linux-kernel - Re: [PATCH v2] hwmon: Add driver for Gigabyte AORUS Waterforce AIO coolers

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: <9f6883fb-d628-4077-bb94-f5ba620bdd57@gmail.com>
Date:   Thu, 7 Dec 2023 13:21:18 +0100
From:   Aleksa Savic <savicaleksa83@...il.com>
To:     Guenter Roeck <linux@...ck-us.net>
Cc:     savicaleksa83@...il.com, linux-hwmon@...r.kernel.org,
        Jean Delvare <jdelvare@...e.com>,
        Jonathan Corbet <corbet@....net>,
        "open list:DOCUMENTATION" <linux-doc@...r.kernel.org>,
        open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2] hwmon: Add driver for Gigabyte AORUS Waterforce AIO
 coolers

On 2023-12-05 15:35:19 GMT+01:00, Guenter Roeck wrote:
> On 12/5/23 06:22, Aleksa Savic wrote:
>> On 2023-12-03 18:36:35 GMT+01:00, Guenter Roeck wrote:
>>> On Sun, Dec 03, 2023 at 01:06:48PM +0100, Aleksa Savic wrote:
>>>> This driver exposes hardware sensors of the Gigabyte AORUS Waterforce
>>>> all-in-one CPU liquid coolers, which communicate through a proprietary
>>>> USB HID protocol. Report offsets were initially discovered in [1] and
>>>> confirmed by me on a Waterforce X240 by observing the sent reports from
>>>> the official software.
>>>>
>>>> Available sensors are pump and fan speed in RPM, as well as coolant
>>>> temperature. Also available through debugfs is the firmware version.
>>>>
>>>> Attaching a fan is optional and allows it to be controlled from the
>>>> device. If it's not connected, the fan-related sensors will report
>>>> zeroes.
>>>>
>>>> The addressable RGB LEDs and LCD screen are not supported in this
>>>> driver and should be controlled through userspace tools.
>>>>
>>>> [1]: https://github.com/liquidctl/liquidctl/issues/167
>>>>
>>>> Signed-off-by: Aleksa Savic <savicaleksa83@...il.com>
>>>> ---
> ...
>>>> +static int waterforce_get_status(struct waterforce_data *priv)
>>>> +{
>>>> +    int ret = 0;
>>>> +
>>>> +    if (!mutex_lock_interruptible(&priv->status_report_request_mutex)) {
>>>> +        if (!time_after(jiffies, priv->updated + msecs_to_jiffies(STATUS_VALIDITY))) {
>>>> +            /* Data is up to date */
>>>> +            goto unlock_and_return;
>>>> +        }
>>>
>>> What is the point of this check ? The calling code already checks it.
>>> Checking it twice, once inside and once outside the lock, seems
>>> excessive.
>>>
>>
>> If there are multiple readers here, only the first one should request the status,
>> so when others enter the mutex they can exit early if the data is there.
>>
> 
> Please change the code to only check once from within the mutex.
> 
> Guenter
> 

Done in v3.

Thanks,
Aleksa