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:   Sat, 24 Sep 2022 12:31:25 -0500
From:   Rob Herring <robh@kernel.org>
To:     Wolfram Sang <wsa+renesas@sang-engineering.com>,
        Asmaa Mnebhi <asmaa@nvidia.com>,
        Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        "devicetree@vger.kernel.org" <devicetree@vger.kernel.org>,
        "linux-i2c@vger.kernel.org" <linux-i2c@vger.kernel.org>,
        Khalil Blaiech <kblaiech@nvidia.com>
Subject: Re: How to remove DT support from a driver? (was Re: [PATCH v5 8/8]
 i2c: i2c-mlxbf.c: Update binding devicetree)
Message-ID: <20220924173125.GA989070-robh@kernel.org>
References: <20220920174736.9766-1-asmaa@nvidia.com>
 <20220920174736.9766-9-asmaa@nvidia.com>
 <20220921065506.6cu6p2cpu3zfhtie@krzk-bin>
 <CH2PR12MB3895572575B5BED5DAFFAA29D74F9@CH2PR12MB3895.namprd12.prod.outlook.com>
 <Yyttt3DiQpMZYejA@shikoro>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <Yyttt3DiQpMZYejA@shikoro>
Precedence: bulk

On Wed, Sep 21, 2022 at 10:01:59PM +0200, Wolfram Sang wrote:
> Hi,
> 
> > I have a question for you and Wolfram, we don’t use device trees and
> > are not planning to use device trees; we only use ACPI tables. But I
> > think when Khalil submitted the first version of the i2c-mlxbf.c
> > driver, it was requested from him to add devicetree support. Do you
> > know why? Is it possible to remove the device tree support and so this
> > doc? or is devicetree support a requirement regardless of the actual
> > implementation? 
> 
> The first version sent from Khalil to the public I2C mailing list already
> had DT bindings [1]. I don't see a sign of someone of the public list
> requesting DT bindings. Maybe it was company internal?
> 
> Technically, there is no requirement to support DT, especially since you
> have working ACPI. I don't know the process, though, of removing DT
> support. You would basically need to be sure that no user made use of
> the DT bindings introduced before. I don't know to what degree you can
> assume that.

There's the whole using DT bindings in ACPI bindings thing, but I have 
little interest (or time) in supporting that. Maybe that's what's 
happening here? I haven't looked. The whole concept is flawed IMO. It 
may work for simple cases of key/value device properties, but the ACPI 
model is quite different in how resources are described and managed.

Rob