• After 15+ years, we've made a big change: Android Forums is now Early Bird Club. Learn more here.

Root [UPDATED 8/8/12]Breath of fresh air....an update

phillip_viado said:
Can we just make a rom or a flashable zip for the leak ics. If we get a working ics rom. I really dont care if we dont get ics update at all, as long theres a ics rom
Click to expand...
The kernel for the leaked update is completely locked down (as in you can't even delete some system files without it locking up). So is the latest one I was able to get my hands on. Making a flashable rom will not work and I'm pretty sure they're proud about finding a way to do it. At this point, our best bet is for the dev's to work their magic on CM9. If an official ICS or CM9 build does not come out by end of year, I foresee myself just flashing the V7 leak and using that till my contract is up next year.
 
I am thinking jb might be a better target. The ics code for aokp is rotting. And isn't the msm86xx the reference platform for jb anyway?
 
There is no point in working on Jelly Bean until we get CM9 booting. F*ck AOKP..... We can't even get a 3.0 kernel booting past the LG Logo. I'm frustrated with this right now. I'd rather get ICS running first. Otherwise we probably don't have a snowball's chance in hell. So IMHO, F*ck Jelly Bean. We don't even know if we're getting ICS officially at this point. Also, I'm trying to get the CM kernel for the p930 building for our device. However I'm stuck. If you want to help great! If not.... then Idk. I need another set of eyes on this crap.

Code:
https://github.com/death2all110/lge-kernel-iproj

build.log:
i_vzw kernel build.log - Pastebin.com

This is what intrigues me
Code: 
config LGE_I_DISP_OV_MUTEX
    bool "fix overlay lock-up issue"
    default y
    ---help---
    select this feature for resolving overlay lock-up issue
The above snippet is in the 3.0 kernel for the Spec AND the CM 2.6.35 for the P930.
Lines 81-85
https://github.com/death2all110/lge...msm-2.6.35/lge/com_device/display/Kconfig#L81
 
death2all110 said:
There is no point in working on Jelly Bean until we get CM9 booting. F*ck AOKP..... We can't even get a 3.0 kernel booting past the LG Logo. I'm frustrated with this right now. I'd rather get ICS running first. Otherwise we probably don't have a snowball's chance in hell. So IMHO, F*ck Jelly Bean. We don't even know if we're getting ICS officially at this point. Also, I'm trying to get the CM kernel for the p930 building for our device. However I'm stuck. If you want to help great! If not.... then Idk. I need another set of eyes on this crap.

Code:
https://github.com/death2all110/lge-kernel-iproj

build.log:
i_vzw kernel build.log - Pastebin.com

This is what intrigues me
Code: 
config LGE_I_DISP_OV_MUTEX
    bool "fix overlay lock-up issue"
    default y
    ---help---
    select this feature for resolving overlay lock-up issue
The above snippet is in the 3.0 kernel for the Spec AND the CM 2.6.35 for the P930.
Lines 81-85
https://github.com/death2all110/lge...msm-2.6.35/lge/com_device/display/Kconfig#L81
Click to expand...


il take a look as soon as i see a "breath of fresh air" on my projects.
 
So I'm not sure if this is helpful but I put some Slog stops in a couple files to see if I could catch where the boot stalls. I compiled CM9 with the prebuilt kernel. Here are the results:

LOGCAT #1:
I/SystemServer( 361): Window Manager
I/InputManager( 361): Initializing input manager
I/InputManager( 361): InputManager.java - After Initializing input manager
W/StatusBarManager.java( 361): Before mContext


StatusBarManager.java:
StatusBarManager(Context context) {
Slog.w("StatusBarManager.java", "Before mContext");
​
mContext = context;
​
Slog.w("StatusBarManager.java", "After mContext");
​
}


LOGCAT #2:
I/InputManager( 359): Initializing input manager
I/InputManager( 359): InputManager.java - After Initializing input manager
W/StatusBarManager.java( 359): Before mContext
W/StatusBarManager.java( 359): After mContext
W/StatusBarManager( 359): warning: no STATUS_BAR_SERVICE


StatusBarManager.java:
private synchronized IStatusBarService getService() {
if (mService == null) {
mService = IStatusBarService.Stub.asInterface(
​
ServiceManager.getService(Context.STATUS_BAR_SERVICE));
​
if (mService == null) {
​
Slog.w("StatusBarManager", "warning: no STATUS_BAR_SERVICE");
​
}
​
}
​
Slog.w("StatusBarManager.java", "Before return mService");
​
return mService;
​
}

The first time it appears to have locked up at "mContext = context;" and it looks like it locked up the second time at my "Slog.w("StatusBarManager.java", "Before return mService");". I'm not sure what this means. Any ideas?
 
Im thinking its kernel related. There's a 'fix' in the 3.0 kernels that fixes overlay hangs. I was trying to build the p930 cm kernel but it kept failing, now im working on the backport of the said fix, and KGSL backport. I'm close, errors still but more manageable.


Update: So close to getting our kernel with KGSL and Overlay Hang fix backported to build. I keep getting
Code: 
drivers/gpu/msm/kgsl.c:52: warning: 'kgsl_ion_client' defined but not used
error, forbidden warning: kgsl.c:52
Which apparently means that
Code: 
warning: 'kgsl_ion_client' defined but not used
is a 'forbidden warning'. I have no idea what the hell that means/whats involved in fixing that.


TL;DR: I *think* I'm close
 
death2all110 said:
Im thinking its kernel related. There's a 'fix' in the 3.0 kernels that fixes overlay hangs. I was trying to build the p930 cm kernel but it kept failing, now im working on the backport of the said fix, and KGSL backport. I'm close, errors still but more manageable.


Update: So close to getting our kernel with KGSL and Overlay Hang fix backported to build. I keep getting
Code: 
drivers/gpu/msm/kgsl.c:52: warning: 'kgsl_ion_client' defined but not used
error, forbidden warning: kgsl.c:52
Which apparently means that
Code: 
warning: 'kgsl_ion_client' defined but not used
is a 'forbidden warning'. I have no idea what the hell that means/whats involved in fixing that.


TL;DR: I *think* I'm close
Click to expand...
Backport ion too ;)
 
Try "Reaver -I mon0 -b 00:1G:CD:37:C7:70 --lock-delay=5 -d 1 -v".. Don't know much about Droid devving but this line got me some wifi's so it might can help here
 
PlayfulGod said:
Backport ion too ;)
Click to expand...
I guess I'll have to, I was hoping I wouldn't have to because then I'd have to recompile cm9 with ION support (i had taken it out). The CM 2.6.35 kernel for the P930 doesn't have it backported. Maybe adding the ion.h file will help.

blaze413 said:
Try "Reaver -I mon0 -b 00:1G:CD:37:C7:70 --lock-delay=5 -d 1 -v".. Don't know much about Droid devving but this line got me some wifi's so it might can help here
Click to expand...

Can't tell if serious, or trolling.....

....

Wtf are you talking about!? I'm talking about kernels and trying to get CM9 to boot, and you're talking about WiFi!? Lol.
 
death2all110 said:
I guess I'll have to, I was hoping I wouldn't have to because then I'd have to recompile cm9 with ION support (i had taken it out). The CM 2.6.35 kernel for the P930 doesn't have it backported. Maybe adding the ion.h file will help.



Can't tell if serious, or trolling.....

....

Wtf are you talking about!? I'm talking about kernels and trying to get CM9 to boot, and you're talking about WiFi!? Lol.
Click to expand...

Lmao
 
Well when your left eye hurts like a bitch because the wind blew a hot cigarette ash into your eye, its kind of hard to read out of it. And my right eye is a stronger prescrip glasses wise, and those are at work in my locker. So..
 
death2all110 said:
Well when your left eye hurts like a bitch because the wind blew a hot cigarette ash into your eye, its kind of hard to read out of it. And my right eye is a stronger prescrip glasses wise, and those are at work in my locker. So..
Click to expand...

Super-Dev's only weakness cigarette ash in the eye.
 
And ION Pukes. awesome. Lovely mess of errors. Easy? Yeah. Okay. Whatever.

Update:
Turned off ion. fixed the forbidden warning error, and most the errors after. got as far as vmlinux.o (right before zImage) and it errored. I think i fixed it, but i'm going to bed.

I should definitely have a kernel build tomorrow night. I'll f*ck with ion later.

This is where it fails at now:
Code: 
 MODPOST vmlinux.o
  GEN     .version
  CHK     include/generated/compile.h
  UPD     include/generated/compile.h
  CC      init/version.o
  LD      init/built-in.o
  LD      .tmp_vmlinux1
[B]drivers/built-in.o: In function `kgsl_ebimem_free':
/home/blake/android/development/LG_Spectrum_Kernel-Devel/drivers/gpu/msm/kgsl_sharedmem.c:414: undefined reference to `free_contiguous_memory_by_paddr'
drivers/built-in.o: In function `_kgsl_sharedmem_ebimem':
/home/blake/android/development/LG_Spectrum_Kernel-Devel/drivers/gpu/msm/kgsl_sharedmem.c:622: undefined reference to `allocate_contiguous_memory_nomap'
make: *** [.tmp_vmlinux1] Error 1[/B]

I edited memory_alloc.h and memory_alloc.c accordingly. Same thing.
 
I think this can be done.

Thanks Death and the others for all your hard work, will be sending funding your way soon.
 
You must log in or register to reply here.
Back
Top Bottom