Re: [PATCH 1/3] irqchip/gic-v4: Clear the domain and fwnode pointers after freeing them

Hi Marc,

On Sun, 23 Aug 2026 at 10:44, Marc Zyngier <[email protected]> wrote:
...
> > The actual patch would factor the shared teardown tail into a helper
> > rather than duplicate it.
> >
> > If you (Marc) agree, I'll send this as a separate patch (after testing
> > it), or as part of a possible respin in a couple of days.
>
> Why isn't this as simple as:
>
> diff --git a/arch/arm64/kvm/vgic/vgic-v4.c b/arch/arm64/kvm/vgic/vgic-v4.c
> index ed236f083f0d7..57a16530801c0 100644
> --- a/arch/arm64/kvm/vgic/vgic-v4.c
> +++ b/arch/arm64/kvm/vgic/vgic-v4.c
> @@ -333,6 +333,8 @@ void vgic_v4_teardown(struct kvm *kvm)
>                 free_irq(irq, vcpu);
>         }
>
> +       /* Make sure we free all VM-wide, per-CPU resources */
> +       its_vm->nr_vpes = aromic_read(&kvm->online_vcpus);
>         its_free_vcpu_irqs(its_vm);
>         kfree(its_vm->vpes);
>         its_vm->nr_vpes = 0;

With the typo fixed, it is ;), but on top of patch 3 of this series.
Without that, a vgic_init() failing after vgic_v4_init() succeeded
leaves the vPE array behind while precreate still admits vCPUs, so
online_vcpus can outgrow the array. If you agree, I'll add it after
patch 3, noting the dependency in the commit message when I respin.

Separately, and I might be missing something: is there anything
stopping KVM_CREATE_VCPU from running concurrently with vgic_init()?
Coming from KVM_DEV_ARM_VGIC_CTRL_INIT it makes the same
created_vcpus/online_vcpus comparison as kvm_vgic_create(), but under
the config_lock alone, where kvm_vgic_create() holds kvm->lock and the
vCPU mutexes too. And precreate admits vCPUs until dist->initialized
is set at the end, so if one lands in the middle, vgic_v4_init() sizes
vpes[] from one read of online_vcpus and fills it from another. If
there isn't, I'll send a fix as a separate patch.

Cheers,
/fuad

> Thanks,
>
>         M.
>
> --
> Jazz isn't dead. It just smells funny.