[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <10e7c2e6-ef7d-7de1-5f02-ea6cb56bfa1a@huawei.com>
Date: Tue, 6 Jan 2026 14:24:46 +0800
From: Jinjie Ruan <ruanjinjie@...wei.com>
To: <alexandre.belloni@...tlin.com>, <linux-rtc@...r.kernel.org>,
<linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] rtc: interface: Fix softlockup in rtc_timer_do_work()
On 2025/12/31 17:23, Jinjie Ruan wrote:
> On kvm qemu with cmos rtc and mc146818 chip, when the read time jump to
> a future time after set the uie timer expire with a current RTC time,
> rtc_timer_do_work() will loop for a while util softlockup because
> the expiration of the uie timer was way before the current
> RTC time and a new timer will be enqueued until the current rtc time
> is reached, as below:
>
> Fix it by voluntarily yield the CPU in the loop in rtc_timer_do_work().
>
> RTC_UIE_ON:
> read now: 2019:04:08:12:32:27, add timer0 (expire: 2019:04:08:12:32:28)
> ^^^^^^^^^^^^^^^^^^^^
> ...
> rtc_timer_do_work() iterate the list in a loop:
> read now: 2033:12:02:07:27:15
Please ignore, this seems to be a bug in QEMU.
> ^^^^^^^^^^^^^^^^^^^
> handle timer0, add timer1 to the list (expire: 2019:04:08:12:32:29)
> handle timer1, add timer2 to the list (expire: 2019:04:08:12:32:30)
> handle timer2, add timer3: 2019:04:08:12:32:31
> ...
> -> softlockup
>
> Signed-off-by: Jinjie Ruan <ruanjinjie@...wei.com>
> ---
> drivers/rtc/interface.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/drivers/rtc/interface.c b/drivers/rtc/interface.c
> index b8b298efd9a9..9ded10e82f4b 100644
> --- a/drivers/rtc/interface.c
> +++ b/drivers/rtc/interface.c
> @@ -964,6 +964,7 @@ void rtc_timer_do_work(struct work_struct *work)
> timer->enabled = 1;
> timerqueue_add(&rtc->timerqueue, &timer->node);
> trace_rtc_timer_enqueue(timer);
> + cond_resched();
> }
> }
>
Powered by blists - more mailing lists