Re: [PATCH RFC] gcc-plugins: Handle GCC version mismatch for OOT modules

From: Masahiro Yamada <masahiroy@kernel.org>
To: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	Michal Marek <michal.lkml@markovi.net>,
	linux-hardening@vger.kernel.org,
	Linux Kbuild mailing list <linux-kbuild@vger.kernel.org>,
	Peter Zijlstra <peterz@infradead.org>,
	Justin Forbes <jforbes@redhat.com>,
	Ondrej Mosnacek <omosnace@redhat.com>,
	Frank Eigler <fche@redhat.com>, Kees Cook <keescook@chromium.org>
Subject: Re: [PATCH RFC] gcc-plugins: Handle GCC version mismatch for OOT modules
Date: Thu, 4 Mar 2021 03:49:35 +0900	[thread overview]
Message-ID: <CAK7LNAReuB5zUq_7S8ZG25+tdQowECDOK1rApYvkPCpHhPjK5w@mail.gmail.com> (raw)
In-Reply-To: <20210302232649.y2tutffhxsblwqlb@treble>

On Wed, Mar 3, 2021 at 8:27 AM Josh Poimboeuf <jpoimboe@redhat.com> wrote:
>
> On Mon, Jan 25, 2021 at 02:42:10PM -0600, Josh Poimboeuf wrote:
> > When building out-of-tree kernel modules, the build system doesn't
> > require the GCC version to match the version used to build the original
> > kernel.  That's probably [1] fine.
> >
> > In fact, for many distros, the version of GCC used to build the latest
> > kernel doesn't necessarily match the latest released GCC, so a GCC
> > mismatch turns out to be pretty common.  And with CONFIG_MODVERSIONS
> > it's probably more common.
> >
> > So a lot of users have come to rely on being able to use a different
> > version of GCC when building OOT modules.
> >
> > But with GCC plugins enabled, that's no longer allowed:
> >
> >   cc1: error: incompatible gcc/plugin versions
> >   cc1: error: failed to initialize plugin ./scripts/gcc-plugins/structleak_plugin.so
> >
> > That error comes from the plugin's call to
> > plugin_default_version_check(), which strictly enforces the GCC version.
> > The strict check makes sense, because there's nothing to prevent the GCC
> > plugin ABI from changing -- and it often does.
> >
> > But failing the build isn't necessary.  For most plugins, OOT modules
> > will otherwise work just fine without the plugin instrumentation.
> >
> > When a GCC version mismatch is detected, print a warning and disable the
> > plugin.  The only exception is the RANDSTRUCT plugin which needs all
> > code to see the same struct layouts.  In that case print an error.
>
> Hi Masahiro,
>
> This problem is becoming more prevalent.  We will need to fix it one way
> or another, if we want to support distro adoption of these GCC
> plugin-based features.
>
> Frank suggested a possibly better idea: always rebuild the plugins when
> the GCC version changes.

That is just another form of the previous patch,
which was already rejected.

- That is a hack just for external modules
- Our consensus is, use the same version for the kernel and external modules

I use Ubuntu, and I do not see such a problem.
(I have never seen it on Debian either, except sid.)

I see Fedora providing GCC whose version is different
from the one used for building the kernel.
That is a problem on the distribution side.

In my ubuntu.

$ grep CC_VERSION_TEXT   /lib/modules/$(uname -r)/build/.config
CONFIG_CC_VERSION_TEXT="gcc (Ubuntu 10.2.0-13ubuntu1) 10.2.0"
$ gcc --version  | head -n1
gcc (Ubuntu 10.2.0-13ubuntu1) 10.2.0

> What do you think?

NACK.

>  Any suggestions on how to
> implement that?  Otherwise I can try to hack something together.
>
> --
> Josh
>
--
Best Regards
Masahiro Yamada