From:   Oleksandr Natalenko <oleksandr@natalenko.name>
To:     linux-kernel@vger.kernel.org, Angelo Haller <lkml@szanni.org>
Cc:     Peter Zijlstra <peterz@infradead.org>,
        Ingo Molnar <mingo@redhat.com>,
        Juri Lelli <juri.lelli@redhat.com>,
        Vincent Guittot <vincent.guittot@linaro.org>,
        Will Deacon <will@kernel.org>, linux-rt-users@vger.kernel.org
Subject: Re: sched: some non-GPL symbols becoming GPL-only with CONFIG_PREEMPT_RT enabled
Date:   Sun, 28 Nov 2021 10:46:13 +0100
Message-ID: <2606453.mvXUDI8C0e@natalenko.name>
In-Reply-To: <5475c3ab-a53c-8728-98c5-98fd948ff556@szanni.org>
References: <5475c3ab-a53c-8728-98c5-98fd948ff556@szanni.org>
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

Hello.

On ned=C4=9Ble 28. listopadu 2021 1:07:49 CET Angelo Haller wrote:
> Greetings. I hope I picked the right mailing list, as this issue might
> be one that affects various subsystems and components:
>=20
> When compiling kernel 5.15 (and 5.16-rc2) with `CONFIG_PREEMPT_RT`
> enabled, some of the symbols being exported as `EXPORT_SYMBOL` suddenly
> become `EXPORT_SYMBOL_GPL` through transitive effects.
>=20
> In particular the symbols `migrate_enable` and `migrate_disable` are
> currently marked as `EXPORT_SYMBOL_GPL` - yet are called from multiple
> functions that are marked as `EXPORT_SYMBOL`.
>=20
> Here an (incomplete?) listing of call sites I came across:
>=20
> kernel/locking/spinlock_rt.c - rt_spin_unlock()
> kernel/locking/spinlock_rt.c - rt_read_unlock()
> kernel/locking/spinlock_rt.c - rt_write_unlock()
> mm/highmem.c - kunmap_local_indexed()
>=20
> The issue I'm facing in particular is kmap_atomic() calling
> `migrate_disable` and therefore suddenly becoming GPL-only. This breaks
> the out-of-tree CDDL licensed module ZFS and has been reported before
> already [0]. The conversation seemingly going nowhere - or patches at
> least not being applied upstream. Downstream issue for reference [1].

What about going the other way around and let ZFS be re-licensed under GPL?

> As the original implementation of `migrate_enable` and `migrate_disable`
> is apparently by Peter Zijlstra [2]. Peter would you be willing to
> re-license both symbols `migrate_enable` and `migrate_disable` as
> `EXPORT_SYMBOL`?
>=20
> The bigger issue I'm seeing though is that there is currently no
> automated test to uncover exported symbols changing their license
> depending on build configuration? I am not intimately familiar with the
> API guarantees the kernel gives, but this seems like a violation. There
> might be other symbols with similar licensing problems.
>=20
> I can open a bugzilla ticket too - if that is preferred.
>=20
> Angelo
>=20
>=20
> [0]
> https://lore.kernel.org/linux-rt-users/20201208212841.694b3022@orivej.ori=
vej
> .org/T/ [1] https://github.com/openzfs/zfs/issues/11097
> [2]
> https://git.kernel.org/pub/scm/linux/kernel/git/rt/linux-rt-devel.git/dif=
f/p
> atches/0009-sched-Add-migrate_disable.patch?h=3Dv5.9-rc8-rt14-patches&id=
=3D9a89b
> fdb3bc77aecdd0ff8cc69b595541c7b50c4


=2D-=20
Oleksandr Natalenko (post-factum)