Re: [Xen-devel] [RFC v01 1/3] arm: omap: introduce iommu module

[Stefano Stabellini <stefano.stabellini@xxxxxxxxxxxxx>]

On Fri, 24 Jan 2014, Andrii Tseglytskyi wrote:
> >> +    writel(*r, mmu->mem_map + ((u32)(info->gpa) - mmu->mem_start));
> >> +
> >> +     res = mmu_trap_write_access(dom, mmu, info);
> >> +     if (res)
> >> +             return res;
> >> +
> >> +    return 1;
> >> +}
> >
> > I wonder if we actually need to trap guest accesses in all cases: if we
> > leave the GPU/IPU in Dom0, mapped 1:1, then we don't need any traps.
> > Maybe we can find a way to detect that so that we can avoid trapping and
> > translating in that case.
> >
> 
> If we leave the GPU/IPU in Dom0 with 1:1 mapping we won't need any
> translation. But for now we need this in DomU, where 1:1 mapping will
> not be available.

Sure, that is fine.

What I meant to say is that in order to make this code more generic, it
would be a good idea to check whether the guest is dom0 or domU and only
trap MMU accesses if the guest is domU (because for dom0 is not
necessary).

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel