Re: [PATCH v3 2/2] KVM: arm64: vgic: Mitigate potential LPI registration failure

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] Fallback XArray node allocation bypasses memcg accounting
--

commit bdc0b53cd314d49c8e7e70017bf5fb300a832c2b
Author: Carlos López <clopez@suse.de>
KVM: arm64: vgic: Mitigate potential LPI registration failure

This commit mitigates a race condition in vgic_add_lpi() where an LPI might
be discarded and re-mapped concurrently. It allows __xa_store() to allocate
a new node under the lock if the previously reserved entry was freed.

> diff --git a/arch/arm64/kvm/vgic/vgic-its.c b/arch/arm64/kvm/vgic/vgic-its.c
> index 577286069368d..ace3e59fff971 100644
> --- a/arch/arm64/kvm/vgic/vgic-its.c
> +++ b/arch/arm64/kvm/vgic/vgic-its.c
[ ... ]
> @@ -121,7 +121,7 @@ static struct vgic_irq *vgic_add_lpi(struct kvm *kvm, u32 intid,
>  		 * from the deferred release path, pending cleanup by
>  		 * vgic_release_deleted_lpis(). Evict and free it if present.
>  		 */
> -		oldirq = __xa_store(&dist->lpi_xa, intid, irq, 0);
> +		oldirq = __xa_store(&dist->lpi_xa, intid, irq, GFP_NOWAIT);

[Severity: Low]
Does this __xa_store() fallback allocation bypass memcg accounting?

The initial allocation in vgic_add_lpi() via xa_reserve_irq() correctly uses
GFP_KERNEL_ACCOUNT. However, the GFP_NOWAIT flag here does not include
__GFP_ACCOUNT.

If a guest races a DISCARD and MAPTI command such that the LPI is released
after xa_reserve_irq() but before xa_lock_irqsave(), this __xa_store()
will allocate a new node. Should this use GFP_NOWAIT | __GFP_ACCOUNT to
ensure this allocation is still charged to the guest's memory cgroup?

>  		ret = xa_err(oldirq);
>  		if (!ret && oldirq)
>  			kfree_rcu(oldirq, rcu);

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260703110130.2609872-2-clopez@suse.de?part=2