Linux 2.6.36: kernel panic when deblobbed

[christophe.jarry at ouvaton.org]

Dear developers,

I have an issue with the kernel Linux 2.6.36-20.

I first used the deblob scripts available on
http://linux-libre.fsfla.org/pub/linux-libre/releases/2.6.36-gnu/ and deblobbed the kernel Linux
2.6.36-20 sources available from Lemote with:

    git clone git://dev.lemote.com/linux-3A.git
    wget http://linux-libre.fsfla.org/pub/linux-libre/releases/2.6.36-gnu/deblob-{2.6.36,check}
    chmod u+x deblob*
    cd linux-3A
    ../deblob-2.6.36 --force
    make mrproper

I used the default config file:

    cp arch/mips/configs/loongson3_config .config

Then I built the kernel image and installed the modules:

    make
    make modules_install

I copied vmlinuz to /dev/sda1. Then I modified boot.cfg in order to boot from this vmlinuz image.
I checked the config from Lemote vmlinuz available on /dev/sda1 and its content is the same as
mine, so the only difference between the kernel from Lemote and the one I built is that mine has
been deblobbed to remove non-free parts from the kernel.

Every time I boot on the deblobbed kernel, I have a kernel panic. Here are the two messages I
obtain that can lead to the panic:

CPU 0 Unable to handle kernel paging request at virtual address 0000000000e00000, epc ==
ffffffff802ddfa0, ra == ffffffff80267fc8.
Oops [#1]:
Cpu0
...
Not tainted
...
Call Trace:
... kfree
... call_usermodehelper_exec
... kobject_uevent_env
... device_add
... device_create_vargs
... device_create
... tty_register_device
... uart_add_one_port
... serial8250_init
... do_one_initcall
... kernel_init
... kernel_thread_helper
Disabling lock debugging due to kernel taint
Kernel panic - not syncing: Attempted to kill init!
panic occured, switching back to text console.

Serial: 8250/16550 driver, 4 ports, IRQ sharing disabled
Unhandled kernel unaligned access[#1]:
Cpu 0
...
Not tainted
...
Call Trace:
... do_ade
... ret_from_exception
... kmem_cache_alloc
... call_usermodhelper_setup
... kobject_uevent_env
... device_add
... device_create_vargs
... device_create
... tty_register_device
... uart_add_one_port
... serial8250_init
... do_one_initcall
... kernel_init
... kernel_thread_helper
Disabling lock debugging due to kernel taint
Kernel panic - not syncing: Attempted to kill init!
panic occured, switching back to text console.
BUG: scheduling while atomic: swapper/1/0x00000002
Modules linked in:
Call Trace:
... dump_stack
... schedule
... __mutex_lock_slowpath
... mutex_lock
... atom_execute_table
... radeon_atom_encoder_dpms
... drm_helper_disable_unused_functions
... drm_crtc_helper_set_config
... drm_fb_helper_force_kernel_mode
... drm_fb_helper_panic
... notifier_call_chain
... __atomic_notifier_call_chain
... panic
... do_exit
... default_cu2_call

So I would like to have your suggestions on those errors. If you need more information, please
tell me.

Thanks,

Christophe