>>Especialy for modules, you must make sure the thread has exited before >>you return from module_done(). And there is only one way to do it -- use >>complete_and_exit() to end the thread and wait_for_completion() in >>module_dome(). That means, that the thread must exit actively. >>Now the question comes down to how to tell the thread to exit. There are >>generaly two ways: >>1) Have a flag somewhere. Wake up the thread (by setting it's state to >> running or using a waitqueue) and have it check that flag and >> complete_and_exit() if it's set. >>2) Use a signal. You can send the thread a signal and it will cause it >> to wake up from interruptible sleep automaticaly. But you must check >> signal_pending() yourself and complete_and_exit() if appropriate bit >> is set. Normaly, signals are handled during return to userspace, but >> kernel threads never return to userspace. Thus you have to check for >> them manualy. I have a doubt regarding the usage of complete_and_exit it makes use of a completion structure. I am initialising it with init_completion() at init starting the thread and Wait_for_completion() at exit module for waiting for the thread a complete_and_exit() i am using at thread and i get the signal sigterm using signal_pending,... But is it not enough to just return a zero from the thread?.Is it just a synchronisation mechanism like waitqueue??.. Thanks in advance Regards Sudharsan -- Kernelnewbies: Help each other learn about the Linux kernel. Archive: http://mail.nl.linux.org/kernelnewbies/ FAQ: http://kernelnewbies.org/faq/
