[PATCH v3] virtio-pci: return IRQ_HANDLED after non-zero ISR

[PATCH v3] virtio-pci: return IRQ_HANDLED after non-zero ISR

[Andrew Stellman]

vp_interrupt() reads the ISR before dispatching config-change and
vring handling. Reading the ISR also clears it, so once the read
returns non-zero the interrupt was from this device and has already
been consumed.

Currently vp_interrupt() returns the result of vp_vring_interrupt().
For a config-change interrupt with no vring work, that can return
IRQ_NONE even though the ISR was non-zero and the interrupt was
handled.

Call vp_vring_interrupt() for any queue work, but once the ISR is
non-zero return IRQ_HANDLED.

Tested with QEMU virtio-blk-pci forced to INTx using vectors=0 and
pci=nomsi. On an idle device, 200 config-change interrupts were
generated using QMP block_resize.

Before this change, irq_handler_exit reported ret=unhandled and
/proc/irq/11/spurious increased from 0 to 200 unhandled interrupts.
After this change, irq_handler_exit reported ret=handled and the
unhandled count remained at 0.

The issue was found during an LLM-assisted Quality Playbook review.

Fixes: 77cf524654a8 ("virtio_pci: split up vp_interrupt")
Suggested-by: Michael S. Tsirkin <mst@redhat.com>
Assisted-by: LLM
Signed-off-by: Andrew Stellman <astellman@stellman-greene.com>
---
Changes from v2:
- Repost as a standalone patch in a new thread, per maintainer request.
- Add LLM-assistance disclosure.
- No code changes.

Changes from v1:
- Return IRQ_HANDLED for any non-zero ISR, as suggested by Michael.
- Add Fixes and Suggested-by tags.
- Test the change with virtio-blk forced to legacy INTx under QEMU.

Full red/green test logs and the tested v2 patch:
https://github.com/andrewstellman/quality-playbook/tree/11ba61d/evidence/virtio-pci-intx

 drivers/virtio/virtio_pci_common.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/virtio/virtio_pci_common.c b/drivers/virtio/virtio_pci_common.c
index 10371ecbc054..b90c174450b2 100644
--- a/drivers/virtio/virtio_pci_common.c
+++ b/drivers/virtio/virtio_pci_common.c
@@ -120,7 +120,9 @@ static irqreturn_t vp_interrupt(int irq, void *opaque)
 	if (isr & VIRTIO_PCI_ISR_CONFIG)
 		vp_config_changed(irq, opaque);
 
-	return vp_vring_interrupt(irq, opaque);
+	vp_vring_interrupt(irq, opaque);
+
+	return IRQ_HANDLED;
 }
 
 static int vp_request_msix_vectors(struct virtio_device *vdev, int nvectors,
base-commit: bc35965f6940a9bf834d54187b6088b8eb09206d

-- 
2.43.0

Re: [PATCH v3] virtio-pci: return IRQ_HANDLED after non-zero ISR

[Michael S. Tsirkin]

On Fri, Sep 04, 2026 at 10:13:18AM -0400, Andrew Stellman wrote:
> vp_interrupt() reads the ISR before dispatching config-change and
> vring handling. Reading the ISR also clears it, so once the read
> returns non-zero the interrupt was from this device and has already
> been consumed.
> 
> Currently vp_interrupt() returns the result of vp_vring_interrupt().
> For a config-change interrupt with no vring work, that can return
> IRQ_NONE even though the ISR was non-zero and the interrupt was
> handled.
> 
> Call vp_vring_interrupt() for any queue work, but once the ISR is
> non-zero return IRQ_HANDLED.
> 
> Tested with QEMU virtio-blk-pci forced to INTx using vectors=0 and
> pci=nomsi. On an idle device, 200 config-change interrupts were
> generated using QMP block_resize.
> 
> Before this change, irq_handler_exit reported ret=unhandled and
> /proc/irq/11/spurious increased from 0 to 200 unhandled interrupts.
> After this change, irq_handler_exit reported ret=handled and the
> unhandled count remained at 0.
> 
> The issue was found during an LLM-assisted Quality Playbook review.
> 
> Fixes: 77cf524654a8 ("virtio_pci: split up vp_interrupt")

this is still the wrong commit to blame.

> Suggested-by: Michael S. Tsirkin <mst@redhat.com>
> Assisted-by: LLM

This is not the correct format. See Documentation.

> Signed-off-by: Andrew Stellman <astellman@stellman-greene.com>
> ---
> Changes from v2:
> - Repost as a standalone patch in a new thread, per maintainer request.
> - Add LLM-assistance disclosure.
> - No code changes.
> 
> Changes from v1:
> - Return IRQ_HANDLED for any non-zero ISR, as suggested by Michael.
> - Add Fixes and Suggested-by tags.
> - Test the change with virtio-blk forced to legacy INTx under QEMU.
> 
> Full red/green test logs and the tested v2 patch:
> https://github.com/andrewstellman/quality-playbook/tree/11ba61d/evidence/virtio-pci-intx
> 
>  drivers/virtio/virtio_pci_common.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/virtio/virtio_pci_common.c b/drivers/virtio/virtio_pci_common.c
> index 10371ecbc054..b90c174450b2 100644
> --- a/drivers/virtio/virtio_pci_common.c
> +++ b/drivers/virtio/virtio_pci_common.c
> @@ -120,7 +120,9 @@ static irqreturn_t vp_interrupt(int irq, void *opaque)
>  	if (isr & VIRTIO_PCI_ISR_CONFIG)
>  		vp_config_changed(irq, opaque);
>  
> -	return vp_vring_interrupt(irq, opaque);
> +	vp_vring_interrupt(irq, opaque);
> +
> +	return IRQ_HANDLED;
>  }
>  
>  static int vp_request_msix_vectors(struct virtio_device *vdev, int nvectors,
> base-commit: bc35965f6940a9bf834d54187b6088b8eb09206d


I already did exactly this in my tree, no further action is necessary,
but I will add your tested by tag.

Thanks!

> -- 
> 2.43.0