Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Sender: Guenter Roeck <groeck7@gmail.com>
Message-ID: <275071f0-d00b-daf5-8232-a0a6a46d685b@roeck-us.net>
Date:   Wed, 10 Aug 2022 11:18:44 -0700
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.11.0
Subject: Re: [PATCH] hwmon: corsair-psu: add reporting of rail mode via
 debugfs
Content-Language: en-US
To:     Wilken Gottwalt <wilken.gottwalt@posteo.net>
Cc:     linux-kernel@vger.kernel.org, Jean Delvare <jdelvare@suse.com>,
        Jonathan Corbet <corbet@lwn.net>, linux-hwmon@vger.kernel.org
References: <YvN4SbnAp3jl+Vzo@monster.localdomain>
 <20220810133012.GB274220@roeck-us.net> <20220810193541.29f95ec5@posteo.net>
From:   Guenter Roeck <linux@roeck-us.net>
In-Reply-To: <20220810193541.29f95ec5@posteo.net>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 8/10/22 10:35, Wilken Gottwalt wrote:
> On Wed, 10 Aug 2022 06:30:12 -0700
> Guenter Roeck <linux@roeck-us.net> wrote:
> 
>>> +static int ocpmode_show(struct seq_file *seqf, void *unused)
>>> +{
>>> +	struct corsairpsu_data *priv = seqf->private;
>>> +	long val;
>>> +	int ret;
>>> +
>>> +	ret = corsairpsu_get_value(priv, PSU_CMD_OCPMODE, 0, &val);
>>> +	if (ret < 0)
>>> +		seq_puts(seqf, "N/A\n");
>>> +	else
>>> +		seq_printf(seqf, "%s\n", (val == 0x02) ? "multi rail" : "single rail");
>>
>> If this is not always available, would it be better not to create the file
>> in the first place ? If that is not feasible, it should at least be
>> documented that the value is not always available to ensure that no one
>> complains about it (or at least no one who read the documentation).
>>
>> Also, is the value strictly 0x02 for multi-rail configurations, or
>> is that possibly just a bit or the number of rails ?
> 
> The mode is actually switchable on fly, similar to the fan. I do not want to
> provide the switching functionality, because also similar to the fan controls,
> it can be used to damage the PSU. It is part of the over current protection
> system (hence the name ocpmode) and people use the RAW interface to switch the
> fans and the ocpmode. This is also the reason I made it that way, because you
> could poll the information right in the process of switching, which can result
> in bogus values. 0x02 is the value for "switching to multi rail was successful",
> every other value is considered "single rail mode". Or you get a malformed
> message which results in "N/A" or unknown. So yes, you are right, I could at
> least add a define for the value and be more clear in the documentation. Would
> that be okay for you?
> 

Define or not is up to you, but it would be nice if at least some of the explanation
above would be documented in the driver.

Guenter