ipaq3870

Members
  • Content count

    92
  • Joined

  • Last visited


Community Reputation

17 Good

About ipaq3870

  • Rank
    Regular

ipaq3870's Activity

  1. ipaq3870 added a post in a topic Kernel Experiment - 4.0 updated 2012.05.28   


    Arrgghhh... there's a bad show-stopper bug cropped up - video like youtube and camera preview don't work with the changed 32bpp.... and the release will be delayed until I am back.



    Try clean the data of talk, market, market updater, android framework, and the like.



    So, it won't turn off when pressing the power/hold button? And the screen stays powered on indefinitely? Does it respond to other button presses? Can you stay use the "voldown+back+menu" kill combo to reboot the phone? And is this reproducible every time you do a reboot via kexec?



    Perhaps in next version.
    • 0
  2. ipaq3870 added a post in a topic Kernel Experiment - 4.0 updated 2012.05.28   

    So much work, so little time!

    I'll be offline until Jul real soon. The code is already in place, I'm just not sure if I have time to release a 5.0 build. Tentative changelog:
    Screen calibration: can calibrate and adjust to one's own device - see special instructions.
    Switch off touch panel on screen blanks - phone no longer wakes up twice when pressing hold button.
    Also "screen turned off delay" of the lockscreen can now be set to anything - it was having accidental wakeup problems.
    24bpp framebuffer now reports itself as 32bpp as expected by Android, though the hardware is still 24bpp.
    Screenshot apps such as AndroSS, and console apps such as Live Logcat now work properly in 24bpp.
    Vibrator now vibrates at 175Hz (strong) and is independent of CPU or AHB frequencies
    User selectable vibrator strength via sysfs file: /sys/class/timed_output/vibrator/freq

    • 0
  3. ipaq3870 added a post in a topic Kernel Experiment - 4.0 updated 2012.05.28   


    If there's anything to save across reboots, it's usually in the config files. That's why I was asking you all to try point 4 of this post - to reset the config file. But somehow none of you tried - or at least none reported trying. Hmmm...



    Finally someone succeeded. Nice.
    • 0
  4. ipaq3870 added a post in a topic Kernel Experiment - 4.0 updated 2012.05.28   



    See Caveat section - it's by design : sacrifice your video memory for something else. :D Choose a build with less free memory (mean more memory for video) helps.



    That means the command "getprop haret.kernel" return nothing. Check step 2 again. Note the setprop line must be placed before the line that reads "class_start default".



    It still baffles me as to why an Andriod reinstallation would work. It's not uncommon for me to switch kernels and modules 20-30 times a day if I am doing development, and still yet to encounter any gsm/wifi problems like you guys described. Anyway, good to know it works for you.
    • 0
  5. ipaq3870 added a post in a topic Kernel Experiment - 4.0 updated 2012.05.28   


    Again, see the first 2 points in the Troubleshooting post (2nd post of this thread).




    Good to know! What exactly you have done to make Wifi work? Did you reinstall the whole OS via the beta2 installer, or just the kernel part? What other things you have done? Perhaps these are useful info for others.
    • 0
  6. ipaq3870 added a post in a topic Kernel Experiment - 4.0 updated 2012.05.28   


    I applaud your persistence. But you still must have typed something wrong. Typing serious commands in terminal is really a challenge and a PITA, I know. I attach the script so that you don't have to type.

    1. Download the ke.txt
    2. Rename it to ke (without the extension)
    3. Put it in say /ke of your phone
    4. Do a "chmod +x /ke"
    5. Run "su"
    6. Disconnect USB
    7. Run "/ke"

    See if this works.
    ke.txt
    • 0
  7. ipaq3870 added a post in a topic Kernel Experiment - 4.0 updated 2012.05.28   


    Check steps 7 - 8 again.



    See if the first 2 points in the Troubleshooting section of this post helps? If necessary, reinstall from scratch may help according to various members who did that successfully.



    Touch? IIRC, I didn't touch the touch part of the kernel, i.e. the touch screen driver. Did you try any prior version(s) of my kernel and encounter this issue? As far as I know, there are calibration data hardcoded in s3c-ts.c. You may want to take a look at it.

    As for GPS, I think I didn't touch that as well, and the problem seems stemmed back at least from at least cm b2. I guess you can find your answer in the stopped b2 thread. I have the feeling that it's an Android build problem though I didn't really look at it.
    • 1
  8. ipaq3870 added a post in a topic Kernel Experiment - 4.0 updated 2012.05.28   


    These are shell programming commands - be very careful about the quotes:
    Double quote --> " (?123 key then c)
    Back quote --> ` (?123 key then ALT key then 2)

    You pasted the command correctly in this post but apparently not in the command terminal. There are non-breaking double quote-back quote or back quote-double quote sequences. And there's no dot between getprop and haret.kernel as well.

    Also remember to unplug USB before trying.
    • 0
  9. ipaq3870 added a post in a topic Kernel Experiment - 4.0 updated 2012.05.28   

    Did you setprop haret.kernel properly, and chown/chmod properly?

    You can verify it by running the manual script. If it does not work and just spits out error messages, then recheck steps 1 - 6.

    If it does work via manual script but not in Android UI, check steps 9 - 11.

    If the Android UI hangs at rebooting phone, check steps 7 - 8.
    • 0
  10. ipaq3870 added a post in a topic Kernel Experiment - 4.0 updated 2012.05.28   


    It's a known issue already stated in the Caveat section. Seems the vibrator somehow depends on the max speed of your CPU.

    EDIT: Okay, it's not difficult to fix - just borrowed some code from the spica forum. :D Download and replaced the original vibrator.ko in /modules/ and you are set. Also a bonus feature: vibration strength can be user tuned via this sysfs parameter /sys/devices/virtual/timed_output/vibrator/freq. This hotfix is from commit 5e743b13 hacked to be used in official 4.0 builds, and will be integrated when a new major release is out.
    vibrator-5e743b13.zip
    • 0
  11. ipaq3870 added a post in a topic Kernel Experiment - 4.0 updated 2012.05.28   


    You guys don't even wait for the docs are posted before trying out. :P libc.so should go to /system/lib/.



    Try doing nothing with your phone for few minutes to let it run into sleep mode itself, wait for another minute after the screen blanks. Wake it up and connect the USB again, and if necessary a couple of times if the first time won't work.



    On the contrary, one AHB overclocked (166MHz) build actually do have more free memory. You need the vanilla build or the one with 141MB free.



    The "free" memory may not be what's actually free. Depending on how it's reported, it may or may not include buffer memory or page cache. So it varies very much. What you should concern most if the total amount of available memory, should be more authoritative.



    Depending on how you executed the zr script, it may or may not have successfully unmounted your old swap partition on the SD card. But it really doesn't matter - as Z-Ram has priority over it and it's almost never used even when you have 2 swap spaces active. If you want to make sure you are not using the old swap, do a "swapoff -a" in terminal manually, and check with the free command again.
    • 0
  12. ipaq3870 added a post in a topic Kernel Experiment - 4.0 updated 2012.05.28   


    It's a known hardware limitation. See the Caveat section. But it's been worked around perfectly in the latest 4.0 builds.



    Ok, what about the command "insmod /modules/libertasspi.ko" (without -f)? Is there any error? What the log shows? Just another thing you can try: switch back to the original cm beta 2 kernels and modules, and see if it works. If not, then you'll know it's not kernel related but rather problems with the android, filesystem, configuration, or even your wifi infrastructure.



    What do you mean by "unlocked"? There's nothing being locked - you can just modify various PLL settings and clock dividers to get different freq. Mind you though the if you are talking about overclocking the CPU, ours S3C6410 was originally designed to only at max 667MHz - and they were already pushing it to the 800MHz that you are using today. Quite a stretch already. The knowledgeable spica guys over the samdroid forum has found its max stable freq is ~890MHz but benchmark don't show much difference.



    I have already edited the script few days ago to silence the error. Go get the latest one.
    • 0
  13. ipaq3870 added a post in a topic Kernel Experiment - 4.0 updated 2012.05.28   

    Arghhh! Finally finished writing the documentation of 4.0 sans the development bits... wasted so much time wrangling with modaco's Wysiwyg editor :angry:

    Will take a much needed rest....
    • 0
  14. ipaq3870 added a post in a topic Kernel Experiment - 4.0 updated 2012.05.28   

    Kexec a.k.a. Never See WinMo Again (Almost)

    These are the files needed for using kexec - for kernel 4.0 or above. See first post for details.
    framework.part1.rar
    framework.part2.rar
    libc.zip
    • 0
  15. ipaq3870 added a post in a topic Kernel Experiment - 4.0 updated 2012.05.28   

    4.0 Even more memory

    CPU: 667Mhz / AHB: 166Mhz / 160MB free memory / 24bpp display
    Build No: a3b8d7f-951

    Note: always watch first post for latest builds in case this build has been superseded.
    4.0-ae49c12-951-667mhz-166mhz-160mb-24bpp.part1.rar
    4.0-ae49c12-951-667mhz-166mhz-160mb-24bpp.part2.rar
    • 0

MoDaCo is part of the MoDaCo.network, © Paul O'Brien 2002-2015. MoDaCo uses IntelliTxt technology.