Re: [PATCH 3/4] virtio-pci: return IRQ_HANDLED for config-change interrupts

["Michael S. Tsirkin" <mst@redhat.com>]

On Tue, Apr 07, 2026 at 08:39:03AM -0400, Andrew Stellman wrote:
> vp_interrupt() unconditionally returns the result of
> vp_vring_interrupt(). When a config-change interrupt fires but no vring
> activity is pending, vp_vring_interrupt() returns IRQ_NONE — even
> though the interrupt was legitimately handled by vp_config_changed().
> Over time this causes the IRQ subsystem to flag the line as spurious.
> 
> Track the return value explicitly: set ret to IRQ_HANDLED when the
> config-change bit is set, OR it with the vring result, and return the
> combined value.
> 
> Signed-off-by: Andrew Stellman <astellman@stellman-greene.com>

Fixes tag?

> ---
>  drivers/virtio/virtio_pci_common.c | 10 ++++++++--
>  1 file changed, 8 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/virtio/virtio_pci_common.c b/drivers/virtio/virtio_pci_common.c
> index 0b9d66b..1d1ab02 100644
> --- a/drivers/virtio/virtio_pci_common.c
> +++ b/drivers/virtio/virtio_pci_common.c
> @@ -106,6 +106,7 @@ static irqreturn_t vp_vring_interrupt(int irq, void *opaque)
>  static irqreturn_t vp_interrupt(int irq, void *opaque)
>  {
>  	struct virtio_pci_device *vp_dev = opaque;
> +	irqreturn_t ret = IRQ_NONE;
>  	u8 isr;
>  
>  	/* reading the ISR has the effect of also clearing it so it's very
> @@ -117,10 +118,15 @@ static irqreturn_t vp_interrupt(int irq, void *opaque)
>  		return IRQ_NONE;
>  
>  	/* Configuration change?  Tell driver if it wants to know. */
> -	if (isr & VIRTIO_PCI_ISR_CONFIG)
> +	if (isr & VIRTIO_PCI_ISR_CONFIG) {
>  		vp_config_changed(irq, opaque);
> +		ret = IRQ_HANDLED;
> +	}
>  
> -	return vp_vring_interrupt(irq, opaque);
> +	if (vp_vring_interrupt(irq, opaque) == IRQ_HANDLED)
> +		ret = IRQ_HANDLED;
> +
> +	return ret;
>  }
>  
>  static int vp_request_msix_vectors(struct virtio_device *vdev, int nvectors,
> -- 
> 2.34.1