premium
welcome aboard..will await version for 4.2.1
demandarin said:
welcome aboard..will await version for 4.2.1
Click to expand...
Click to collapse
+1, would appreciate multirom kexec compatibility as well.
bonzer2u said:
+1, would appreciate multirom kexec compatibility as well.
Click to expand...
Click to collapse
I'm not familiar with that. Can you point me to what you mean?
intersectRaven said:
I'm not familiar with that. Can you point me to what you mean?
Click to expand...
Click to collapse
People want this in every kernel
http://forum.xda-developers.com/showthread.php?t=2011403
Also, Welcome aboard. It's always nice to have new stuff from familiar names (i stalked the kindle fire forums before buying a n7)
I'll definitely be trying this out soon
Sent from my Infected ViperVivo
intersectRaven said:
I'm not familiar with that. Can you point me to what you mean?
Click to expand...
Click to collapse
You can find more info at this link
markj338 said:
People want this in every kernel
http://forum.xda-developers.com/showthread.php?t=2011403
Also, Welcome aboard. It's always nice to have new stuff from familiar names (i stalked the kindle fire forums before buying a n7)
I'll definitely be trying this out soon
Sent from my Infected ViperVivo
Click to expand...
Click to collapse
rifraf1 said:
You can find more info at this link
Click to expand...
Click to collapse
Thanks! I'll compile a test kernel with the needed changes and replace the test kernel with one with kexec hardboot support.
intersectRaven said:
Thanks! I'll compile a test kernel with the needed changes and replace the test kernel with one with kexec hardboot support.
Click to expand...
Click to collapse
I see you already got the technical info needed and look forward to your next release, thanks.....
bonzer2u said:
I see you already got the technical info needed and look forward to your next release, thanks.....
Click to expand...
Click to collapse
It's already done in the TEST kernel section. I haven't tested the kexec functionality if it works though. I just followed the instructions and integrated the patch. Test if you know what you're doing.
intersectRaven said:
It's already done in the TEST kernel section. I haven't tested the kexec functionality if it works though. I just followed the instructions and integrated the patch. Test if you know what you're doing.
Click to expand...
Click to collapse
Im running stock JB 4.2.1 so I assume that the 1226 test kernels are still only for CM10 and 10.1 accordingly.
bonzer2u said:
Im running stock JB 4.2.1 so I assume that the 1226 test kernels are still only for CM10 and 10.1 accordingly.
Click to expand...
Click to collapse
I see...hmmm...I'll try to get a stock ramdisk sometime this week for stock.
intersectRaven said:
I see...hmmm...I'll try to get a stock ramdisk sometime this week for stock.
Click to expand...
Click to collapse
And maybe everything to use your great kernel with rasbean jelly's rom.please.
Thanks it will works on stock ROM 4.2.1?
Sent from my Asus Nexus 7 3g 32GB
denzel09 said:
Thanks it will works on stock ROM 4.2.1?
Sent from my Asus Nexus 7 3g 32GB
Click to expand...
Click to collapse
Not yet. CM based ROMs only for now.
Sent from my Nexus 4 using xda app-developers app
intersectRaven said:
Not yet. CM based ROMs only for now.
Sent from my Nexus 4 using xda app-developers app
Click to expand...
Click to collapse
Thanks.
Will be there overclock for stock ROM?
Sent from my Asus Nexus 7 3g 32GB
Anyone experiencing sleep issues with the latest TEST kernel? I've had to long-press reboot my N7 a couple of times and reverted to my previous TEST kernel without the interactive patches.
*just need verification if anyone else experienced it...
intersectRaven said:
Anyone experiencing sleep issues with the latest TEST kernel? I've had to long-press reboot my N7 a couple of times and reverted to my previous TEST kernel without the interactive patches.
*just need verification if anyone else experienced it...
Click to expand...
Click to collapse
Yesterday i got my new N7 and installed directly ur kernel.
As u wrote, i have experienced a SoD yesterday while using BFQ, with CFQ none until now (might check later).
If you need some kind of last_kmsg for debugging just tell me (if your kernel has the dmesg/kmsg active)
PS: using latest beta with PA 4.2.1 Beta9
Inviato dal mio Nexus 7 con Tapatalk 2
Related
This is my own personally compiled kernel based on the latest kernel from Cyanogen's Github repository with Kmobs' undervolt modifications, CodeAurora's AVS code, pershoot and rotohammer's audio gain mod and several compiler optimizations based on initial idea from psyq.
Only major releases will be advertised here.
All changes since 05/05 can be found at my Euroskank host:
http://intersectraven.euroskank.com/kernels/
*Thanks to RyanMacG for the free hosting!
Old uploads with minor changes can be found at my MediaFire folder or my Bitpad folder:
http://www.bitpad.co.uk/intersectraven
http://www.mediafire.com/intersectRaven
Major features:
- based on latest Cyanogen Mod kernel source from his GitHub repository
- numerous compiler optimizations with a custom compiler by redstar3894
- all CPU power governors for user dependent tweaking of power saving method
- Hybrid AVS (Adaptive Voltage Scaling combined with Static Voltage Scaling) support for maximum possible power savings dependent on CPU requirements and a customizable version (CAVS) for people who like to tweak how far their N1s can go
- universal update.zip template made by Koush
Instructions:
1.) Reboot to recovery and flash the update.zip directly.
OR
Instructions for zImage and bcm4329.ko driver extracted from the update.zip(from command line):
1.) adb remount
2.) adb push bcm4329.ko /system/lib/modules
3.) adb reboot bootloader
4.) fastboot flash zimage zImage
5.) fastboot reboot
OR
Use ADB GUI by minooch found here:
http://forum.xda-developers.com/showthread.php?t=666964
*please note the instructions...push the wifi driver BEFORE rebooting for flashing zImage...if your wifi is turned on when you reboot before you pushed the wifi driver for the kernel, there is a chance that you will go into a bootloop due to the incompatible wifi driver!
Changelog:
20120108_2143:
- just merged pershoot's commits
20111203_11XX:
- enabled MSM EHCI
20111114_23XX:
- integrated CM's commits (mainly bluetooth and WiFi fixes)
20111111_19XX:
- compiled using updated Mjolnir/Linaro compiler hybrid (having problems with our Mjolnir GCC)
- enabled SYN_COOKIES as requested
- some tweaks to the VFS settings
- switched network scheduler to SFB
- switched TCP congestion to Veno from YeaH (seems it's better for devices with a greater chance of random drops of packets)
20111010_11XX:
- disable CleanCache for YAFFS (too complex to change)
- more proper reapplication of changes from 3.0
20111009_22XX:
- enabled CleanCache for YAFFS, EXT3 and EXT4 (experimental)
20111008_23XX:
- ported BogoMIPS calibration from 3.0
- added CleanCache from 3.0
- switched to SIO from BFQ
- block IO batching from 3.0
- activate_pages batching from 3.0
20110904_07XX:
- added SmartAssV2 CPU governor
20110828_13XX:
- fix AVS to actually work (see previous latest post for apology... )
*this should restore the instability on some devices that can't handle AVS
20110828_00XX:
- prevent excessive suspend attempts
- optimized sha1 implementation
20110813_15XX:
- increased NAND buffer to 8k similar to codeaurora's version
- enabled 8-bit transfers for when the MMC card supports it
20110724_20XX:
- SmartAss improvement from the test kernels
- AVS code improvement also from the test kernels (hopefully improved the stability)
- RCU optimizations
- updated Mjolnir compiler
20110713_10XX:
- re-hauled SmartAss governor:
* interactive threads instead of workqueues to improve responsiveness when ramping-up frequencies
* reduced stepping frequency to use lower frequencies more
20110707_08XX:
- fixed the Voice Search problem
20110706_09XX:
- code cleanup
- additional tweaks
20110628_09XX:
- ARM improvements to memcpy and memmove operations from Wildfire (arco's kernel source)
- cherry-picked serial number commit from CM kernel
20110608_21XX:
- rebased everything together with removal of worthless commits
- added 2 new governors from SavagedZen kernel (SavagedZen & InteractiveX)
- updated code of Smartass governor to the one in SavagedZen since it seems more updated than the one I found
20110527_21XX:
- new method for addressing slow writes on USB from CodeAurora (although it's still slow using native USB mount and I didn't test using another mounter)
- some SIRC potential bug fixes
- input event handling modification from Google
20110523_08XX:
- compiled using updated Mjolnir
20110522_11XX:
- upgraded to v2.6.38.7
20110521_15XX:
- interrupt masking
- smd_tty buffer limit implementation
20110519_22XX:
- fix for potential bug and power leak improvement in DSP driver
- GPIO tweaks
20110516_19XX:
- increased DMA zone to 14MB (may speed some things or may not)
- timer workarounds have been removed as they're unneccessary on Scorpion
- prevent reading from write-only registers (just silly)
- used relaxed access functions for some functions
- remove extra interrupts sent from the SMD channel
20110515_21XX:
- added another commit from android unmerged which implements a watchdog to catch lockups during device resume
- fix for wakelocks which addresses the problem where while being connected to a computer, any attempt to power up will result in display immediately shutting off with touchscreen buttons still on
- uses an updated Mjolnir compiler
20110514_17XX:
- rebased everything
- removed some commits which were useless on the N1
- more zen branches merged
- WiFi-Fast patch has been integrated in all kernels since it seems to have no effect on battery (no more separate WiFi-Fast release)
20110511_22XX:
- reverted a change made to PMEM driver since the commit it was reliant to was reverted (sorry! didn't notice this...I wasn't too critical of my earlier cherry-picks... )
Link to a file which contains all kernels:
http://hotfile.com/dl/117478444/390f688/intersectR_-_20110511_22XX.zip.html
20110510_18XX:
- updated to 2.6.38.6
- committed some more video driver commits from CodeAurora
Link to a file which contains all kernels:
http://hotfile.com/dl/117350839/0a9a504/intersectR_-_20110510_18XX.zip.html
20110509_14XX:
- merged some commits from Android repositories that were still unmerged yet may prove useful for Ashmem and RPC
Link to a file which contains all kernels:
http://hotfile.com/dl/117209820/5b7fe7b/intersectR_-_20110509_14XX.zip.html
20110507_21XX:
- even more improvements from CodeAurora
Link to a file which contains all kernels:
http://hotfile.com/dl/117048553/0831385/intersectR_-_20110507_21XX.zip.html
20110506_15XX-16XX:
- enabled cache error reporting as this is indicative of how tolerant your N1 is to AVS undervolting
- smartass governor (from Temasek)
*this was mistakenly included in the previous release
- a minor kernel scheduling statistic commit
Link to a file which contains all kernels:
http://hotfile.com/dl/116922979/5fb2ec0/intersectR_-_20110506_15XX-16XX.zip.html
20110505_16XX-17XX:
- updated Mjolnir compiler
Link to a file which contains all kernels:
http://hotfile.com/dl/116827854/5839b43/intersectR_-_20110505_16XX-17XX.zip.html
20110505_08XX-09XX:
- AVS and CAVS now both allow changing of voltages on-the-fly. The only difference now is that AVS is undervolted by default while CAVS is undervolted to the same voltages that CM uses in his SVS kernel
- uses eviollet's on-the-fly voltage modification system instead of the previous one I had which is a lot more flexible
Link to a file which contains all kernels:
http://hotfile.com/dl/116798881/0d5b5bf/intersectR_-_20110505_08XX-09XX.zip.html
20110503_09XX:
- updated to version 2.6.38.5
Link to a file which contains all kernels:
http://hotfile.com/dl/116600585/b76a4b6/intersectR_-_20110503_09XX.zip.html
20110502_10XX:
- synced with pershoot's latest modifications which mirror CM's latest addition with regards to USB accessory function (not too important I think since it seems to be for future ADB use)
- uses an updated Mjolnir compiler (20110429)
Link to a file which contains all kernels:
http://hotfile.com/dl/116528049/1865aa6/intersectR_-_20110502_10XX.zip.html
20110424_11XX:
- reverted WiFi driver to same version CM uses for mainline kernel to fix channel 11 issues with the newest one
20110422_12XX:
- updated to 2.6.38.4
- compiler updated
20110421_15XX-16XX:
- first files to be hosted by Bitpad (http://www.bitpad.co.uk/intersectraven/)
*Thanks to MajorProbes
- just a minor release since I only updated the compiler
20110416_20XX:
- several compiler optimizations enabled (loop unrolling, peeling, etc.)
- zen-kernel cherry-picks for memory and fs optimization
20110415_23XX:
- updated to 2.6.38.3
- compiled using latest Mjolnir with an experimental merge by redstar
20110409_23XX:
- integrated a bluetooth fix and MMC quirks improvement from official Google repositories
20110407_13XX-14XX:
- integrate CM commits on futex optimization and removal of dodgy optimizations
20110403_12XX:
- fix for USB transfer speed (should now hold at 1MB/s without dropping)
20110401_22XX-23XX:
- first 2.6.38.2 release based on pershoot's 2.6.38
- added the usual mix (AVS, SLQB, CodeAurora patches, etc.)
- compiled using Mjolnir GCC 4.6.1
- changed FPU optimization to NEON
20110328_08XX-09XX:
- updated to 2.6.37.6
20110328_07XX:
- compiled using Mjolnir GCC4.6.0
- enabled Link Time Optimization and Graphite Optimization (use Google for definitions)
20110326_08XX:
- updated BFQ to v2-r1
20110324_18XX:
- updated to 2.6.37.5
20110320_17XX:
- merged Nick Piggin's RCU patches which were originally for 2.6.38 (one of the things Linux was excited about according to Phoronix)
20110319_09XX:
- merged latest CM kernel commits which enables the ff:
- enabled RCU boost
- enabled touchscreen filter (reduce CPU load made by touchscreen)
20110315_22XX:
- updated to 2.6.37.4
20110312_23XX:
- round 2 of CM's wonk fix attempt integrated
- toolchain update
*for links, go to my MediaFire folder as specified above
20110311_1623:
- integrated cyan's wonk fix attempt
- VPN "fix" (I don't like this one since it's just a backport of the old PPP interfaces)
*for links, go to my MediaFire folder as specified above
20110308_2246:
- updated to .37.3
- based on CM's latest kernel source
- with SLQB and BFQ v2
- regular and customizable AVS
- some CodeAurora patches
*for links, go to my MediaFire folder as specified above
20110213_1506:
- corrected minimum voltage value to 800mV
CFS-HAVS-CM7-NOBOOST -> http://www.mediafire.com/?m32mi1744ksb55m
20110213_1035:
- test release for new AVS-CUSTOMIZEABLE build which allows for runtime customization of AVS minimum and maximum limiters for more flexible AVS voltages depending on your CPU tolerance (/sys/module/avs/parameters/avs_adjust)
- AVS debugging outputs can also be toggled in runtime (/sys/module/avs/parameters/avs_debug -> set to 0 to not display, 1 to display)
CFS-HAVS-CM7-NOBOOST -> http://www.mediafire.com/?bt7mmtlmg407ne1
*format for avs_adjust is:
frequency,minimum voltage,maximum voltage
e.g.
echo 245000,925,975 > avs_adjust
**AVS debugging output will be enabled by default when you modify the limits
Finally created a github to store all of my kernel modifications:
http://github.com/intersectRaven/
To follow me for updates on Twitter:
http://www.twitter.com/intersectRaven
[Kernel] [.35.10, .37] intersectRaven's Kernel (HAVS-AXI-FM-720p-Zen)1/9/2011 10:13
Reserved for intersectRaven - Added 2post to allow OP to have more place for futur update
Finally a thread!
been using your kernels for some time now, good work!
---------------------------------------------------------------
Updates: I thought I'll use this space to provide some info since its right next to the OP
BFS:
BFS is the Brain **** Scheduler. It was designed to be forward looking only,
make the most of lower spec machines, and not scale to massive hardware. ie
it is a desktop orientated scheduler, with extremely low latencies for
excellent interactivity by design rather than "calculated", with rigid
fairness, nice priority distribution and extreme scalability within normal
load levels.
http://ck.kolivas.org/patches/bfs/bfs-faq.txt
CFS
Completely Fair Scheduler is the name of a task scheduler which was merged into the 2.6.23 release of the Linux kernel. It handles CPU resource allocation for executing processes, and aims to maximize overall CPU utilization while maximizing interactive performance.
http://en.wikipedia.org/wiki/Completely_Fair_Scheduler
Comparison of BFS vs CFS: http://www.cs.unm.edu/~eschulte/data/bfs-v-cfs_groves-knockel-schulte.pdf
Conclusion
The results indicate that CFS outperformed BFS with minimizing turnaround time but that BFS
outperformed CFS for minimizing latency. This indicates that BFS is better for interactive tasks
that block on I/O or user input and that CFS is better for batch processing that is CPU bound.
Click to expand...
Click to collapse
can i use this kernel along with the desire camera app?
britoso said:
Finally a thread!
been using your kernels for some time now, good work!
Click to expand...
Click to collapse
Had my dog press the enter key to prevent me from chickening out.
jblazea50 said:
can i use this kernel along with the desire camera app?
Click to expand...
Click to collapse
I haven't tested it so I don't know if it needs something from the kernel to work.
@intersectRaven - in the http://forum.xda-developers.com/showthread.php?p=6003800#post6003800Audio Mod thread rotohammer incorporated settings that increased both the bluetooth and in-call/earpiece volume. is your audio fix for a headset/bluetooth headset and/or does it incorporate higher values for in-call volume as well? thanks!
cortez.i said:
@intersectRaven - in the http://forum.xda-developers.com/showthread.php?p=6003800#post6003800Audio Mod thread rotohammer incorporated settings that increased both the bluetooth and in-call/earpiece volume. is your audio fix for a headset/bluetooth headset and/or does it incorporate higher values for in-call volume as well? thanks!
Click to expand...
Click to collapse
Hmmm...thanks for pointing that out. I'll probably release a new version later which will have this. (loading my Nexus Compilation Environment VM now...)
*Edit: I recompiled with the .min and .max settings he specified but reading further back it seems he changed some other things so I can't be sure if just this will provide the volume increase desired.
cortez.i said:
@intersectRaven - in the http://forum.xda-developers.com/showthread.php?p=6003800#post6003800Audio Mod thread rotohammer incorporated settings that increased both the bluetooth and in-call/earpiece volume. is your audio fix for a headset/bluetooth headset and/or does it incorporate higher values for in-call volume as well? thanks!
Click to expand...
Click to collapse
Hmmm...thanks for pointing that out. I'll probably release a new version later which will have this. (loading my Nexus Compilation Environment VM now...)
*Edit: I recompiled with the .min and .max settings he specified but reading further back it seems he changed some other things so I can't be sure if just this will provide the volume increase desired.
intersectRaven said:
I haven't tested it so I don't know if it needs something from the kernel to work.
Click to expand...
Click to collapse
It works. I thought you incorporated the audio fix yesterday too!
britoso said:
It works. I thought you incorporated the audio fix yesterday too!
Click to expand...
Click to collapse
I just compiled one with higher settings based on rotohammer's research. I think I'll stick with the default ones in my previous version though so its okay to skip this latest one.
For those who wanted to try out the modified values:
http://www.mediafire.com/?dqynw1inh3y <= 2.6.32
http://www.mediafire.com/?mmjwxzjlmmw <= 2.6.33
I'm working on a mod that will allow us to change the min_vol and max_vol from a file. Kernel SOP dictates this not advised, but I don't have time to learn how to implement a procfs hook.
I also used the voltages specified by kmobs so there should be no difference with the voltages. .32 is from the android kernel source while .33 is from cyanogen's. .32 is the stable kernel while .33 is still in its experimental stage.
Thanks a lot I've been looking for something like this!
intersectRaven, any chance to release Overclock kernel ?
1267Mhz on the nexus one will be great. hahahaha
rheza02 said:
intersectRaven, any chance to release Overclock kernel ?
1267Mhz on the nexus one will be great. hahahaha
Click to expand...
Click to collapse
haha...I'll think about it...I don't have the courage to try an OCed version though so you'll have to test it yourself if ever!
intersectRaven said:
I just compiled one with higher settings based on rotohammer's research. I think I'll stick with the default ones in my previous version though so its okay to skip this latest one.
For those who wanted to try out the modified values:
http://www.mediafire.com/?dqynw1inh3y <= 2.6.32
http://www.mediafire.com/?mmjwxzjlmmw <= 2.6.33
Click to expand...
Click to collapse
i'm currently using MoDaCo Alpha 19, which reports using a 2.6.33.1 kernel. so i would use the 2.6.33 version, correct? just to confirm, are your volume values the same as used by rotohammer? also, where is the zImage stored on my device, that is if it's actually saved at all (yes, i am a newbie, lol). thanks!
cortez.i said:
i'm currently using MoDaCo Alpha 19, which reports using a 2.6.33.1 kernel. so i would use the 2.6.33 version, correct? just to confirm, are your volume values the same as used by rotohammer? also, where is the zImage stored on my device, that is if it's actually saved at all (yes, i am a newbie, lol). thanks!
Click to expand...
Click to collapse
yes u would use the .33 if ever...the links u quoted does use rotohammer's values...and lastly, the zimage is flashed onto the image/kernel area in your device's boot partition...oh and no shame being a noob...we were all noobs once...
This kernel crash on my nexus one with desire camera when I try to see a video.
Anyone can confirm this?
www1 said:
This kernel crash on my nexus one with desire camera when I try to see a video.
Anyone can confirm this?
Click to expand...
Click to collapse
I just recorded a video with the desire camera app and played it back fine.
Okay folks, no fancy graphics or information here. Searching this site will tell you any of the generic information you may be interested in. I am not going to clutter up things any more than I normally do. Which is already enough.
I need to make something very clear here. I'm sure it's just me, but I feel like I need to say it out loud, rather than just leaving it in my source. At least 90% of what I give you is other peoples original work, modified by me to run on whatever device I have. I don't do it to look smart, or be important, I do it 9 times out of 10 because there isn't something already out there. The vast majority of the options I give you are from amazing coders like show-p1948 and faux123. There are random snippets from others too. This time Zarboz contributed a lot in the beginning. I just wanted to make that clear. If you ever wanna know exactly who did something I put into a kernel, go look at the source. If I cannot get the original commit, with them as author, they are listed in the comment. If nobody is there but me, I actually did it.
So, here you go:
Droid DNA Cubed Kernel 3.1.3
Installation:
1. Unzip the file you downloaded.
2. Copy the Modules zip to the SDCard.
3. adb reboot bootloader
4. fastboot flash boot <file> (This will depend on where you unzipped to.)
5. Use menu on phone to go to recovery.
6. Flash Modules file in recovery.
7. Reboot
8. Profitt!
Please Note: This kernel removes some stock files you will need to use other kernels. Please flash this file before flashing anything else to return everything to stock.
Change Log:
12.01.12: Initial release, 1.0.1
SIO I/O Scheduler
Linaro 4.7 GCC Toolchain
-O3 Build
Force Fast Charge from Chad0989 - This is on by default. I will create a zip to turn it off and post later tonight.
Swap enabled
Charging from unknown sources enabled
12.02.12: Increase version to 1.0.2
Added SLQB
Added Sweep2Wake
Added SYSFS to thermald
Added GPU OC
12.03.12: Increase version to 1.0.3
Fixed S2W in call touchscreen issue - Zarboz
12.05.12: Increase version to 1.2.3
Various code patches and updates
Dynamic FSync
OC to 1.89Ghz
Re-did GPU OC
Disabled Perflock
Added VDD
12.09.12: Increased version to 1.2.5
Fixed freqs sticking
Fixed hsci wakelock
Added Intellidemand Gov
Tweaked thermal settings
12.12.12: Increased version to 1.4.3
Various fixes/updates
Added FIOPS IO Sched
Added ROW IO Sched
Added Smartass V2 Governor
Added Interactive Governor
Added Lionheart Governor
Added Badass Governor
Added SavagedZen Governor
12.13.12: Increased version to 1.4.9
Updated Intellidemand, thanks Faux123
Added AssWax Governor
Added BrazilianWax Givernor
Various tweaks/fixes
12.22.12: Increased version to 2.2.1
Added CIFS
Added NTFS
Various fixes/tweaks
Updated thermal - showp
Added kernel based mpdecision - showp
Added InteractiveX V2 Governor
Enabled Conservative Governor
Added SuperBad Governor
Added Virtuous Governor
12.23.12: Increased version to 2.2.3
Added BFQ I/O Scheduler
Enabled write to system partition
12.26.12: Increased version to 2.3.1
Added "Faster" PVCS bin
Fix for thermal
Fixes for Intellidemand
Fixes for Ondemand
Fixes for sched
Fixes for lockups/reboots
Fixes for Interactive
12.27.12: Increased version to 2.3.2
Removed sched fix
12.28.12: Increased version to 2.3.3
Restored sched fix
Tweaked FASTER table
12.29.12: Increased version to 2.3.5
Raised max bus speed to 600
Various fixes/tweaks
12.30.12: Increased version to 2.3.9
Updates to Intellidemand - Faux
USB fixes
HSIC Wakelock patch
Remove my changes for HSIC Wakelock
12.31.12: Increased version to 2.4.0
Removed a few of the patches added in 2.3.9
01.01.13: Increased version to 2.4.1
Removed the rest of the USB patches
01.02.12: Increased version to 2.4.2
Put back all the USB fixes
Removed Intellidemand update
Few other fixes.
01.31.13: Increased version to 2.5.1
Removed BFQ
Removed SLQB
Removed a few other tweaks
Tweaked thermal settings
Tweaked freq tables
02.01.13: Increased version to 2.5.2
Tweaked thermal
02.04.13: Increased version to 2.5.4
Completed -O3 compliance
Tweaked GPU to fix low scores
02.07.13: Increased version to 3.1.1
Patched up to Linux 3.4.29
02.13.13: Increased version to 3.1.3
Fixes to Linaro
Fixes to BadAss
Various tweaks
Downloads:
Droid DNA Cubed Kernel 3.1.3
MD5: f4ffcf4d75bf50f2c8458b10e6f6d2a9
Droid DNA Cubed Kernel 3.1.1
MD5: b1cfca9364d652b32827abe348f7bcb3
Droid DNA Cubed Kernel 2.5.4
MD5: f5bedb1d659e7d968e5c44c562cc8425
Enjoy!
D
.
My Git
Excited for the dev scene here on DNA I'm gonna flash this asap, will it work fine on an unlocked, rooted, stock odex rom??
Oh and does chad have a DNA??
Sent from my HTC6435LVW using xda premium
x70xchallengerx said:
Excited for the dev scene here on DNA I'm gonna flash this asap, will it work fine on an unlocked, rooted, stock odex rom??
Oh and does chad have a DNA??
Sent from my HTC6435LVW using xda premium
Click to expand...
Click to collapse
Will work fine.
I have no idea what Chad has. I just ported the code.
D
.
dsb9938 said:
Will work fine.
I have no idea what Chad has. I just ported the code.
D
.
Click to expand...
Click to collapse
Sorry for the noob question, what is charging from unknown sources mean ?
From my DROID DNA
So far so good... Thanks again!
dsb9938 said:
Okay folks, no fancy graphics or information here. Searching this site will tell you any of the generic information you may be interested in. I am not going to clutter up things any more than I normally do. Which is already enough.
So, here you go:
Droid DNA Cubed Kernel 1.0.1
Okay, it's still very basic. I've been watching/talking to Zarboz as he has been making his. I will be catching up a lot of his mods here in the next few days.
Installation:
1. Unzip the file you downloaded.
2. Copy the Modules zip to the SDCard.
3. adb reboot bootloader
4. fastboot flash boot <file> (This will depend on where you unzipped to.)
5. Use menu on phone to go to recovery.
6. Flash Modules file in recovery.
7. Reboot
8. Profitt!
Change Log:
12.01.12: Initial release, 1.0.1
SIO I/O Scheduler
Linaro 4.7 GCC Toolchain
-O3 Build
Force Fast Charge from Chad0989 - This is on by default. I will create a zip to turn it off and post later tonight.
Swap enabled
Charging from unknown sources enabled
Downloads:
Droid DNA Cubed Kernel 1.0.1
MD5: 3726b1673cea267e10e91bca8ead36cb
Enjoy!
D
.
Click to expand...
Click to collapse
Source code, please.
Rammstein1313 said:
Sorry for the noob question, what is charging from unknown sources mean ?
From my DROID DNA
Click to expand...
Click to collapse
The stereo in my car, a ufo....one time I tried to use a wall charger with the same port but from a different manufacturer and it wouldnt do it. doesnt happen so much anymore tho thanks to this..
adrynalyne said:
Source code, please.
Click to expand...
Click to collapse
LOL, added.
D
.
Rammstein1313 said:
Sorry for the noob question, what is charging from unknown sources mean ?
From my DROID DNA
Click to expand...
Click to collapse
It means that instead of assuming an unrecognized device is USB, it will assume AC, thus charging faster.
D
.
dsb9938 said:
It means that instead of assuming an unrecognized device is USB, it will assume AC, thus charging faster.
D
.
Click to expand...
Click to collapse
Lol, thought it meant I could use my power mat without the cords
From my DROID DNA
hows this kernel wats the highest cpu can u set it to and is it better or on par w/the stock kernel?
holla420 said:
hows this kernel wats the highest cpu can u set it to and is it better or on par w/the stock kernel?
Click to expand...
Click to collapse
just using linaro and -O3 optimization will make this above stock kernel even if the OC hasn't been added yet
Good stuff DSB can't wait to see what you come up with i always loved your rezound work
its been a while since I had to flash in Fastboot... need to clean out the spiderwebs in my ADB closet... flashed great.. thanks allot... any plans on a senseless cubed rom?
Added a few things, see the change log.
D
.
It all works for me thank you.
Just for the heck of it I threw up a benchmark I don't know what to think about antutu tho....
Sent from my HTC6435LVW using xda premium
Thanks again DSB! V2 Up and running great...
I dont care about overclock. This stock speeds when default? Like the sweep function...
Sent from my HTC6435LVW using Tapatalk 2
i flashed this kernel and now my wifi doesnt shows up did i do something wrong?
holla420 said:
i flashed this kernel and now my wifi doesnt shows up did i do something wrong?
Click to expand...
Click to collapse
Did you flash the modules zip?
BotsOne said:
Did you flash the modules zip?
Click to expand...
Click to collapse
yea i extracted the zip and flash that
Features:• Based on SEMC 2011 ICS sources, re-based on Linux kernel 2.6.32.61
• Built with linaro 4.7.4 toolchain
• CWM-v6.0.2.8
• Idiot-proof feature: will boot into recovery if an unsupported or no android version is found
• Custom mtd partitioning scheme suitable for JB
• Modular recovery flashing support
• Much more RAM than stock for improved multitasking experience (shakira: 186MB, robyn/mimmi: 187MB)
• Patched to latest 2.6.32.61 release
• kgsl: SEMC 2011
• ARM: many backported fixes and optimizations from CAF 3.x
• Netfilter: many backported fixes and optimizations from CAF 3.x.
• net/iptables: Fully working networking and iptables support. Fully working traffic control via data limits menu of JB
• Backported tiny-rcu
• Superior AHB overclocking - improved responsiveness and performance. 'Buttery-smooth' experience
• improvements: memcpy, kmalloc, jhash, sha1, scheduler, optimized CRC32, LZO, vmalloc, pm_suspend
• Two-way call recording
• binder, lowmemorykiller fixes backported from CAF 3.x
• WEXT tiwlan
• pmem: backported from CAF 2.6.35
• modular undervolting
• tripple buffering support
• Supported IO Schedulers: NOOP, BFQ, AS, DEADLINE, CFQ, VR, SIO
• net: Stochastic Fair Blue scheduler
• ZRAM + google snappy compression backported from CAF 3.x
• ZCACHE backported from CAF 3.x
• Cleancache backported from CAF 3.x
• Kernel compression: LZO for best kernel booting times
• Makefile: optimization flags
• dcache: vfs_cache_pressure -> 25
• Proportional Rate Reduction for TCP
• fsync enable/disable toggle at: /sys/module/sync/parameters/fsync_enabled (enabled by default)
• Many many more fixes and improvements available @ github
*Optimize with -O3 flag
*fix deadlock when system is low on memory (thank you @xelyt)
*backport lowmemorykiller from CAF msm-3.4
download
http://yairpatch.com/blog/2013/10/yairpatch-kernel-for-xperia-x8/
version 1.0 Final - http://yairpatch.com/blog/2013/11/yairpatch-kernel-for-xperia-x8-version-1-0/
KITKAT and CM10.2 KERNEL - http://yairpatch.com/blog/2014/03/kitkat-kernel-for-xperia-x8alpha/
Install:•Backup
•Unlock your bootloader as described here: http://forum.xda-developers.com/show...77&postcount=3 or here: http://forum.xda-developers.com/show....php?t=1254225
•Download the latest kernel ftf file from here and move it to Flashtoolfirmwares so that Flashtool can pick it up
•Flash the kernel via Flashtool
•Reboot and you should see the coolest kernel boot animation
• (First time only) Get into recovery and format /system /cache and /data manually
Uninstall:•
Flash back to stock using SEUS
OR
• Format /system /data and /cache via CWM and THEN flash another kernel via Flashtool
Any other way is completely unsupported, don't ask how to unbrick if you can't read the instructions
*thx goes to nAa for his great
Source:
https://github.com/yairpatch/msm7x27-2.6.32.x-nAa
XDA:DevDB Information
[KERNEL] yairpatch-JB[LINARO 4.7.4][31-10-13] , a Kernel for the Sony Ericsson XPERIA X8
Contributors
afi1982
Kernel Special Features:
Version Information
Status: Stable
Created 2013-10-31
Last Updated 2014-03-27
So Test, and report ; D
Howilion said:
So Test, and report ; D
Click to expand...
Click to collapse
already tested by me now it's your turn
I will test for now give you thanks and to amazing work ;D
This only work with your ROM ? Or another JB ...
Sent from my E15i using xda app-developers app
Sir Roven said:
I will test for now give you thanks and to amazing work ;D
This only work with your ROM ? Or another JB ...
Sent from my E15i using xda app-developers app
Click to expand...
Click to collapse
every jb and GB rom
Everything seems to be stable and working on the first version for me..
Just one thing and it's not a bug
You know on the YairPatch kernel logo there seems to be a bit of an overlapping of the kernel version at the bottom.
No big deal but would look much better if it could be sorted out.
CtrlAltDelIrl said:
Everything seems to be stable and working on the first version for me..
Just one thing and it's not a bug
You know on the YairPatch kernel logo there seems to be a bit of an overlapping of the kernel version at the bottom.
No big deal but would look much better if it could be sorted out.
Click to expand...
Click to collapse
yeah i know but for 0.2 and 0.3 i'm using nAa ramdisk so thus "bug" is gone
Sorry for this question but... can you try to make something for poor 2.3 users? If you would find some time to fix bluetooth problems it'll be great.
iks8 said:
Sorry for this question but... can you try to make something for poor 2.3 users? If you would find some time to fix bluetooth problems it'll be great.
Click to expand...
Click to collapse
Why don't you upgrade to 4.1.2? Everything works and the speed is better in my opinion.
------------------------------------------------------------------------------------------------------------------------------------
Reflashed to 3rd kernel. I can really feel a speed difference. Really good job Yair!
@up: maybe it is but it's unusable for me: most of time I'm using my phone as my mp3 player and everybody knows how music on 4.1 is annoying :/
iks8 said:
@up: maybe it is but it's unusable for me: most of time I'm using my phone as my mp3 player and everybody knows how music on 4.1 is annoying :/
Click to expand...
Click to collapse
With Google Play Music it's great!
CtrlAltDelIrl said:
Why don't you upgrade to 4.1.2? Everything works and the speed is better in my opinion.
------------------------------------------------------------------------------------------------------------------------------------
Reflashed to 3rd kernel. I can really feel a speed difference. Really good job Yair!
Click to expand...
Click to collapse
Did you notice any trace of the "hard to wake" bug ?
afi1982 said:
Did you notice any trace of the "hard to wake" bug ?
Click to expand...
Click to collapse
None whatsoever!
CtrlAltDelIrl said:
None whatsoever!
Click to expand...
Click to collapse
Good to hear
afi1982 said:
Good to hear
Click to expand...
Click to collapse
I'm gonna restore my MiniCM7 backup and see how that goes. Just an experiment but I'm sure it interests you
Show logo nAa it´s normal ?
Sir Roven said:
Show logo nAa it´s normal ?
Click to expand...
Click to collapse
Yes
CtrlAltDelIrl said:
Yes
Click to expand...
Click to collapse
i honestly flashed a couple of times using different versions of flashtool. thought something was wrong lol.
on topic: i tried checking yair's site, as well as the op, and i can't seem to find the differences between 02 and 03 kernel. i've already tried both, and 03 seems to be much smoother.
cascabel said:
i honestly flashed a couple of times using different versions of flashtool. thought something was wrong lol.
on topic: i tried checking yair's site, as well as the op, and i can't seem to find the differences between 02 and 03 kernel. i've already tried both, and 03 seems to be much smoother.
Click to expand...
Click to collapse
That made me laugh. Thank you!
The differences are only in the compile flag optimisations. But yes, 03 is very smooth.
CtrlAltDelIrl said:
Why don't you upgrade to 4.1.2? Everything works and the speed is better in my opinion.
Click to expand...
Click to collapse
Little bit off-topic, but I remember when the minicm10 came out with the .32 kernel, you was happy with the incredible performance. And after a longer test period you always went back to froyopro or minicm7 due the better performance...
Okay, go back to the thread.
@cascabel: you can't find the differences between the o2 and o3 versions of the kernels because the two kernels source code is same, nothing different. The compiler settings is invisible, this is not add or remove features, the difference is only the increased/decreased performance/stability.
If you interested what is the difference and how to work the code optimization in gcc, use google and you can find lot of information about how to work the o2 and o3 optimization, about the strict-aliasing and tons of other optimization flags.
@afi1982: can you post the gcc flags what you used? Or simple changed to -O3 without additional flags? (now playing with the .32 code to build a gb only kernel).
Build compiled from our CM branch sources with various optimizations / kernel features
Final build for 4.4.4 released (20th April 2015)
Many of the new features can be configured by kernel tools like Synapse. Install synapse app from market and UKM zip from downloads section.
We have moved on to Lollipop thread.
V7 [Changelog]
KERNEL:
- New hotplugs - ALUCARD, MSM
- MSM_limiter
- Fix synapse stuff - GPU etc..
- Governor tweaks
ROM
- Latest CM 11 sources.
- Fix chrome issues
- Fix button lighting-up when screen is off
Recommeded: Alucard governor with Alucard hotplug.
This is my synapse profile: http://goo.gl/3EbMUS
Note: For V6 and above (and specially for governors like zzmoove / alucard) , using Trickster/Synapse:
1. Disable MP-Decision
2. Enable intelli-plug
3. Enable intelli-thermal
4. And set to restore kernel settings on boot.
Old changelogs
V6 [Changelog]
- ROM sources updated to 20th November.
- Switch from O3 to Ofast [Even more compiler optimizations] and Re-enable Justarchi compiler optimizations [We have given up on exfat]
- GPU should now idle to 27MHz [instead of earlier 200Mhz]
- GPU overclocked to 535Mhz [earlier 500Mhz]
- CPU should now idle to 268Mhz (instead of earlier 300Mhz)
- Port few of my favorite governors from Oneplus one kernel these days - zzmoove and alucard: Both have been awesome for me (credits go to Oneplus one kernel developers - @Lord Boeffla, @ZaneZam, @alucard_24) - I will eventually start removing some of the older CPU governors. Looks like we have lots of them now.
- Various kernel patches - arch power, cpu boost, gentle fair sleepers, software crc is off
- Upgrade BFQ scheduler to BFQ-v7r6
V5.1 [Changelog]
- Newer sources from 20141107
- fsync is now disabled (Defers file writes but risk of data loss with crashes etc. Can be enabled in trickster)
- tweak intelliactive (hispeed freq was 2457, calm it down to 1985)
- voltage tweaks for speed 1 pvs 2
- default IO schd BFQ
- Frequency throttle limit increased slightly (Thermal policies can now be controlled through Trickster)
V5 [Changelog]
- Newer sources from 20141102 and use SaberMOD Toolchain, GCC 4.8.2 for kernel as well as entire build.
- Add a new hotplug engine - IntelliPlug (developed and used in famous Franco's Kernel) - I recommend to use this one, and switch off MP-Decision. (Enable in Trickster)
- Add a new thermal engine - Intellithermal (again from Franco) - Recommend to use this one. (Enable in Trickster)
- Add USB Fast charge (Enable in Trickster)
- Add another performance centric governor - SmartMAX (Combines features of OnDemand and SmartASSv2), Use it as gaming driver.
- Tweak intelliactive - Recommend to use this as default (Interactive is battery friendly but a bit laggy)
- Add powersuspend
- Reduce VM Swappiness to 10 (We have enough RAM)
- Some GPU tweaks
Other features
- CPU Overclocks of upto - 2457Mhz
- GPU overclocks of upto - 500Mhz
- Voltage control (UV)
- Additional CPU governors - Intelliactive, Intellidemand, InteractiveX
- Additional IO Schedulers - SIO, FIOPS, ZEN, VR
- Tweaked thermal policies (aggressive by default)
- Lower GPU clocks by default
- Reduced logging
- Net tweaks - Lock initial TCP window size to 64k
- FS tweaks disable atime and diratime
- Some mixer_paths patches (for higher audio) and button light fixes (to not glow on screen touch)
- Compiler optimizations
Branches(Github):
GCC4.8/Linaro/O3 related fixes/tweaks:
https://github.com/vikrant82/android_build
https://github.com/vikrant82/android_external_valgrind
https://github.com/vikrant82/android_external_fuse
https://github.com/vikrant82/android_external_exfat
https://github.com/vikrant82/android_art
Device tweaks
https://github.com/vikrant82/android_device_samsung_mondrianwifi
Kernel
https://github.com/vikrant82/android_kernel_samsung_mondrianwifi
Buttons Fix
https://github.com/vikrant82/android_frameworks_base
Downloads
Final release for 4.4.4
V444Final-20150419-SM482-Ofast-mondrianwifi.zip
UKM zip - UKM 3.6.1
Old releases
V6 ROM - V6-ROM-20141120-SM482-Ofast-mondrianwifi.zip
V6 Kernel - V6-Kernel-20141120-SM482-Ofast-mondrianwifi.zip
V5.1 ROM - V5.1-cm-11-20141107-SaberMOD-mondrianwifi.zip
V5.1 Kernel - V5.1-Kernel-20141107-Kernel-SaberMOD-mondrianwifi.zip
V5 ROM - V5-cm-11-20141102-SaberMOD-mondrianwifi.zip
V5 Kernel - V5-Kernel-20141102-Kernel-SaberMOD-mondrianwifi.zip
V4.1 (CPU 2457Mhz, GPU 500Mhz) - V4.1-cm-11-20140803-Linaro-O3-OC-mondrianwifi.zip
V4.1 Kernel (CPU 2457Mhz, GPU 500Mhz) - V4.1-cm-11-20140803-Kernel-Linaro-O3-OC-mondrianwifi.zip
V4.2 (CPU 2457Mhz, GPU 500Mhz) - V4.2-cm-11-20140809-Linaro-O3-OC-mondrianwifi.zip
V4.2 Kernel (CPU 2457Mhz, GPU 500Mhz) - V4.2-cm-11-20140809-Kernel-Linaro-O3-OC-mondrianwifi.zip
V4.3 (CPU 2457Mhz, GPU 500Mhz) - V4.3-cm-11-20140829-Linaro-O3-OC-mondrianwifi.zip
V4.3 Kernel (CPU 2457Mhz, GPU 500Mhz) - V4.3-cm-11-20140829-Kernel-Linaro-O3-OC-mondrianwifi.zip
Thanks @crpalmer @franciscofranco
XDA:DevDB Information
[SM-T320][ROM][KERNEL][4.4.4][-Ofast][SaberMod]CM11 [OC UV], ROM for the Samsung Galaxy Tab Pro 12.2, 10.1, 8.4
Contributors
vikrant82
ROM OS Version: 4.4.x KitKat
ROM Kernel: Linux 3.4.x
Based On: CyanogenMod
Version Information
Status: Testing
Created 2014-07-17
Last Updated 2015-04-19
excellent!
I'll try it later.
I always flash PurePerformances™X mod after every nightlie and it feels somewhat faster after it. I did flashed it on all my devices.
Maybe you can insert it in your future builds.
Keep them coming !
LE: mod link fixed
Thanks for this... Will give it a shot!
For anyone wanting to try this.. So far so good here. Hasn't been running long enough to make any major comparisons yet. Clean flashed, so not sure about dirty flashing.
I installed it over nightly and it runs pretty well but the playstore icons bug is pretty bad.
For now I moved on to the next nightly but if you can fix that bug I will surely flash it again.
csjneek said:
I installed it over nightly and it runs pretty well but the playstore icons bug is pretty bad.
For now I moved on to the next nightly but if you can fix that bug I will surely flash it again.
Click to expand...
Click to collapse
Should be fixed with Linaro 4.9 (thats next). Trying out a few more optimizations for next build (expect to post one on weekend).
vikrant82 said:
Should be fixed with Linaro 4.9 (thats next). Trying out a few more optimizations for next build (expect to post one on weekend). Target is to beat benchmark scores from stock.
Click to expand...
Click to collapse
Don't pay attention to benchmarks....
If you want to meet or beat stock, just turn off thermal throttling, enable all 4 cores and set performance as the governor. Make it a fair comparison...
Finally a linaro build!! Keep up the great work! Could you modify the kernel too? Overclock cpu and gpu would be awesome...
Gesendet von meinem SM-T320 mit Tapatalk
crpalmer said:
Don't pay attention to benchmarks....
If you want to meet or beat stock, just turn off thermal throttling, enable all 4 cores and set performance as the governor. Make it a fair comparison...
Click to expand...
Click to collapse
@crpalmer
I suppose we are talking about /etc/thermal-engine-8974.conf, /etc/thermal-engine.conf (symlinked to thermal-engine-8960.conf) and /etc/thermald.conf (symlinked to thermal-8960.conf) on cm builds. They seem to be using different set of values from stock.
Can't we not use these values/files directly-from/similar-to stock builds ? As the thermal engine libs and binaries we are directly picking from stocks ?
Is stock too aggressive or CM too defensive ?
Also noticed that these linked files are broken as well. (no thermal-engine-8960.conf and thermal-8960.conf). Some build script issues ?
vikrant82 said:
@crpalmer
I suppose we are talking about /etc/thermal-engine-8974.conf, /etc/thermal-engine.conf (symlinked to thermal-engine-8960.conf) and /etc/thermald.conf (symlinked to thermal-8960.conf) on cm builds. They seem to be using different set of values from stock.
Can't we not using these values/files directly-from/similar-to stock builds ? As the thermal engine libs and binaries we are directly picking from stocks ?
Is stock too aggressive or CM too defensive ?
Click to expand...
Click to collapse
I wasn't suggesting that you turn that off permanently, that's just what the stock ROM is most likely (automatically) doing when you run a benchmark. That's how you get a higher score, not how you have a better ROM...
My point being, don't worry about whether or not the scores are better, worry about the whole experience being better (which is very much is). On a phone of mine, CM was way more performant even though it scored less than half of stock's Quadrant score.
vikrant82 said:
@crpalmer
I suppose we are talking about /etc/thermal-engine-8974.conf, /etc/thermal-engine.conf (symlinked to thermal-engine-8960.conf) and /etc/thermald.conf (symlinked to thermal-8960.conf) on cm builds. They seem to be using different set of values from stock.
Can't we not using these values/files directly-from/similar-to stock builds ? As the thermal engine libs and binaries we are directly picking from stocks ?
Is stock too aggressive or CM too defensive ?
Also noticed that these linked files are broken as well. (no thermal-engine-8960.conf and thermal-8960.conf). Some build script issues ?
Click to expand...
Click to collapse
How is the new build? Linaro 4.9?
Gesendet von meinem SM-T320 mit Tapatalk
crpalmer said:
I wasn't suggesting that you turn that off permanently, that's just what the stock ROM is most likely (automatically) doing when you run a benchmark. That's how you get a higher score, not how you have a better ROM...
My point being, don't worry about whether or not the scores are better, worry about the whole experience being better (which is very much is). On a phone of mine, CM was way more performant even though it scored less than half of stock's Quadrant score.
Click to expand...
Click to collapse
I agree benchmarks are just academic (and just nice-to-have). I was just suggesting that we could use same thermal configuration files as found in stock builds. Are these even the files that would affect benchmarks (thermal-engine*.conf and thermald*.conf) ? Are you suggesting that stock build just shuts off any kind of throttling as soon as it detects benchmark tests ? I thought only some devices did that and they got caught.
I am building with linaro 4.9 and have swapped out those conf files from stock. Lets see how it goes.
vikrant82 said:
I agree benchmarks are just academic (and just nice-to-have). I was just suggesting that we could use same thermal configuration files as found in stock builds. Are these even the files that would affect benchmarks ?
I am building with linaro 4.9 and have swapped out those conf files from stock. Lets see how it goes.
Click to expand...
Click to collapse
Are the messed up icons fixed in play store?
iamnotkurtcobain said:
Are the messed up icons fixed in play store?
Click to expand...
Click to collapse
Expecting them to get fixed with this build. Lets see.
vikrant82 said:
Expecting them to get fixed with this build. Lets see.
Click to expand...
Click to collapse
I'll test as soon as it is available
iamnotkurtcobain said:
I'll test as soon as it is available
Click to expand...
Click to collapse
Unfortunately, Linaro 4.9 is breaking stuff. Getting into a bootloop n all. Need more time.
vikrant82 said:
Unfortunately, Linaro 4.9 is breaking stuff. Getting into a bootlock n all. Need more time.
Click to expand...
Click to collapse
Hope you can fix it!
iamnotkurtcobain said:
Hope you can fix it!
Click to expand...
Click to collapse
Been struggling with some adreno crashes on boot when using 4.9 toolchain. Earlier I thought it's the compiler optimizations not playing well with 4.9. But I run into same without any customizations -
Here's the logcat if anyone wants to take a look.
http://pastebin.com/eS0eHvzL
If I can't get around these in a day or two, we'll come back to 4.8 and carry on with other optimizations. I read somewhere the gibberish icons are fixed on newer play store.
vikrant82 said:
Been struggling with some adreno crashes on boot when using 4.9 toolchain. Earlier I thought it's the compiler optimizations not playing well with 4.9. But I run into same without any customizations -
Here's the logcat if anyone wants to take a look.
http://pastebin.com/eS0eHvzL
If I can't get around these in a day or two, we'll come back to 4.8 and carry on with other optimizations. I read somewhere the gibberish icons are fixed on newer play store.
Click to expand...
Click to collapse
Its probably the adreno libs built using older tool chains.
vikrant82 said:
Its probably the adreno libs built using older tool chains.
Click to expand...
Click to collapse
Just posted another build with today's sources with both CPU and GPU overclocks. Had to go back to older 4.8 toolchain. The overclocks are running surprisingly stable on device.
The playstore icons issue is still not resolved. Its being discussed here. However, I am not too worried by it.
Mod Edit: Thread Closed per OP request
Welcome to the N0G33K edition Kernel
This kernel is built on the idea of having optimized battery life and performance. A well made balance between these two is what you get.
Read this first
Please do not post requests, because this is my personal kernel and I won't add any stuff I do not want myself (i.e D2W, S2W, AOD, sound control etc). Also I won't add any other support for other devices. Last but not least: I don't reply to questions like: "which governor/scheduler is the best" etc. The choice is all yours.
PS: You need to root yourself (Magisk on Nougat or SuperSU on Marshmallow)
Features
- Based on A810FXXU2CRH7 source
- Unified kernel that supports G920/G925 F/I/S/K/L/T/W8 models
- Upstreamed to Linux version 3.10.108
- Extra governors & I/O schedulers
- Linaro 6.5 compiler
- Always on display (AOD) support
- Wireguard support
- Knox 0x0 with safetynet bypass
- Permissive mode by default
- Synapse Control for OC/UC, voltages and a lot more
- LZ4 1.8.3
- Other misc optimizations
Download:
- V4 Marshmallowhttps://mega.nz/#!xxAAwBTA!2IL8I5Dx5IZsZp_uwrIT_DKTabzFu8_P6NlkD2EFORk
- Final Nougat https://mega.nz/#!RxhTHaaR!OLsnZ7yulPnaWd-mfSB1JCfMql3M_PjaUWyiFTMi1_U
Changelog:
V1
- Reworked whole kernel
- Optimized battery and performance
V1.1
- Reverted some stuff to test 3
- Fixed Fading LED
V1.5
- Tweaked LMK
- Optimized Workqueue
- Added battery efficiency patches
- AES put to x4 for more speed
- Readahead tweaked
- Vnswap to fast device
- HMP adjusted for more smoothness / battery efficiency
V1.8
- Reverted LMK and Workqueue edits.
- Enabled powereffici�nt Workqueue
- Tweaked Interactive governor
- Added Powersuspend
- Added more Binder changes for UI smoothness
- HMP and Cpuidle improvements.
- LZ4 Compressed
1.9
- Went back to UBERTC 5.3 because N0G33K V1.8 had kernel panics.
- Removed wlan wakelock control and reduced it myself.
- Some other edits and reverts
V2.0
- Reverted ALOT of changes I made in previous versions.
- Updated Kernel source to the latest DPJ2 source
- Compiled with Sabermod 7.0 Toolchain with CCache (should be stable now)
- Updated Linux to 3.10.104
- Backport latest Linux 4.8 random driver (= like super Seeder)
- Backport latest crypto algorithm Chacha20
- Added new battery efficiency patches
- Tweaked interactive governor for better performance/battery.
- Added new governors: Yankactive, Smartmax and Smartmax_eps
- Updated BFQ I/O
- Added dual sim support
- Fixed Invisible CPUset (should actually work now)
- Screen on hotplug disabled (Dynamic hotplug wasnt good enough)
- Added wakelock controls again.
- Updated Synapse
- Patched "Dirty Cow" (CVE-2016-5195) exploit
- Zswap improvements
- Disabled big cores on screen off
- Improved CPUIdle
- Powersuspend hooks added
- and more
V3
Reworked the whole kernel:
- Based on U5DPL4 source
- Powersuspend is out, State Notifier is in.
- Replaced Linux 4.8 random driver with Frandom
- Interactive governor: merged Galaxy S7 Interactive changes (android N changes wont boot, so kept it MM)
- Merged S7 zswap/zpool and zsmalloc android N changes
- Added Impulse governor
- Tweaked I/O Schedulers for our SSD + Zen V2 update
- Xbox control support added and updated to latest upstream (can be enabled in Synapse)
- Noatime as default for better performance
- SafetyNet patched (soon there probably will be a workaround for it...so enjoy as long it lasts)
- More wakelocks toggles added
- Sync on suspend disabled
- Improved Idle drain
- Android OS drain should be better now.
- Removed some other stuff that caused yank like invisible cpusets
V4
- Updated Linux kernel to 3.10.105
- Updated Safetynet patch
- Binder updated to latest Linux 4.6 (smoother UI)
- Backported Linux 4.1.38 random driver and removed Frandom
- DVFS decision mode in Synapse added
- DVFS: scale to 1k instead of 1.2k GHz
- Backported latest CPUidle driver from Linux 4.7 (should give better idle experience throughout the day)
- Other misc optimizations
V7
- Rebased kernel and kept it quite stock/light
- Doesn't use recycle commits (= commits that have been using in all S6 kernels for the last 4 years, without anything new).
- Based on the A810FXXU2CRH7 source code, which means that its newer than any G920x source out there.
- Upstreamed to 3.10.108
- Unification added to work for nearly all S6(e) devices (No A & P model support)
- Improved touch latency
- LZ4 fast 1.8.3
- Compiled with Linaro 6.5
- Added Wireguard support
- Great performance and battery life -> Ask @ananjaser1211 who copied the kernel
The kernel wont be updated anymore (no oreo support/updates), therefore enjoy this last kernel while it lasts. If something is broken in Synapse, then so be it.
Credits:
- CODEG33K - Dear friend that I partially named kernel after
- Aria.A97 - Intensive daily testing
- Everyone else I cherry-picked from.
Source Code: https://github.com/Noxxxious/S6
Want to use this in your rom?: Don't, just forward to the thread.
Disclaimer: I am not responsible for all the damages followed by or to your phone after flashing this kernel. If it bricks, wipes, explodes and you will hold me responsible I will only laugh
Known issues:
- Probably Synapse "backup profile" doesnt work.
F.A.Q. - Frequently Asked Questions
Q: Synapse says no UCI detected
A: Install Busybox to system/xbin
Q: I had a reboot
A: Only report it here if you didn't undervolt and do it with a log.
A2: If you had random reboot without undervolting, then make a proper log please. Which means I need a last_kmsg from right after the reboot.
You create the log by using this command in adb shell or a terminal or use any 3rd party app like Syslog.
Code:
cat /proc/last_kmsg > /sdcard/last_kmsg.txt
Good luck .
Maybe you will add for g925F support to .
senzatiOn- said:
Good luck .
Maybe you will add for g925F support to .
Click to expand...
Click to collapse
+1
That will be awesome
@senzatiOn- @Sommerrain
Read my "Read this first". Already said that I wont.
But never say never, who knows.
So if rhis works with 920f does that mean it work with 920i
panda968 said:
So if rhis works with 920f does that mean it work with 920i
Click to expand...
Click to collapse
Yes, 920F/I same device, just different regions.
Happy flashing :good:
Quintz said:
@senzatiOn- @Sommerrain
Read my "Read this first". Already said that I wont.
But never say never, who knows.
Click to expand...
Click to collapse
I have read this.... i just ask... MAYBE
GL & HF
I lost RIL (signal) after flashing V3.
Unknown IMEI, but efs partition intact.
Going back to Arter v22 Grace, fixes this issue.
OP is aware of this, and will look into it, so have a nandroid backup in your hands, and be sure you have a fresh efs backup, prior to flashing V3.
It might not happen to you, but since some users have V3 installed already without issues, consider your self warned.
-----
senzatiOn- said:
I have read this.... i just ask... MAYBE
GL & HF
Click to expand...
Click to collapse
Thanks !
CODEG33K said:
I lost RIL (signal) after flashing V3.
Unknown IMEI, but efs partition intact.
Going back to Arter v22 Grace, fixes this issue.
OP is aware of this, and will look into it, so have a nandroid backup in your hands, and be sure you have a fresh efs backup, prior to flashing V3.
It might not happen to you, but since some users have V3 installed already without issues, consider your self warned.
-----
Click to expand...
Click to collapse
Yes like I stated the N7 compatibility is still in beta. So far this is the first case where there is no signal. I am taking a look into it.
Also people you dont need an EFS backup or something, flashing back to stock or arter N7 kernel will just be fine.
I'm happy to see you finally here ?
Inviato dal mio SM-G925F utilizzando Tapatalk
Quintz said:
Thanks !
Also people you dont need an EFS backup or something, flashing back to stock or arter N7 kernel will just be fine.
Click to expand...
Click to collapse
Couldn't hurt
It's refreshing seeing a new kernel here, thank you for your efforts and your development, I shall surely be trying this out.
like u my bro and thanks for hard work
gave glory make perfect kernel for us
keep going im with you:fingers-crossed:
CODEG33K said:
Yes, 920F/I same device, just different regions.
Happy flashing :good:
Click to expand...
Click to collapse
If thats the case if i find a rom that only says 920f it will work for my phone
Great work!
Quintz said:
Thanks !
Yes like I stated the N7 compatibility is still in beta. So far this is the first case where there is no signal. I am taking a look into it.
Also people you dont need an EFS backup or something, flashing back to stock or arter N7 kernel will just be fine.
Click to expand...
Click to collapse
I can confirm, using tamerlan n7 port ROM, I lost signal and my IMEI
And thanks for your hard work , nice to see other developers on our device.
Netmain36 said:
I can confirm, using tamerlan n7 port ROM, I lost signal and my IMEI
And thanks for your hard work , nice to see other developers on our device.
Click to expand...
Click to collapse
Lol I'm using tamerlan and is all good
Cesarcami53 said:
Lol I'm using tamerlan and is all good
Click to expand...
Click to collapse
Yes wait, I help my friend to fix this
AlexisXDA said:
Yes wait, I help my friend to fix this
Click to expand...
Click to collapse
Alexis signal works the problem is after every reboots takes like 2 minutes to get signal again