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

Root Kernel Modification

rm50

rm50

Well-Known Member
To Hroark and Team,
What hinders you from permission to access and modify the prevail kernal?
Is it missing script or possibly a code....or maybe both?
Ive been searching out Samsung codes and there is quite a list of them!
Im thinking that maybe we dont have proper permission ie: code to access where we need to get????
If Im on the wrong tract of thought please enlighten me

Ric
 
rm50 said:
To Hroark and Team,
What hinders you from permission to access and modify the prevail kernal?
Is it missing script or possibly a code....or maybe both?
Ive been searching out Samsung codes and there is quite a list of them!
Im thinking that maybe we dont have proper permission ie: code to access where we need to get????
If Im on the wrong tract of thought please enlighten me

Ric
Click to expand...

We can not compile a kernel zImage because we do not have complete source code from Samsung, they do release source code for the kernel but it is incomplete and when you compile the zImage it does not boot
 
hroark13 said:
We can not compile a kernel zImage because we do not have complete source code from Samsung, they do release source code for the kernel but it is incomplete and when you compile the zImage it does not boot
Click to expand...

GUess we have to wait for a leak! ;)

Script wise....Its probably a small but,important line of script in a vault and guarded 24/7 with alot of other goodies!
Samsung seems to be big with numerical codes,you dont think that has anything to do with it?
Ric
 
rm50 said:
GUess we have to wait for a leak! ;)

Script wise....Its probably a small but,important line of script in a vault and guarded 24/7 with alot of other goodies!
Samsung seems to be big with numerical codes,you dont think that has anything to do with it?
Ric
Click to expand...

Thanks for your response,Ive been reading alot about this subject and Im trying to understand it!
Ric
 
In case anyone else has something to add, here is what I've discovered from my experimentation.

Kernels compiled from the Samsung source technically run, they simply are unable to boot the system.

What I've done to determine this is attempt to connect with ADB. Depending on the error, I tweak the Ramdisk Image.

From what I can tell, the Kernel starts, but for some reason, cannot mount the /system partition.

I came pretty close to having a kernel that would give shell access through ADB but stopped before my company took it's Christmas break, and I've been lazy/swamped with work stuff since and haven't messed with it.

:)
 
Koumajutsu said:
In case anyone else has something to add, here is what I've discovered from my experimentation.

Kernels compiled from the Samsung source technically run, they simply are unable to boot the system.

What I've done to determine this is attempt to connect with ADB. Depending on the error, I tweak the Ramdisk Image.

From what I can tell, the Kernel starts, but for some reason, cannot mount the /system partition.

I came pretty close to having a kernel that would give shell access through ADB but stopped before my company took it's Christmas break, and I've been lazy/swamped with work stuff since and haven't messed with it.

:)
Click to expand...

did u check drivers/staging/android/logger.c, when compiling I saw there was a few errors in that file too.. but I am no pro by no means.
 
QuietStorm1785 said:
did u check drivers/staging/android/logger.c, when compiling I saw there was a few errors in that file too.. but I am no pro by no means.
Click to expand...

I get lots of warnings that things are defined and not used, or are uninitialized properly. I can't remember what it was, but there is one warning that comes up a LOT that, when I googled it, I found out that it was something that always comes up because it is no longer supported but all of the linux source still has the #ifdef for it
 
hmmm!

Soooo at that point it throws the error...does the rest of the script run or are there more errors down the line?
what version of python are you using?
are you using hroarks ramdisk packer?

Ric
 
rm50 said:
hmmm!

Soooo at that point it throws the error...does the rest of the script run or are there more errors down the line?
what version of python are you using?
are you using hroarks ramdisk packer?

Ric
Click to expand...
Ive googleing the heck out of this topic to learn and be of some help....LOL!
It helps when there is interaction on this subject!
to have an active conversation about it sometimes helps to find answers.
 
I think the only way we be able to compile a working kernel (if Samsung does not release working source code) would be to port another kernels source code that is known to work


This would be hard to do, and there would be a lot of bugs to work out.


A good example of this is, is how TomGiordano took the Samsung Crespo Kernel and ported it the ZTE Blade

Kwiboo Samsung Crespo Kernel (before port or unmodified)
https://github.com/Kwiboo/kernel_samsung_crespo

TomGiordano ports Kwiboo Samsung Crespo kernel to ZTE Blade
https://github.com/TomGiordano/kernel_zte_blade

You can look threw TomGiordano's history and commits you can see what he was doing. I may try this with the Prevail but I have not had too much time latley.

This seems like a good Kouma Kernel 3.0 project, hint, hint, nudge, nudge.

You can also take a look at how t0mm13b took TomGiordano's ZTE Blade kernel and added support for ICS here
https://github.com/t0mm13b/ics4blade_kernel
 
damn hroark you always find a solution, i didnt even know that was possible lol.
which kernel will you be porting? if you start working on this.
 
ron580 said:
damn hroark you always find a solution, i didnt even know that was possible lol.
which kernel will you be porting? if you start working on this.
Click to expand...


I would start with Kwiboo Samsung Crespo and do exactly what TomGiordano did but I would just use the files and settings from the Prevail kernel source code

I actually started doing this, but dropped it to work more on CM7, and now I am just too busy to do anything.


Once I have some time, I would like to finish CM7 and then maybe work on this (if someone has not done so already) cough cough Kouma cough
 
hroark13 said:
I would start with Kwiboo Samsung Crespo and do exactly what TomGiordano did but I would just use the files and settings from the Prevail kernel source code

I actually started doing this, but dropped it to work more on CM7, and now I am just too busy to do anything.


Once I have some time, I would like to finish CM7 and then maybe work on this (if someone has not done so already) cough cough Kouma cough
Click to expand...

Thanks for the links Hroark I got some googleing too do!
Ric
 
Whewee ! Just wanted to jump in and say it's good to see ya back in here regularly , Shabby . Hope all is well w/ Da Lil One . = )
 
SlasherVRGR said:
O.o I guess I've never seen it utilized properly :P
Click to expand...

Scripting languages go and out of vogue. Part of the natural cycle and way of things.

My company fields multi-language solutions for testing computer chips during manufacturing that are used by semiconductor manufacturers worldwide. It wouldn't be unreasonable for me to claim that the machine you used to post with or the chips you carry in your pocket were tested using Python programming.

The skills of devs trump languages every time.

It's for this reason that AndroidForums.com supports devs and why I will publicly defend them, every chance I get. :) ;)
 
EarlyMon said:
Scripting languages go and out of vogue. Part of the natural cycle and way of things.

My company fields multi-language solutions for testing computer chips during manufacturing that are used by semiconductor manufacturers worldwide. It wouldn't be unreasonable for me to claim that the machine you used to post with or the chips you carry in your pocket were tested using Python programming.

The skills of devs trump languages every time.

It's for this reason that AndroidForums.com supports devs and why I will publicly defend them, every chance I get. :) ;)
Click to expand...


I knew that python was being used some where in the process!
THANKS!
 
EarlyMon said:
Scripting languages go and out of vogue. Part of the natural cycle and way of things.

My company fields multi-language solutions for testing computer chips during manufacturing that are used by semiconductor manufacturers worldwide. It wouldn't be unreasonable for me to claim that the machine you used to post with or the chips you carry in your pocket were tested using Python programming.

The skills of devs trump languages every time.

It's for this reason that AndroidForums.com supports devs and why I will publicly defend them, every chance I get. :) ;)
Click to expand...

So, is it a possibility that samsung changed scripting languages when they started designing this kernel...or changed over in the middle of it because they thought some other language would be better?
Reason being someone said that when they tried to compile the kernel it looked unstable and threw an error in a certain area.....(he didn't say where)?
 
rm50 said:
So, is it a possibility that samsung changed scripting languages when they started designing this kernel...or changed over in the middle of it because they thought some other language would be better?
Reason being someone said that when they tried to compile the kernel it looked unstable and threw an error in a certain area.....(he didn't say where)?
Click to expand...


it has nothing to do with the scripting language


Samsung has security crap in the kernel that they dont want us to see,
the FSR drivers are missing, the stuff that that contains the CIQ crap,(yes there are elements of CIQ in the kernel), and other security stuff

If you compile the source code that they make available on their open source site, the resulting zImage is smaller than the zImage we get with the phone and it does not boot the phone

They are leaving stuff out because they don't want us seeing it

They have to post their source code for the kernel according to the Open Source License Agreement, so they are putting out this incomplete code to satisfy that requirement, but they are not releasing their proprietary parts of that code, which leaves it incomplete, and not able to build a working image.

It has nothing to do with Python, and Python is not even needed to compile a kernel. The build tools needed to compile the kernel are EABI or something like that, and can be found easily on the Internet
 
hroark13 said:
it has nothing to do with the scripting language


Samsung has security crap in the kernel that they dont want us to see,
the FSR drivers are missing, the stuff that that contains the CIQ crap,(yes there are elements of CIQ in the kernel), and other security stuff

If you compile the source code that they make available on their open source site, the resulting zImage is smaller than the zImage we get with the phone and it does not boot the phone

They are leaving stuff out because they dont want us seeing it

They have to post their source code for the kernel according to the Open Source License Agreement, so they are putting out this incomplete code to satisfy that requirement, but they are not releasing their propietary parts of that code, which leaves it incomplete, and not able to build a working image.

It has nothing to do with Python, and Python is not even needed to compile a kernel. The build tools needed to compile the kernel are EABI or something like that, and can be found easily on the internet
Click to expand...

Ok.......That settles it then......Sorry to beat your head off with this! Or anyone else!
Just trying to get some things clarified!
 
You must log in or register to reply here.
Back
Top Bottom