linux-kernel - RE: [PATCH] leds/trigger: system can't enter suspend.

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

Date:   Sun, 7 May 2017 10:55:13 +0000
From:   Pete Zhang <pete.zhang@....com>
To:     Pavel Machek <pavel@....cz>
CC:     "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "linux-leds@...r.kernel.org" <linux-leds@...r.kernel.org>,
        "j.anaszewski@...sung.com" <j.anaszewski@...sung.com>,
        "rpurdie@...ys.net" <rpurdie@...ys.net>
Subject: RE: [PATCH] leds/trigger: system can't enter suspend.

Please see my comments inline.

-----Original Message-----
From: Pavel Machek [mailto:pavel@....cz] 
Sent: Sunday, May 07, 2017 4:30 PM
To: Pete Zhang
Cc: linux-kernel@...r.kernel.org; linux-leds@...r.kernel.org; j.anaszewski@...sung.com; rpurdie@...ys.net
Subject: Re: [PATCH] leds/trigger: system can't enter suspend.

On Sun 2017-05-07 10:01:21, sanshan zhang wrote:
> From: zhang sanshan <sanshan.zhang@....com>
> 
> system can't enter suspend when enable led heartbeat.
> 
> pm will call heartbeat_pm_notifier when suspend.
> system will prepare led states, and led_trigger_unregister will use 
> led_trigger_set to set trigger.
> kobject_uevent_env will send event,it will call ep_poll_callback to 
> hold a wakeup source if we enable wakelock in kernel.
> So that system can't enter suspend.
> 
> If we set led device tigger to NONE, there is not need to send 
> notification to user space.

Does this bug also happen on mainline kernels?

[Sanshan]: We found this bug in kernel 4.9, But I think it also can reproduce in mainline
 
Are you sure noone relies on the "none" trigger behaviour? We usually don't change user interface without good reason.

[Sanshan]:  From android's perspective, we do not use 'none' trigger. If we set the trigger to 'none', the sys node have been deregistered, 
            There is no mean to notify userland to change the permission of sys node. 

Actually.. is someone using those notifications at all? Telling userland that userland changed LED trigger is somehow... interesting.
[sanshan] : Yes, maybe the effect of this patch is userland can't know current trigger is 'none'

[Sanshan]: I don't know whether we can just turn off the LED rather than registering the led trigger when suspend. 

 

								

> 
> Change-Id: I9f7ee5764d7e31b9a225dae5517cb36137675f9d
> Signed-off-by: zhang sanshan <sanshan.zhang@....com>
> Signed-off-by: Zhang Bo <bo.zhang@....com>
> ---
>  drivers/leds/led-triggers.c | 7 ++++---
>  1 file changed, 4 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/leds/led-triggers.c b/drivers/leds/led-triggers.c 
> index 431123b..f03c2bd1 100644
> --- a/drivers/leds/led-triggers.c
> +++ b/drivers/leds/led-triggers.c
> @@ -108,13 +108,10 @@ void led_trigger_set(struct led_classdev *led_cdev, struct led_trigger *trig)
>  	unsigned long flags;
>  	char *event = NULL;
>  	char *envp[2];
> -	const char *name;
>  
>  	if (!led_cdev->trigger && !trig)
>  		return;
>  
> -	name = trig ? trig->name : "none";
> -	event = kasprintf(GFP_KERNEL, "TRIGGER=%s", name);
>  
>  	/* Remove any existing trigger */
>  	if (led_cdev->trigger) {
> @@ -138,6 +135,10 @@ void led_trigger_set(struct led_classdev *led_cdev, struct led_trigger *trig)
>  			trig->activate(led_cdev);
>  	}
>  
> +	if (NULL != trig)
> +		event = kasprintf(GFP_KERNEL, "TRIGGER=%s", trig->name);
> +	else
> +		return;
>  	if (event) {
>  		envp[0] = event;
>  		envp[1] = NULL;
> --
> 1.9.1

--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html