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;
Date:   Wed, 26 Apr 2023 13:16:33 +0000 (UTC)
From:   James Seo <james@equiv.tech>
Subject: Re: [PATCH v3] hwmon: add HP WMI Sensors driver
Message-ID: <ZEkkLggFLCGlvq8f@equiv.tech>
References: <20230424100459.41672-1-james@equiv.tech>
 <cd81a7d6-4b81-f074-1f28-6d1b5300b937@gmx.de>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <cd81a7d6-4b81-f074-1f28-6d1b5300b937@gmx.de>
To:     Armin Wolf <W_Armin@gmx.de>
Cc:     James Seo <james@equiv.tech>, Jean Delvare <jdelvare@suse.com>,
        Guenter Roeck <linux@roeck-us.net>,
        linux-hwmon@vger.kernel.org, platform-driver-x86@vger.kernel.org,
        linux-kernel@vger.kernel.org
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Mon, Apr 24, 2023 at 11:13:36PM +0200, Armin Wolf wrote:
> Am 24.04.23 um 12:05 schrieb James Seo:
> 
>> +	for (i = 0; i < HP_WMI_MAX_INSTANCES; i++, pevents++) {
> 
> Hi,
> 
> the WMI driver core already knows how many instances of a given WMI object are available.
> Unfortunately, this information is currently unavailable to drivers. Would it be convenient
> for you to access this information? I could try to implement such a function if needed.
> 
>> +	for (i = 0; i < HP_WMI_MAX_INSTANCES; i++, info++) {
> 
> Same as above.
> 

Hello,

Having the WMI object instance count wouldn't make much difference to
me for now. The driver has to iterate through all instances during
init anyway. If I were forced to accommodate 50+ sensors, I'd rewrite
some things and I think I'd want such a function then, but I picked
the current arbitrary limit of 32 because even that seems unlikely.

So, maybe don't worry about it unless you want to. Or am I missing
something?

>> +	err = wmi_install_notify_handler(HP_WMI_EVENT_GUID,
>> +					 hp_wmi_notify, state);
> 
> As a side note: the GUID-based interface for accessing WMI devices is deprecated.
> It has known problems handling WMI devices sharing GUIDs and/or notification IDs. However,
> the modern bus-based WMI interface (currently) does not support such aggregate devices well,
> so i think using wmi_install_notify_handler() is still the best thing you can currently do.
> 

Interesting. Of course I had no idea. Though, for some strange
reason, it does look like some documentation to that effect has
emerged on the topic since the last time I checked ;)

>> +	if (err) {
>> +		dev_info(dev, "Failed to subscribe to WMI event\n");
>> +		return false;
>> +	}
>> +
>> +	err = devm_add_action(dev, hp_wmi_devm_notify_remove, NULL);
>> +	if (err) {
>> +		wmi_remove_notify_handler(HP_WMI_EVENT_GUID);
>> +		return false;
>> +	}
> 
> Maybe use devm_add_action_or_reset() here?

Will do.

Thanks for reviewing/writing.

James