Re: [PATCH v4 10/17] KVM: arm64: Add a shrinker for pKVM

On Mon, 24 Aug 2026 at 17:38, Vincent Donnefort <[email protected]> wrote:
>
> On Tue, Aug 18, 2026 at 04:28:38PM +0100, Fuad Tabba wrote:
> > Hi Vincent,
> >
> > On Fri, 31 Jul 2026 at 15:36, 'Vincent Donnefort' via kernel-team
> > <[email protected]> wrote:
> > >
> > > Integrate the pKVM memory reclaim interface with the host's memory
> > > management subsystem.
> > >
> > > This allows the host to automatically recover unused memory fom the
> > > hypervisor's heap allocator when the host is under memory pressure.
> > >
> > > Tested-by: Fuad Tabba <[email protected]>
> > > Signed-off-by: Vincent Donnefort <[email protected]>
> > >
> > > diff --git a/arch/arm64/kvm/pkvm.c b/arch/arm64/kvm/pkvm.c
> > > index d28422f5c3d6..bfbb1266491d 100644
> > > --- a/arch/arm64/kvm/pkvm.c
> > > +++ b/arch/arm64/kvm/pkvm.c
> > > @@ -115,7 +115,7 @@ static int pkvm_hyp_topup(enum pkvm_topup_id id, unsigned long nr_pages)
> > >         return ret;
> > >  }
> > >
> > > -static __maybe_unused unsigned long pkvm_hyp_reclaim(enum pkvm_topup_id id, unsigned long target)
> > > +static unsigned long pkvm_hyp_reclaim(enum pkvm_topup_id id, unsigned long target)
> >
> > This is the first caller of these, so it is where reclaim starts
> > running against a concurrent top-up.
> >
> > Nothing marks the pages a top-up just put in allocator->mc as spoken
> > for, and hyp_allocator_reclaim() ends with an unbounded drain of it,
> > so a shrink with target 1 hands back the lot. Land that between a
> > top-up and the retry it was for, and the retry asks again, and
> > pkvm_call_hyp_req() goes round.
>
> Sorry, I am not sure I follow here.
>
> IIRC, the shrinker will only reclaim half of what is available. So the pressure
> should be proportional to what is available and limit races with topup!

It's the ordering, not the amount.

topup and its retry are separate hypercalls, lock dropped between
them, so a shrink on another CPU can slip in, right? .
hyp_allocator_reclaim() drains allocator->mc, where the topup pages
sit, before any chunk, so half still comes out of them first: a target
of 1 fails the retry.

>
> However now looking at it. I wonder if I don't want to ratelimit here the number
> of pages reclaimed in one go to limit the time spent at EL2. Especially we do
> all that with the allocator lock taken...

Ratelimiting would cap the time under the lock, but it wouldn't stop a
concurrent shrink from taking the topup pages, would it?

Cheers,
/fuad


>
> >
> > Both are driven by memory pressure, so they are busiest together.
> > Worth holding back what a pending request asked for?
> >
> > >  {
> > >         struct kvm_hyp_memcache mc;
> > >         struct arm_smccc_res res;
> > > @@ -133,7 +133,7 @@ static __maybe_unused unsigned long pkvm_hyp_reclaim(enum pkvm_topup_id id, unsi
> > >         return reclaimed;
> > >  }
> > >
> > > -static __maybe_unused unsigned long pkvm_hyp_reclaimable(enum pkvm_topup_id id)
> > > +static unsigned long pkvm_hyp_reclaimable(enum pkvm_topup_id id)
> > >  {
> > >         return kvm_call_hyp_nvhe(__pkvm_hyp_reclaimable, id);
> > >  }
> > > @@ -342,8 +342,19 @@ void __init pkvm_selftests(void)
> > >  #endif
> > >  }
> > >
> > > +static unsigned long pkvm_shrinker_count(struct shrinker *shrink, struct shrink_control *sc)
> > > +{
> > > +       return pkvm_hyp_reclaimable(PKVM_TOPUP_HYP_ALLOC) ?: SHRINK_EMPTY;
> > > +}
> > > +
> > > +static unsigned long pkvm_shrinker_scan(struct shrinker *shrink, struct shrink_control *sc)
> > > +{
> > > +       return pkvm_hyp_reclaim(PKVM_TOPUP_HYP_ALLOC, sc->nr_to_scan);
> > > +}
> >
> > Returning 0 rather than SHRINK_STOP when reclaim comes back empty
> > leaves do_shrink_slab() calling this until total_scan runs out, and
> > each call is an HVC. count_objects() covers the case where there is
> > nothing at all, but not the one where it goes stale between the two
> > calls.
> >
> > Cheers,
> > /fuad
>
> Ha right, that SHRINK_STOP seems indeed to be the convention for other users.
>
> --
> Vincent
> >
> > > +
> > >  static int __init finalize_pkvm(void)
> > >  {
> > > +       struct shrinker *pkvm_shrinker;
> > >         int ret;
> > >
> > >         if (!is_protected_kvm_enabled() || !is_kvm_arm_initialised())
> > > @@ -359,10 +370,21 @@ static int __init finalize_pkvm(void)
> > >         kmemleak_free_part_phys(hyp_mem_base, hyp_mem_size);
> > >
> > >         ret = pkvm_drop_host_privileges();
> > > -       if (ret)
> > > +       if (ret) {
> > >                 pr_err("Failed to finalize Hyp protection: %d\n", ret);
> > > +               return ret;
> > > +       }
> > >
> > > -       return ret;
> > > +       pkvm_shrinker = shrinker_alloc(0, "pkvm");
> > > +       if (pkvm_shrinker) {
> > > +               pkvm_shrinker->count_objects = pkvm_shrinker_count;
> > > +               pkvm_shrinker->scan_objects = pkvm_shrinker_scan;
> > > +               shrinker_register(pkvm_shrinker);
> > > +       } else {
> > > +               kvm_err("Failed to register shrinker for pKVM\n");
> > > +       }
> > > +
> > > +       return 0;
> > >  }
> > >  device_initcall_sync(finalize_pkvm);
> > >
> > > --
> > > 2.55.0.508.g3f0d502094-goog
> > >
> > > To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
> > >
> >
> > To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
> >