linux-kernel - RFC: get rid of #ifdef CONFIG

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 for Android: free password hash cracker in your pocket

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

Message-Id: <1559303690-8108-1-git-send-email-info@metux.net>
Date:   Fri, 31 May 2019 13:54:47 +0200
From:   "Enrico Weigelt, metux IT consult" <info@...ux.net>
To:     linux-kernel@...r.kernel.org
Cc:     axboe@...nel.dk, herbert@...dor.apana.org.au, davem@...emloft.net,
        linux-block@...r.kernel.org, linux-crypto@...r.kernel.org
Subject: RFC: get rid of #ifdef CONFIG_OF's around of match tables

Hi folks,

we've got many places where code declaring oftree match tables is
enclodes by #ifdef CONFIG_OF, so it can also be built w/o oftree
support.

To make this easier to read, I'm proposing a new helper macro
MODULE_OF_TABLE(foo) which just calls MODULE_DEVICE_TABLE(of, foo)
when CONFIG_OF enabled, otherwise just noop. Along w/ of_match_ptr(),
we can get rid most of these #ifdef CONFIG_OF cases. I believe, the
compiler can automatically optimize-away the unused tables.
(correct me if I'm wrong)

This queue just introduces the macro and converts two random examples.
I'll post more patches for the rest of the tree, if you folks aggree
to this approach. And also I'd like to do the same w/ ACPI, PCI, etc.

Another idea I'm currently thinking about is moving the whole table
definition into a macro call, so the tables won't even get defined
when CONFIG_OF isn't defined, and we've got even less to type. It
then would look like this:

    MODULE_DECLARE_OF_MATCH(foo_of_match,
        { .compatible = "acme,foo" },
        { .compatible = "acme,bar" })

(note that the macro also defines the sentinel on its own).

What do you think about this idea ?

--mtx