Ghokanmoral and some kernel devs over the i9100 suspected the following:
- with latest JB ROMs, /data partition gets corrupted on unexpected shutdown
- changing the init binary with older ones will let the system boot even if /data is corrupted
- some ROMs seem immune to the issue, what ever kernel/init binary combo is used
- changing the init, will not fix the /data corruption issue
- many parts were suspected like bootloader, some binaries related to toolbox..., none showed to be the reason
The only person I know that can give a definite explanation for this is Ghokanmoral , but his dev over the i9100 is for now on hold because of personal reasons
In my case, using an older JB ROM with latest kernels, repacked or stock, will make all reboots work. Using any newer ROM with a non repacked kernel will stuck in the bootloop
I sadly have no time at all to focus on this. As I said, I stick to stock kernels and I do not plan any soon to start compiling kernels from source. I only focus on coding for CWM and maintaining ramdisk changes for the kernel whenever it is needed or a fix proven to be effective
- with latest JB ROMs, /data partition gets corrupted on unexpected shutdown
- changing the init binary with older ones will let the system boot even if /data is corrupted
- some ROMs seem immune to the issue, what ever kernel/init binary combo is used
- changing the init, will not fix the /data corruption issue
- many parts were suspected like bootloader, some binaries related to toolbox..., none showed to be the reason
The only person I know that can give a definite explanation for this is Ghokanmoral , but his dev over the i9100 is for now on hold because of personal reasons
In my case, using an older JB ROM with latest kernels, repacked or stock, will make all reboots work. Using any newer ROM with a non repacked kernel will stuck in the bootloop
I sadly have no time at all to focus on this. As I said, I stick to stock kernels and I do not plan any soon to start compiling kernels from source. I only focus on coding for CWM and maintaining ramdisk changes for the kernel whenever it is needed or a fix proven to be effective