Skip to contentSkip to conversation

Fixes for virtio PCI interrupt handling

PATCH Applied Keir Fraser 2 participants 5 messages Sep 12, 2023 – Sep 18, 2023 KVM

lore
Share & tools
lore
#0 lore raw

cover Fixes for virtio PCI interrupt handling

Please consider these simple fixes and cleanups to legacy interrupt handling for virtio PCI devices.

Keir Fraser (3):
virtio/pci: Level-trigger the legacy IRQ line in all cases virtio/pci: Treat PCI ISR as a set of bit flags virtio/pci: Use consistent naming for the PCI ISR bit flags

 virtio/pci-legacy.c | 2 +-
 virtio/pci.c        | 9 ++++++---
 2 files changed, 7 insertions(+), 4 deletions(-)
Signature · 1 line
2.42.0.283.g2d96d420d3-goog
#1 lore raw

PATCH 1/3 virtio/pci: Level-trigger the legacy IRQ line in all cases

The PCI legacy IRQ line is level triggered, but is treated as edge triggered via kvm__irq_trigger() for signalling of config changes.

Fix this by using kvm__irq_level(), as for queue signalling.

---
 virtio/pci.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
virtio/pci.c +1 -1
index 701f456..7a206be 100644
@@ -264,7 +264,7 @@ int virtio_pci__signal_config(struct kvm *kvm, struct virtio_device *vdev)
 			kvm__irq_trigger(kvm, vpci->config_gsi);
 	} else {
 		vpci->isr = VIRTIO_PCI_ISR_CONFIG;
-		kvm__irq_trigger(kvm, vpci->legacy_irq_line);
+		kvm__irq_line(kvm, vpci->legacy_irq_line, VIRTIO_IRQ_HIGH);
 	}
 
 	return 0;
Signature · 1 line
2.42.0.283.g2d96d420d3-goog
#2 lore raw

PATCH 2/3 virtio/pci: Treat PCI ISR as a set of bit flags

In reply to Keir Fraser

The PCI ISR is defined in the virtio spec as a set of flags which can be bitwise ORed together. Therefore we should avoid clearing previously-set flags.

---
 virtio/pci.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
virtio/pci.c +2 -2
index 7a206be..74bc9a4 100644
@@ -239,7 +239,7 @@ int virtio_pci__signal_vq(struct kvm *kvm, struct virtio_device *vdev, u32 vq)
 		else
 			kvm__irq_trigger(kvm, vpci->gsis[vq]);
 	} else {
-		vpci->isr = VIRTIO_IRQ_HIGH;
+		vpci->isr |= VIRTIO_IRQ_HIGH;
 		kvm__irq_line(kvm, vpci->legacy_irq_line, VIRTIO_IRQ_HIGH);
 	}
 	return 0;
@@ -263,7 +263,7 @@ int virtio_pci__signal_config(struct kvm *kvm, struct virtio_device *vdev)
 		else
 			kvm__irq_trigger(kvm, vpci->config_gsi);
 	} else {
-		vpci->isr = VIRTIO_PCI_ISR_CONFIG;
+		vpci->isr |= VIRTIO_PCI_ISR_CONFIG;
 		kvm__irq_line(kvm, vpci->legacy_irq_line, VIRTIO_IRQ_HIGH);
 	}
 
Signature · 1 line
2.42.0.283.g2d96d420d3-goog
#3 lore raw

PATCH 3/3 virtio/pci: Use consistent naming for the PCI ISR bit flags

In reply to Keir Fraser

Avoid using VIRTIO_IRQ_{HIGH,LOW} which belong to a different namespace. Instead define VIRTIO_PCI_ISR_QUEUE as a logical extension of the VIRTIO_PCI_ISR_* namespace. Since this bit flag is missing from a header imported verbatim from Linux, define it directly in pci.c.

---
 virtio/pci-legacy.c | 2 +-
 virtio/pci.c        | 5 ++++-
 2 files changed, 5 insertions(+), 2 deletions(-)
virtio/pci-legacy.c +1 -1
index 5804796..02a8f8c 100644
@@ -61,7 +61,7 @@ static bool virtio_pci__data_in(struct kvm_cpu *vcpu, struct virtio_device *vdev
 	case VIRTIO_PCI_ISR:
 		ioport__write8(data, vpci->isr);
 		kvm__irq_line(kvm, vpci->legacy_irq_line, VIRTIO_IRQ_LOW);
-		vpci->isr = VIRTIO_IRQ_LOW;
+		vpci->isr = 0;
 		break;
 	default:
 		ret = virtio_pci__specific_data_in(kvm, vdev, data, size, offset);
virtio/pci.c +4 -1
index 74bc9a4..8a34cec 100644
@@ -14,6 +14,9 @@
 #include <assert.h>
 #include <string.h>
 
+/* The bit of the ISR which indicates a queue change. */
+#define VIRTIO_PCI_ISR_QUEUE	0x1
+
 int virtio_pci__add_msix_route(struct virtio_pci *vpci, u32 vec)
 {
 	int gsi;
@@ -239,7 +242,7 @@ int virtio_pci__signal_vq(struct kvm *kvm, struct virtio_device *vdev, u32 vq)
 		else
 			kvm__irq_trigger(kvm, vpci->gsis[vq]);
 	} else {
-		vpci->isr |= VIRTIO_IRQ_HIGH;
+		vpci->isr |= VIRTIO_PCI_ISR_QUEUE;
 		kvm__irq_line(kvm, vpci->legacy_irq_line, VIRTIO_IRQ_HIGH);
 	}
 	return 0;
Signature · 1 line
2.42.0.283.g2d96d420d3-goog
Applied
#4 lore raw
In reply to Keir Fraser
Quoted 9 lines · Keir Fraser

Please consider these simple fixes and cleanups to legacy interrupt handling for virtio PCI devices.

Keir Fraser (3):
virtio/pci: Level-trigger the legacy IRQ line in all cases virtio/pci: Treat PCI ISR as a set of bit flags virtio/pci: Use consistent naming for the PCI ISR bit flags

[...]

Applied to kvmtool (master), thanks!

[1/3] virtio/pci: Level-trigger the legacy IRQ line in all cases

      https://git.kernel.org/will/kvmtool/c/353fa0d89e29
[2/3] virtio/pci: Treat PCI ISR as a set of bit flags
      https://git.kernel.org/will/kvmtool/c/292144845000
[3/3] virtio/pci: Use consistent naming for the PCI ISR bit flags
      https://git.kernel.org/will/kvmtool/c/d560235f4556

Cheers,

Signature · 5 lines
Will

https://fixes.arm64.dev
https://next.arm64.dev
https://will.arm64.dev