On Fri, Jun 04, 2021 at 01:51:38PM +0100, Steven Price wrote:
> On 04/06/2021 12:36, Catalin Marinas wrote:
> > On Fri, Jun 04, 2021 at 11:42:11AM +0100, Steven Price wrote:
> >> On 03/06/2021 17:00, Catalin Marinas wrote:
> >>> On Mon, May 24, 2021 at 11:45:09AM +0100, Steven Price wrote:
> >>>> @@ -971,8 +1007,13 @@ static int user_mem_abort(struct kvm_vcpu *vcpu, phys_addr_t fault_ipa,
> >>>> if (writable)
> >>>> prot |= KVM_PGTABLE_PROT_W;
> >>>>
> >>>> - if (fault_status != FSC_PERM && !device)
> >>>> + if (fault_status != FSC_PERM && !device) {
> >>>> + ret = sanitise_mte_tags(kvm, pfn, vma_pagesize);
> >>>> + if (ret)
> >>>> + goto out_unlock;
> >>>
> >>> Maybe it was discussed in a previous version, why do we need this in
> >>> addition to kvm_set_spte_gfn()?
> >>
> >> kvm_set_spte_gfn() is only used for the MMU notifier path (e.g. if a
> >> memslot is changed by the VMM). For the initial access we will normally
> >> fault the page into stage 2 with user_mem_abort().
> >
> > Right. Can we move the sanitise_mte_tags() call to
> > kvm_pgtable_stage2_map() instead or we don't have the all the
> > information needed?
>
> I tried that before: kvm_pgtable_stage2_map() is shared with the
> hypervisor so sadly we can't go poking around in the host as this breaks
> on nVHE. I mentioned it in the v12 cover letter but it was in a wall of
> text:
Ah, I missed this in the cover letter (haven't read it).
So, apart from the nitpick with the early return for less indentation,
feel free to add:
Reviewed-by: Catalin Marinas <catalin.marinas@xxxxxxx>
_______________________________________________
kvmarm mailing list
kvmarm@xxxxxxxxxxxxxxxxxxxxx
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
