lkml.org 
[lkml]   [2020]   [Mar]   [6]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH v2 0/7] vfio/pci: SR-IOV support
On Fri, 6 Mar 2020 11:35:21 +0800
Jason Wang <jasowang@redhat.com> wrote:

> On 2020/3/6 上午1:14, Alex Williamson wrote:
> > On Tue, 25 Feb 2020 14:09:07 +0800
> > Jason Wang <jasowang@redhat.com> wrote:
> >
> >> On 2020/2/25 上午10:33, Tian, Kevin wrote:
> >>>> From: Alex Williamson
> >>>> Sent: Thursday, February 20, 2020 2:54 AM
> >>>>
> >>>> Changes since v1 are primarily to patch 3/7 where the commit log is
> >>>> rewritten, along with option parsing and failure logging based on
> >>>> upstream discussions. The primary user visible difference is that
> >>>> option parsing is now much more strict. If a vf_token option is
> >>>> provided that cannot be used, we generate an error. As a result of
> >>>> this, opening a PF with a vf_token option will serve as a mechanism of
> >>>> setting the vf_token. This seems like a more user friendly API than
> >>>> the alternative of sometimes requiring the option (VFs in use) and
> >>>> sometimes rejecting it, and upholds our desire that the option is
> >>>> always either used or rejected.
> >>>>
> >>>> This also means that the VFIO_DEVICE_FEATURE ioctl is not the only
> >>>> means of setting the VF token, which might call into question whether
> >>>> we absolutely need this new ioctl. Currently I'm keeping it because I
> >>>> can imagine use cases, for example if a hypervisor were to support
> >>>> SR-IOV, the PF device might be opened without consideration for a VF
> >>>> token and we'd require the hypservisor to close and re-open the PF in
> >>>> order to set a known VF token, which is impractical.
> >>>>
> >>>> Series overview (same as provided with v1):
> >>> Thanks for doing this!
> >>>
> >>>> The synopsis of this series is that we have an ongoing desire to drive
> >>>> PCIe SR-IOV PFs from userspace with VFIO. There's an immediate need
> >>>> for this with DPDK drivers and potentially interesting future use
> >>> Can you provide a link to the DPDK discussion?
> >>>
> >>>> cases in virtualization. We've been reluctant to add this support
> >>>> previously due to the dependency and trust relationship between the
> >>>> VF device and PF driver. Minimally the PF driver can induce a denial
> >>>> of service to the VF, but depending on the specific implementation,
> >>>> the PF driver might also be responsible for moving data between VFs
> >>>> or have direct access to the state of the VF, including data or state
> >>>> otherwise private to the VF or VF driver.
> >>> Just a loud thinking. While the motivation of VF token sounds reasonable
> >>> to me, I'm curious why the same concern is not raised in other usages.
> >>> For example, there is no such design in virtio framework, where the
> >>> virtio device could also be restarted, putting in separate process (vhost-user),
> >>> and even in separate VM (virtio-vhost-user), etc.
> >>
> >> AFAIK, the restart could only be triggered by either VM or qemu. But
> >> yes, the datapath could be offloaded.
> >>
> >> But I'm not sure introducing another dedicated mechanism is better than
> >> using the exist generic POSIX mechanism to make sure the connection
> >> (AF_UINX) is secure.
> >>
> >>
> >>> Of course the para-
> >>> virtualized attribute of virtio implies some degree of trust, but as you
> >>> mentioned many SR-IOV implementations support VF->PF communication
> >>> which also implies some level of trust. It's perfectly fine if VFIO just tries
> >>> to do better than other sub-systems, but knowing how other people
> >>> tackle the similar problem may make the whole picture clearer. 😊
> >>>
> >>> +Jason.
> >>
> >> I'm not quite sure e.g allowing userspace PF driver with kernel VF
> >> driver would not break the assumption of kernel security model. At least
> >> we should forbid a unprivileged PF driver running in userspace.
> > It might be useful to have your opinion on this series, because that's
> > exactly what we're trying to do here. Various environments, DPDK
> > specifically, want a userspace PF driver. This series takes steps to
> > mitigate the risk of having such a driver, such as requiring this VF
> > token interface to extend the VFIO interface and validate participation
> > around a PF that is not considered trusted by the kernel.
>
>
> I may miss something. But what happens if:
>
> - PF driver is running by unprivileged user
> - PF is programmed to send translated DMA request
> - Then unprivileged user can mangle the kernel data

ATS is a security risk regardless of SR-IOV, how does this change it?
Thanks,

Alex

> > We also set
> > a driver_override to try to make sure no host kernel driver can
> > automatically bind to a VF of a user owned PF, only vfio-pci, but we
> > don't prevent the admin from creating configurations where the VFs are
> > used by other host kernel drivers.
> >
> > I think the question Kevin is inquiring about is whether virtio devices
> > are susceptible to the type of collaborative, shared key environment
> > we're creating here. For example, can a VM or qemu have access to
> > reset a virtio device in a way that could affect other devices, ex. FLR
> > on a PF that could interfere with VF operation. Thanks,
>
>
> Right, but I'm not sure it can be done only via virtio or need support
> from transport (e.g PCI).
>
> Thanks
>
>
> >
> > Alex
> >

\
 
 \ /
  Last update: 2020-03-06 17:25    [W:0.108 / U:24.356 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site