On Sat, Feb 10, 2018 at 01:12:28AM +0800, Chao Gao wrote:
> On Fri, Feb 09, 2018 at 04:27:54PM +0000, Roger Pau Monné wrote:
> >On Fri, Nov 17, 2017 at 02:22:14PM +0800, Chao Gao wrote:
> >> +    if ( !vvtd )
> >> +        return ENOMEM;
> >> +
> >> +    vvtd_reset(vvtd);
> >> +    vvtd->base_addr = viommu->base_address;
> >
> >I think it would be good to have some check here, so that the vIOMMU
> >is not for example positioned on top of a RAM region. Ideally you
> >should check that the gfns [base_address, base_address + size) are
> >unpopulated.
> 
> Yes. Except some checks here, this page should be reserved in guest e820,
> which implies some work in qemu or tool stack.

Right... I guess since the toolstack is the one that actually
populates memory &c it should be the one to position the vIOMMU, so
just leave this as-is for the time being, let's see other's people
opinion.

Out of curiosity, is the IOMMU on real hardware always at the same
memory address?

Thanks, Roger.

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

Reply via email to