linux-kernel - Re: [PATCH] cpuidle: not unset the driver if it already exist

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <CAJZ5v0iaWxu1vL08zyEfx6_Eku9yK5kh-HK4ZdbBtqZvSVd6tw@mail.gmail.com>
Date:   Tue, 22 Oct 2019 18:35:31 +0200
From:   "Rafael J. Wysocki" <rafael@...nel.org>
To:     Zhenzhong Duan <zhenzhong.duan@...cle.com>
Cc:     Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        "Rafael J. Wysocki" <rjw@...ysocki.net>,
        Daniel Lezcano <daniel.lezcano@...aro.org>,
        Linux PM <linux-pm@...r.kernel.org>,
        Marcelo Tosatti <mtosatti@...hat.com>
Subject: Re: [PATCH] cpuidle: not unset the driver if it already exist

On Mon, Oct 14, 2019 at 2:43 AM Zhenzhong Duan
<zhenzhong.duan@...cle.com> wrote:
>
> __cpuidle_set_driver() check if there is an already exist driver and
> unset it before return with -EBUSY. The next call will succeed as it's
> just unset. This is strange as we should either keep old driver and
> return -EBUSY or unset and set new driver and return 0.
>
> Signed-off-by: Zhenzhong Duan <zhenzhong.duan@...cle.com>
> ---
>  drivers/cpuidle/driver.c | 1 -
>  1 file changed, 1 deletion(-)
>
> diff --git a/drivers/cpuidle/driver.c b/drivers/cpuidle/driver.c
> index 80c1a83..a200304 100644
> --- a/drivers/cpuidle/driver.c
> +++ b/drivers/cpuidle/driver.c
> @@ -74,7 +74,6 @@ static inline int __cpuidle_set_driver(struct cpuidle_driver *drv)
>         for_each_cpu(cpu, drv->cpumask) {
>
>                 if (__cpuidle_get_cpu_driver(cpu)) {
> -                       __cpuidle_unset_driver(drv);
>                         return -EBUSY;
>                 }

There is a bug here, but it needs to be fixed differently.

The function should check if any of the CPUs in the mask have a driver
(which is different from drv) already and if so return -EBUSY before
updating any cpuidle_drivers per-CPU pointers.