| Message ID | 20250311230623.1264283-1-alex.williamson@redhat.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | vfio/pci: Handle INTx IRQ_NOTCONNECTED | expand |
> From: Alex Williamson <alex.williamson@redhat.com> > Sent: Wednesday, March 12, 2025 7:06 AM > > Some systems report INTx as not routed by setting pdev->irq to > IRQ_NOTCONNECTED, resulting in a -ENOTCONN error when trying to > setup eventfd signaling. Include this in the set of conditions > for which the PIN register is virtualized to zero. > > Additionally consolidate vfio_pci_get_irq_count() to use this > virtualized value in reporting INTx support via ioctl and sanity > checking ioctl paths since pdev->irq is re-used when the device > is in MSI mode. > > The combination of these results in both the config space of the > device and the ioctl interface behaving as if the device does not > support INTx. > > Signed-off-by: Alex Williamson <alex.williamson@redhat.com> Reviewed-by: Kevin Tian <kevin.tian@intel.com>
diff --git a/drivers/vfio/pci/vfio_pci_config.c b/drivers/vfio/pci/vfio_pci_config.c index 94142581c98c..14437396d721 100644 --- a/drivers/vfio/pci/vfio_pci_config.c +++ b/drivers/vfio/pci/vfio_pci_config.c @@ -1814,7 +1814,8 @@ int vfio_config_init(struct vfio_pci_core_device *vdev) cpu_to_le16(PCI_COMMAND_MEMORY); } - if (!IS_ENABLED(CONFIG_VFIO_PCI_INTX) || vdev->nointx) + if (!IS_ENABLED(CONFIG_VFIO_PCI_INTX) || vdev->nointx || + vdev->pdev->irq == IRQ_NOTCONNECTED) vconfig[PCI_INTERRUPT_PIN] = 0; ret = vfio_cap_init(vdev); diff --git a/drivers/vfio/pci/vfio_pci_core.c b/drivers/vfio/pci/vfio_pci_core.c index 586e49efb81b..c857630f447b 100644 --- a/drivers/vfio/pci/vfio_pci_core.c +++ b/drivers/vfio/pci/vfio_pci_core.c @@ -727,15 +727,7 @@ EXPORT_SYMBOL_GPL(vfio_pci_core_finish_enable); static int vfio_pci_get_irq_count(struct vfio_pci_core_device *vdev, int irq_type) { if (irq_type == VFIO_PCI_INTX_IRQ_INDEX) { - u8 pin; - - if (!IS_ENABLED(CONFIG_VFIO_PCI_INTX) || - vdev->nointx || vdev->pdev->is_virtfn) - return 0; - - pci_read_config_byte(vdev->pdev, PCI_INTERRUPT_PIN, &pin); - - return pin ? 1 : 0; + return vdev->vconfig[PCI_INTERRUPT_PIN] ? 1 : 0; } else if (irq_type == VFIO_PCI_MSI_IRQ_INDEX) { u8 pos; u16 flags; diff --git a/drivers/vfio/pci/vfio_pci_intrs.c b/drivers/vfio/pci/vfio_pci_intrs.c index 8382c5834335..565966351dfa 100644 --- a/drivers/vfio/pci/vfio_pci_intrs.c +++ b/drivers/vfio/pci/vfio_pci_intrs.c @@ -259,7 +259,7 @@ static int vfio_intx_enable(struct vfio_pci_core_device *vdev, if (!is_irq_none(vdev)) return -EINVAL; - if (!pdev->irq) + if (!pdev->irq || pdev->irq == IRQ_NOTCONNECTED) return -ENODEV; name = kasprintf(GFP_KERNEL_ACCOUNT, "vfio-intx(%s)", pci_name(pdev));
Some systems report INTx as not routed by setting pdev->irq to IRQ_NOTCONNECTED, resulting in a -ENOTCONN error when trying to setup eventfd signaling. Include this in the set of conditions for which the PIN register is virtualized to zero. Additionally consolidate vfio_pci_get_irq_count() to use this virtualized value in reporting INTx support via ioctl and sanity checking ioctl paths since pdev->irq is re-used when the device is in MSI mode. The combination of these results in both the config space of the device and the ioctl interface behaving as if the device does not support INTx. Signed-off-by: Alex Williamson <alex.williamson@redhat.com> --- Testing is welcome on systems that actually implement this form of reporting an unrouted INTx, I don't have access to such as system and have only tested mimicking the behavior with a quirk to write the LINE register to 0xff and hack to mask the GSI info for a device. drivers/vfio/pci/vfio_pci_config.c | 3 ++- drivers/vfio/pci/vfio_pci_core.c | 10 +--------- drivers/vfio/pci/vfio_pci_intrs.c | 2 +- 3 files changed, 4 insertions(+), 11 deletions(-)