Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <20201208164821.2686082-1-paul@crapouillou.net>
 <20201208164821.2686082-2-paul@crapouillou.net> <CAFr9PX=EgQSXeATLn++DSHkkQar35rpLGh978J5Lnw9jS8XMrw@mail.gmail.com>
In-Reply-To: <CAFr9PX=EgQSXeATLn++DSHkkQar35rpLGh978J5Lnw9jS8XMrw@mail.gmail.com>
From:   Arnd Bergmann <arnd@kernel.org>
Date:   Wed, 9 Dec 2020 12:04:53 +0100
Message-ID: <CAK8P3a37rQz6ihEJBZNkOARJXJE7U9TX7pCUhyQjQn6fJw+jUQ@mail.gmail.com>
Subject: Re: [PATCH 2/2] pinctrl: ingenic: Only support SoCs enabled in config
To:     Daniel Palmer <daniel@0x0f.com>
Cc:     Paul Cercueil <paul@crapouillou.net>,
        Linus Walleij <linus.walleij@linaro.org>, od@zcrc.me,
        linux-arm-kernel <linux-arm-kernel@lists.infradead.org>,
        "open list:BROADCOM NVRAM DRIVER" <linux-mips@vger.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Wed, Dec 9, 2020 at 11:13 AM Daniel Palmer <daniel@0x0f.com> wrote:
>
> Hi Paul and others,
>
> Sorry to hijack this but I actually want to do something similar to
> this in some other drivers.
> The targets I'm working with have only 64MB of ram so I want to remove
> code wherever possible.
> Is there any reason to do it like this instead of wrapping the whole
> unneeded of_device_id struct in an #ifdef?
> For example there is a rule that the compatible strings have to be
> present even if the driver isn't usable or something?

No, there is no such rule, but adding lots of #ifdef checks in this
file would be much less readable and more error-prone, as you'd
have to make sure the two #ifdef blocks around the structure
match the one for the ID table, and any function that is called
by more than one SoC has the correct combination of A || B || D
checks, and nobody ever gets that right.

      Arnd