On Thu, Sep 11, 2025 at 08:30:44PM +0800, wangtao (EQ) wrote:
在 2025/9/11 20:20, Greg KH 写道:
On Mon, Sep 08, 2025 at 08:42:30AM +0000, Wang Tao wrote:
When CPU 1 enters the nohz_full state, and the kworker on CPU 0 executes the function sched_tick_remote, holding the lock on CPU1's rq and triggering the warning WARN_ON_ONCE(delta > (u64)NSEC_PER_SEC * 3). This leads to the process of printing the warning message, where the console_sem semaphore is held. At this point, the print task on the CPU1's rq cannot acquire the console_sem and joins the wait queue, entering the UNINTERRUPTIBLE state. It waits for the console_sem to be released and then wakes up. After the task on CPU 0 releases the console_sem, it wakes up the waiting console_sem task. In try_to_wake_up, it attempts to acquire the lock on CPU1's rq again, resulting in a deadlock.
The triggering scenario is as follows:
CPU 0 CPU1 sched_tick_remote WARN_ON_ONCE(delta > (u64)NSEC_PER_SEC * 3)
report_bug con_write printk
console_unlock do_con_write console_lock down(&console_sem) list_add_tail(&waiter.list, &sem->wait_list); up(&console_sem) wake_up_q(&wake_q) try_to_wake_up __task_rq_lock _raw_spin_lock
This patch fixes the issue by deffering all printk console printing during the lock holding period.
Fixes: d84b31313ef8 ("sched/isolation: Offload residual 1Hz scheduler tick") Signed-off-by: Wang Tao wangtao554@huawei.com
kernel/sched/core.c | 2 ++ 1 file changed, 2 insertions(+)
diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 40f40f359c5d..fd2c83058ec2 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -4091,6 +4091,7 @@ static void sched_tick_remote(struct work_struct *work) goto out_requeue; rq_lock_irq(rq, &rf);
- printk_deferred_enter(); curr = rq->curr; if (cpu_is_offline(cpu)) goto out_unlock;
@@ -4109,6 +4110,7 @@ static void sched_tick_remote(struct work_struct *work) calc_load_nohz_remote(rq); out_unlock:
- printk_deferred_exit(); rq_unlock_irq(rq, &rf); out_requeue:
-- 2.34.1
Sorry, we initially discovered the issue while testing the stable branch, and it seems that the mainline has the same problem, but I haven't submitted a patch yet.
That is required for us to be able to take a patch into the stable trees, thanks.
greg k-h