Does this mean we can flash the stock kernel without the dual core mod over a ROM that comes with the dual core mod, and it'll actually run at factory specs?
[Kernel]3/14/12 StockMOD 2.17.651.5 Core Duo/ScreenOff Script/NoCIQ/DimLights
- Thread starter freeza
- Start date
Does this mean we can flash the stock kernel without the dual core mod over a ROM that comes with the dual core mod, and it'll actually run at factory specs?
Does your WiFi randomly disconnect when you are using a bluetooth headset in-call?
Well, I think I have fixed it.
Kernel update coming! Pending testing, of course.
Yes but not while in call, after a min. Or so it does automatically. Can't wait. As usual, I'm open to test.
Via My HTC Evo 3D On The Now Network From Sprint.
Thank you much. Appreciate the feedback.
Okay, the WiFi problem wasn't eliminated 100%, BUT it did substantially improve WiFi performance for me. I was on the phone for a good hour with my bt headset before the WiFI cut out. Normally this would have happened 5 minutes into the conversation. Releasing updated kernel today.
3/14/12
Changed loglevel to 0. Afterall, this is a stock kernel. The less logging the better.
Removed wimaxDumpKmsg
Removed wimaxDumpLastKmsg
Removed wimaxDumpLogcat
Removed bugreport
Removed loading and references to those files from the Kernel ramdisk. Don't worry, WiMax still works. If anything it will work just a bit better now that it doesn't need to worry about dumping any information to logcat and kmsg.
Updated unmodified kernel package to include missing files for those who want to go back to pure stock after flashing one of these.
Last edited:
I have faith in you that the issue will soon be 100% resolved.
How about the issue where Bluetooth itself cuts out? I'm thinking they're both the same issue, as they happen at the same time. (WiFi and Bluetooth cut out)
As always, you da man!
Via My HTC Evo 3D On The Now Network From Sprint.
How about the issue where Bluetooth itself cuts out? I'm thinking they're both the same issue, as they happen at the same time. (WiFi and Bluetooth cut out)
As always, you da man!
Via My HTC Evo 3D On The Now Network From Sprint.
I have faith in you that the issue will soon be 100% resolved.
How about the issue where Bluetooth itself cuts out? I'm thinking they're both the same issue, as they happen at the same time. (WiFi and Bluetooth cut out)
As always, you da man!
Via My HTC Evo 3D On The Now Network From Sprint.
You know......it's so weird that that's happening to you. I use my bluetooth headphones (Cardo S2) all the time and wouldn't be able to stand it if that were happening to me. I will definitely try my hardest to figure that out.
for reals lol. i was ready to smash this shiny new stereo until i found out it was my phone. ;/ can you try a 100% stock kernel with init.d enabled temp.? i know i asked before, but this is exciting. #nerdyways
I'll report in the a.m. did you get anything smoother than earlier?
Via My HTC Evo 3D On The Now Network From Sprint.
Via My HTC Evo 3D On The Now Network From Sprint.
He'll get it, just give him some time. He's been hella busy.
Sent Using My HTC Evo 3D, On The Now Network From Sprint!
Freakin sweet. Now I gotta get it working with PGM lol
EDIT: It's fast and stable and everything functions exactly as stated in the OP. Big thanks.
Last edited:
Kernels will be getting a slight update:
-remove scaling_min_frequency from ramdisk. It was set to 192mhz. Some phones don't like to idle at 192mhz, so I will remove it. scaling_min_frequency will now be handled by init.post_boot.sh like it used to be.
-re-add mounting of debugfs /sys/kernel/debug for tweaks that rely on it and don't mount by themselves.
-remove scaling_min_frequency from ramdisk. It was set to 192mhz. Some phones don't like to idle at 192mhz, so I will remove it. scaling_min_frequency will now be handled by init.post_boot.sh like it used to be.
-re-add mounting of debugfs /sys/kernel/debug for tweaks that rely on it and don't mount by themselves.
Similar threads
Top Liked Posts
-
There are no posts matching your filters.
-
119Here's the modified stock kernel from software ver 2.17.651.5, flashable through recovery, of course.
As of 1/13/12, CIQ back and front ends have been removed by HTC.
Includes init.d support.
Changelog:
8/21/11
On boot the following happens:
-Removes files from /data/system/dropbox
-Removes files from /data/system/appusagestats
-Removes files from /data/system/usagestats
-Removes files from /data/tombstones
-Modifies permissions on /data/system/dropbox to read-only
-Modifies permissions on /data/system/appusagestats to read-only
-Modifies permissions on /data/system/usagestats to read-only
-Removes permissions on /data/system/userbehavior.db
-Removes permissions on /data/system/userbehavior.db
-Removes permissions on /data/system/userbehavior.db-shm
-Removes permissions on /data/system/userbehavior.db-wal
-Removes permissions on /data/system/userbehavior.xml
8/30/11
Thank you to mike1986 of the awesome Sensation forum and the Android Revolution HD ROM for the initial idea, and smokin1337 for bringing it in mod form to the EVO 3D forums.
-DualCore support (my own twist). It sets the second core to run at half the speed of the first core as well as drops the minimum frequency to 192MHz on both cores.
-Compatible with scripts/apps that don't bother the second core.
-PowerSave script is not compatible or needed with this kernel.
-For Quadrant braggers, your scores will take a slight dip since Quadrant isn't optimized for dual cores. However, performance does increase.
8/31/11
-Minor ramdisk tweaks.
-Removed /system/etc/init.post_boot.sh from zip. No longer needed.
-A2SD support.
-594MHz second core now runs with a threshold of 40. (should be slightly better performing)
-Update4: 1.2GHz second core runs in sync with first core at 75 up-threshold.
9/1/11
-Fixed stuck-at-splash-screen.
-Removed A2SD for now.
9/6/11
Removed HTC IPC Server from ramdisk.
10/26/11
Updated kernels with HTC's ramdisk edits.
Added TrevE's ro.lb=unknown and ro.bootloader=unknown values to default.prop. TAKE THAT GOOGLE! Thanks TrevE!
Removed references to htc_ebdlogd
Set ANR (app not responding) history to 0
Set dirty_expire_centisecs to 100
Set capacitive lights to dim (2)
10/27/11
Added ScreenOff script
Added Uninstall ScreenOff script
1/13/12
Updated kernel
Added mpdecision (Thanks TrevE)
Removed insmod /system/lib/modules/qce.ko
3/12/12
Removed mpdecision and references to it in the kernel ramdisk
Updated unmodified and stock dual core kernels to include stock mpdecision so that users of the modified kernels can get mpdecision back when flashing unmodified kernels.
3/14/12
Changed loglevel to 0. Afterall, this is a stock kernel. The less logging the better.
Removed wimaxDumpKmsg
Removed wimaxDumpLastKmsg
Removed wimaxDumpLogcat
Removed bugreport
Removed loading and references to those files from the Kernel ramdisk. Don't worry, WiMax still works. If anything it will work just a bit better now that it doesn't need to worry about dumping any information to logcat and kmsg.
Updated unmodified kernel package to include missing files for those who want to go back to pure stock after flashing one of these.
Here are the changes from 2.08.651.3 ramdisk to 2.17.651.5:
Code:init.rc Removed: mount debugfs /sys/kernel/debug /sys/kernel/debug Added: #set projector permissions chown system system /sys/devices/platform/projector/interface/enable Added: on property:crypto.driver.load=1 insmod /system/lib/modules/qce.ko
Not exactly sure why insmod /system/lib/modules/qce.ko is in there. This file is not in the OTA or full ROM package. But it looks like it only gets called if crypto.driver.load=1, but even so, the file isn't there. Strange. Gonna go ahead and remove that.
HTC has made the following ramdisk edits with 2.08.651.3 but actual kernel version hasn't changed.
Here are the edits:
Code:init.shooter.rc: Added: # to enable low charge current in voice call chown radio radio /sys/class/power_supply/battery/charger_control chmod 0660 /sys/class/power_supply/battery/charger_control Removed: # for vnc service androidvncserver /system/bin/androidvncserver -a disabled oneshot service usbnet /system/bin/usbnet on disabled oneshot Removed (in bold): on property:init.svc.bootanim=stopped [b] start thermald[/b] start mpdecision Changed: [b]FROM[/b] service mpdecision /system/bin/mpdecision --Nw 5.4 --Ts 480 --no_sleep --avg_comp --single_core_while_panel_off [b]TO[/b] service mpdecision /system/bin/mpdecision --Nw 5.4 --Ts 350 --no_sleep --avg_comp --single_core_while_panel_off
Kernel information:
freEVO 3D-StockKernel 2.17.651.5.zip: Dual Cores active. 2nd core max frequency 1.2GHz, Up-threshold at 90 (Stock setting). Minimum freq at 192MHz.
freEVO 3D-StockKernel 2.17.651.5b.zip: Dual Cores active. 2nd core max frequency 1.2GHz, Up-threshold at 75. Minimum freq at 192MHz.
freEVO 3D-StockKernel 2.17.651.5c.zip: Dual Cores active. 2nd core max frequency 594MHz, Up-threshold at 90 (Stock setting). Minimum freq at 192MHz.
freEVO 3D-StockKernel 2.17.651.5d.zip: Dual Cores active. 2nd core max frequency 594MHz, Up-threshold at 75. Minimum freq at 192MHz.
Unmodified:
freEVO 3D-StockKernel 2.17.651.5ORIG.zip
Modified (dual cores left at stock settings):
freEVO 3D-StockKernel 2.17.651.5e.zip
Dual Core ScreenOff script:
freEVO-3D-DualCoreScreenOffscript.zip (1.2GHz version):
freEVO-3D-DualCoreScreenOffscript594.zip (594MHz version):
This will install the 70cputweak file in /system/etc/init.d/. To check whether or not it's working (and if your kernel supports init.d), with your phone connected to ADB, simply type or copy: "cat /sys/devices/system/cpu/cpu1/online" If it returns a 0 while the screen is off, that means it is working. You can also do "cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq" while the screen is off, and if it returns a 384000 then it is working. Keep in mind that you may also edit this file with root explorer or notepad++ to your liking, but you will need to reboot for it to take effect.
Uninstall ScreenOff script
freEVO-3D-UninstallScreenOffScript.zip: This will simply delete the 70cputweak file in /system/etc/init.d. You can accomplish the same thing from adb or root explorer.
NOTE: Please make sure that you don't have any other CPU controlling app such as SetCPU, virtuous OC/OC daemon, CPU Tuner/CPU Master, 98dualcore (from ziggy kernel/synergy) installed if you decide to use this, as they will most definitely interfere with this script and I haven't coded a failsafe to disable it in their presence. You can essentially accomplish what this script does if you use any of the above mentioned apps, but this is mainly for those who want to 'set it and forget it.' Also, SetCPU is most definitely not compatible with this method of dual cores, script or no script. It will return cpu1's online state to default and will render this script as well as the kernel useless.
NOTE 2: If you are going to install another kernel and would like to uninstall dual cores active always, you will need to flash either the unmodified or modified without dual cores kernel first with a reboot afterward, as the ramdisk method employed here will survive many kernel replacements that use Koush's anykernel updater. I've thought about using init.d scripts rather than ramdisk method but eventually decided on this way. Sorry for any inconvenience!
ENOUGH CHOICES?
HBOOT 1.5 users? or just want an easier way to flash things? Joeykrim has created a fantastic app that will allow you to flash kernels, among other things, directly from Android. Check it out here: http://xdaforums.com/showthread.php?p=161299165I'm sure most of you know this, but since I made the mistake, ill share it. Don't flash the kernel unless your on a 2.3.4 rom. It will break the camera and wifi, among other things. If you did flash, just reflash the old one.3Maybe it will help someone, maybe it wont. we went through the framework and logging tied into services.jar/framework.jar/com.htc.framework.jar. There are 3 services
Dropbox Service (easily disabled)
UserBehaviorLogging (easily disabled)
HTCs App Usage stats (not easily shut off)
The first 2 services apear to be googles stuff. I believe thse were just in services.jar. The HTC stuff was all over the place, the easiest way i found is to reverse the check where it goes to write data and checks that its a reasonable time and make it so itll error each time. You can also 400 the directories to stop it from writing.
There is also a /data/system/userbehavior.db that gets written out. This has the IP address in a policy(iirc) table. I made it so it writes 127.0.0.1 in the framework right to this db. It actually makes a raw socket connect right from framework, nothing in app/bin/apks.
I had intentions of diffing all my changes (i dont even remember) and documenting, i just got behind on another project....
