[dpdk-users] DPDK on Alternate Architectures
>Hi all, > >> > >I am trying to get started with DPDK. Just wanted to know if DPDK is ported to >any other arch other than Intel. I saw that DPDK has been ported to Cavium >Nitrox, but I am not really sure on the architecture of Nitrox cores so any >info on the architecture of Nitrox or any alternate architecture >implementation of DPDK would help(ideally MIPS port). DPDK runs on IA, ARM and PPC architectures. Please have a look at the Docs for DPDK. http://dpdk.readthedocs.org/en/v2.2.0/ > >Thank you > >+--+ >Sirshak Das >Research Assistant @ Indiana University >Sent via Outlook > > Regards, Keith
[dpdk-users] DPDK on Alternate Architectures
Hi all, <> I am trying to get started with DPDK. Just wanted to know if DPDK is ported to any other arch other than Intel. I saw that DPDK has been ported to Cavium Nitrox, but I am not really sure on the architecture of Nitrox cores so any info on the architecture of Nitrox or any alternate architecture implementation of DPDK would help(ideally MIPS port). Thank you +--+ Sirshak Das Research Assistant @ Indiana University Sent via Outlook
[dpdk-users] DPDK ip-pipeline error when using virtual function interface
Hi, I would like to try ip-pipline sample application and I'm getting the below error. The virtual functions work fine with l2fwd application though. I find this thread is close to my problem, however that solution didn't workout for me. http://dpdk.org/ml/archives/users/2015-November/72.html My server is Ubuntu 14.04 with kernel 3.19. dpdk_nic_bind.py --status Network devices using DPDK-compatible driver :82:10.0 '82599 Ethernet Controller Virtual Function' drv=vfio-pci unused= :82:10.2 '82599 Ethernet Controller Virtual Function' drv=vfio-pci unused= :82:10.4 '82599 Ethernet Controller Virtual Function' drv=vfio-pci unused= :82:10.6 '82599 Ethernet Controller Virtual Function' drv=vfio-pci unused= :82:11.0 '82599 Ethernet Controller Virtual Function' drv=vfio-pci unused= This is the error when I run ./build/ip_pipeline -p 0x01 AL: probe driver: 8086:10ed rte_ixgbevf_pmd EAL: Not managed by a supported kernel driver, skipped EAL: PCI device :82:1e.4 on NUMA socket 1 EAL: probe driver: 8086:10ed rte_ixgbevf_pmd EAL: Not managed by a supported kernel driver, skipped EAL: PCI device :82:1e.6 on NUMA socket 1 EAL: probe driver: 8086:10ed rte_ixgbevf_pmd EAL: Not managed by a supported kernel driver, skipped [APP] Initializing MEMPOOL0 ... [APP] Initializing LINK0 (0) (1 RXQ, 1 TXQ) ... PMD: ixgbevf_dev_configure(): Configured Virtual Function port id: 0 PMD: ixgbevf_dev_configure(): VF can't disable HW CRC Strip PMD: ixgbe_dev_rx_queue_setup(): sw_ring=0x7fbdbffef6c0 sw_sc_ring=0x7fbdbffef180 hw_ring=0x7fbdbffefc00 dma_addr=0xc7ffefc00 PMD: ixgbe_dev_tx_queue_setup(): sw_ring=0x7fbdbffdcfc0 hw_ring=0x7fbdbffdf000 dma_addr=0xc7ffdf000 PMD: ixgbe_set_tx_function(): Using simple tx code path PMD: ixgbe_set_tx_function(): Vector tx enabled. PMD: ixgbe_set_rx_function(): Vector rx enabled, please make sure RX burst size no less than 4 (port=0). PANIC in app_link_up_internal(): LINK0 (0): PMD set up error -95 7: [./build/ip_pipeline() [0x42fd83]] 6: [/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf5) [0x7fc25546cec5]] 5: [./build/ip_pipeline(main+0x5f) [0x42e92f]] 4: [./build/ip_pipeline(app_init+0xcf9) [0x43e6c9]] 3: [./build/ip_pipeline(app_link_up_internal+0x52f) [0x43d1cf]] 2: [./build/ip_pipeline(__rte_panic+0xc9) [0x4293ad]] 1: [./build/ip_pipeline(rte_dump_stack+0x1a) [0x4b54da]] Aborted Thanks in advance Murad
[dpdk-users] L2fwd very slow throughput
Hi everyone, I went through previous messages on the mailing list to look for people with a similar issue, even though I found some users with slower than expected throughput it is still significantly higher than what I?m getting (http://dpdk.org/ml/archives/dev/2014-July/004114.html). Therefore I would like to know if some people have experienced similar problems or would be able to point out what might be wrong with the setup. The setup is fairly standard, I have two hosts with Intel X710 quad port 10G NICs. One host is running l2fwd and the other one is running MoonGen with the rfc2544 throughput test. The two hosts are connected with two SFP+ cables on port 0 and 1 of the NICs. Both machines are similar, with an Intel 6700K Skylake processor (4 cores at 4GHz with hyperthreading) and 32GB of DDR4 RAM (clocked at 3GHz) running Linux with a kernel 4.4.3 and 8, 1G hugepages and finally on both hosts DPDK 2.2.0 is used and the uio_pci_generic driver is loaded. I think that?s it for the setup. If I run the throughput test using the Linux bridge with 64 bytes packet size, I achieve a slow but expected throughput of 1505.14Mbit/s maximal rate for packetsize 64: 2.24 Mpps, 1146.77 MBit/s, 1505.14 MBit/s wire rate Running the same test with l2fwd instead of the Linux bridge ends up with a very slow throughput maximal rate for packetsize 64: 1.07 Mpps, 546.08 MBit/s, 716.74 MBit/s wire rate L2Fwd is started using the command below, I tried to give more or less cores, change the number of memory channel and also the number of RX queues but the result is always pretty much exactly the same. sudo ./build/l2fwd -c 0xff -n 3 -- -p 3 I tried to use testpmd in iofwd and macfwd mode with the default configuration and with exactly the same test I reach the 10Gbit/s without problem however with the l2fwd, l3fwd the performances are slow as described above. I tried to change the rx buffer size in the l2fwd example from 32 to 64 but the result is the same. Finally I tried the netmap_compat bridge and the result is half the speed of l2fwd maximal rate for packetsize 64: 0.75 Mpps, 382.26 MBit/s, 501.71 MBit/s wire rate Am I missing something obvious? Best regards, Simon
[dpdk-users] Not able to access uio0/device/config from LXC
I have followed the steps mentioned in this thread: to create a uio0, but EAL complains about the following: -- EAL: Cannot open /sys/class/uio/uio0/device/config: Permission denied EAL: Error - exiting with code: 1 Cause: Requested device :00:19.0 cannot be used -- even though I run as a root user and have all the rw permissions. Can anybody tell me what else I could be missing? Thanks, -nagp
[dpdk-users] Minimal dpdk configuration for 2 hosts
Hello Harold, I did investigation and one directional forwarding through 2 hosts connected by one port NIC each is possible indeed. Testpmd with following command line arguments can do that: ? On host1 (as sender): ./testpmd -c 0x3 -n4 -- -i --forward-mode=txonly --port-topology=chained On host2 (as reciever): ./testpmd -c 0x3 -n4 -- -i --forward-mode=rxonly --port-topology=chained ? Anyway still I don't how to run two directional communication. Harold, could you provide commands/command line options for this? Also It seems that? 2 port NIC on every host is required to run basicfwd, rxtx_callbacks and other examples (dpdk gives me print that number of ports must be even). ? Regards, Dawid ? W dniu 2016-03-06 10:30:51 u?ytkownik Harold Demure napisa?: Hello Dawid, I am no expert but a single port should be able to take care of both TX and RX queues. For example, I am currently running two hosts with only one port each and they are able to both send and receive messages. Regards, Harold 2016-03-04 21:47 GMT+01:00 dawid_jurek : Hello, I wonder what is the minimal configuration (in sense on number of NIC ports) to run basic dpdk examples like basicfwd, rxtx_callbacks or forwarding by testpmd for 2 hosts connected directly by Ethernet. Is it possible to perform one directional transmission for some kind of sender-reciever scenario (2 hosts, every host with one port)? It seems that for every kind of transmission between 2 machines I need at least 4 ports (because every port may take care of TX or RX but not both of them at the same time). Is it correct? Regards, Dawid ?