Red Hat Bugzilla – Bug 1509270
[downstream clone - 4.1.8] host_nic_vfs_config is not populated if parent pci device of nic is shared with another device
Last modified: 2018-04-30 04:21:09 EDT
+++ This bug is a downstream clone. The original bug is: +++ +++ bug 1506887 +++ ====================================================================== Description of problem: In case the parent PCI device of a SR-IOV NIC is shared with another device, like a scsi_host, the engine fails to fill the host_nic_vfs_config table. This prevents the user from using the NIC for SR-IOV as he is unable to configure the number of VFs from the Administration Portal. For example, these entries from HostdevListByCaps triggers the problem: device_name | parent_device_name | capability | total_vfs | net_iface_name ------------------------+--------------------+------------+-----------+---------------- pci_0000_22_00_0 | pci_0000_00_03_0 | pci | 7 | scsi_host20 | pci_0000_22_00_0 | scsi_host | | net_enp22s0f0 | pci_0000_22_00_0 | net | | enp22s0f0 Version-Release number of selected component (if applicable): rhevm-4.1.6.2-0.1.el7.noarch How reproducible: 100%, use the attached hooks. IMPORTANT: to reproduce the problem, the host needs to be installed with the hook already in place. If the host_nic_vfs_config table is already populated with the unmodified fakesriov hook, the problem does not reproduce. So please remove and re-install the host to reproduce (and QA). Actual results: No SR-IOV icon on the NIC. Expected results: SR-IOV icon on the NIC. (Originally by Germano Veit Michel)
*** Bug 1505873 has been marked as a duplicate of this bug. *** (Originally by Germano Veit Michel)
Created attachment 1344155 [details] VDSM hooks to reproduce the problem after_get_caps: 10_fakesriov # place this hook to reproduce the problem after_hostdev_list_by_caps: 10_fakesriov # place this hook to reproduce the problem 20_filter_sriov # hook to correct the problem on vdsm side (Originally by Germano Veit Michel)
Germano, would the customer agree to take a hotfix for this bug and help us verify it? We currently have no access to the "offending" hardware.
Tested and verified on - 4.1.8.2-0.1.el7 and vdsm-4.19.41-1.el7ev Using Germano's vdsm fakesriov hooks only. As this bug blocking the errata process before the release we decided to not wait for the customer's reply on the hot-fix and we still don't have any access to the 'offending' HW.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHSA-2017:3427