| |
|
 |
|
 |
|
|
|
| |
|
|
xen-devel
RE: [Xen-devel] swiotlb and Xen query?
> -----Original Message-----
> From: pradeep singh rautela [mailto:rautelap@xxxxxxxxx]
> Sent: 03 May 2007 10:25
> To: Petersson, Mats
> Cc: xen-devel
> Subject: Re: [Xen-devel] swiotlb and Xen query?
>
>
>
> On 5/3/07, Petersson, Mats <Mats.Petersson@xxxxxxx> wrote:
>
> [snip]
> >
> >
> > That was really very informative Mats.
> > Thanks you very much for the help.
> > So, i guess swiotlb comes into picture only if there is a DMA
> > transfer from a domU whose p2m mapping turns out to be above
> > the 4GB( assuming device to be 32 bit) of the machine memory.
> > Other wise just normal p2m mapping is enough to find the
> > actual machine address, which can be handed over to the
> > device for DMAing. Am i right?
>
> Yes, that's pretty much how it works. In theory, it
> could be that
> devices only work on a 31, 28 or some other number of
> bit-range, but
> yes, that's the principle.
>
>
> Ok.Got this correct.
> I am sure this can cause problems while migrating live
> domains across cpus.
> yes/no/thoughts?
Yes, you can't live-migrate (or migrate at all, I think) driver domains,
as there is no way to determine if a driver has outstanding DMA
requests. You need to shut it down and restart it elsewhere if you want
to move it from one machine to another. [Of course not across CPU's, but
if you actually move across physical machine boundary].
[I had a quick look, but I couldn't actually find any checks for this,
but I still believe it holds true that driver domains can't be
saved/restored/migrated].
--
Mats
>
> Thank you
> ~psr
>
>
>
> --
> Mats
> >
> > Thank you once again for the patience.
> >
> > Regards
> > ~psr
> >
> [snip]
>
>
>
>
>
>
>
> --
> ---
> pradeep singh rautela
>
> "Genius is 1% inspiration, and 99% perspiration" - not me :)
>
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel
|
|
|
| |
|
Home