https://bugs.dpdk.org/show_bug.cgi?id=1750 Bug ID: 1750 Summary: ixgbe: X553 link remains down (NO-CARRIER) Product: DPDK Version: 24.11 Hardware: x86 OS: Linux Status: UNCONFIRMED Severity: normal Priority: Normal Component: ethdev Assignee: dev@dpdk.org Reporter: clinton@netgate.com Target Milestone: --- Ethernet Connection X553 10 GbE SFP+ link can get stuck in a DOWN/NO-CARRIER state when starting an application (testpmd or VPP). The interface remains in a DOWN/NO-CARRIER state, even when unbinding the NIC from vfio-pci followed by binding to the ixgbe kernel driver. The only way I've been able to bring the interface back up is by rebooting the system. I've tested with multiple systems, SFP+ modules, switches, etc... I initially ran into this when running CSIT tests against VPP. After some time, failures would start occurring because of the link issue. I found that starting/stopping VPP multiple times eventually resulted in the interface state getting stuck down. I replicated this with testpmd as well. Steps to Reproduce: 1. Bind X553 interfaces to vfio-pci driver 2. Compile testpmd application if not already compiled 3. Create a file (ie: testpmd-flap.txt) with the following contents: port stop all set link-up port 0 set link-up port 1 port reset all port start all 4. Create a script to rerun the dpdk-testpmd application while true; do echo | ./dpdk-testpmd -a 0000:03:00.0 -a 0000:03:00.1 -l 1-4 -- --disable-link-check --disable-device-start --cmdline=testpmd-flap.txt sleep .5 done 5. Run the script. The time it takes to fail can vary...sometimes it only takes a minute or so and other times it can take 30+ minutes. You'll need to stop the script to check the port state (testpmd> show port info all). I'm using two interfaces and generally when this happens only one of the two interfaces shows in the down state. -- You are receiving this mail because: You are the assignee for the bug.