Praveen, It looks to me that the kernel thread is not really killed and the address is invalid after unloading the modules (But the kernel thread still being scheduled) ! Just a guess. Can you add a new global variable checked in the kernel thread and if it is set, just exit from the kernel thread rather than killing from out side. From the module cleanup function, set the kill variable and exit from the cleanup function only after kernel thread is killed itself. This is just my thought ... kill_proc() working properly would have been nice ! Aboo -----Original Message----- From: kernelnewbies-bounce@xxxxxxxxxxxx [mailto:kernelnewbies-bounce@xxxxxxxxxxxx] On Behalf Of Praveen Nair Sent: Monday, October 11, 2004 2:25 PM To: kernelnewbies@xxxxxxxxxxxx Subject: Problem killing a kernel_thread. Hi, I am writting a kernel module with a thread that waits for accepting socket connections (sock->ops->accept()). This thread is spawned from the init function and i am trying to kill the thread from the exit function using kill_proc(tid, SIGKILL,1). But as soon as i unload the module i get the following error. Unable to handle kernel paging request at virtual address d296e337 printing eip: d296e337 *pde = 01bae067 *pte = 00000000 Oops: 0000 i810_audio ac97_codec soundcore agpgart parport_pc lp parport nfsd lockd sunrpc autofs ds yenta_socket pcmcia_core e100 keybdev mousedev hid input usb-uhci eh CPU: 0 EIP: 0060:[<d296e337>] Tainted: GF EFLAGS: 00010246 EIP is at __insmod_i810_audio_S.data_L776 [i810_audio] 0x6077 (2.4.20-28.9) eax: fffffe00 ebx: 00000000 ecx: cb1918e0 edx: fffffe00 esi: 00000000 edi: 00000000 ebp: c6449fec esp: c6449f84 ds: 0068 es: 0068 ss: 0068 Process insmod (pid: 3953, stackpage=c6449000) Stack: cb1918c4 c64a5374 00000000 c0329da0 00000000 00000000 00000000 00000000 00000000 00000000 00000000 cff46b00 cff46b00 ced29e60 c0108d2e cff46b00 d296e254 00000000 00000000 00000000 703d0002 0100007f 00000000 00000068 Call Trace: [<c0108d2e>] ret_from_fork [kernel] 0x6 (0xc6449fbc)) [<c0106fc1>] kernel_thread_helper [kernel] 0x5 (0xc6449ff0)) Code: Bad EIP value. If i don't kill the thread then the unload happens without issues. Please do help me out. Regards, - Praveen -- Kernelnewbies: Help each other learn about the Linux kernel. Archive: http://mail.nl.linux.org/kernelnewbies/ FAQ: http://kernelnewbies.org/faq/
