[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [Xen-users] SVM not detected on Opteron 1220


  • To: "Brian P. Martin" <Brian@xxxxxxxxxxxxxxxxxxxx>
  • From: <djmagee@xxxxxxxxxxxx>
  • Date: Tue, 28 Jul 2009 08:20:58 -0400
  • Cc: xen-users@xxxxxxxxxxxxxxxxxxx
  • Delivery-date: Tue, 28 Jul 2009 05:24:18 -0700
  • List-id: Xen user discussion <xen-users.lists.xensource.com>
  • Thread-index: AcoO5hG0ek+2x6ULTwqlt2XLATjNjQAEHYrgAAFOa6AAH9GiUA==
  • Thread-topic: [Xen-users] SVM not detected on Opteron 1220

No problem.

Your DomUs will only see the emulated hardware.  To see your real
hardware inside a DomU you will have to pass it through.  Without the
appropriate hardware, you'll only be able to pass-through hardware to a
PV guest.  Since you're on AMD, I'm going to have to assume you don't
have IOMMU hardware, as they haven't sold any yet that I know.  Google
searches for the right keywords will reveal more than enough info you
can piece together to achieve this, but here's a short summary:

Make sure pciback is available in your Dom0 kernel (either built-in or
as a module).
Identify PCI device you want to hide, likely using lspci
If pciback is builtin, then you'd have to add a 'pciback.hide=(...)' to
the Dom0 kernel line in grub and restart
Otherwise, you can hide the device after startup as described here:
http://wiki.xensource.com/xenwiki/Assign_hardware_to_DomU_with_PCIBack_a
s_module
Assign the device to your DomU with a pci = [ '...' ] line the domain's
configuration file
 
I don't have any experience with PV passthrough, but it should be
easier, though less stable/secure than HVM passthrough.  Also, lately,
xen pci pass-through development has focused mostly on iommu backed
passthrough, and I've read messages suggesting there are more
restrictions to pv passthrough in more recent versions of xen.

Good luck,
Doug Magee

-----Original Message-----
From: xen-users-bounces@xxxxxxxxxxxxxxxxxxx
[mailto:xen-users-bounces@xxxxxxxxxxxxxxxxxxx] On Behalf Of Brian P.
Martin
Sent: Monday, July 27, 2009 4:55 PM
To: djmagee@xxxxxxxxxxxx
Cc: xen-users@xxxxxxxxxxxxxxxxxxx
Subject: RE: [Xen-users] SVM not detected on Opteron 1220

Awesome, Doug!  I've been on that wild goose chase for quite a while.  

I've been trying to figure out why a DomU guest can't see a parallel
port
(either the one on the mobo or another on a PCI card).  I thought it was
because I didn't have SVM active for some reason.  I'm still stuck, but
at
least now I'm stuck on the right problem.  

The few minutes you spent responding has saved me many hours of wasted
effort.  Thanks a lot.

                               -Brian M.

> -----Original Message-----
> From: djmagee@xxxxxxxxxxxx [mailto:djmagee@xxxxxxxxxxxx]
> Sent: Monday, July 27, 2009 1:18 PM
> To: Brian P. Martin
> Cc: xen-users@xxxxxxxxxxxxxxxxxxx
> Subject: RE: [Xen-users] SVM not detected on Opteron 1220
> 
> I believe this is normal (though I don't have any AMD hardware, it
> should be the same as vmx flag on intel).  Xen masks certain bits in
> the
> cpu information, and Dom0 does not have direct access to the physical
> CPUs, only the virtual CPUs assigned by the hypervisor.
> 
> To verify you have svm enabled, you will want to look at the output of
> "xm info", specifically the line that starts "xen_caps".  If you have
> entries there that start with "hvm" then your hardware supports
> virtualization.
> 
> Doug Magee
> 
> -----Original Message-----
> From: xen-users-bounces@xxxxxxxxxxxxxxxxxxx
> [mailto:xen-users-bounces@xxxxxxxxxxxxxxxxxxx] On Behalf Of Brian P.
> Martin
> Sent: Monday, July 27, 2009 2:14 PM
> To: xen-users@xxxxxxxxxxxxxxxxxxx
> Subject: [Xen-users] SVM not detected on Opteron 1220
> 
> I have SuperMicro mainboard, H8SMi-2 rev 2.01, with a dual core
Opteron
> 1220
> processor.  I've verified that the BIOS had "Secure Virtual Machine"
> enabled.  This appears to be the only virtualization option in the
> BIOS.
> I'm running xen 3.3.1 and kernel-xen-base 2.6.27.  When I boot up the
> "svm"
> flag doesn't appear in the /proc/cpuinfo flags area (full cpuinfo
shown
> at
> bottom).  In essence, it seems that the BIOS says virtualization is
> enabled,
> but Linux says it isn't.
> 
> I tried updating the BIOS to the latest version, but that didn't
change
> anything.
> Does anyone have any recommendations as to how to proceed?  Any advice
> would
> be greatly appreciated.
> 
>                                          -Brian M.
> 
> /proc/cpuinfo:
> 
> processor       : 0
> vendor_id       : AuthenticAMD
> cpu family      : 15
> model           : 67
> model name      : Dual-Core AMD Opteron(tm) Processor 1220
> stepping        : 3
> cpu MHz         : 2814.456
> cache size      : 1024 KB
> physical id     : 0
> siblings        : 1
> core id         : 0
> cpu cores       : 1
> apicid          : 0
> initial apicid  : 0
> fpu             : yes
> fpu_exception   : yes
> cpuid level     : 1
> wp              : yes
> flags           : fpu de tsc msr pae cx8 apic mtrr cmov pat clflush
mmx
> fxsr
> sse sse2 ht syscall nx mmxext fxsr_opt lm 3dnowext 3dnow rep_good nopl
> pni
> cx16 lahf_lm cmp_legacy extapic cr8_legacy
> bogomips        : 5630.48
> TLB size        : 1024 4K pages
> clflush size    : 64
> cache_alignment : 64
> address sizes   : 40 bits physical, 48 bits virtual
> power management: ts fid vid ttp tm stc
> 
> processor       : 1
> vendor_id       : AuthenticAMD
> cpu family      : 15
> model           : 67
> model name      : Dual-Core AMD Opteron(tm) Processor 1220
> stepping        : 3
> cpu MHz         : 2814.456
> cache size      : 1024 KB
> physical id     : 0
> siblings        : 1
> core id         : 1
> cpu cores       : 1
> apicid          : 1
> initial apicid  : 1
> fpu             : yes
> fpu_exception   : yes
> cpuid level     : 1
> wp              : yes
> flags           : fpu de tsc msr pae cx8 apic mtrr cmov pat clflush
mmx
> fxsr
> sse sse2 ht syscall nx mmxext fxsr_opt lm 3dnowext 3dnow rep_good nopl
> pni
> cx16 lahf_lm cmp_legacy extapic cr8_legacy
> bogomips        : 5630.48
> TLB size        : 1024 4K pages
> clflush size    : 64
> cache_alignment : 64
> address sizes   : 40 bits physical, 48 bits virtual
> power management: ts fid vid ttp tm stc
> 
> 
> _______________________________________________
> Xen-users mailing list
> Xen-users@xxxxxxxxxxxxxxxxxxx
> http://lists.xensource.com/xen-users


_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.