Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   Vincent Guittot <vincent.guittot@linaro.org>
To:     mingo@redhat.com, peterz@infradead.org, juri.lelli@redhat.com,
        dietmar.eggemann@arm.com, rostedt@goodmis.org, bsegall@google.com,
        mgorman@suse.de, bristot@redhat.com, linux-kernel@vger.kernel.org,
        odin@uged.al, sachinp@linux.vnet.ibm.com, riel@surriel.com
Cc:     Vincent Guittot <vincent.guittot@linaro.org>
Subject: [PATCH] sched/fair: Ensure that the cfs parent is added after unthrottling
Date:   Mon, 21 Jun 2021 19:43:30 +0200
Message-Id: <20210621174330.11258-1-vincent.guittot@linaro.org>
Precedence: bulk

From: Rik van Riel <riel@surriel.com>

Ensure that a cfs will be in the list whenever one of its child is also
in the list.

A warning on rq->tmp_alone_branch != &rq->leaf_cfs_rq_list has been
reported while running LTP test cfs_bandwidth01.

Odin found the root cause:

$ tree /sys/fs/cgroup/ltp/ -d --charset=ascii
/sys/fs/cgroup/ltp/
|-- drain
`-- test-6851
    `-- level2
        |-- level3a
        |   |-- worker1
        |   `-- worker2
        `-- level3b
            `-- worker3

Timeline (ish):
- worker3 gets throttled
- level3b is decayed, since it has no more load
- level2 get throttled
- worker3 get unthrottled
- level2 get unthrottled
  - worker3 is added to list
  - level3b is not added to list, since nr_running==0 and is decayed

Fixes: a7b359fc6a37 ("sched/fair: Correctly insert cfs_rq's to list on unthrottle")
Reported-by: Sachin Sant <sachinp@linux.vnet.ibm.com>
Signed-off-by: Rik van Riel <riel@surriel.com>
Suggested-by: Vincent Guittot <vincent.guittot@linaro.org>
[Rebased and updated to fix for the reported warning]
Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Acked-by: Odin Ugedal <odin@uged.al>
Tested-by: Sachin Sant <sachinp@linux.vnet.ibm.com>
---

 kernel/sched/fair.c | 28 ++++++++++++++++++++++++++++
 1 file changed, 28 insertions(+)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 7b8990fd4896..8cc27b847ad8 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -3255,6 +3255,31 @@ static inline void cfs_rq_util_change(struct cfs_rq *cfs_rq, int flags)
 
 #ifdef CONFIG_SMP
 #ifdef CONFIG_FAIR_GROUP_SCHED
+/*
+ * Because list_add_leaf_cfs_rq always places a child cfs_rq on the list
+ * immediately before a parent cfs_rq, and cfs_rqs are removed from the list
+ * bottom-up, we only have to test whether the cfs_rq before us on the list
+ * is our child.
+ * If cfs_rq is not on the list, test whether a child needs its to be added to
+ * connect a branch to the tree  * (see list_add_leaf_cfs_rq() for details).
+ */
+static inline bool child_cfs_rq_on_list(struct cfs_rq *cfs_rq)
+{
+	struct cfs_rq *prev_cfs_rq;
+	struct list_head *prev;
+
+	if (cfs_rq->on_list) {
+		prev = cfs_rq->leaf_cfs_rq_list.prev;
+	} else {
+		struct rq *rq = rq_of(cfs_rq);
+
+		prev = rq->tmp_alone_branch;
+	}
+
+	prev_cfs_rq = container_of(prev, struct cfs_rq, leaf_cfs_rq_list);
+
+	return (prev_cfs_rq->tg->parent == cfs_rq->tg);
+}
 
 static inline bool cfs_rq_is_decayed(struct cfs_rq *cfs_rq)
 {
@@ -3270,6 +3295,9 @@ static inline bool cfs_rq_is_decayed(struct cfs_rq *cfs_rq)
 	if (cfs_rq->avg.runnable_sum)
 		return false;
 
+	if (child_cfs_rq_on_list(cfs_rq))
+		return false;
+
 	/*
 	 * _avg must be null when _sum are null because _avg = _sum / divider
 	 * Make sure that rounding and/or propagation of PELT values never
-- 
2.17.1