Message-ID: <4c0696c0f02cf5da747a88a40d3f29ba597482ea.camel@linux.ibm.com>
Subject: Re: [PATCH v2 1/4] PCI: s390: Fix use-after-free of PCI resources
 with per-function hotplug
From:   Niklas Schnelle <schnelle@linux.ibm.com>
To:     Bjorn Helgaas <helgaas@kernel.org>
Cc:     Bjorn Helgaas <bhelgaas@google.com>,
        Lukas Wunner <lukas@wunner.de>,
        Gerd Bayer <gbayer@linux.ibm.com>,
        Matthew Rosato <mjrosato@linux.ibm.com>,
        linux-s390@vger.kernel.org, linux-kernel@vger.kernel.org,
        linux-pci@vger.kernel.org
Date:   Thu, 09 Mar 2023 17:39:13 +0100
In-Reply-To: <20230308231449.GA1057317@bhelgaas>
References: <20230308231449.GA1057317@bhelgaas>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.46.4 (3.46.4-1.fc37) 
MIME-Version: 1.0
Precedence: bulk

On Wed, 2023-03-08 at 17:14 -0600, Bjorn Helgaas wrote:
> On Mon, Mar 06, 2023 at 04:10:11PM +0100, Niklas Schnelle wrote:
> > On s390 PCI functions may be hotplugged individually even when they
> > belong to a multi-function device. In particular on an SR-IOV device VF=
s
> > may be removed and later re-added.
> >=20
> > In commit a50297cf8235 ("s390/pci: separate zbus creation from
> > scanning") it was missed however that struct pci_bus and struct
> > zpci_bus's resource list retained a reference to the PCI functions MMIO
> > resources even though those resources are released and freed on
> > hot-unplug. These stale resources may subsequently be claimed when the
> > PCI function re-appears resulting in use-after-free.
> >=20
> > One idea of fixing this use-after-free in s390 specific code that was
> > investigated was to simply keep resources around from the moment a PCI
> > function first appeared until the whole virtual PCI bus created for
> > a multi-function device disappears. The problem with this however is
> > that due to the requirement of artificial MMIO addreesses (address
> > cookies) extra logic is then needed to keep the address cookies
> > compatible on re-plug. At the same time the MMIO resources semantically
> > belong to the PCI function so tying their lifecycle to the function
> > seems more logical.
> >=20
> > Instead a simpler approach is to remove the resources of an individuall=
y
> > hot-unplugged PCI function from the PCI bus's resource list while
> > keeping the resources of other PCI functions on the PCI bus untouched.
> >=20
> > This is done by introducing pci_bus_remove_resource() to remove an
> > individual resource. Similarly the resource also needs to be removed
> > from the struct zpci_bus's resource list. It turns out however, that
> > there is really no need to add the MMIO resources to the struct
> > zpci_bus's resource list at all and instead we can simply use the
> > zpci_bar_struct's resource pointer directly.
> >=20
> > Fixes: a50297cf8235 ("s390/pci: separate zbus creation from scanning")
> > Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com>
>=20
> Acked-by: Bjorn Helgaas <bhelgaas@google.com>
>=20
> The meat of this is mostly in s390, so I think it makes more sense to
> merge via that tree.  But let me know if you'd rather that I take it.
>=20
>=20

Thanks for taking a look and the valuable suggestions. I'll coordinate
with Vasily to take this via the s390 tree. As for the locking I agree
it is out of scope for this series. Meant more that the resource
handling might be a good place to start splitting up the
pci_rescan_remove_lock and that I might take a look at that if I find
the time which of course we're all lacking.

Regards,
Niklas