[Devel] Re: [RFC][PATCH 5/6] core changes for group fairness

Srivatsa Vaddagiri vatsa at linux.vnet.ibm.com
Thu Jun 14 05:06:05 PDT 2007


On Wed, Jun 13, 2007 at 10:56:06PM +0200, Dmitry Adamushko wrote:
> >+static int balance_tasks(struct rq *this_rq, int this_cpu, struct rq 
> >*busiest,
> >+                     unsigned long max_nr_move, unsigned long 
> >max_load_move,
> >+                     struct sched_domain *sd, enum idle_type idle,
> >+                     int *all_pinned, unsigned long *load_moved,
> >+                     int this_best_prio, int best_prio, int 
> >best_prio_seen,
> >+                     void *iterator_arg,
> >+                     struct task_struct *(*iterator_start)(void *arg),
> >+                     struct task_struct *(*iterator_next)(void *arg));
> 
> IMHO, it looks a bit frightening :) 

I agree :) It is taking (ooops) 15 args (8 perhaps was the previous record
in sched.c (move_tasks)!

> maybe it would be possible to
> create a structure that combines some relevant argumens .. at least,
> the last 3 ones.

How does this look?

struct balance_tasks_args {
	struct rq *this_rq, struct rq *busiest;
	unsigned long max_nr_move, unsigned long max_load_move;
	struct sched_domain *sd, enum idle_type idle;
	int this_best_prio, best_prio, best_prio_seen;
	int *all_pinned;
	unsigned long *load_moved;
	void *iterator_arg;
	struct task_struct *(*iterator_start)(void *arg);
	struct task_struct *(*iterator_next)(void *arg));
};

static int balance_tasks(struct balance_tasks_args *arg);

[ down to one argument now! ]

?

I will try this in my next iteration ..


> >-static int move_tasks(struct rq *this_rq, int this_cpu, struct rq 
> >*busiest,
> >+static int balance_tasks(struct rq *this_rq, int this_cpu, struct rq 
> >*busiest,
> >                      unsigned long max_nr_move, unsigned long 
> >                      max_load_move,
> >                      struct sched_domain *sd, enum idle_type idle,
> >-                     int *all_pinned)
> >+                     int *all_pinned, unsigned long *load_moved,
> >+                     int this_best_prio, int best_prio, int 
> >best_prio_seen,
> >+                     void *iterator_arg,
> >+                     struct task_struct *(*iterator_start)(void *arg),
> >+                     struct task_struct *(*iterator_next)(void *arg))
> 
> I think, there is a possible problem here. If I'm not complete wrong,
> this function (move_tasks() in the current mainline) can move more
> 'load' than specified by the 'max_load_move'..

Yes I think you are right. I will tackle this in next iteration.

Thanks for all your review so far!

-- 
Regards,
vatsa
_______________________________________________
Containers mailing list
Containers at lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers




More information about the Devel mailing list