linux-kernel - Re: [RESEND PATCH V2] leds: sc27xx: Move mutex

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:   Thu, 12 Oct 2023 19:43:54 +0800
From:   Baolin Wang <baolin.wang@...ux.alibaba.com>
To:     Lee Jones <lee@...nel.org>
Cc:     Chunyan Zhang <chunyan.zhang@...soc.com>,
        Pavel Machek <pavel@....cz>, linux-leds@...r.kernel.org,
        Orson Zhai <orsonzhai@...il.com>,
        Chunyan Zhang <zhang.lyra@...il.com>,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: [RESEND PATCH V2] leds: sc27xx: Move mutex_init() to the end of
 probe



On 10/12/2023 5:16 PM, Lee Jones wrote:
> On Thu, 12 Oct 2023, Baolin Wang wrote:
> 
>>
>>
>> On 10/12/2023 11:47 AM, Chunyan Zhang wrote:
>>> Move the mutex_init() to avoid redundant mutex_destroy() calls after
>>> that for each time the probe fails.
>>>
>>> Signed-off-by: Chunyan Zhang <chunyan.zhang@...soc.com>
>>> ---
>>> Rebased onto linux-next.
>>>
>>> V2:
>>> - Move the mutex_init() to the end of .probe() instead of adding
>>> mutex_destroy() according to Lee's comments.
>>> ---
>>>    drivers/leds/leds-sc27xx-bltc.c | 9 ++++-----
>>>    1 file changed, 4 insertions(+), 5 deletions(-)
>>>
>>> diff --git a/drivers/leds/leds-sc27xx-bltc.c b/drivers/leds/leds-sc27xx-bltc.c
>>> index af1f00a2f328..ef57e57ecf07 100644
>>> --- a/drivers/leds/leds-sc27xx-bltc.c
>>> +++ b/drivers/leds/leds-sc27xx-bltc.c
>>> @@ -296,7 +296,6 @@ static int sc27xx_led_probe(struct platform_device *pdev)
>>>    		return -ENOMEM;
>>>    	platform_set_drvdata(pdev, priv);
>>> -	mutex_init(&priv->lock);
>>>    	priv->base = base;
>>>    	priv->regmap = dev_get_regmap(dev->parent, NULL);
>>>    	if (!priv->regmap) {
>>> @@ -309,13 +308,11 @@ static int sc27xx_led_probe(struct platform_device *pdev)
>>>    		err = of_property_read_u32(child, "reg", &reg);
>>>    		if (err) {
>>>    			of_node_put(child);
>>> -			mutex_destroy(&priv->lock);
>>>    			return err;
>>>    		}
>>>    		if (reg >= SC27XX_LEDS_MAX || priv->leds[reg].active) {
>>>    			of_node_put(child);
>>> -			mutex_destroy(&priv->lock);
>>>    			return -EINVAL;
>>>    		}
>>> @@ -325,9 +322,11 @@ static int sc27xx_led_probe(struct platform_device *pdev)
>>>    	err = sc27xx_led_register(dev, priv);
>>>    	if (err)
>>> -		mutex_destroy(&priv->lock);
>>> +		return err;
>>> -	return err;
>>> +	mutex_init(&priv->lock);
>>
>> I think it is better to prepare all the required resources before
>> registering the led device, what I mean is moving mutex_init() before
>> calling sc27xx_led_register().
> 
> Is the mutex used before this point?
> 
> If not, I don't see any reason to initialise it sooner.

When inserting the led module, after registering the led device, users 
can set the led brightness or pattern trigger before initializing the 
mutex, which will crash the system. I know this may not be an actual 
scenario, but this patch opens a small race window, that's what I concerned.