Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:5 as permitted sender) client-ip=2620:137:e000::3:5;
Date:   Tue, 17 Oct 2023 17:09:53 +0800 (CST)
Message-ID: <202310171709530660462@zte.com.cn>
In-Reply-To: <87lec9l3mi.ffs@tglx>
References: 202309161037416349176@zte.com.cn,87lec9l3mi.ffs@tglx
Mime-Version: 1.0
From:   <yang.yang29@zte.com.cn>
To:     <tglx@linutronix.de>
Cc:     <peterz@infradead.org>, <linux-kernel@vger.kernel.org>,
        <jiang.xuexin@zte.com.cn>
Subject: =?UTF-8?B?W1BBVENIIFYyXSBjcHUvaG90cGx1ZzogZG9udCBvZmZsaW5lIHRoZSBsYXN0IG5vbi1pc29sYXRlZCBDUFU=?=
Content-Type: text/plain;
        charset="UTF-8"
Precedence: bulk

From: Ran Xiaokai <ran.xiaokai@zte.com.cn>

If system has some isolate cpus with "isolcpus=" parameter,
and user try to offline all the non-isolated CPUs, kernel will
first report a warning and then panic when taking the last
non-isolated CPU offline.

When offline the last non-isolate cpu,
cpuset_hotplug_workfn()
  rebuild_sched_domains_locked()
    ndoms = generate_sched_domains(&doms, &attr);
      cpumask_and(doms[0], top_cpuset.effective_cpus,
                  housekeeping_cpumask(HK_FLAG_DOMAIN));

As isolated cpus are excluded, "doms" passes to partition_sched_domains_locked()
will be an empty cpumask.

WARNING: CPU: 4 PID: 80 at kernel/sched/topology.c:2366 build_sched_domains+0x120c/0x1408
Call trace:
 build_sched_domains+0x120c/0x1408
 partition_sched_domains_locked+0x234/0x880
 rebuild_sched_domains_locked+0x37c/0x798
 rebuild_sched_domains+0x30/0x58
 cpuset_hotplug_workfn+0x2a8/0x930
 process_scheduled_works+0x178/0x3e0
 worker_thread+0x174/0x2f0
---[ end trace 0000000000000000 ]---
Unable to handle kernel paging request at virtual address fffe80027ab37080
 partition_sched_domains_locked+0x318/0x880
 rebuild_sched_domains_locked+0x37c/0x798
 rebuild_sched_domains+0x30/0x58
 cpuset_hotplug_workfn+0x2a8/0x930
 process_scheduled_works+0x178/0x3e0
 worker_thread+0x174/0x2f0
 kthread+0x10c/0x128
 ret_from_fork+0x10/0x20
Code: 1a850042 b9441883 f862dae0 8b000021 (f945003b)

Kernel should not panic when user managing CPUs with this
meaningless behavior, so we add a check for this, which
prevent user taking the last non-isolated CPU offline.

Signed-off-by: Ran Xiaokai <ran.xiaokai@zte.com.cn>
Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Yang Yang <yang.yang29@zte.com.cn>
Cc: Xuexin Jiang <jiang.xuexin@zte.com.cn>
---
 kernel/cpu.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/kernel/cpu.c b/kernel/cpu.c
index 6de7c6bb74ee..94430ea8b4a5 100644
--- a/kernel/cpu.c
+++ b/kernel/cpu.c
@@ -1515,11 +1515,14 @@ static int cpu_down_maps_locked(unsigned int cpu, enum cpuhp_state target)
 	/*
 	 * Ensure that the control task does not run on the to be offlined
 	 * CPU to prevent a deadlock against cfs_b->period_timer.
+	 * Also keep at least one housekeeping cpu onlined to avoid generating
+	 * an empty sched_domain span.
 	 */
-	cpu = cpumask_any_but(cpu_online_mask, cpu);
-	if (cpu >= nr_cpu_ids)
-		return -EBUSY;
-	return work_on_cpu(cpu, __cpu_down_maps_locked, &work);
+	for_each_cpu_and(cpu, cpu_online_mask, housekeeping_cpumask(HK_TYPE_DOMAIN)) {
+		if (cpu != work.cpu)
+			return work_on_cpu(cpu, __cpu_down_maps_locked, &work);
+	}
+	return -EBUSY;
 }

 static int cpu_down(unsigned int cpu, enum cpuhp_state target)
-- 
2.15.2