Commit b517ab06 authored by Tomasz Zawadzki's avatar Tomasz Zawadzki Committed by Tomasz Zawadzki
Browse files

test/nvmf: add multipath to ns hotplug tests 



Adds additional paths for the bdev_nvme during ns hotplug tests.

While adding async AER handling I've found it useful to catch
assert(nbdev != NULL) at bdev_nvme.c:5141. Not possible in
current master.

Change-Id: I01e3877b94601c66de7a1f55bebf9629f464349e
Signed-off-by: default avatarTomasz Zawadzki <tomasz.zawadzki@nutanix.com>
Reviewed-on: https://review.spdk.io/c/spdk/spdk/+/26398


Reviewed-by: default avatarJim Harris <jim.harris@nvidia.com>
Reviewed-by: default avatarBen Walker <ben@nvidia.com>
Reviewed-by: default avatarJacek Kalwas <jacek.kalwas@nutanix.com>
Tested-by: default avatarSPDK Automated Test System <spdkbot@gmail.com>
parent bb97032c

test/nvmf/host/ns_hotplug_stress.sh

+13 −4
Original line number Diff line number Diff line
@@ -11,6 +11,12 @@ rpc_py="$rootdir/scripts/rpc.py" 
tgt_sock="/var/tmp/tgt.sock"

tgt_rpc="$rpc_py -s $tgt_sock"



function attach_controller() {

	for ((i = 0; i < io_paths_nr; i++)); do

		$rpc_py bdev_nvme_attach_controller -t "$TEST_TRANSPORT" -a "$NVMF_FIRST_TARGET_IP" -f IPv4 -s "$((NVMF_PORT + i))" -n "$NVME_SUBNQN" -b nvme0

	done

}



function get_resize_count() {

	# Always use id zero to get notifications from the begining of time

	local notify_id=0
@@ -47,9 +53,13 @@ add_remove() { 
nvmftestinit

DEFAULT_RPC_ADDR="$tgt_sock" nvmfappstart -r "$tgt_sock" -m 0x1



io_paths_nr=10



$tgt_rpc nvmf_create_transport $NVMF_TRANSPORT_OPTS

$tgt_rpc nvmf_create_subsystem "$NVME_SUBNQN" -a -s SPDK00000000000001 -m 512

$tgt_rpc nvmf_subsystem_add_listener "$NVME_SUBNQN" -t "$TEST_TRANSPORT" -a "$NVMF_FIRST_TARGET_IP" -s "$NVMF_PORT"

for ((i = 0; i < io_paths_nr; i++)); do

	$tgt_rpc nvmf_subsystem_add_listener "$NVME_SUBNQN" -t "$TEST_TRANSPORT" -a "$NVMF_FIRST_TARGET_IP" -s "$((NVMF_PORT + i))"

done



"${SPDK_APP[@]}" -m 0x2 "${NO_HUGE[@]}" &

spdk_app_pid=$!
@@ -66,8 +76,7 @@ blk_size=4096 


# Instead of default 10ms, use 1 second timeout

$rpc_py bdev_nvme_set_options --nvme-adminq-poll-period-us 1000000



$rpc_py bdev_nvme_attach_controller -t "$TEST_TRANSPORT" -a "$NVMF_FIRST_TARGET_IP" -f IPv4 -s "$NVMF_PORT" -n "$NVME_SUBNQN" -b nvme0

attach_controller



for ((i = 0; i < nthreads; ++i)); do

	# Every thread can use ns_per_thread NSIDs starting at specific offset.
@@ -80,7 +89,7 @@ wait "${pids[@]}" 
# Reattach controller with restored admin queue poll period to 10ms

$rpc_py bdev_nvme_detach_controller nvme0

$rpc_py bdev_nvme_set_options --nvme-adminq-poll-period-us 10000

$rpc_py bdev_nvme_attach_controller -t "$TEST_TRANSPORT" -a "$NVMF_FIRST_TARGET_IP" -f IPv4 -s "$NVMF_PORT" -n "$NVME_SUBNQN" -b nvme0

attach_controller



# Test case 2 - add/remove ns constantly