Bug 10
Summary: | [Testpmd] NUMA, speed issue | ||
---|---|---|---|
Product: | DPDK | Reporter: | Anas (nounoussma) |
Component: | testpmd | Assignee: | Anas (nounoussma) |
Status: | CONFIRMED --- | ||
Severity: | normal | CC: | ajit.khaparde, qian.q.xu, xiao.w.wang |
Priority: | Normal | ||
Version: | unspecified | ||
Target Milestone: | --- | ||
Hardware: | x86 | ||
OS: | Linux |
Looks like only speed issue, how do u know NUMA and hugepage issue? "1024 hugepages of size 2097152 reserved, but no mounted hugetlbfs found for that size" didn't mean huge page issue, the app will check default folder for the huge page mounted folder. If huge page has issue, you can't run up the app. As to numa, since you are using lcore 1-3, then it will use socket 0 for memory allocation and core allocation. How about the fm10k device firmware and dpdk version? fm10k can be used in PICeX8 or X4 mode, the speed is ~50Gbps per Gen3 x8 PCIe interface, and ~25Gbps per x4 interface. And the speed can be restricted by TestPoint (Switch manager). Currently the driver leaves 'speed' field as 0 because it's not aware of the x8 or x4 configuration. nounoussma@hotmail.com, Are you satisfied with the explanation? Do you anything need more on this issue? If not, can you close it? Thanks Ajit |
Hello, I need help to manage packet using dpdk under xeon intel chip. When I launch testpmd, I'm wondering if output traces below are blocking to check bandwith: >./testpmd -l 0-3 -n 4 -- -i --portmask=0x1 --nb-cores=2 EAL: Detected 8 lcore(s) EAL: 1024 hugepages of size 2097152 reserved, but no mounted hugetlbfs found for that size EAL: Probing VFIO support... EAL: cannot open /proc/self/numa_maps, consider that all memory is in socket_id 0 EAL: PCI device 0000:01:00.0 on NUMA socket 0 EAL: probe driver: 8086:15a4 net_fm10k EAL: PCI device 0000:02:00.0 on NUMA socket 0 EAL: probe driver: 8086:15a4 net_fm10k EAL: PCI device 0000:04:00.0 on NUMA socket 0 EAL: probe driver: 8086:15ab net_ixgbe EAL: PCI device 0000:04:00.1 on NUMA socket 0 EAL: probe driver: 8086:15ab net_ixgbe EAL: PCI device 0000:04:10.1 on NUMA socket 0 EAL: probe driver: 8086:15a8 net_ixgbe_vf EAL: PCI device 0000:04:10.3 on NUMA socket 0 EAL: probe driver: 8086:15a8 net_ixgbe_vf EAL: PCI device 0000:04:10.5 on NUMA socket 0 EAL: probe driver: 8086:15a8 net_ixgbe_vf EAL: PCI device 0000:06:00.0 on NUMA socket 0 EAL: probe driver: 8086:15a4 net_fm10k EAL: PCI device 0000:08:00.0 on NUMA socket 0 EAL: probe driver: 8086:1533 net_e1000_igb Interactive-mode selected previous number of forwarding ports 2 - changed to number of configured ports 1 USER1: create a new mbuf pool <mbuf_pool_socket_0>: n=171456, size=2240, socket=0 Warning! Cannot handle an odd number of ports with the current port topology. Configuration must be changed to have an even number of ports, or relaunch application with --port-topology=chained Configuring Port 0 (socket 0) PMD: fm10k_dev_configure(): fm10k always strip CRC Port 0: 00:A0:C9:23:45:69 Configuring Port 1 (socket 0) PMD: fm10k_dev_configure(): fm10k always strip CRC Port 1: 00:A0:C9:23:45:6A Checking link statuses... Port 0 Link Up - speed 0 Mbps - full-duplex Port 1 Link Up - speed 0 Mbps - full-duplex On one side, traces show that there is NUMA, speed and hupepage issue. Have you a idea ? Thank you