linux-kernel - Re: [PATCH] hwmon: Remove checks for validity of dev

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: <d7686623-c5ab-ce2a-386c-f35ec167ca81@collabora.com>
Date:   Fri, 28 Jan 2022 21:59:39 +0500
From:   Muhammad Usama Anjum <usama.anjum@...labora.com>
To:     Guenter Roeck <linux@...ck-us.net>,
        Jean Delvare <jdelvare@...e.com>
Cc:     usama.anjum@...labora.com, kernel@...labora.com,
        kernel-janitors@...r.kernel.org, linux-hwmon@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] hwmon: Remove checks for validity of dev

On 1/28/22 8:05 PM, Guenter Roeck wrote:
> On 1/28/22 04:59, Muhammad Usama Anjum wrote:
>> dev is being dereferenced in device_property_present() which means that
>> it is valid. Don't check its validity again and simplify the code.
>>
>> Signed-off-by: Muhammad Usama Anjum <usama.anjum@...labora.com>
>> ---
>>   drivers/hwmon/hwmon.c | 4 ++--
>>   1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/hwmon/hwmon.c b/drivers/hwmon/hwmon.c
>> index e36ea82da1474..aec32abd0a89f 100644
>> --- a/drivers/hwmon/hwmon.c
>> +++ b/drivers/hwmon/hwmon.c
>> @@ -822,7 +822,7 @@ __hwmon_device_register(struct device *dev, const
>> char *name, void *drvdata,
>>       hwdev->name = name;
>>       hdev->class = &hwmon_class;
>>       hdev->parent = dev;
>> -    hdev->of_node = dev ? dev->of_node : NULL;
>> +    hdev->of_node = dev->of_node;
>>       hwdev->chip = chip;
>>       dev_set_drvdata(hdev, drvdata);
>>       dev_set_name(hdev, HWMON_ID_FORMAT, id);
>> @@ -834,7 +834,7 @@ __hwmon_device_register(struct device *dev, const
>> char *name, void *drvdata,
>>         INIT_LIST_HEAD(&hwdev->tzdata);
>>   -    if (dev && dev->of_node && chip && chip->ops->read &&
>> +    if (dev->of_node && chip && chip->ops->read &&
>>           chip->info[0]->type == hwmon_chip &&
>>           (chip->info[0]->config[0] & HWMON_C_REGISTER_TZ)) {
>>           err = hwmon_thermal_register_sensors(hdev);
> 
> Wrong fix, sorry. While I would love to make dev mandatory, the function
> is called with dev == NULL from at least one place, and the check is
> (still)
> needed. Even if/when it is removed we would have to add an early check
> and return -EINVAL if it is NULL.
> 
Thank you for the reply. I've looked at the code again. You are correct.

> Guenter