linux-kernel - Re: [PATCH v2 2/2] watchdog: max63xx_wdt: Add support for specifying WDI logic via GPIO

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 for Android: free password hash cracker in your pocket

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

Date:   Mon, 2 May 2022 21:37:16 -0700
From:   Guenter Roeck <linux@...ck-us.net>
To:     Tzung-Bi Shih <tzungbi@...nel.org>,
        Pali Rohár <pali@...nel.org>
Cc:     Wim Van Sebroeck <wim@...ux-watchdog.org>,
        Rob Herring <robh+dt@...nel.org>,
        linux-watchdog@...r.kernel.org, devicetree@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 2/2] watchdog: max63xx_wdt: Add support for specifying
 WDI logic via GPIO

On 5/2/22 20:57, Tzung-Bi Shih wrote:
> On Fri, Apr 29, 2022 at 03:13:49PM +0200, Pali Rohár wrote:
>> @@ -27,6 +27,7 @@
>>   #include <linux/io.h>
>>   #include <linux/slab.h>
>>   #include <linux/property.h>
>> +#include <linux/gpio/consumer.h>
> 
> It would be better to keep them alphabetically.  Anyway, they aren't sorted
> originally...
> 
>> +static void max63xx_gpio_ping(struct max63xx_wdt *wdt)
>> +{
>> +	spin_lock(&wdt->lock);
> 
> Does it really need to acquire the lock?  It looks like the lock is to prevent
> concurrent accesses to the mmap in max63xx_mmap_ping() and max63xx_mmap_set().
> 

Actually, that doesn't work at all. spin_lock() directly contradicts
with gpiod_set_value_cansleep().

>> +	gpiod_set_value_cansleep(wdt->gpio_wdi, 1);
>> +	udelay(1);
> 
> Doesn't it need to include <linux/delay.h> for udelay()?
> 
>> @@ -225,10 +240,19 @@ static int max63xx_wdt_probe(struct platform_device *pdev)
>>   		return -EINVAL;
>>   	}
>>   
>> +	wdt->gpio_wdi = devm_gpiod_get(dev, NULL, GPIOD_FLAGS_BIT_DIR_OUT);
>> +	if (IS_ERR(wdt->gpio_wdi) && PTR_ERR(wdt->gpio_wdi) != -ENOENT)
> 
> Use devm_gpiod_get_optional() to make the intent clear.  Also, it gets rid of
> the check for -ENOENT.
> 
>> +		return dev_err_probe(dev, PTR_ERR(wdt->gpio_wdi),
>> +				     "unable to request gpio: %ld\n",
>> +				     PTR_ERR(wdt->gpio_wdi));
> 
> It doesn't need to again print for PTR_ERR(wdt->gpio_wdi).  dev_err_probe()
> prints the error.
> 
>>   	err = max63xx_mmap_init(pdev, wdt);
>>   	if (err)
>>   		return err;
>>   
>> +	if (!IS_ERR(wdt->gpio_wdi))
>> +		wdt->ping = max63xx_gpio_ping;
> 
> Thus, the max63xx_gpio_ping() overrides max63xx_mmap_ping() if the GPIO was
> provided?  It would be better to mention the behavior in the commit message.
> 
> Also, could both the assignments of `wdt->gpio_wdi` and `wdt->ping` happen
> after max63xx_mmap_init()?