Hi... On Mon, May 11, 2009 at 4:22 PM, Sukanto Ghosh <sukanto.cse.iitb@xxxxxxxxx> wrote: > Hi, > > I was adding a field to the 3 trace_mark() calls in kernel/sched.c. > The calls are at: i) context_switch(prev, next, ... ) ii) > try_to_wake_up(p, ...) iii) wake_up_new_task(p, ...) functions > > The field is 'task'->mm->pfrate. (I have added the pfrate field in > mm_struct), where 'task' is a placeholder for prev/next/p/rq->curr in > those trace_mark() calls. I found that always either the next/p > pointers are NULL or task->mm is NULL at that particular point. Is it > supposed to be so ? Why ? > > PS: The trace_mark() in kernel/sched.c calls are for adding entries to > the trace-file of sched_switch tracer. Always? hmm strange. sure you did the deferencing to task_struct correctly? perhaps you could share to us the change you made? in a form of diff -u perhaps? regards, Mulyadi. -- To unsubscribe from this list: send an email with "unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx Please read the FAQ at http://kernelnewbies.org/FAQ
