TCT Meeting/October 2013 Minutes: Difference between revisions

From Xen
Jump to navigationJump to search
No edit summary
Line 26: Line 26:
* xend
* xend
** {{Action|Matt & Konrad}} to reply to the 4.4 planning thread with their BLOCKERS. {{NotDone}} (Konrad: absent)
** {{Action|Matt & Konrad}} to reply to the 4.4 planning thread with their BLOCKERS. {{NotDone}} (Konrad: absent)
** For Matt's, list discussion about SXP -- better to change the higher level stack. Matt will post this to xen-devel
*** For Matt's, list discussion about SXP -- better to change the higher level stack. Matt will post this to xen-devel
** {{Action|Ian}} to write patch to warn when xm is used {{Tick}}
** {{Action|Ian}} to write patch to warn when xm is used {{Tick}}



Revision as of 15:45, 4 November 2013

Attendees

  • Ian Campbell
  • Matt Wilson
  • Jan Beulich
  • Boris Ostrovsky
  • Ravi Suravee
  • Daniel De Graaf
  • Anthony Liguori


Review ACTION items.

  • Recurring need for pings on VMX/SVM/IOMMU patches

    • Pictogram voting comment 15px.png Action Jun: to figure out the focal point. Will Auld
  • xend

    • Pictogram voting comment 15px.png Action Matt & Konrad: to reply to the 4.4 planning thread with their BLOCKERS. (Konrad: absent)
      • For Matt's, list discussion about SXP -- better to change the higher level stack. Matt will post this to xen-devel

    • Pictogram voting comment 15px.png Action Ian: to write patch to warn when xm is used

Current technical challenges.

  • PVH next steps
    • Right people not on this call
    • Ian suggested to George to have a separate call
    • Jan would like to be invited

Coordination on current and future work

  • Konrad absent. All delayed to next time.
    • Proposals for interships program for summer or winter in Xen ecosystem. (Konrad) see AB_Votes/Outreach_Program_For_Women_Round_7
    • SR-IOV, PCI passthrough and all that fun. (Konrad)
    • SR-IOV, PCI passthrough gaps. (Konrad)
  • Boris. How much testing did we have for NMI injection into PV domains.
    • Jan, only recently implemented in pvops, but has worked for a long time in classis/SuSE kernel
    • Works for internal self NMI
    • Only dom0 might receive a real hardware NMI forwarded
    • Jan: Not a lot of testing of a high rate NMI (e.g. performance monitoring). Currently NMI is a rare event like a machine check.