If someone wants to analyse or track how the processes are scheduled under SMP machine, a possible way is to use relayfs to report log information. For logging purposes sometimes is wothwhile to create a struct that represents your log data.You shouldn't call a function that might sleep (such as kmalloc())
from schedule().
You may get away with it if you use GFP_ATOMIC or similar, but still
it sounds like a really bad idea.
Why do you think you need to call kmalloc() from schedule anyway?
Basically the malloc should be performed in schedule() function or in the sched_info_switch() function called by schedule() as:
my_data = (struct log *)kmalloc(sizeof(struct log), GFP_ATOMIC);
But even using the GFP_ATOMIC does not work in sched_info_switch().
Are there any difference to use kmalloc() in schedule() or schedule_info_switch() ?
BR,
Mauricio Lin.
Yahoo! Search
Dê uma espiadinha e saiba tudo sobre o Big Brother Brasil.
