[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <5f43e80a-7f04-65c0-9c24-02c8330d691a@infradead.org>
Date: Wed, 20 Jan 2021 14:11:38 -0800
From: Randy Dunlap <rdunlap@...radead.org>
To: "Zhou Ti (x2019cwm)" <x2019cwm@...x.ca>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] fix the issue that the tick_nohz_get_sleep_length()
function could return a negative value
On 1/20/21 10:58 AM, Zhou Ti (x2019cwm) wrote:
> Fix the issue that the tick_nohz_get_sleep_length() function could return
> a negative value.
>
> The variable "next_event" has a small possibility to be smaller than the
> variable "now" during running. Since both the menu idle governor and the
> teo idle governor use u64 to store the return value of the function, this
> may result in an extremely large and invalid value.
>
> One can easily detect the existence of this issue by using printk to
> output a warning.
>
> Signed-off-by: Ti Zhou <x2019cwm@...x.ca>
> ---
> --- tip/kernel/time/tick-sched.c.orig 2021-01-20 05:34:25.151325912 -0400
> +++ tip/kernel/time/tick-sched.c 2021-01-20 07:09:15.060980886 -0400
> @@ -1156,6 +1156,9 @@ ktime_t tick_nohz_get_sleep_length(ktime
> next_event = min_t(u64, next_event,
> hrtimer_next_event_without(&ts->sched_timer));
>
> + if (unlikely(next_event < now))
> + next_event = now;
> +
> return ktime_sub(next_event, now);
> }
>
Hi,
You need to send this to some maintainer who can apply/merge it.
Just sending it to a mailing list is not sufficient.
thanks.
--
~Randy
Powered by blists - more mailing lists