Attilla Danko(deleted)
03/10/2008 6:10 PM
post5669
|
Primarily in:
trunk/services/system/ker/nano_sched.c
Search for SCHED_FIFO and SCHED_RR
-ad
chris PANAGIOTOU wrote:
> thanks for, the reply.
> Ok, something else now, some functions related with the scheduler
> receive as arguments the policies implemented as SCHED_FIFO, SCHED_RR
> etc. The only related references i've noticed are some defines . My
> query is, where the implementation of these algorithms (fifo, round
> robin, sporadic...) is located in the source......
>
> thanks!!!
>
> _______________________________________________
> OSTech
> http://community.qnx.com/sf/go/post5668
>
|
|
|
Thomas Fletcher
03/10/2008 7:05 PM
post5672
|
On Mon, Mar 10, 2008 at 5:49 PM, chris PANAGIOTOU <chpanag@gmail.com> wrote:
> thanks for, the reply.
> Ok, something else now, some functions related with the scheduler receive
> as arguments the policies implemented as SCHED_FIFO, SCHED_RR etc. The only
> related references i've noticed are some defines . My query is, where the
> implementation of these algorithms (fifo, round robin, sporadic...) is
> located in the source......
>
Yes, but if you mean in a manner more centralized than nano_sched.c, then
no. Since scheduling policies
are mixed in the priority queue (discarding AP for a moment), then the
action to take for each of the policies
on ready/running/block operations are interspersed in the code at the
appropriate points.
To be honest, the only policy that actually has an 'algorithm' is sporadic
and the algorithm for it is
documented by the POSIX spec. If you are looking for something similar to
the Linux scheduler
algorithms, then you won't find them (unless you start looking at AP which
has a bit more complexity)
due to the pure priority driven nature of an RTOS.
Thomas
|
|
|