JetKernel 2.6.29 testing release 2011-02-01

JetKernel testing releases will be posted here by the DevTeam. This forum is open for feedback from test users.
Forum rules
Dev Team members can post testing releases here. Please give a link to the source base that was used for a release or post patches / changelogs along with your binary packages.

Registered users can reply with their test feedback. Please only post test feedback here. For general (e.g. install related questions) please use the support Q&A section.

Re: JetKernel 2.6.29 testing release 2011-02-01

Postby MadRocker » Tue Feb 08, 2011 9:44 pm

Flickering is a video problem, not yet fixed.

I think my previous statement might be incorrect about this kernel as I might have contaminated the testing.
Why:
I have found issues with libChunkAlloc.so and gralloc.default.so that is causing some PMEM and ioctl issues.
This files needs to remain the same as was part of the original android-eclair platform or android-froyo platform build and could cause issues if you use another not part of the original platform build.

New Testing:
Cleaned then, Dopi install script install + yes to Gapps + this kernel nothing else added.
Normal chmod safely remove.
1St boot, gets into android gui then hangs and kills itself.
2nd boot,

2nd Testing:
No Gapps only Dopi install script + 1Feb Kernel.
Menu button shows keyboard if you hold it for 2 secs, press it quickly then normal menu bar. Calibration / touchscreen not happy can not click anything specific, but registers a click and removes content from screen.
On power off click power = yes orange, 2nd power off click calibration a bit off, then restarts the android animation back into the android gui.

Android.log = gralloc video mem allocation error see notes above for gralloc + chunk .
Code: Select all
W/SurfaceFlinger( 1817): mBuffers[0] is null (mWidth=480, mHeight=430)
E/gralloc ( 1817): gralloc failed err=Out of memory
W/        ( 1817): alloc(480, 430, 1, 00000133, ...) failed -12 (Out of memory)
D/        ( 1817): Allocated buffers:
D/        ( 1817):   0x121750:  750.00 KiB |  480 x  800 |  4 | 0x00000133


The Gapps version does not have this as it adds some needed files even though this Gapps is unstable and causes a lot of issues.

Both kernel.log's looks good no apparent issues, Note: missing lib files triggers more devices / data in kernel log.

Modules:
I get the same when this modules using KB 29 Jan kernel, keymaps and calibrations is unstable. No /boot folder. See test 1 and test 2 phone cloned for details.

Might need this for init.rc cpu speed governing. "up for review"
Code: Select all
############################# MadRocker 6 Feb ###############  start  Cpu speed govrn. + Compass needs correcting to JetHardware ################
# for ctrl cpu freq.
    chown system system /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor
    chmod 0666 /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor
    write /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor "conservative"

#magnetic sensor yamaha
    chown compass system /dev/i2c-12
    chmod 0777 /dev/i2c-12
    chown compass system /dev/i2c-5
    chmod 0777 /dev/yamaha_compass
    chown system system /dev/light
    chmod 0777 /dev/light
    chown system system /dev/proximity
    chmod 0777 /dev/proximity
############################# MadRocker 6 Feb ###############  end  ################


Conclusion:
Test 2 appears much more stable fix = add /boot/append-jet and it works nicely.
The 2D/3D drivers is still using android.so software renderer even though the modules load it is not triggered as default video driver.
Working on this. I presume somehow with a egl.cfg we need to specify the "s3c 3d" driver as default or loaded as supplement to android.so same as what happens with android.so and fimg.so . Power vg androids load it similar to our fimg loading.

Gapps is unstable irrespective of kernel or modules. Might need to re-build a lite Gapps version.

-- Modules:
Wifi dhd.ko connects, problem when adding wep / wpa password the numbers does not show up, possible android.so or a ../lib file missing.

-- Memory:
Will need to get memory sorted out as there is a bit of lag but could also be because of video driver android.so . Menu button wallpaper is nice and responsive with test2 and /boot/append-jet
Bluetooth might be active but would need a .conf and some init.rc possibly , same with other modules.
MadRocker
Dev Team
Dev Team
 
Posts: 528
Joined: Sun Dec 12, 2010 7:15 am

Advertisement

       

Re: JetKernel 2.6.29 testing release 2011-02-01

Postby Dopi » Wed Feb 09, 2011 12:42 pm

MadRocker wrote:Conclusion:
Test 2 appears much more stable fix = add /boot/append-jet and it works nicely.
The 2D/3D drivers is still using android.so software renderer even though the modules load it is not triggered as default video driver.

Hi MadRocker,

thanks for your detailed feedback. Can I take this as a sign that this JetKernel testing release (including all modules) is working for you?

The calibration (/boot/append-jet) is of course optional. The current settings that I included in the kernel work fine for me. We should find values that work for the majority of users.

Cheers,
Dopi
Ad banners support the JetDroid forum. Please consider clicking them once in a while.
User avatar
Dopi
Dev Team
Dev Team
 
Posts: 926
Joined: Sun Aug 22, 2010 9:47 pm

Re: JetKernel 2.6.29 testing release 2011-02-01

Postby HenriA » Wed Feb 09, 2011 12:59 pm

or run the jetdroid calibration at first run. just like gapps. that would fix the calibration issue, or not?
The jet will take-off to meet the android people! :)
HenriA
Forum Moderator
Forum Moderator
 
Posts: 203
Joined: Thu Dec 09, 2010 5:45 pm
Location: Netherlands

Re: JetKernel 2.6.29 testing release 2011-02-01

Postby Kaesar » Wed Feb 09, 2011 2:01 pm

joopykoopy wrote:or run the jetdroid calibration at first run. just like gapps. that would fix the calibration issue, or not?


The gapps calibration has nothing to do with the touchscreen driver im afraid. But the default screen calibration is already included in the kernel.

Dopi, does the append-jet file override the defaulted values in the kernel if the file exists?
Kaesar
Dev Team
Dev Team
 
Posts: 114
Joined: Thu Oct 21, 2010 4:27 pm

Re: JetKernel 2.6.29 testing release 2011-02-01

Postby Dopi » Wed Feb 09, 2011 3:29 pm

[quote="Kaesar"
Dopi, does the append-jet file override the defaulted values in the kernel if the file exists?[/quote]
Yes, it should. This was the reason I put the default value into the kernel. My first idea was to have default uncalibrated values ( absolute min, absolute max for X and Y ) to enable proper calibration. But as the default values seemed to be okay for many users and you could apply these min/max values for calibration via append-jet I made the switch to real calibration values as default.

Cheers,
Dopi
Ad banners support the JetDroid forum. Please consider clicking them once in a while.
User avatar
Dopi
Dev Team
Dev Team
 
Posts: 926
Joined: Sun Aug 22, 2010 9:47 pm

Re: JetKernel 2.6.29 testing release 2011-02-01

Postby HenriA » Wed Feb 09, 2011 6:59 pm

Kaesar wrote:
joopykoopy wrote:or run the jetdroid calibration at first run. just like gapps. that would fix the calibration issue, or not?


The gapps calibration has nothing to do with the touchscreen driver im afraid. But the default screen calibration is already included in the kernel.



I did not say that. There are people having trouble with the default callibration of jetdroid. If we could make the jetdroid calibration utility to run at the first run, it is fixed. People can calibrate themselves. Or am I thinking the wrong way?
The jet will take-off to meet the android people! :)
HenriA
Forum Moderator
Forum Moderator
 
Posts: 203
Joined: Thu Dec 09, 2010 5:45 pm
Location: Netherlands

Re: JetKernel 2.6.29 testing release 2011-02-01

Postby MadRocker » Wed Feb 09, 2011 7:19 pm

Dopi I would need to do more testing on the modules + fimg as I had issues in the past.
Let me get back to you on this. Then we can know for sure unless someone else beats me too it LOL.

Come on guys replicate actions then test then reports finding, you all need to help out.

Calibration = Jet-Ginger and the 29Jan kernel then there is no need of calibration so something else interferes with this calibration in the kernel or something else is needed.

appent-jet overrides the kernel options form what I have seen so far.
MadRocker
Dev Team
Dev Team
 
Posts: 528
Joined: Sun Dec 12, 2010 7:15 am

Re: JetKernel 2.6.29 testing release 2011-02-01

Postby craigbeat » Wed Feb 09, 2011 10:08 pm

You guys are doing a fantastic job so far. One thing I was thinking about, though, in terms of speed. My girlfriend has an HTC Hero, and recently I've stuck a custom 2.2 Rom on it. With it, it brought Apps2SD. I have noticed (as have others) that when apps are run from the SD card, they do run slower compared to phone storage. I figure that since we run the entire system from SD, this could be part of our speed problems. Another thing to notice is that if you take the battery cover off the Jet and touch the SD Card, it gets extremely hot when doing lots in Jetdroid.

Anyway, I was just browsing the xda dev forums, and spotted this:
http://forum.xda-developers.com/showthread.php?t=948001

I have no idea whether it helps or not, so apologies up front if it wastes any time.
craigbeat
Freshman
 
Posts: 1
Joined: Sat Oct 23, 2010 6:46 am

Re: JetKernel 2.6.29 testing release 2011-02-01

Postby alfonzobonzo » Wed Feb 09, 2011 11:07 pm

MadRocker wrote:Come on guys replicate actions then test then reports finding, you all need to help out.


Sorry mate, I'm just completely lost now what version of what go's with which version of that..... :roll:

I tried testingearlier but just ended up screwing my install
Do Androids Dream of Electric Sheep?
User avatar
alfonzobonzo
Dev Team
Dev Team
 
Posts: 143
Joined: Thu Oct 21, 2010 7:00 pm
Location: UK

Re: JetKernel 2.6.29 testing release 2011-02-01

Postby MadRocker » Thu Feb 10, 2011 4:12 am

Alfonz, no problem, the default pre0.3 or Dopi install script is what you can use for testing.
Froyo and Ginger is separate and for the future so can post their testing on them in their own forums.

Yes for testing you will need to make a desktop folder then copy sdcard ext3 content to this new desktop folder, then sudo rm -rf /media/1.5gb/* or similar or manual delete and then put new files on the now clean ext3 sdcard space for JetDroid.

Dopi good news 1Feb kernel + 1Feb modules = stable.
Not having append-jet causes a mem alloc error when sensors.S8000.so is on sdcard.

RTC is still at 2x normal, current Freq: 400MHZ , see cpu freq govern. in 8Feb init.rc , running quadrant will push the speed to 800mhz and possibly 1:1 seconds count ratio ?

Easter egg's:
Tested with latest 8Feb video patch and works nicely, screen brightness works, rotation works, vibrate works without modules with 29Jan kernel or 1Feb kernel with the 8Feb video patch init.rc .
evdev accelerometer, saw it last with quadrant.apk and system info option on the bottom of this system info page, might have to do something to see it again.

MEM:
Available: 77MB + 37MB in 10 Other: 25MB in 3

Error:
Only error with this is normal calendar gui error + android.log egl swap mem error.
8Feb init.rc wifi and auto mount needs to be fixed still when we get time.

How to replicate Final test.:
Dopi installer script + (no Gapps) + 19Dec sensors patch + 1Feb kernel + 1Feb modules + 8Feb video patch = running nicely.

Thanks all for the hard work.
MadRocker
Dev Team
Dev Team
 
Posts: 528
Joined: Sun Dec 12, 2010 7:15 am

PreviousNext

Return to JetKernel Testing

Who is online

Users browsing this forum: No registered users and 1 guest

  • Advertisement
cron