Re: [PULL 40/78] pcie_sriov: register user created virtual function before realizing it

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

On Fri, Oct 30, 2026 at 09:28:50AM +0300, Michael Tokarev wrote:
> On 10/28/26 20:34, Daniel Paziyski wrote:
> > Hello Michael,
> > 
> > I think you have queued this patch for pulling on accident. There is a bug here,
> > which is yet to be solved, as you have pointed out in
> > 
> > https://lore.kernel.org/qemu-devel/20261028074508-mutt-send-email-mst@kernel.org/T/#mca344ac148a62107cdc298e97e6713b6657b5c55
> > 
> > Please, hold off this patch for a while, until I come up with a fix soon.
> Am I right this buggy version of the patch has just landed in the master branch,
> instead of the fixed v2 one?
> 
> Here's the difference between what's in current master and the v2:
> 
> diff --git a/hw/pci/pci.c b/hw/pci/pci.c
> index d4187293b1f..6242204a52c 100644
> --- a/hw/pci/pci.c
> +++ b/hw/pci/pci.c
> @@ -2310,10 +2310,8 @@ static void pci_qdev_realize(DeviceState *qdev, Error **errp)
>          pci_dev->cap_present |= QEMU_PCIE_CAP_CXL;
>      }
> 
> -    pci_dev = do_pci_register_device(pci_dev,
> -                                     object_get_typename(OBJECT(qdev)),
> -                                     pci_dev->devfn, errp);
> -    if (pci_dev == NULL) {
> +    if (!do_pci_register_device(pci_dev, object_get_typename(OBJECT(qdev)),
> +                                pci_dev->devfn, errp)) {
>          pci_unregister_acpi_index(pci_dev);
>          return;
>      }
> @@ -2332,6 +2330,11 @@ static void pci_qdev_realize(DeviceState *qdev, Error **errp)
>          }
>      }
> 
> +    if (!pcie_sriov_check_after_realize(pci_dev, errp)) {
> +        pci_qdev_unrealize(DEVICE(pci_dev));
> +        return;
> +    }
> +
>      /*
>       * A PCIe Downstream Port that do not have ARI Forwarding enabled must
>       * associate only Device 0 with the device attached to the bus


Yes sorry, pls post a fix as a patch.

> ~
> 
> Thanks,
> 
> /mjt