Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Waiman Long <longman@redhat.com>
To:     Tejun Heo <tj@kernel.org>, Zefan Li <lizefan.x@bytedance.com>,
        Johannes Weiner <hannes@cmpxchg.org>,
        Christian Brauner <brauner@kernel.org>
Cc:     cgroups@vger.kernel.org, linux-kernel@vger.kernel.org,
        Juri Lelli <juri.lelli@redhat.com>,
        Dietmar Eggemann <dietmar.eggemann@arm.com>,
        =?UTF-8?q?Michal=20Koutn=C3=BD?= <mkoutny@suse.com>,
        gscrivan@redhat.com, Waiman Long <longman@redhat.com>
Subject: [PATCH v3 1/4] cgroup/cpuset: Wake up cpuset_attach_wq tasks in cpuset_cancel_attach()
Date:   Thu,  6 Apr 2023 10:04:01 -0400
Message-Id: <20230406140404.2718574-2-longman@redhat.com>
In-Reply-To: <20230406140404.2718574-1-longman@redhat.com>
References: <20230406140404.2718574-1-longman@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Precedence: bulk

After a successful cpuset_can_attach() call which increments the
attach_in_progress flag, either cpuset_cancel_attach() or cpuset_attach()
will be called later. In cpuset_attach(), tasks in cpuset_attach_wq,
if present, will be woken up at the end. That is not the case in
cpuset_cancel_attach(). So missed wakeup is possible if the attach
operation is somehow cancelled. Fix that by doing the wakeup in
cpuset_cancel_attach() as well.

Fixes: e44193d39e8d ("cpuset: let hotplug propagation work wait for task attaching")
Signed-off-by: Waiman Long <longman@redhat.com>
Reviewed-by: Michal Koutný <mkoutny@suse.com>
---
 kernel/cgroup/cpuset.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
index bc4dcfd7bee5..066689a7dcc3 100644
--- a/kernel/cgroup/cpuset.c
+++ b/kernel/cgroup/cpuset.c
@@ -2507,11 +2507,15 @@ static int cpuset_can_attach(struct cgroup_taskset *tset)
 static void cpuset_cancel_attach(struct cgroup_taskset *tset)
 {
 	struct cgroup_subsys_state *css;
+	struct cpuset *cs;
 
 	cgroup_taskset_first(tset, &css);
+	cs = css_cs(css);
 
 	percpu_down_write(&cpuset_rwsem);
-	css_cs(css)->attach_in_progress--;
+	cs->attach_in_progress--;
+	if (!cs->attach_in_progress)
+		wake_up(&cpuset_attach_wq);
 	percpu_up_write(&cpuset_rwsem);
 }
 
-- 
2.31.1