Re: [PATCH v3 01/10] KVM: arm64: vgic: Store LPIs in an xarray

[Zenghui Yu]

On 2024/2/21 1:15, Oliver Upton wrote:
> Hi Zenghui,
>
> On Wed, Feb 21, 2024 at 12:30:24AM +0800, Zenghui Yu wrote:
> > On 2024/2/17 02:41, Oliver Upton wrote:
> > > Using a linked-list for LPIs is less than ideal as it of course requires
> > > iterative searches to find a particular entry. An xarray is a better
> > > data structure for this use case, as it provides faster searches and can
> > > still handle a potentially sparse range of INTID allocations.
> > >
> > > Start by storing LPIs in an xarray, punting usage of the xarray to a
> > > subsequent change.
> > >
> > > Signed-off-by: Oliver Upton <oliver.upton@xxxxxxxxx>
> >
> > [..]
> >
> > > diff --git a/arch/arm64/kvm/vgic/vgic.c b/arch/arm64/kvm/vgic/vgic.c
> > > index db2a95762b1b..c126014f8395 100644
> > > --- a/arch/arm64/kvm/vgic/vgic.c
> > > +++ b/arch/arm64/kvm/vgic/vgic.c
> > > @@ -131,6 +131,7 @@ void __vgic_put_lpi_locked(struct kvm *kvm, struct vgic_irq *irq)
> > >  		return;
> > >  	list_del(&irq->lpi_list);
> > > +	xa_erase(&dist->lpi_xa, irq->intid);
> >
> > We can get here *after* grabbing the vgic_cpu->ap_list_lock (e.g.,
> > vgic_flush_pending_lpis()/vgic_put_irq()).  And as according to vGIC's
> > "Locking order", we should disable interrupts before taking the xa_lock
> > in xa_erase() and we would otherwise see bad things like deadlock..
>
> Nice catch!
>
> Yeah, the general intention was to disable interrupts outside of the
> xa_lock, however:
>
> > It's not a problem before patch #10, where we drop the lpi_list_lock and
> > start taking the xa_lock with interrupts enabled.  Consider switching to
> > use xa_erase_irq() instead?
>
> I don't think this change is safe until #10, as the implied xa_unlock_irq()
> would re-enable interrupts before the lpi_list_lock is dropped. Or do I
> have wires crossed?

No, you're right.  My intention was to fix it in patch #10.  And as
you've both pointed out, using xa_erase_irq() can hardly be the correct
fix.  My mistake :-( .

Thanks,
Zenghui