[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <ae49ad23-f577-d5d7-7541-42a5c056151e@arm.com>
Date: Tue, 6 Apr 2021 13:42:27 +0200
From: Dietmar Eggemann <dietmar.eggemann@....com>
To: Zeng Tao <prime.zeng@...ilicon.com>, mingo@...hat.com,
peterz@...radead.org, juri.lelli@...hat.com,
vincent.guittot@...aro.org
Cc: linuxarm@...wei.com, Steven Rostedt <rostedt@...dmis.org>,
Ben Segall <bsegall@...gle.com>, Mel Gorman <mgorman@...e.de>,
Daniel Bristot de Oliveira <bristot@...hat.com>,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH] sched/cpupri: fix the task priority BUG_ON checks
On 26/03/2021 04:15, Zeng Tao wrote:
> The BUG_ON checks are intended to check if the task priotity is valid,
> but in the function convert_prio, if the task priority is not valid, it
> will be converted to an uninitialized stack variable. Fix it by moving
> the BUG_ON checks to the default branch of convert_prio.
>
> Fixes: 934fc3314b39 ("sched/cpupri: Remap CPUPRI_NORMAL to MAX_RT_PRIO-1")
> Signed-off-by: Zeng Tao <prime.zeng@...ilicon.com>
> ---
> kernel/sched/cpupri.c | 6 ++----
> 1 file changed, 2 insertions(+), 4 deletions(-)
>
> diff --git a/kernel/sched/cpupri.c b/kernel/sched/cpupri.c
> index ec9be78..c5a0e6e 100644
> --- a/kernel/sched/cpupri.c
> +++ b/kernel/sched/cpupri.c
> @@ -60,6 +60,8 @@ static int convert_prio(int prio)
> case MAX_RT_PRIO:
> cpupri = CPUPRI_HIGHER; /* 100 */
> break;
> + default:
> + BUG();
> }
>
> return cpupri;
> @@ -148,8 +150,6 @@ int cpupri_find_fitness(struct cpupri *cp, struct task_struct *p,
> int task_pri = convert_prio(p->prio);
> int idx, cpu;
>
> - BUG_ON(task_pri >= CPUPRI_NR_PRIORITIES);
> -
> for (idx = 0; idx < task_pri; idx++) {
>
> if (!__cpupri_find(cp, p, lowest_mask, idx))
> @@ -215,8 +215,6 @@ void cpupri_set(struct cpupri *cp, int cpu, int newpri)
>
> newpri = convert_prio(newpri);
>
> - BUG_ON(newpri >= CPUPRI_NR_PRIORITIES);
> -
> if (newpri == oldpri)
> return;
>
>
LGTM. Currently, convert_prio() is never called with prio <
CPUPRI_INVALID (-1) or prio > MAX_RT_PRIO (100) but in case it would
then this would be caught much easier with this patch.
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@....com>
Powered by blists - more mailing lists