Re: [PATCH V2] sched: fix autogroup nice tune on UP

[Pekka Enberg]

Hi,

On Mon, Jan 24, 2011 at 8:11 AM, Yong Zhang <yong.zhang0@xxxxxxxxx> wrote:
> On Mon, Jan 24, 2011 at 1:54 PM, Pekka Enberg <penberg@xxxxxxxxxx> wrote:
>> On Mon, Jan 24, 2011 at 7:40 AM, Yong Zhang <yong.zhang0@xxxxxxxxx> wrote:
>>> +static void update_cfs_shares(struct cfs_rq *cfs_rq, long weight_delta)
>>> +{
>>> +       struct task_group *tg;
>>> +       struct sched_entity *se;
>>> +
>>> +       if (!cfs_rq)
>>> +               return;
>>> +
>>> +       tg = cfs_rq->tg;
>>> +       se = tg->se[0];
>>> +       if (!se)
>>> +               return;
>>> +       if (likely(se->load.weight == tg->shares))
>>> +               return;
>>> +       reweight_entity(cfs_rq_of(se), se, tg->shares);
>>> +}
>>
>> Wouldn't it be cleaner if we'd separate the shares calculation in a
>> separate helper function that's just
>>
>>  return tg->shares;
>
> I'm not sure I get your point correctly.
> You mean the two tg->shares above, right?
>
> If so, yeah, we can declare a variable for that.
>
>>
>> for UP and extract the current logic for the SMP version?
>>
>
> This is the UP specific version, I don't touch SMP version.
> On SMP, update_cfs_shares() is more complex.

I proposed extracting the shares calculation logic in
update_cfs_shares() to reduce code duplication for the SMP and UP
patch. So something like this:

#ifdef CONFIG_SMP
static long calc_cfs_shares(struct cfs_rq *cfs_rq, struct task_group
*tg, long weight_delta)
{
        long load_weight, load, shares;

        load = cfs_rq->load.weight + weight_delta;

        load_weight = atomic_read(&tg->load_weight);
        load_weight -= cfs_rq->load_contribution;
        load_weight += load;

        shares = (tg->shares * load);
        if (load_weight)
                shares /= load_weight;

        if (shares < MIN_SHARES)
                shares = MIN_SHARES;
        if (shares > tg->shares)
                shares = tg->shares;

        return shares;
}
#else
static long calc_cfs_shares(struct cfs_rq *cfs_rq, struct task_group
*tg, long weight_delta)
{
        return tg->shares;
}
#endif

static void update_cfs_shares(struct cfs_rq *cfs_rq, long weight_delta)
{
        struct task_group *tg;
        struct sched_entity *se;
        long shares;

        if (!cfs_rq)
                return;

        tg = cfs_rq->tg;
        se = tg->se[cpu_of(rq_of(cfs_rq))];
        if (!se)
                return;

#ifndef CONFIG_SMP
        if (likely(se->load.weight == tg->shares))
                return;
#else

        shares = calc_cfs_shares(cfs_rq, tg, weight_delta);

        reweight_entity(cfs_rq_of(se), se, shares);
}
--
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/