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

Re: [Xen-API] XCP 1.5: Missing IOMMU- support

Thanks James,

After a thorough check, I found that I had been mixing up the specs for i7-920 and i7-920-XM - the latter one supporting VT-d at a lower clock speed and a different socket. Seems that Intel does not support VT-d on i7 using the 3-channel socket 1366, and only one socket-2011: i7-3820.

Just a final issue: is the Intel TXT (Trusted Execution Technology) important for XEN HVM/PVM? Am planning to try out GPU-passthrough to Windows. That should conclude my speculations at this time.

Kind regards,


Þann 6.9.2012 16:38, skrifaði James Bulpin:
Johann wrote:
Bios menus:
I was wrong, the new bios (f13s) changed the menu item to
"Virtualization Technology" rather than the older Smart Virtualization.
The descriptive text is "When enabled, a VMM can utilize the additional
hardware capabilities provided by virtualization technology" (is this
text the only upgrade from f13 to f13s? ;-)
OK, that's VT-x (VMX) then.

I assume you can have VMX support without having full VT-d support? At
least, this is what I will assume in my new request to Gigabyte.
Correct. VMX has been around for longer than VT-d. VMX is necessary to run HVM 
guests on Xen, but it only handles trapping certain privileged instructions. 
VT-d adds support for DMA remapping and protection necessary for HVM guests to 
utilise passed-through PCI hardware devices. Without VT-d you can still run HVM 
guests with VMX but you won't be able to pass through PCI devices to them, 
you'll have to use paravirtualised I/O.


No virus found in this message.
Checked by AVG - www.avg.com
Version: 2012.0.2197 / Virus Database: 2437/5252 - Release Date: 09/06/12

Xen-api mailing list



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