Allow TX/RX descriptor ring size to be configurable. VMXNET3 pmd defa…

[vincentmli]

…ult minimal TX

descriptor ring size is 512, this change allows user to configure TX ring size to support mTCP runs in ESXi VM, for example:

in epwget.conf:

num_tx = 512 num_rx = 128

then run epwget in ESXi VM with VMXNET3 PMD:

(Hardware checksum also needs to be disabled in ESXi VM when compiling mTCP in ESXi VM) ./configure --with-dpdk-lib=$RTE_SDK/$RTE_TARGET --disable-hwcsum

/usr/src/mtcp# ./apps/example/epwget 10.1.72.68 160000000 -f /etc/mtcp/config/epwget.conf -N 1 -c 160 Configuration updated by mtcp_setconf(). Application configuration: URL: / Concurrency: 160

Loading mtcp configuration from : /etc/mtcp/config/epwget.conf Loading interface setting EAL: Detected 16 lcore(s) EAL: Detected 1 NUMA nodes EAL: Auto-detected process type: PRIMARY EAL: Multi-process socket /var/run/dpdk/rte/mp_socket EAL: Probing VFIO support... EAL: PCI device 0000:0b:00.0 on NUMA socket -1 EAL: Invalid NUMA socket, default to 0 EAL: probe driver: 15ad:7b0 net_vmxnet3 Total number of attached devices: 1 Interface name: dpdk0 EAL: Auto-detected process type: PRIMARY Configurations: Number of CPU cores available: 1 Number of CPU cores to use: 1 Number of TX ring descriptor: 512 Number of RX ring descriptor: 128 Number of source ip to use: 8 Maximum number of concurrency per core: 1000000 Maximum number of preallocated buffers per core: 1000000 Receive buffer size: 1024 Send buffer size: 1024 TCP timeout seconds: 30 TCP timewait seconds: 0 NICs to print statistics: dpdk0

Interfaces: name: dpdk0, ifindex: 0, hwaddr: 00:50:56:86:10:76, ipaddr: 10.1.72.28, netmask: 255.255.0.0 Number of NIC queues: 1

Loading routing configurations from : config/route.conf Routes: Destination: 10.1.0.0/16, Mask: 255.255.0.0, Masked: 10.1.0.0, Route: ifdx-0

Loading ARP table from : config/arp.conf ARP Table: IP addr: 10.1.72.68, dst_hwaddr: 00:50:56:86:22:BA

Initializing port 0... Ethdev port_id=0 tx_queue_id=0, new added offloads 0x8011 must be within pre-queue offload capabilities 0x0 in rte_eth_tx_queue_setup()

done: rte_eth_dev_flow_ctrl_get: Function not supported [dpdk_load_module: 765] Failed to get flow control info! rte_eth_dev_flow_ctrl_set: Function not supported [dpdk_load_module: 772] Failed to set flow control info!: errno: -95

Checking link statusdone Port 0 Link Up - speed 10000 Mbps - full-duplex Configuration updated by mtcp_setconf(). CPU 0: initialization finished. [mtcp_create_context:1359] CPU 0 is now the master thread. [CPU 0] dpdk0 flows: 0, RX: 26(pps) (err: 0), 0.00(Gbps), TX: 0(pps), 0.00(Gbps) [ ALL ] dpdk0 flows: 0, RX: 26(pps) (err: 0), 0.00(Gbps), TX: 0(pps), 0.00(Gbps) [CPU 0] dpdk0 flows: 0, RX: 12(pps) (err: 0), 0.00(Gbps), TX: 0(pps), 0.00(Gbps) [ ALL ] dpdk0 flows: 0, RX: 12(pps) (err: 0), 0.00(Gbps), TX: 0(pps), 0.00(Gbps) Thread 0 handles 160000000 flows. connecting to 10.1.72.68:80 rte_eth_stats_reset: Function not supported Response size set to 86 [ ALL ] connect: 8334, read: 0 MB, write: 0 MB, completes: 8174 (resp_time avg: 2890, max: 6412 us) [CPU 0] dpdk0 flows: 247, RX: 95984(pps) (err: 0), 0.09(Gbps), TX: 120431(pps), 0.11(Gbps) [ ALL ] dpdk0 flows: 247, RX: 95984(pps) (err: 0), 0.09(Gbps), TX: 120431(pps), 0.11(Gbps) rte_eth_stats_reset: Function not supported [ ALL ] connect: 33905, read: 2 MB, write: 3 MB, completes: 33905 (resp_time avg: 2512, max: 12273 us) [CPU 0] dpdk0 flows: 233, RX: 137128(pps) (err: 0), 0.12(Gbps), TX: 172068(pps), 0.15(Gbps) [ ALL ] dpdk0 flows: 233, RX: 137128(pps) (err: 0), 0.12(Gbps), TX: 172068(pps), 0.15(Gbps)

[vincentmli]

ok, I will submit through devel branch and add the comments in *.conf files

[vincentmli]

Hi Asim, could you please give me git commands tips on how to create PR with upstream mTCP "devel" branch? the github workflow I got from other project is to fork the repo from upstream github repo, clone the fork to local workstation, checkout development branch, making/commit changes, push to github, then create PR from the development branch.

[ajamshed]

Sorry.. got around to this very, very late. But I think you figured it out yourself! :)