linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jason Gunthorpe <jgg@nvidia.com>
To: Jason Wang <jasowang@redhat.com>
Cc: mst <mst@redhat.com>, linux-kernel <linux-kernel@vger.kernel.org>,
	kvm <kvm@vger.kernel.org>,
	virtualization <virtualization@lists.linux-foundation.org>,
	netdev <netdev@vger.kernel.org>,
	Maxime Coquelin <maxime.coquelin@redhat.com>,
	"Liang, Cunming" <cunming.liang@intel.com>,
	zhihong.wang@intel.com, rob.miller@broadcom.com,
	Xiao W Wang <xiao.w.wang@intel.com>,
	Zhu Lingshan <lingshan.zhu@intel.com>,
	eperezma <eperezma@redhat.com>, Cindy Lu <lulu@redhat.com>,
	Parav Pandit <parav@mellanox.com>,
	"Tian, Kevin" <kevin.tian@intel.com>,
	Stefan Hajnoczi <stefanha@redhat.com>,
	Randy Dunlap <rdunlap@infradead.org>,
	Christoph Hellwig <hch@infradead.org>,
	Ariel Adam <aadam@redhat.com>,
	jiri@mellanox.com, shahafs@mellanox.com,
	Harpreet Singh Anand <hanand@xilinx.com>,
	mhabets@solarflare.com, Gautam Dawar <gdawar@xilinx.com>,
	Saugat Mitra <saugatm@xilinx.com>,
	vmireyno@marvell.com, zhangweining@ruijie.com.cn,
	Tiwei Bie <tiwei.bie@intel.com>,
	Lu Baolu <baolu.lu@linux.intel.com>
Subject: Re: [PATCH V9 7/9] vhost: introduce vDPA-based backend
Date: Tue, 2 Nov 2021 12:56:11 -0300	[thread overview]
Message-ID: <20211102155611.GL2744544@nvidia.com> (raw)
In-Reply-To: <CACGkMEtbs3u7J7krpkusfqczTU00+6o_YtZjD8htC=+Un9cNew@mail.gmail.com>

On Tue, Nov 02, 2021 at 11:52:20AM +0800, Jason Wang wrote:
> On Mon, Nov 1, 2021 at 10:11 PM Jason Gunthorpe <jgg@nvidia.com> wrote:
> >
> > On Thu, Mar 26, 2020 at 10:01:23PM +0800, Jason Wang wrote:
> > > From: Tiwei Bie <tiwei.bie@intel.com>
> > >
> > > This patch introduces a vDPA-based vhost backend. This backend is
> > > built on top of the same interface defined in virtio-vDPA and provides
> > > a generic vhost interface for userspace to accelerate the virtio
> > > devices in guest.
> > >
> > > This backend is implemented as a vDPA device driver on top of the same
> > > ops used in virtio-vDPA. It will create char device entry named
> > > vhost-vdpa-$index for userspace to use. Userspace can use vhost ioctls
> > > on top of this char device to setup the backend.
> > >
> > > Vhost ioctls are extended to make it type agnostic and behave like a
> > > virtio device, this help to eliminate type specific API like what
> > > vhost_net/scsi/vsock did:
> > >
> > > - VHOST_VDPA_GET_DEVICE_ID: get the virtio device ID which is defined
> > >   by virtio specification to differ from different type of devices
> > > - VHOST_VDPA_GET_VRING_NUM: get the maximum size of virtqueue
> > >   supported by the vDPA device
> > > - VHSOT_VDPA_SET/GET_STATUS: set and get virtio status of vDPA device
> > > - VHOST_VDPA_SET/GET_CONFIG: access virtio config space
> > > - VHOST_VDPA_SET_VRING_ENABLE: enable a specific virtqueue
> > >
> > > For memory mapping, IOTLB API is mandated for vhost-vDPA which means
> > > userspace drivers are required to use
> > > VHOST_IOTLB_UPDATE/VHOST_IOTLB_INVALIDATE to add or remove mapping for
> > > a specific userspace memory region.
> > >
> > > The vhost-vDPA API is designed to be type agnostic, but it allows net
> > > device only in current stage. Due to the lacking of control virtqueue
> > > support, some features were filter out by vhost-vdpa.
> > >
> > > We will enable more features and devices in the near future.
> >
> > [..]
> >
> > > +static int vhost_vdpa_alloc_domain(struct vhost_vdpa *v)
> > > +{
> > > +     struct vdpa_device *vdpa = v->vdpa;
> > > +     const struct vdpa_config_ops *ops = vdpa->config;
> > > +     struct device *dma_dev = vdpa_get_dma_dev(vdpa);
> > > +     struct bus_type *bus;
> > > +     int ret;
> > > +
> > > +     /* Device want to do DMA by itself */
> > > +     if (ops->set_map || ops->dma_map)
> > > +             return 0;
> > > +
> > > +     bus = dma_dev->bus;
> > > +     if (!bus)
> > > +             return -EFAULT;
> > > +
> > > +     if (!iommu_capable(bus, IOMMU_CAP_CACHE_COHERENCY))
> > > +             return -ENOTSUPP;
> > > +
> > > +     v->domain = iommu_domain_alloc(bus);
> > > +     if (!v->domain)
> > > +             return -EIO;
> > > +
> > > +     ret = iommu_attach_device(v->domain, dma_dev);
> > > +     if (ret)
> > > +             goto err_attach;
> > >
> >
> > I've been looking at the security of iommu_attach_device() users, and
> > I wonder if this is safe?
> >
> > The security question is if userspace is able to control the DMA
> > address the devices uses? Eg if any of the cpu to device ring's are in
> > userspace memory?
> >
> > For instance if userspace can tell the device to send a packet from an
> > arbitrary user controlled address.
> 
> The map is validated via pin_user_pages() which guarantees that the
> address is not arbitrary and must belong to userspace?

That controls what gets put into the IOMMU, it doesn't restrict what
DMA the device itself can issue.

Upon investigating more it seems the answer is that
iommu_attach_device() requires devices to be in singleton groups, so
there is no leakage from rouge DMA

Jason

  reply	other threads:[~2021-11-02 15:56 UTC|newest]

Thread overview: 49+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-03-26 14:01 [PATCH V9 0/9] vDPA support Jason Wang
2020-03-26 14:01 ` [PATCH V9 1/9] vhost: refine vhost and vringh kconfig Jason Wang
2020-04-01 11:21   ` Christian Borntraeger
2020-04-01 12:50     ` Jason Wang
2020-04-01 12:56       ` Christian Borntraeger
2020-04-01 13:02         ` Christian Borntraeger
2020-04-01 14:01           ` Michael S. Tsirkin
2020-04-01 14:13           ` Jason Wang
2020-04-01 14:18             ` Michael S. Tsirkin
2020-04-01 14:36               ` Jason Wang
2020-04-01 14:27             ` Michael S. Tsirkin
2020-04-01 14:50               ` Jason Wang
2020-04-01 15:57                 ` Michael S. Tsirkin
2020-04-01 18:53                   ` Christian Borntraeger
2020-04-01 13:22   ` Michael S. Tsirkin
2020-04-01 14:08     ` Jason Wang
2020-04-01 14:13       ` Michael S. Tsirkin
2020-04-01 14:29         ` Jason Wang
2020-04-01 14:35           ` Michael S. Tsirkin
2020-04-01 14:39             ` Jason Wang
2020-04-01 14:39           ` Michael S. Tsirkin
2020-04-01 14:43             ` Jason Wang
2020-04-01 16:05           ` Michael S. Tsirkin
2020-04-01 16:08           ` Michael S. Tsirkin
2020-04-02  3:22             ` Jason Wang
2020-04-02 14:03               ` Michael S. Tsirkin
2020-04-02 14:23                 ` Jason Wang
2020-04-02 14:38                   ` Michael S. Tsirkin
2020-04-02 14:56                     ` Jason Wang
2020-03-26 14:01 ` [PATCH V9 2/9] vhost: allow per device message handler Jason Wang
2020-03-26 14:01 ` [PATCH V9 3/9] vhost: factor out IOTLB Jason Wang
2020-03-26 14:01 ` [PATCH V9 4/9] vringh: IOTLB support Jason Wang
2020-03-26 14:01 ` [PATCH V9 5/9] vDPA: introduce vDPA bus Jason Wang
2020-03-26 14:01 ` [PATCH V9 6/9] virtio: introduce a vDPA based transport Jason Wang
2020-03-26 14:01 ` [PATCH V9 7/9] vhost: introduce vDPA-based backend Jason Wang
2021-11-01 14:11   ` Jason Gunthorpe
2021-11-02  3:52     ` Jason Wang
2021-11-02 15:56       ` Jason Gunthorpe [this message]
2021-11-03  7:34         ` Jason Wang
2020-03-26 14:01 ` [PATCH V9 8/9] vdpasim: vDPA device simulator Jason Wang
2020-04-10  7:45   ` Geert Uytterhoeven
2020-04-10  8:23     ` Jason Wang
2020-03-26 14:01 ` [PATCH V9 9/9] virtio: Intel IFC VF driver for VDPA Jason Wang
2020-04-09 10:41   ` Arnd Bergmann
2020-04-09 12:43     ` Jason Wang
2020-04-09 12:49       ` Zhu, Lingshan
2020-04-09 20:25     ` Michael S. Tsirkin
2020-04-10  3:15       ` Zhu, Lingshan
2020-03-29 11:07 ` [PATCH V9 0/9] vDPA support Michael S. Tsirkin

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20211102155611.GL2744544@nvidia.com \
    --to=jgg@nvidia.com \
    --cc=aadam@redhat.com \
    --cc=baolu.lu@linux.intel.com \
    --cc=cunming.liang@intel.com \
    --cc=eperezma@redhat.com \
    --cc=gdawar@xilinx.com \
    --cc=hanand@xilinx.com \
    --cc=hch@infradead.org \
    --cc=jasowang@redhat.com \
    --cc=jiri@mellanox.com \
    --cc=kevin.tian@intel.com \
    --cc=kvm@vger.kernel.org \
    --cc=lingshan.zhu@intel.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=lulu@redhat.com \
    --cc=maxime.coquelin@redhat.com \
    --cc=mhabets@solarflare.com \
    --cc=mst@redhat.com \
    --cc=netdev@vger.kernel.org \
    --cc=parav@mellanox.com \
    --cc=rdunlap@infradead.org \
    --cc=rob.miller@broadcom.com \
    --cc=saugatm@xilinx.com \
    --cc=shahafs@mellanox.com \
    --cc=stefanha@redhat.com \
    --cc=tiwei.bie@intel.com \
    --cc=virtualization@lists.linux-foundation.org \
    --cc=vmireyno@marvell.com \
    --cc=xiao.w.wang@intel.com \
    --cc=zhangweining@ruijie.com.cn \
    --cc=zhihong.wang@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).