1AF_VSOCK test suite 2------------------- 3These tests exercise net/vmw_vsock/ host<->guest sockets for VMware, KVM, and 4Hyper-V. 5 6The following tests are available: 7 8 * vsock_test - core AF_VSOCK socket functionality 9 * vsock_diag_test - vsock_diag.ko module for listing open sockets 10 11The following prerequisite steps are not automated and must be performed prior 12to running tests: 13 141. Build the kernel, make headers_install, and build these tests. 152. Install the kernel and tests on the host. 163. Install the kernel and tests inside the guest. 174. Boot the guest and ensure that the AF_VSOCK transport is enabled. 18 19Invoke test binaries in both directions as follows: 20 21 # host=server, guest=client 22 (host)# $TEST_BINARY --mode=server \ 23 --control-port=1234 \ 24 --peer-cid=3 25 (guest)# $TEST_BINARY --mode=client \ 26 --control-host=$HOST_IP \ 27 --control-port=1234 \ 28 --peer-cid=2 29 30 # host=client, guest=server 31 (guest)# $TEST_BINARY --mode=server \ 32 --control-port=1234 \ 33 --peer-cid=2 34 (host)# $TEST_BINARY --mode=client \ 35 --control-port=$GUEST_IP \ 36 --control-port=1234 \ 37 --peer-cid=3 38 39vsock_perf utility 40------------------- 41'vsock_perf' is a simple tool to measure vsock performance. It works in 42sender/receiver modes: sender connect to peer at the specified port and 43starts data transmission to the receiver. After data processing is done, 44it prints several metrics(see below). 45 46Usage: 47# run as sender 48# connect to CID 2, port 1234, send 1G of data, tx buf size is 1M 49./vsock_perf --sender 2 --port 1234 --bytes 1G --buf-size 1M 50 51Output: 52tx performance: A Gbits/s 53 54Output explanation: 55A is calculated as "number of bits to send" / "time in tx loop" 56 57# run as receiver 58# listen port 1234, rx buf size is 1M, socket buf size is 1G, SO_RCVLOWAT is 64K 59./vsock_perf --port 1234 --buf-size 1M --vsk-size 1G --rcvlowat 64K 60 61Output: 62rx performance: A Gbits/s 63total in 'read()': B sec 64POLLIN wakeups: C 65average in 'read()': D ns 66 67Output explanation: 68A is calculated as "number of received bits" / "time in rx loop". 69B is time, spent in 'read()' system call(excluding 'poll()') 70C is number of 'poll()' wake ups with POLLIN bit set. 71D is B / C, e.g. average amount of time, spent in single 'read()'. 72