Thx Matthew for the quick test.

I've now submitted the kernel SRU for the remaining fix for focal:
https://lists.ubuntu.com/archives/kernel-team/2020-December/thread.html#115495
and changed the status to 'In Progress' for focal.

-- 
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to linux in Ubuntu.
https://bugs.launchpad.net/bugs/1907265

Title:
  [UBUNTU 21.04] s390/pci: vfio-pci mmio being disabled erroneously

Status in Ubuntu on IBM z Systems:
  In Progress
Status in linux package in Ubuntu:
  In Progress
Status in linux source package in Focal:
  In Progress
Status in linux source package in Groovy:
  Fix Released
Status in linux source package in Hirsute:
  In Progress

Bug description:
  Description:   s390/pci: vfio-pci mmio being disabled erroneously
  Symptom:       PCI virtual functions passed through via vfio-pci are unusable 
 
  Problem:       The fix for CVE-2020-12888 'abafbc551fdd vfio-pci: Invalidate
                 mmaps and block MMIO access on disabled memory' introduced an
                 issue which prevented PCI Virtual Functions from being passed-
                 through via vfio-pci as VFs are not allowed to have the
                 PCI_COMMAND_MEMORY bit enabled per spec.  This issue was
                 initially thought to be fixed via 'ebfa440ce38b vfio/pci: Fix
                 SR-IOV VF handling with MMIO blocking' which removes the
                 PCI_COMMAND_MEMORY requirement for VFs but this did not solve
                 the issue for virtual functions on s390 that are first passed
                 through to an LPAR without their associated physical function
                 as they are not identified as VFs in the traditional sense
                 (there is no PF available to the host kernel to link to).  As a
                 result, when passing these devices through to a guest via
                 vfio-pci, these devices do not trigger the check added by
                 ebfa440ce38b which results in MMIO access from the guest being
                 blocked because the PCI_COMMAND_MEMORY bit is off.
  Solution:      Identify devices beyond traditional linked VFs that vfio must
                 consider memory-enabled. 
  Reproduction:  Pass a PCI Virtual Function to a qemu-kvm guest via vfio-pci
  Upstream-ID:   12856e7acde4702b7c3238c15fcba86ff6aa507f
                 08b6e22b850c28b6032da1e4d767a33116e23dfb
                 515ecd5368f1510152fa4f9b9ce55b66ac56c334

  These patches need to be applied for 20.10 and 20.04.
  The git-commits will apply cleanly on the dedicated kernels.

  Manx thx

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu-z-systems/+bug/1907265/+subscriptions

-- 
Mailing list: https://launchpad.net/~kernel-packages
Post to     : kernel-packages@lists.launchpad.net
Unsubscribe : https://launchpad.net/~kernel-packages
More help   : https://help.launchpad.net/ListHelp

Reply via email to