Andrew Morton wrote:
quoted text
> On Tue, 19 Aug 2008 22:23:41 +0500
> Constantin Baranov <const@const.mimas.ru> wrote:
>> --- linux-2.6.27-rc3/drivers/leds/leds-alix.c	1970-01-01 04:00:00.000000000 +0400
>> +++ linux-2.6.27-rc3-alix/drivers/leds/leds-alix.c	2008-08-19 21:59:05.207153570 +0500
...
quoted text
>> +static int __init alix_led_probe(struct platform_device *pdev)
>> +{
>> +	int i;
>> +	int ret = 0;
>> +	
>> +	for (i = 0; i < ARRAY_SIZE(alix_leds) && ret >= 0; i++)
>> +		ret = led_classdev_register(&pdev->dev, &alix_leds[i].cdev);
>> +
>> +	if (ret < 0) {
>> +		for (i = i - 2; i >= 0; i--)
> 
> this is off-by-one, surely.
> 
> If we get here with i==1, we'll loop 4 billion times.
> 
>> +			led_classdev_unregister(&alix_leds[i].cdev);
>> +	}
>> +
>> +	return ret;
>> +}
...
quoted text
> How's this look?
...
quoted text
> @@ -92,7 +95,7 @@ static int __init alix_led_probe(struct 
>  		ret = led_classdev_register(&pdev->dev, &alix_leds[i].cdev);
>  
>  	if (ret < 0) {
> -		for (i = i - 2; i >= 0; i--)
> +		while (--i >= 0)
>  			led_classdev_unregister(&alix_leds[i].cdev);
>  	}


Really?  Constantin's code looks correct to me.  He increments i once 
more after failure.

Perhaps write it easier to read; i.e. in the same way as most error 
return checks everywhere in the kernel:

	for (i = 0; i < ARRAY_SIZE(alix_leds); i++) {
		ret = led_classdev_register(...etc...);
		if (ret < 0)
			break;
	}

	if (ret < 0)
		while (i--)
			led_classdev_unregister(&alix_leds[i].cdev);


Or		while (--i >= 0)

or		for (i--; i >= 0; i--)
-- 
Stefan Richter
-=====-==--- =--- ==---
http://arcgraph.de/sr/
--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/