Received-SPF: pass (google.com: domain of linux-nfs-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   Zhang Xiaoxu <zhangxiaoxu5@huawei.com>
To:     <zhangxiaoxu5@huawei.com>, <trond.myklebust@hammerspace.com>,
        <anna.schumaker@netapp.com>, <bfields@fieldses.org>,
        <chuck.lever@oracle.com>, <linux-nfs@vger.kernel.org>
Subject: [PATCH 2/2] SUNRPC: Should wake up the privileged task firstly.
Date:   Sat, 26 Jun 2021 15:50:42 +0800
Message-ID: <20210626075042.805548-3-zhangxiaoxu5@huawei.com>
In-Reply-To: <20210626075042.805548-1-zhangxiaoxu5@huawei.com>
References: <20210626075042.805548-1-zhangxiaoxu5@huawei.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 7BIT
Content-Type:   text/plain; charset=US-ASCII
Precedence: bulk

When find a task from wait queue to wake up, a non-privileged task may
be found out, rather than the privileged. This maybe lead a deadlock
same as commit dfe1fe75e00e ("NFSv4: Fix deadlock between nfs4_evict_inode()
and nfs4_opendata_get_inode()"):

Privileged delegreturn task is queued to privileged list because all
the slots are assigned. If there has no enough slot to wake up the
non-privileged batch tasks(session less than 8 slot), then the privileged
delegreturn task maybe lost waked up because the found out task can't
get slot since the session is on draining.

So we should treate the privileged task as the emergency task, and
execute it as for as we can.

Reported-by: Hulk Robot <hulkci@huawei.com>
Fixes: 5fcdfacc01f3 ("NFSv4: Return delegations synchronously in evict_inode")
Cc: stable@vger.kernel.org
Signed-off-by: Zhang Xiaoxu <zhangxiaoxu5@huawei.com>
---
 net/sunrpc/sched.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/net/sunrpc/sched.c b/net/sunrpc/sched.c
index 37cd09574628..e08ce3b791fd 100644
--- a/net/sunrpc/sched.c
+++ b/net/sunrpc/sched.c
@@ -591,6 +591,15 @@ static struct rpc_task *__rpc_find_next_queued_priority(struct rpc_wait_queue *q
 	struct list_head *q;
 	struct rpc_task *task;
 
+	/*
+	 * Service the privileged queue.
+	 */
+	q = &queue->tasks[RPC_NR_PRIORITY - 1];
+	if (queue->maxpriority > RPC_PRIORITY_PRIVILEGED && !list_empty(q)) {
+		task = list_first_entry(q, struct rpc_task, u.tk_wait.list);
+		goto out;
+	}
+
 	/*
 	 * Service a batch of tasks from a single owner.
 	 */
-- 
2.25.4