Thanks for answer me. Ok, it is better manipulate the functions where
task sleeps to wait on IO. But, there is a lot of kinds of IO... So, I need
to modify every IO function inside the kernel? I really don't know how
to do this because I even don't know this list of functions... Where
can I start? Is there another way to this task?
2005/6/28, Rik van Riel <riel@xxxxxxxxxxx>:
On Tue, 28 Jun 2005, Eduardo Bezerra wrote:
> 1. Is there any implementation of this task (getting IO time of the
> processes) already done in the kernel?
Not as far as I know, but I suspect this statistic will
be useful.
It will be good if the statistics are also exported
through /proc somehow.
> 2. If not, Is this idea correct? Adding a new field in the sched_info to
> hold IO spent and manipulate it in
> functions like __add_wait_queue and __remove_wait_queue or active_task and
> deactivate_task.
Better manipulate this directly in the functions where the
task sleeps to wait on IO, not in the wait queue functions,
because those also get called from eg. the semaphore code.
> 3. How do activate_task and deactivate_task work?
You should not have to modify these.
--
"Debugging is twice as hard as writing the code in the first place.
Therefore, if you write the code as cleverly as possible, you are,
by definition, not smart enough to debug it." - Brian W. Kernighan
--
Kernelnewbies: Help each other learn about the Linux kernel.
Archive: http://mail.nl.linux.org/kernelnewbies/
FAQ: http://kernelnewbies.org/faq/
--
Eduardo Bezerra Valentin
Bacharelado em Ciência da Computação - DCC - UFAM
Laboratório de VoIP - GRSD - UFAM
