Re: Subject: [PATCH] sched: fixed erroneous all_pinned logic.

[Ken Chen]

On Fri, Apr 8, 2011 at 3:57 AM, Vladimir Davydov wrote:
>
> As far as I understand, this patch sets the all_pinned flag if
> and only if we fail to move any tasks during the load balance.
> However, the migration can fail because e.g. all tasks are
> cache hot on their cpus (can_migrate_task() returns 0 in this
> case), and in this case we shouldn't treat all tasks as cpu
> bound, should we?

all_pinned logic is not restricted for tasks that have hard CPU affinity.
In balance_tasks(), e.g. if all tasks have high weight, it would also trip
all_pinned to 1.  See ( p->se.load.weight >> 1) > rem_load_move).  I don't
see much difference whether a task has high weight, or is cache hot, or
has hard cpu affinity.

The effect of classifying cache_hot as pinned task or not is difficult to
measure, I think.  Because that state is more temporal compare to the
other two conditions.

I don't have enough data to indicate which way is better though.

- Ken
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/