linux-kernel - Re: [PATCH v2 0/4] kbuild: resurrect generic header check facility

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]

Message-ID: <CAHk-=wgXb5ZVd+=b49957gm2Uo58h8EHBX5_mZfX4R_MJFqd0w@mail.gmail.com>
Date: Tue, 8 Apr 2025 12:48:13 -0700
From: Linus Torvalds <torvalds@...ux-foundation.org>
To: Jani Nikula <jani.nikula@...el.com>
Cc: Jason Gunthorpe <jgg@...dia.com>, Masahiro Yamada <masahiroy@...nel.org>, linux-kernel@...r.kernel.org, 
	David Airlie <airlied@...il.com>, Simona Vetter <simona.vetter@...ll.ch>, linux-kbuild@...r.kernel.org, 
	dri-devel@...ts.freedesktop.org, intel-xe@...ts.freedesktop.org, 
	intel-gfx@...ts.freedesktop.org
Subject: Re: [PATCH v2 0/4] kbuild: resurrect generic header check facility

On Tue, 8 Apr 2025 at 01:28, Jani Nikula <jani.nikula@...el.com> wrote:
>
> Your goal may be to make everything self-contained, but AFAICS there is
> no agreement on that goal.

Yeah, absolutely not.

I'm not interested in making some general rule that all headers should
be self-contained.

We already have some fairly obvious and clear exceptions to that, in
how some headers are special and get included early on and headers are
*not* supposed to include them themselves (ie things like
compiler-version.h and kconfig.h)

And while those are *really* special and end up being done
automatically by our compiler flags, they are by no means the only
special cases.

Quite a *lot* of our headers have things like

    # error "Please do not include this file directly."

because those headers are literally *designed* to not be
self-sufficient. And that is absolutely not a mistake. These are
headers that are meant to be included in very specific situations by
specific other header files.

So no. The whole "everything is going to be self-contained" is simply
not going to happen. It's not even worth discussing. It's a
no-starter, and limits our header file design much too much.

Honestly, I think the whole "headers have to be self-contained in
general" thing is a mistake. But I think it's fine for people to mark
their "generic" headers for some kind of checking.

I think it's a bad bad idea to make it a rule, though.

              Linus