Re: [Qemu-devel] [RESEND PATCH 2/6] memory: introduce AddressSpaceOps and IOMMUObject

[Liu, Yi L]

On Wed, Dec 20, 2017 at 10:01:10PM +1100, David Gibson wrote:
> On Wed, Dec 20, 2017 at 02:32:42PM +0800, Liu, Yi L wrote:
> > On Mon, Dec 18, 2017 at 10:22:18PM +1100, David Gibson wrote:
> > > On Mon, Dec 18, 2017 at 05:17:35PM +0800, Liu, Yi L wrote:
> > > > On Mon, Dec 18, 2017 at 05:14:42PM +1100, David Gibson wrote:
> > > > > On Thu, Nov 16, 2017 at 04:57:09PM +0800, Liu, Yi L wrote:
[snip]
> > > Partly.  Each PE has an address space which all devices in the PE see.
> > > Only some of that address space is mapped to system memory though,
> > > other parts are occupied by devices, others are unmapped.
> > > 
> > > Only the parts mapped by the IOMMU vary between PEs - the other parts
> > > of the address space will be identical for all PEs on the host
> > 
> > Thx, this comment addressed me well. This is different from what we have
> > on VT-d.
> 
> Really?  That's hard to believe.  I'm pretty sure the VT-d IOMMU must
> have a range < 2^64, and anything on the bus outside that range I
> expect would be common between all domains.  In particular I'd expect
> the BARs for other devices not to be remapped by the IOMMU (though
> they may be inaccessible on PCI-E due peer to peer transactions being
> blocked).  As well as things above the IOMMU's range, I'd expect the
> region for 32-bit BARs to be common between all domains.

Sorry I misunderstood you. In each IOVA space, there is reserved range
, it is the BARs MMIO range. Such reservation is to avoid un-expected
Peer-To-Peer transaction. So regards to the IOVA space, all vendors should
be similar. So you are right~

Thanks,
Yi L