linux-kernel - Re: [PATCH 3/5] mfd: tps65217: Add a dependency on OF

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

Date:   Wed, 7 Jun 2017 16:10:46 +0200
From:   Javier Martinez Canillas <javier@...hile0.org>
To:     Keerthy <j-keerthy@...com>
Cc:     Enric Balletbo Serra <eballetbo@...il.com>,
        Lee Jones <lee.jones@...aro.org>,
        Tony Lindgren <tony@...mide.com>, Milo Kim <milo.kim@...com>,
        "linux-omap@...r.kernel.org" <linux-omap@...r.kernel.org>,
        linux-kernel <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 3/5] mfd: tps65217: Add a dependency on OF

Hello Keerthy,

On Wed, Jun 7, 2017 at 3:45 PM, Keerthy <j-keerthy@...com> wrote:

[snip]

>>>>>>>>
>>>>>>>
>>>>>>> I think you can remove the of_match_device checks in some drivers too
>>>>>>>
>>>>>>> i.e:
>>>>>>>
>>>>>>> http://elixir.free-electrons.com/linux/latest/source/drivers/mfd/tps65217.c#L330
>>>>>>
>>>>>> Yes that and removal of unused i2c_device_id. I will follow it up once
>>>>>> this OF dependency is in.
>>>>>
>>>>> The of_match_device() checks should be removed with the OF patch.
>>>
>>> Lee Jones/ Enric,
>>>
>>> IIUC of_match_device call is still needed to obtain a match and in case
>>> there are multiple compatibles with different match data then this call
>>> is definitely needed.
>>>

That's correct...

>>
>> Not sure if I follow you. My understanding is that with DT the probe
>> of this driver is only called if there is a node with the compatible =
>> "ti,tps65217" string. So if probe is called there is always a match
>> and the call to of_match_device is redundant.
>
> How will you get the matching data?
>
> For the tps65217 case you mentioned we need the match pointer to get the
> chip_id right?
>
> chip_id = (unsigned long)match->data;
>

...but this particular driver only has a single entry in the OF table
and so you can just do:

tps->id = TPS65217;

Later if there's a variant for this chip, then you can add the logic
to query the struct of_device_id .data. But for now I think that's
better to just remove as Enric proposes and also remove the .data
field from the struct of_device_id entry.

Best regards,
Javier