[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Xen-users] Gigabit Interface - XEN Full Virtualization
No, I mean YOU MUST do it on the domU not on the dom0.
The domU must load the unmodified compiled drivers from the xen fedora
distribution src rpm that yuo compiled by hand.
The dom0 is unaffected by this procedure.
So get the xen source rpm according to your domU distribution,
compile the unmodifided drivers on the domU and load the kernel module.
remember depmod -a to register the new installed modules
Riccardo
Emre Erenoglu ha scritto:
Hi Riccardo, I tried to understand how you do it. I'm also trying to
use my HVM domU with PV drivers. If I understand right, the important
point is:
- use the source code of the XEN in the Dom0 (don't download from
Xensource etc.). Is it true?
So, for example, if I'm using Fedora 8 as my Dom0, I have to download
the source rpm of the xen-hypervisor that's running in Fedora 8. then
go inside the unmodified_drivers and do the compilation. right?
I did like this for my system, unfortunately, either the system
crashes completely or network stops working after some time... I guess
there's a bug somewhere. For example, when unloading the xen-vnif
module, I receive an error in dmesg talking about some "memory leak".
Br,
Emre
On Dec 2, 2007 5:08 PM, Riccardo Veraldi
<Riccardo.Veraldi@xxxxxxxxxxxx <mailto:Riccardo.Veraldi@xxxxxxxxxxxx>>
wrote:
if you compile the netfront drivers on your guest kernel and load them
as kernel module on your guests, you will have gigabit speed on a
fully
virtualized
virtual machine, hust use netfront and not ioemu then load
the kernel drivers on the virtual machine.
I have written something about it but it is in italian.
If you understand italian...
https://calcolo.infn.it/wiki/doku.php?id=network_overbust_compilare_e_installare_il_kernel_module_con_il_supporto_netfront
you should understand the way to do it even if it is written in
italian.
I can translate in english if really it is needed
Riccardo
Artur Baruchi ha scritto:
> Hi Mark,
>
> Unfortunately, I did some tests, and even my nic being a giga and my
> lan supporting giga, my full virtualized domain wasnt able to pass
> 100Mb/sec. My paravirtualized guests had a very good performance in
> this environment.
>
> Thanks,
>
> Att.
> Artur Baruchi
>
>
> On Dec 2, 2007 12:34 AM, Mark Williamson <
mark.williamson@xxxxxxxxxxxx
<mailto:mark.williamson@xxxxxxxxxxxx>> wrote:
>
>>>> I noted that the lspci output are different, and I would like
to know,
>>>> how can I make my xen virtual machine see the same Hardware,
so I can
>>>> force a 1000 Full speed.
>>>>
>>> You can't get Gigabit speed for your NIC in HVM. Whatever
network cards
>>> you have on the physical system, it's emulated as a 10/100
mbps Realtek
>>> card to the HVM.
>>>
>>> Use paravirtualized drivers for HVMs to enhance the network card
>>> performance and throughput, but don't expect 1000 mbps. :-)
>>>
>> The virtual machine won't have hardware matching the physical
system; it just
>> sees whatever hardware Xen is emulating.
>>
>> Xen doesn't emulate a GigE network card so you won't be able to
see one in the
>> guest. However, I suspect that the 10/100mbps card will be
able to transmit
>> above that speed if your machine can emulate the card fast
enough: because
>> there's not a real hardware 100mbps ethernet card there, there
actually
>> shouldn't be anything limiting the speed to that.
>>
>> As Sadique mentioned, if you use paravirtualised drivers in
your guest (either
>> by using a PV guest or by installing paravirt network drivers
in an HVM
>> guest) you could potentially increase the throughput and
decrease the CPU
>> overhead further.
>>
>> Cheers,
>> Mark
>>
>> --
>> Dave: Just a question. What use is a unicyle with no seat? And
no pedals!
>> Mark: To answer a question with a question: What use is a
skateboard?
>> Dave: Skateboards have wheels.
>> Mark: My wheel has a wheel!
>>
>>
>
> _______________________________________________
> Xen-users mailing list
> Xen-users@xxxxxxxxxxxxxxxxxxx <mailto:Xen-users@xxxxxxxxxxxxxxxxxxx>
> http://lists.xensource.com/xen-users
<http://lists.xensource.com/xen-users>
>
_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx <mailto:Xen-users@xxxxxxxxxxxxxxxxxxx>
http://lists.xensource.com/xen-users
--
Emre Erenoglu
erenoglu@xxxxxxxxx <mailto:erenoglu@xxxxxxxxx>
_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users
|
