6.11-stable review patch. If anyone has any objections, please let me know. ------------------ From: Marc Zyngier <maz@xxxxxxxxxx> commit d561491ba927cb5634094ff311795e9d618e9b86 upstream. Make sure we filter out non-LPI invalidation when handling writes to GICR_INVLPIR. Fixes: 4645d11f4a553 ("KVM: arm64: vgic-v3: Implement MMIO-based LPI invalidation") Reported-by: Alexander Potapenko <glider@xxxxxxxxxx> Tested-by: Alexander Potapenko <glider@xxxxxxxxxx> Signed-off-by: Marc Zyngier <maz@xxxxxxxxxx> Cc: stable@xxxxxxxxxxxxxxx Link: https://lore.kernel.org/r/20241117165757.247686-2-maz@xxxxxxxxxx Signed-off-by: Oliver Upton <oliver.upton@xxxxxxxxx> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> --- arch/arm64/kvm/vgic/vgic-mmio-v3.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) --- a/arch/arm64/kvm/vgic/vgic-mmio-v3.c +++ b/arch/arm64/kvm/vgic/vgic-mmio-v3.c @@ -530,6 +530,7 @@ static void vgic_mmio_write_invlpi(struc unsigned long val) { struct vgic_irq *irq; + u32 intid; /* * If the guest wrote only to the upper 32bit part of the @@ -541,9 +542,13 @@ static void vgic_mmio_write_invlpi(struc if ((addr & 4) || !vgic_lpis_enabled(vcpu)) return; + intid = lower_32_bits(val); + if (intid < VGIC_MIN_LPI) + return; + vgic_set_rdist_busy(vcpu, true); - irq = vgic_get_irq(vcpu->kvm, NULL, lower_32_bits(val)); + irq = vgic_get_irq(vcpu->kvm, NULL, intid); if (irq) { vgic_its_inv_lpi(vcpu->kvm, irq); vgic_put_irq(vcpu->kvm, irq);