------- Comment From ale...@au1.ibm.com 2016-10-06 03:23 EDT-------
The message means that KVM cannot accelerate TCE operations on one of DMA 
windows, in this particular case - the huge DMA window. Such support is not 
present in the HostOS as it is not in upstream kernel so there is a message.

Commit 178a7875 tried adding VFIO KVM device but it did not on PPC64
book3s KVM due to a mistake, This -
https://git.kernel.org/cgit/linux/kernel/git/paulus/powerpc.git/commit/?h
=kvm-ppc-next&id=4b3d173d0440d37534906b6d93c02dfb577c68ce - does add it
but it is not in upstream yet.

-- 
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/1630554

Title:
  QEMU throws failure msg while booting guest with SRIOV VF

Status in linux package in Ubuntu:
  New

Bug description:
  == Comment: #10 - VIPIN K. PARASHAR <vipar...@in.ibm.com> - 2016-10-04 
02:39:30 ==
  (In reply to comment #9)
  > I took a quick look to this one and also noticed these 2 links:
  > https://bugzilla.redhat.com/show_bug.cgi?id=1237034
  > http://marc.info/?l=linux-kernel&m=143737274332400&w=2
  > and it looks like to avoid these message maybe just need to have this in
  > config file:
  > CONFIG_KVM_VFIO=y
  > I think we have that in powerkvm but not in Ubuntu KVM.

  On x86 running Ubuntu 16.04
  ==================
  $ lsb_release -a
  No LSB modules are available.
  Distributor ID:       Ubuntu
  Description:  Ubuntu 16.04.1 LTS
  Release:      16.04
  Codename:     xenial
  $ head /proc/cpuinfo 
  processor     : 0
  vendor_id     : GenuineIntel
  cpu family    : 6
  model         : 61
  model name    : Intel(R) Core(TM) i5-5300U CPU @ 2.30GHz
  stepping      : 4
  microcode     : 0x22
  cpu MHz               : 2300.179
  cache size    : 3072 KB
  physical id   : 0
  $ uname -a
  Linux Workstation 4.4.0-38-generic #57-Ubuntu SMP Tue Sep 6 15:42:33 UTC 2016 
x86_64 x86_64 x86_64 GNU/Linux
  $ grep CONFIG_KVM_VFIO /boot/config-4.4.0-38-generic 
  CONFIG_KVM_VFIO=y
  $

  On PowerPC
  ======== 
  # lsb_release -a
  No LSB modules are available.
  Distributor ID:       Ubuntu
  Description:  Ubuntu Yakkety Yak (development branch)
  Release:      16.10
  Codename:     yakkety
  # head /proc/cpuinfo
  processor     : 0
  cpu           : POWER8E (raw), altivec supported
  clock         : 2061.000000MHz
  revision      : 2.1 (pvr 004b 0201)

  processor     : 1
  cpu           : POWER8E (raw), altivec supported
  clock         : 2061.000000MHz
  revision      : 2.1 (pvr 004b 0201)

  # uname -a
  Linux c158f2u09os 4.8.0-17-generic #19 SMP Thu Sep 29 12:03:21 CDT 2016 
ppc64le ppc64le ppc64le GNU/Linux
  # grep CONFIG_KVM_VFIO /boot/config-4.8.0-17-generic 
  #

  I see that CONFIG_KVM_VFIO is enabled on x86 running Ubuntu 16.04
  while ppc64 running 16.10 doesn't have it enabled.

  == Comment: #9 - Carol L. Soto <cls...@us.ibm.com> - 2016-09-30 16:25:32 ==
  I took a quick look to this one and also noticed these 2 links:
  https://bugzilla.redhat.com/show_bug.cgi?id=1237034
  http://marc.info/?l=linux-kernel&m=143737274332400&w=2
  and it looks like to avoid these message maybe just need to have this in 
config file:
  CONFIG_KVM_VFIO=y
  I think we have that in powerkvm but not in Ubuntu KVM.

  == Comment: #6 - VIPIN K. PARASHAR <vipar...@in.ibm.com> - 2016-09-23 
06:50:22 ==
  $ git log 178a787502123b0 -1
  commit 178a787502123b01499c5a4617b94bb69ad49dd5
  Author: David Gibson <da...@gibson.dropbear.id.au>
  Date:   Mon Feb 1 11:14:15 2016 +1100

      vfio: Enable VFIO device for powerpc
      
      ec53500f "kvm: Add VFIO device" added a special KVM pseudo-device which is
      used to handle any necessary interactions between KVM and VFIO.
      
      Currently that device is built on x86 and ARM, but not powerpc, although
      powerpc does support both KVM and VFIO.  This makes things awkward in
      userspace
      
      Currently qemu prints an alarming error message if you attempt to use VFIO
      and it can't initialize the KVM VFIO device.  We don't want to remove the
      warning, because lack of the KVM VFIO device could mean coherency problems
      on x86.  On powerpc, however, the error is harmless but looks disturbing,
      and a test based on host architecture in qemu would be ugly, and break if
      we do need the KVM VFIO device for something important in future.
      
      There's nothing preventing the KVM VFIO device from being built for
      powerpc, so this patch turns it on.  It won't actually do anything, since
      we don't define any of the arch_*() hooks, but it will make qemu happy and
      we can extend it in future if we need to.
      
      Signed-off-by: David Gibson <da...@gibson.dropbear.id.au>
      Reviewed-by: Eric Auger <eric.au...@linaro.org>
      Signed-off-by: Paul Mackerras <pau...@samba.org>
  $ 

  $ git describe --contains 178a787502123b
  v4.6-rc1~141^2~49^2~20
  $ 

  
  Commit (178a7875) seems to have enabled VFIO devices for powerpc.
  But this commit should already be available since 4.6 kernel and ubuntu is
  using 4.7 kernel here, so this should be already in.

  
  == Comment: #0 - SRIKANTH B. AITHAL <bssrika...@in.ibm.com> - 2016-09-15 
12:31:49 ==
  ---Problem Description---
  QEMU throwing false failure message while booting guest with SRIOV VF device

   
  ---uname output---
  Linux c158f2u09os 4.7.0unofficial #5 SMP Mon Sep 5 08:53:38 EDT 2016 ppc64le 
ppc64le ppc64le GNU/Linux
   
  ---Additional Hardware Info---
  Mellanox CX4 

   
  Machine Type = 8247-22L 
   
  ---Debugger---
  A debugger is not configured
   
  ---Steps to Reproduce---
   1. Edit guest ubuntu1610_srik and add below VF device [cold plug]
      <hostdev mode='subsystem' type='pci' managed='yes'>
        <source>
          <address domain='0x0005' bus='0x01' slot='0x00' function='0x3'/>
        </source>
        <address type='pci' domain='0x0000' bus='0x00' slot='0x01' 
function='0x0'/>
      </hostdev>

  2. Start the guest: virsh start ubuntu1610_srik

  3. In the /var/log/libvirt/qemu/ubuntu1610_srik.log we will notice
  below failure message.

  "-device vfio-pci,host=0005:01:00.3,id=hostdev0,bus=pci.0,addr=0x1:
  Failed to create KVM VFIO device: No such device"

  But VF would have been passthroughed and probed succesfully inside the
  guest.. I could configure it and transfer data over it.. and VF got
  reprobed succesfully in host after I took down 'ubuntu1610_srik'.

  This bug is to see why QEMU is printing false failure message.

  =====================
  Environment details:
  =====================

  Host : 
        Ubuntu 16.10
        kernel version: 
        Linux c158f2u09os 4.7.0unofficial #5 SMP Mon Sep 5 08:53:38 EDT 2016 
ppc64le ppc64le ppc64le GNU/Linux
  Guest: 
  1. ubuntu1610_srik
        kernel version: 4.7.0unofficial
  2. ubuntu160401_srik
        kernel version: 4.4.0-38-generic

  MOFED versions:
        MOFED version in Host as well Guest: 3.4-OFED.3.4.0.1.0.1
        CX4 Firmware version: 12.17.0222

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1630554/+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