Bug 2361921
| Summary: | Network Regression for Fedora 41 and 42 Cloud images on libvirt | ||
|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Scott Williams <vwfoxguru> |
| Component: | kernel | Assignee: | Kernel Maintainer List <kernel-maint> |
| Status: | NEW --- | QA Contact: | Fedora Extras Quality Assurance <extras-qa> |
| Severity: | high | Docs Contact: | |
| Priority: | unspecified | ||
| Version: | 42 | CC: | acaringi, adscvr, airlied, bskeggs, hdegoede, hpa, josef, kernel-maint, linville, masami256, mchehab, ptalbert, steved, suraj.ghimire7 |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | x86_64 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | --- | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | Type: | --- | |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
|
Description
Scott Williams
2025-04-23 19:23:21 UTC
Libvirt hosts are SUSE Harvester v1.4.1 (released January 2025) running SLE Micro kernel 5.14.21-150500.55.88-default on kubevirt v1.2.2. Test it in a newer hypervisor environment: Harvester v1.4.2 (released March 11, 2025) that is based on kuvebirt v1.3.1, so I suspect that somewhere between at least v.1.2.2 and v.1.3.1 of kubevirt, there's something that is not playing well with 6.14. I also reproduced this on OpenSUSE Tumbleweed (6.14.2) and Ubuntu 25.04 (6.14.0), so it certainly seems like an upstream kernel issue. Also filed a bug with OpenSUSE Tumbleweed: https://bugzilla.opensuse.org/show_bug.cgi?id=1241662 Beyond the different kubevirt versions, the physical NICs are also different: Affected hypervisor NICs (the NetXtreme-E's are the relevant ones here): ``` 21:00.0 Ethernet controller: Broadcom Inc. and subsidiaries BCM57508 NetXtreme-E 10Gb/25Gb/40Gb/50Gb/100Gb/200Gb Ethernet (rev 12) 21:00.1 Ethernet controller: Broadcom Inc. and subsidiaries BCM57508 NetXtreme-E 10Gb/25Gb/40Gb/50Gb/100Gb/200Gb Ethernet (rev 12) 63:00.0 Ethernet controller: Intel Corporation I350 Gigabit Network Connection (rev 01) 63:00.1 Ethernet controller: Intel Corporation I350 Gigabit Network Connection (rev 01) ``` Unaffected hypervisor NICs: ``` 01:00.0 Ethernet controller: Broadcom Inc. and subsidiaries NetXtreme BCM5720 Gigabit Ethernet PCIe 01:00.1 Ethernet controller: Broadcom Inc. and subsidiaries NetXtreme BCM5720 Gigabit Ethernet PCIe 17:00.0 Ethernet controller: Broadcom Inc. and subsidiaries BCM57416 NetXtreme-E Dual-Media 10G RDMA Ethernet Controller (rev 01) 17:00.1 Ethernet controller: Broadcom Inc. and subsidiaries BCM57416 NetXtreme-E Dual-Media 10G RDMA Ethernet Controller (rev 01) ``` After doing some more testing in different clusters and versions, it's not related to the kubevirt version, but a regression for the Ethernet controller: Broadcom Inc. and subsidiaries BCM57508 NetXtreme-E 10Gb/25Gb/40Gb/50Gb/100Gb/200Gb Ethernet (rev 12) NIC. The Harvester/Kubervirt version didn't matter. On another cluster with mixed NICs, I was able to reproduce it and then migrate to a different host in the same cluster with a BCM57416 10G NIC and then it worked. Tested with a few different Broadcom NICs, but the BCM57508 was the only problematic one. Per the cross-reported OpenSUSE ticket, it appears to be related to this upstream: https://lore.kernel.org/lkml/1d388413ab9cfd765cd2c5e05b5e69cdb2ec5a10.camel@webked.de/ |