Sorry for simple question: What techonology is better for HVM-using: VT or AMD-V. Thank''s a lot if any can take answer. -- Best Regards, alex.faq8@gmail.com _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
VT currently because you get PCI Pass-through support. New DDR3 Opterons will offer PCI Passthrough later this year but I assume you are buying now. From: xen-users-bounces@lists.xensource.com [mailto:xen-users-bounces@lists.xensource.com] On Behalf Of Andrey Varsavsky Sent: 07 August 2009 15:33 To: xen-users@lists.xensource.com Subject: [Xen-users] AMD-V or VT Sorry for simple question: What techonology is better for HVM-using: VT or AMD-V. Thank''s a lot if any can take answer. -- Best Regards, alex.faq8@gmail.com The SAQ Group Registered Office: 18 Chapel Street, Petersfield, Hampshire GU32 3DZ SAQ is the trading name of SEMTEC Limited. Registered in England & Wales Company Number: 06481952 http://www.saqnet.co.uk AS29219 SAQ Group Delivers high quality, honestly priced communication and I.T. services to UK Business. Broadband : Domains : Email : Hosting : CoLo : Servers : Racks : Transit : Backups : Managed Networks : Remote Support. ISPA Member _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Correction, for clarity: VT and AMD-V give you hardware virtualization, I am not familiar with the various caveats each may or may not entail. VT-D is necessary for PCI passthrough in an Intel system, IOMMU is necessary for PCI passthrough in an AMD system. AMD may name IOMMU support something else when they officially start supporting it, or it may be their official name, I dont know. Intel systems that dont support VT can still be purchased, as well as Intel systems that support VT, but not VT-D (both the motherboard and the processor must support the given technology, this will probably be true of AMD systems as well, but I dont know). Dustin From: xen-users-bounces@lists.xensource.com [mailto:xen-users-bounces@lists.xensource.com] On Behalf Of Robert Dunkley Sent: Friday, August 07, 2009 10:45 To: Andrey Varsavsky; xen-users@lists.xensource.com Subject: RE: [Xen-users] AMD-V or VT VT currently because you get PCI Pass-through support. New DDR3 Opterons will offer PCI Passthrough later this year but I assume you are buying now. From: xen-users-bounces@lists.xensource.com [mailto:xen-users-bounces@lists.xensource.com] On Behalf Of Andrey Varsavsky Sent: 07 August 2009 15:33 To: xen-users@lists.xensource.com Subject: [Xen-users] AMD-V or VT Sorry for simple question: What techonology is better for HVM-using: VT or AMD-V. Thank''s a lot if any can take answer. -- Best Regards, alex.faq8@gmail.com The SAQ Group Registered Office: 18 Chapel Street, Petersfield, Hampshire GU32 3DZ SAQ is the trading name of SEMTEC Limited. Registered in England & Wales Company Number: 06481952 http://www.saqnet.co.uk AS29219 SAQ Group Delivers high quality, honestly priced communication and I.T. services to UK Business. Broadband : Domains : Email : Hosting : CoLo : Servers : Racks : Transit : Backups : Managed Networks : Remote Support. ISPA Member _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
I can''t tell you the answer to that (also, define
''better'', etc), but what I can tell is that it seemed to me
that all mid desktop AMDs supported its virt technology, whereas with intel, it
was clear that most mid-entry chips desktop-orientated chips didn''t.
From that I would say that Intel ''charge'' you for it by
forcing you to use certain chips, whereas AMD ''give it away'',
if you see what I mean.
________________________________
From: Andrey Varsavsky <alex.faq8@gmail.com>
To: "xen-users@lists.xensource.com"
<Xen-users@lists.xensource.com>
Sent: Friday, 7 August, 2009 15:33:26
Subject: [Xen-users] AMD-V or VT
Sorry for simple question:
What techonology is better for HVM-using: VT or AMD-V. Thank''s a lot if
any can take answer.
--
Best Regards,
alex.faq8@gmail.com
_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users
On Aug 7, 2009, at 11:10 AM, Ian Murray <murrayie@yahoo.co.uk> wrote:> I can''t tell you the answer to that (also, define ''better'', etc), > but what I can tell is that it seemed to me that all mid desktop > AMDs supported its virt technology, whereas with intel, it was clear > that most mid-entry chips desktop-orientated chips didn''t.More frequently it''s the mainboard that doesn''t support the virtualization extensions. It can be remedied with a bios update but many time you wait forever for it or the manufacturer never releases. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
> > VT currently because you get PCI Pass-through support. New DDR3 > Opterons > will offer PCI Passthrough later this year but I assume you are buying > now. >Does that mean that current and older Opterons *can''t* pass a PCI device through to a guest, or that they don''t do it as efficiently as on the Intel VT machines? I''ve been pulling my hair out trying to get a fully virtualized guest access to a PCI card on an Opteron 1220. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
If you search the archives, you will find that it is a bit complicated. I don''t know if I''m getting the details right, but I think basically you can pass PCI through to HVM guests without VT-D, but you need VT-D to pass through to PV guests. If this is accurate, it may also apply to AMD-V without IOMMU. I could be all out of whack, though, and it may be that VT-D doesn''t even work for PV without hacking, I don''t remember the details, and I haven''t done it. Dustin -----Original Message----- From: xen-users-bounces@lists.xensource.com [mailto:xen-users-bounces@lists.xensource.com] On Behalf Of Brian P. Martin Sent: Friday, August 07, 2009 13:32 To: xen-users@lists.xensource.com Subject: RE: [Xen-users] AMD-V or VT> > VT currently because you get PCI Pass-through support. New DDR3 > Opterons > will offer PCI Passthrough later this year but I assume you are buying > now. >Does that mean that current and older Opterons *can''t* pass a PCI device through to a guest, or that they don''t do it as efficiently as on the Intel VT machines? I''ve been pulling my hair out trying to get a fully virtualized guest access to a PCI card on an Opteron 1220. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Dustin Henning wrote:> If you search the archives, you will find that it is a bit > complicated. I don''t know if I''m getting the details right, but I think > basically you can pass PCI through to HVM guests without VT-D, but you need > VT-D to pass through to PV guests. If this is accurate, it may also apply > to AMD-V without IOMMU. I could be all out of whack, though, and it may be > that VT-D doesn''t even work for PV without hacking, I don''t remember the > details, and I haven''t done it. > DustinIt''s exactly the other way around, you''ll need VT-d (which is a chipset feature, NOT a cpu feature) to passthrough PCI(e) device to HVM guests. For PV guests VT-d is not needed (not even VT for that matter)... Best regards, Christian _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
> > > It''s exactly the other way around, you''ll need VT-d (which is a chipset > feature, NOT a cpu feature) to passthrough PCI(e) device to HVM guests. > For PV guests VT-d is not needed (not even VT for that matter)... >Just to make sure I''m clear on this ... you''re saying that unless I have VT-d, there is no way I''m going to give a fully virtualized guest direct access to a parallel port of any kind (either on the mobo or on a PCI card). Do I have that right? In a separate thread I''ve been trying to figure out why my fully virtualized guest can''t access a license dongle on a parallel port, and if this is correct it sounds like the answer is "You can''t." -Brian M. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Hi Brian, For what its worth see if there are "parallel network servers". These exist for USB and serial ports. cheers, Jan Brian P. Martin wrote:>>> It''s exactly the other way around, you''ll need VT-d (which is a chipset >>> >> feature, NOT a cpu feature) to passthrough PCI(e) device to HVM guests. >> For PV guests VT-d is not needed (not even VT for that matter)... >> >> > > Just to make sure I''m clear on this ... you''re saying that unless I have > VT-d, there is no way I''m going to give a fully virtualized guest direct > access to a parallel port of any kind (either on the mobo or on a PCI card). > Do I have that right? > > In a separate thread I''ve been trying to figure out why my fully virtualized > guest can''t access a license dongle on a parallel port, and if this is > correct it sounds like the answer is "You can''t." > > -Brian M. > > > > _______________________________________________ > Xen-users mailing list > Xen-users@lists.xensource.com > http://lists.xensource.com/xen-users > >_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Brian P. Martin wrote:> Just to make sure I''m clear on this ... you''re saying that unless I have > VT-d, there is no way I''m going to give a fully virtualized guest direct > access to a parallel port of any kind (either on the mobo or on a PCI card). > Do I have that right?Not quite, for a PCI card yes, you would definitively need VT-d, but there are some exceptions for special devices like serial ports and usb devices. I''m not aware of a special parallel passthrough option but I may have missed that. If there isn''t one have you thought about a usb to parallel adapter and passing through the usb device? Afaik James Harper is currently working on an extension of his GPLPV drivers that will cover PVUSB... Best regards, Christian _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Jan Bakuwel wrote:>For what its worth see if there are "parallel network servers". These >exist for USB and serial ports.Christian Tramnitz wrote:>If there isn''t one have you thought about a usb to parallel adapter >and passing through the usb device?In the specific application mentioned (a security dongle key for a software package), neither of these are likely to work. Dongles often do "non standard" things to interface and aren''t likely to work through some abstracted standard interface. -- Simon Hobson Visit http://www.magpiesnestpublishing.co.uk/ for books by acclaimed author Gladys Hobson. Novels - poetry - short stories - ideal as Christmas stocking fillers. Some available as e-books. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Hi Simon, There is really no way of doing "non standard" operations with a parallel port. The actual key most likely holds a crypto IC that is going to either generate some proprietary codes, or are decrypting/encrypting some values passed by by the software. You can take an USB parallel adapter and use it like any parallel port. Think of the newer (after year 2000) laptops - these don''t have a serial port. And post likely the software runs on these machines also. Have phun, Adrian Simon Hobson wrote:> Jan Bakuwel wrote: > >> For what its worth see if there are "parallel network servers". These >> exist for USB and serial ports. > > > Christian Tramnitz wrote: > >> If there isn''t one have you thought about a usb to parallel adapter >> and passing through the usb device? > > > In the specific application mentioned (a security dongle key for a > software package), neither of these are likely to work. Dongles often > do "non standard" things to interface and aren''t likely to work > through some abstracted standard interface. >-- Deac Mihai-Adrian W: www.mikesoftware.com P: +40-745-256.364 _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Current AMD chips support nested paging, which will improve memory performance for virtualized guests. I believe Xen has had support for nested paging for some time. Intel''s next generation chips support this, but I think the only chips available with this feature right now is their Core i7 line. Obviously, check my facts before making a purchase, but I no one else pointed this out so I thought I''d say something. From: xen-users-bounces@lists.xensource.com [mailto:xen-users-bounces@lists.xensource.com] On Behalf Of Andrey Varsavsky Sent: Friday, August 07, 2009 10:33 AM To: xen-users@lists.xensource.com Subject: [Xen-users] AMD-V or VT Sorry for simple question: What techonology is better for HVM-using: VT or AMD-V. Thank''s a lot if any can take answer. -- Best Regards, alex.faq8@gmail.com _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Adrian, Are you sure on this? I have definitely seen a lot of things that don''t work on a USB serial port, and I have heard about a lot of things that don''t work on a USB parallel adapter. I can think of no good solid technical reason other than bad engineering, cheap manufacturing, or poor driver programming, but I would say there are definitely USB serial and parallel ports that will not work with some devices. Regarding usb parallel ports, I also know that in shopping for them, many of them specifically say they won''t support some standard functions or functionalities, or they say they were designed for printers and only printers. That isn''t to say that there aren''t USB serial and parallel ports that would function fully, just that there certainly are those that won''t. I suppose it is possible that my experiences with serial ports and those I have read about with parallel ports are the fault of Windows, but it seems unlikely. Dustin -----Original Message----- From: xen-users-bounces@lists.xensource.com [mailto:xen-users-bounces@lists.xensource.com] On Behalf Of Ady Deac Sent: Sunday, August 09, 2009 15:47 To: xen-users@lists.xensource.com Subject: Re: [Xen-users] Re: AMD-V or VT Hi Simon, There is really no way of doing "non standard" operations with a parallel port. The actual key most likely holds a crypto IC that is going to either generate some proprietary codes, or are decrypting/encrypting some values passed by by the software. You can take an USB parallel adapter and use it like any parallel port. Think of the newer (after year 2000) laptops - these don''t have a serial port. And post likely the software runs on these machines also. Have phun, Adrian Simon Hobson wrote:> Jan Bakuwel wrote: > >> For what its worth see if there are "parallel network servers". These >> exist for USB and serial ports. > > > Christian Tramnitz wrote: > >> If there isn''t one have you thought about a usb to parallel adapter >> and passing through the usb device? > > > In the specific application mentioned (a security dongle key for a > software package), neither of these are likely to work. Dongles often > do "non standard" things to interface and aren''t likely to work > through some abstracted standard interface. >-- Deac Mihai-Adrian W: www.mikesoftware.com P: +40-745-256.364 _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users