sched/fair: Fix calc_cfs_shares() fixed point arithmetics width confusion

Commit:

  fde7d22e01aa ("sched/fair: Fix overly small weight for interactive group entities")

did something non-obvious but also did it buggy yet latent.

The problem was exposed for real by a later commit in the v4.7 merge window:

  2159197d6677 ("sched/core: Enable increased load resolution on 64-bit kernels")

... after which tg->load_avg and cfs_rq->load.weight had different
units (10 bit fixed point and 20 bit fixed point resp.).

Add a comment to explain the use of cfs_rq->load.weight over the
'natural' cfs_rq->avg.load_avg and add scale_load_down() to correct
for the difference in unit.

Since this is (now, as per a previous commit) the only user of
calc_tg_weight(), collapse it.

The effects of this bug should be randomly inconsistent SMP-balancing
of cgroups workloads.

Reported-by: Jirka Hladky <jhladky@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Fixes: 2159197d6677 ("sched/core: Enable increased load resolution on 64-bit kernels")
Fixes: fde7d22e01aa ("sched/fair: Fix overly small weight for interactive group entities")
Signed-off-by: Ingo Molnar <mingo@kernel.org>

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index cc48bef40cca708f62c4faba64e03567476b9661..c8c5d2d484249048cca7dd62fdfc38fbe7de2883 100644 (file)
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -2497,28 +2497,22 @@ account_entity_dequeue(struct cfs_rq *cfs_rq, struct sched_entity *se)
 
 #ifdef CONFIG_FAIR_GROUP_SCHED
 # ifdef CONFIG_SMP
-static inline long calc_tg_weight(struct task_group *tg, struct cfs_rq *cfs_rq)
+static long calc_cfs_shares(struct cfs_rq *cfs_rq, struct task_group *tg)
 {
-       long tg_weight;
+       long tg_weight, load, shares;
 
        /*
-        * Use this CPU's real-time load instead of the last load contribution
-        * as the updating of the contribution is delayed, and we will use the
-        * the real-time load to calc the share. See update_tg_load_avg().
+        * This really should be: cfs_rq->avg.load_avg, but instead we use
+        * cfs_rq->load.weight, which is its upper bound. This helps ramp up
+        * the shares for small weight interactive tasks.
         */
-       tg_weight = atomic_long_read(&tg->load_avg);
-       tg_weight -= cfs_rq->tg_load_avg_contrib;
-       tg_weight += cfs_rq->load.weight;
-
-       return tg_weight;
-}
+       load = scale_load_down(cfs_rq->load.weight);
 
-static long calc_cfs_shares(struct cfs_rq *cfs_rq, struct task_group *tg)
-{
-       long tg_weight, load, shares;
+       tg_weight = atomic_long_read(&tg->load_avg);
 
-       tg_weight = calc_tg_weight(tg, cfs_rq);
-       load = cfs_rq->load.weight;
+       /* Ensure tg_weight >= load */
+       tg_weight -= cfs_rq->tg_load_avg_contrib;
+       tg_weight += load;
 
        shares = (tg->shares * load);
        if (tg_weight)
@@ -2537,6 +2531,7 @@ static inline long calc_cfs_shares(struct cfs_rq *cfs_rq, struct task_group *tg)
        return tg->shares;
 }
 # endif /* CONFIG_SMP */
+
 static void reweight_entity(struct cfs_rq *cfs_rq, struct sched_entity *se,
                            unsigned long weight)
 {