Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
From:   Valentin Schneider <valentin.schneider@arm.com>
To:     linux-kernel@vger.kernel.org, cgroups@vger.kernel.org
Cc:     lizefan@huawei.com, tj@kernel.org, hannes@cmpxchg.org,
        mingo@kernel.org, peterz@infradead.org, vincent.guittot@linaro.org,
        Dietmar.Eggemann@arm.com, morten.rasmussen@arm.com,
        qperret@google.com,
        =?UTF-8?q?Michal=20Koutn=C3=BD?= <mkoutny@suse.com>
Subject: [PATCH v2] sched/topology, cpuset: Account for housekeeping CPUs to avoid empty cpumasks
Date:   Mon,  4 Nov 2019 00:39:06 +0000
Message-Id: <20191104003906.31476-1-valentin.schneider@arm.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Michal noted that a cpuset's effective_cpus can be a non-empy mask, but
because of the masking done with housekeeping_cpumask(HK_FLAG_DOMAIN)
further down the line, we can still end up with an empty cpumask being
passed down to partition_sched_domains_locked().

Do the proper thing and don't just check the mask is non-empty - check
that its intersection with housekeeping_cpumask(HK_FLAG_DOMAIN) is
non-empty.

Fixes: cd1cb3350561 ("sched/topology: Don't try to build empty sched domains")
Reported-by: Michal Koutný <mkoutny@suse.com>
Signed-off-by: Valentin Schneider <valentin.schneider@arm.com>
---
 kernel/cgroup/cpuset.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
index c87ee6412b36..e4c10785dc7c 100644
--- a/kernel/cgroup/cpuset.c
+++ b/kernel/cgroup/cpuset.c
@@ -798,8 +798,14 @@ static int generate_sched_domains(cpumask_var_t **domains,
 		    cpumask_subset(cp->cpus_allowed, top_cpuset.effective_cpus))
 			continue;
 
+		/*
+		 * Skip cpusets that would lead to an empty sched domain.
+		 * That could be because effective_cpus is empty, or because
+		 * it's only spanning CPUs outside the housekeeping mask.
+		 */
 		if (is_sched_load_balance(cp) &&
-		    !cpumask_empty(cp->effective_cpus))
+		    cpumask_intersects(cp->effective_cpus,
+				       housekeeping_cpumask(HK_FLAG_DOMAIN)))
 			csa[csn++] = cp;
 
 		/* skip @cp's subtree if not a partition root */
-- 
2.22.0