On 23.04.2025 05:31, Chen, Jiqian wrote: > On 2025/4/23 00:01, Jan Beulich wrote: >> On 21.04.2025 08:18, Jiqian Chen wrote: >>> @@ -759,10 +759,11 @@ static int vpci_init_capability_list(struct pci_dev >>> *pdev) >>> PCI_CAP_ID_MSI, >>> PCI_CAP_ID_MSIX, >>> }; >>> + const unsigned int *caps = is_hwdom ? NULL : supported_caps; >>> + const unsigned int n = is_hwdom ? 0 : ARRAY_SIZE(supported_caps); >>> >>> next = pci_find_next_cap_ttl(pdev->sbdf, PCI_CAPABILITY_LIST, >>> - supported_caps, >>> - ARRAY_SIZE(supported_caps), &ttl); >>> + caps, n, &ttl); >> >> As per the v3 adjustment to patch 02, you can pass supported_caps here in >> all cases. Only n needs to be zero for the hwdom case. > Oh, right. I will change in next version.
And, at the risk of stating the obvious, a brief comment might be a good idea here. Jan
