No. It is not about JFS or ext4 but about RFS or rather the wrongly set mount options in ULTK kernels. So no change in lagfix perspectives. JFS is still "unsafe" (or as safe as it ever was)
- Status
That is what I thought. Thanks for clarifying
I had been running JFS overkill with binds before; never had an issue even if I drained the battery to 0 a number of times and no issue with the time zone neither. After hearing though the potential dangers associated with JFS, I decided to go to ext4 instead.vukandric
Senior Member
Hi zenkinz,
Just wanted to say that I love your work. Just one thing that I noticed, in ver. 30 and 31. there is more battery drain than in ver. 23 (at least for me). I'm now using ver 31 because of the /system bug, but I must admit that I loved ver 23.
I use Ultimate Juice, because SuperPower didn't work for me (I might try it again just to be sure) and wanted to ask what is your opinion about that?
Don't get me wrong, I'm not complaining, just wanted to give some user feedback.
Thanks again for your work!
Just wanted to say that I love your work. Just one thing that I noticed, in ver. 30 and 31. there is more battery drain than in ver. 23 (at least for me). I'm now using ver 31 because of the /system bug, but I must admit that I loved ver 23.
I use Ultimate Juice, because SuperPower didn't work for me (I might try it again just to be sure) and wanted to ask what is your opinion about that?
Don't get me wrong, I'm not complaining, just wanted to give some user feedback.
Thanks again for your work!
Here it is.
IMPORTANT: disable lagfix first, flash #31, reboot system, boot to recovery, enable lagfix.
Last edited:
Hi zenkinz,
Just wanted to say that I love your work. Just one thing that I noticed, in ver. 30 and 31. there is more battery drain than in ver. 23 (at least for me). I'm now using ver 31 because of the /system bug, but I must admit that I loved ver 23.
I use Ultimate Juice, because SuperPower didn't work for me (I might try it again just to be sure) and wanted to ask what is your opinion about that?
Don't get me wrong, I'm not complaining, just wanted to give some user feedback.
Thanks again for your work!
what superpower settings did you try? also, the HZ values have changed from 23 to 31. i'm getting 2-3 days with normal usage and superpower installed. i don't have setcpu or any voltage app installed either.
YES, disable lagfix first, flash #31, reboot system, boot to recovery, enable lagfix.
BTW< it sounds like you have the ULK RFS corruption...
Last edited:
Right after running Supercurios RFS fix (ran successfully) and flashing YA-OK build 31, these lines still appear in user.log. Still, phone boots perfectly normal, so this does NOT seem to be related to a RFS corruption or the Force Close i feel during boot when lagfixed.
Reapplied Voodoo lagfix from recovery, and immediately Force Closes during boot returned. So i will have to investigate further.
vukandric
Senior Member
I tried older version, so I installed 0.75 few minutes ago to try it again. I have setcpu and have profile for screen off (100-400, have no slow wake up or similar problems) There are far to many settings in superpower, but I'm quite sure that my settings should be good because most of them are similar to settings in Juice Defender. Maybe it's because of doc's v.8.0.1. which I started to use few days ago...
EDIT: My battery lasted for about two days earlier, now one day only. Huge difference.
yes, flashed it this morning. i have the RFS problem so was waiting for this to come to out. first i flashed a new barebone, then did voodoo's fix, then flashed the new zenkinz, and all good.
remove setcpu and do not do any sort of cpu scaling. just run it as is and trust the system to manage the cpu for you. it will run lightning fast, be very stable and use less power. yes, superpower is complicated with it's settings and you have to set stuff up to trigger in the right order too but once it's done it's perfect.
vukandric
Senior Member
You have corrupted /system..check out supercurio's thread,ora flash a stock rom,and then the new kernel (the latest kernels that are coming out have only fix for these "Ultimate-lagfix-kernel-bug",as long as you use the bugged versions everything works fine,but it gets corrupted over time and denies you the option to flash a different "kernel family",for example stock or fixed ones)
- Status
Similar threads
Top Liked Posts
-
There are no posts matching your filters.
-
13Yes, this is yet another (shameless) attempt to optimise the kernel for our Galaxy-S (i9000) by taking tweaks implemented by various xda developers. The aim at this moment is to 'integrate' all the desirable tweaks, since I'm still grasping the ins and outs of Linux kernel to experiment some of the possible tweaks myself.
Not sure how far I would bring this thread forward, because I will be working on a project (my day job) very soon again, so I'm going to enjoy/do while I can.
Key Implementation
- based on JPM kernel
- sztupy's Universal Lag Fix (v0.3)
- hardcore's Free Memory config (341MB)
- raspdeep's Overclock (v1.1)
- hardcore's SpeedMod (Tweaks + Compiler Optimisation) with Jhash update
- supercurio Voodoo Display tweak (Gamma Correction and Improve Sharpness)
- neldar's Backlight Notification (BLN, v2.2, enabled via ULF in recovery mode)
- Budget Fair Queuing (BFQ) I/O Scheduler
I also made some minor modification to other codes to experiment the overclock enhancement. It's in this build but may not have any effect.
Please refer to this link for the latest download
See this video of my Galaxy S running on Ya-OK kernel. Video playback, internet browsing, gaming, UI navigation.
attached is the screenshot of quadrant, on a JPM-based ROM with EXT4 on all partition.
CREDITS
in no order of priority.
- supercurio
- sztupy
- raspdeep
- hardcore
- neldar
- and rest of folks in xda-developers who have contributed one way or another
2DOWNLOAD HISTORY
v1.50 (build 38) - 8th December
- Updated Overclock to 1.12GHz
- Updated Kernel Scheduler parameter (sched_min_granularity_ns & sched_wakeup_granularity_ns) to give more priority to foreground application, therefore improving the fluidity of the UI experience. From brief research, these are the key parameters used to tune kernel between a server (therefore higher values) and a desktop.
- CWM version available here
v1.45 (build 35) - 5th December
- Fixed stability issue in v1.40 by reverting to the voltage profile used in v1.2b.
v1.40 (build 34) - 28th November
- Added BFQ (Budget Fair Queuing) I/O Scheduler and set BFQ as the default I/O Scheduler. The aim is to boost I/O throughput, reduce the latency. Please refer to this site for more information
v1.31 (build 31) - 24th November
- Bug fix on /system mount to prevent filesystem corruption in /system
- Does not include fix, please refer to supercurio's thread on how to correct the corrupted filesystem if you are encountering problem with your device.
v1.3 (build 30) - 20th November
- Refined voltage level for cpu frequency 100/200 MHz. Effectively, reduce 75mV for 100/200, and 25mV for 400/800 (same as 1.2b). Again not user configurable for now, though I may implement it if people are finding the UV feature is causing unstability.
- Upgraded Backlight Notification (BLN) to v2.2 (Please test and feedback)
- Minor optimisation on kernel configuration (PREEMPT_VOLUNTARY, UACCESS_WITH_MEMCPY)
v1.2b - 2nd November (2585 downloads)
- Refined voltage level for cpu frequency 100/200 MHz. Effectively, reduce 50mV for 100/200, and 25mV for 400/800. Again not user configurable for now, and once raspdeep released his UV code I may implement it if people are finding the UV feature is causing unstability.
v1.2a - 31st October (668 downloads)
- Implemented lower voltage for cpu frequency 100/200/400/800MHz.
Note: Not user configurable
v1.1 - 30th October (1168 downloads)
- Implemented Voodoo Display Sharpness & Color Gamma
v1.0 - 29th October (759 downloads)
- Implemented Universal Lagfix, Overclock, Memory optimization, etc
- Additional minior OC tweaks.
1Cool! But how about the OC? Does it works well too? I'm interested in this kernal, cuz of the OC function actually
OC works very well. No FC, no freezing what so ever. Can't fault it.1
you can't undervolt, overclock, optimise core of kernel through compiling, etc w/o the source.
if you like the other kernel that are compatible with JPU, then stick with it.1I really like your kernel, the performance of the phone is awesome and it has so far been the best kernel I've ever used.
There's only a slight problem I have (not sure about others)
My GPS is somewhat useless with this kernel! No sat lock even after about 10mins!
Takes about 4 or 5 mins to even find 3-4 sats! Use to have ~30sec locks indoors with Hardcore's kernel.
Don't really know if it's me or others are having the same issue...
Thank you so much for this kernel!
Keep up the good work.
that's not really just the kernel's fault, it's the combination of rom, modem and kernel. i use insanity rom, ya-ok kernel and zsjpe modem and i get lock perfectly with great accuracy down to 3m.
i tried 7 different modems to get the right one for my rom/kernel combo.
