Re: [PATCH v15 0/7] MTE support for KVM guest

[Marc Zyngier <maz@kernel.org>]

On Thu, 17 Jun 2021 14:24:25 +0100,
Steven Price <steven.price@arm.com> wrote:
> 
> On 17/06/2021 14:15, Marc Zyngier wrote:
> > On Thu, 17 Jun 2021 13:13:22 +0100,
> > Catalin Marinas <catalin.marinas@arm.com> wrote:
> >>
> >> On Mon, Jun 14, 2021 at 10:05:18AM +0100, Steven Price wrote:
> >>> I realise there are still open questions[1] around the performance of
> >>> this series (the 'big lock', tag_sync_lock, introduced in the first
> >>> patch). But there should be no impact on non-MTE workloads and until we
> >>> get real MTE-enabled hardware it's hard to know whether there is a need
> >>> for something more sophisticated or not. Peter Collingbourne's patch[3]
> >>> to clear the tags at page allocation time should hide more of the impact
> >>> for non-VM cases. So the remaining concern is around VM startup which
> >>> could be effectively serialised through the lock.
> >> [...]
> >>> [1]: https://lore.kernel.org/r/874ke7z3ng.wl-maz%40kernel.org
> >>
> >> Start-up, VM resume, migration could be affected by this lock, basically
> >> any time you fault a page into the guest. As you said, for now it should
> >> be fine as long as the hardware doesn't support MTE or qemu doesn't
> >> enable MTE in guests. But the problem won't go away.
> > 
> > Indeed. And I find it odd to say "it's not a problem, we don't have
> > any HW available". By this token, why should we merge this work the
> > first place, or any of the MTE work that has gone into the kernel over
> > the past years?
> > 
> >> We have a partial solution with an array of locks to mitigate against
> >> this but there's still the question of whether we should actually bother
> >> for something that's unlikely to happen in practice: MAP_SHARED memory
> >> in guests (ignoring the stage 1 case for now).
> >>
> >> If MAP_SHARED in guests is not a realistic use-case, we have the vma in
> >> user_mem_abort() and if the VM_SHARED flag is set together with MTE
> >> enabled for guests, we can reject the mapping.
> > 
> > That's a reasonable approach. I wonder whether we could do that right
> > at the point where the memslot is associated with the VM, like this:
> > 
> > diff --git a/arch/arm64/kvm/mmu.c b/arch/arm64/kvm/mmu.c
> > index a36a2e3082d8..ebd3b3224386 100644
> > --- a/arch/arm64/kvm/mmu.c
> > +++ b/arch/arm64/kvm/mmu.c
> > @@ -1376,6 +1376,9 @@ int kvm_arch_prepare_memory_region(struct kvm *kvm,
> >  		if (!vma)
> >  			break;
> >  
> > +		if (kvm_has_mte(kvm) && vma->vm_flags & VM_SHARED)
> > +			return -EINVAL;
> > +
> >  		/*
> >  		 * Take the intersection of this VMA with the memory region
> >  		 */
> > 
> > which takes the problem out of the fault path altogether? We document
> > the restriction and move on. With that, we can use a non-locking
> > version of mte_sync_page_tags().
> 
> Does this deal with the case where the VMAs are changed after the
> memslot is created? While we can do the check here to give the VMM a
> heads-up if it gets it wrong, I think we also need it in
> user_mem_abort() to deal with a VMM which mmap()s over the VA of the
> memslot. Or am I missing something?

No, you're right. I wish the memslot API wasn't so lax... Anyway, even
a VMA flag check in user_mem_abort() will be cheaper than this new BKL.

> But if everyone is happy with the restriction (just for KVM) of not
> allowing MTE+VM_SHARED then that sounds like a good way forward.

Definitely works for me.

	M.

-- 
Without deviation from the norm, progress is not possible.