Spdk/test/common/autotest_common.sh

1318 lines
32 KiB
Bash
Raw Normal View History

#!/usr/bin/env bash
function xtrace_disable() {
if [ "$XTRACE_DISABLED" != "yes" ]; then
PREV_BASH_OPTS="$-"
if [[ "$PREV_BASH_OPTS" == *"x"* ]]; then
XTRACE_DISABLED="yes"
fi
set +x
elif [ -z $XTRACE_NESTING_LEVEL ]; then
XTRACE_NESTING_LEVEL=1
else
XTRACE_NESTING_LEVEL=$((++XTRACE_NESTING_LEVEL))
fi
}
xtrace_disable
set -e
shopt -s expand_aliases
source "$rootdir/test/common/applications.sh"
if [[ -e $rootdir/test/common/build_config.sh ]]; then
source "$rootdir/test/common/build_config.sh"
elif [[ -e $rootdir/mk/config.mk ]]; then
build_config=$(< "$rootdir/mk/config.mk")
source <(echo "${build_config//\?=/=}")
else
source "$rootdir/CONFIG"
fi
# Dummy function to be called after restoring xtrace just so that it appears in the
# xtrace log. This way we can consistently track when xtrace is enabled/disabled.
function xtrace_enable() {
# We have to do something inside a function in bash, and calling any command
# (even `:`) will produce an xtrace entry, so we just define another function.
function xtrace_dummy() { :; }
}
# Keep it as alias to avoid xtrace_enable backtrace always pointing to xtrace_restore.
# xtrace_enable will appear as called directly from the user script, from the same line
# that "called" xtrace_restore.
alias xtrace_restore='if [ -z $XTRACE_NESTING_LEVEL ]; then
if [[ "$PREV_BASH_OPTS" == *"x"* ]]; then
XTRACE_DISABLED="no"; PREV_BASH_OPTS=""; set -x; xtrace_enable;
fi
else
XTRACE_NESTING_LEVEL=$((--XTRACE_NESTING_LEVEL));
if [ $XTRACE_NESTING_LEVEL -eq "0" ]; then
unset XTRACE_NESTING_LEVEL
fi
fi'
: ${RUN_NIGHTLY:=0}
export RUN_NIGHTLY
# Set defaults for missing test config options
: ${SPDK_AUTOTEST_DEBUG_APPS:=0}
export SPDK_AUTOTEST_DEBUG_APPS
: ${SPDK_RUN_VALGRIND=0}
export SPDK_RUN_VALGRIND
: ${SPDK_RUN_FUNCTIONAL_TEST=0}
export SPDK_RUN_FUNCTIONAL_TEST
: ${SPDK_TEST_UNITTEST=0}
export SPDK_TEST_UNITTEST
: ${SPDK_TEST_AUTOBUILD=0}
export SPDK_TEST_AUTOBUILD
: ${SPDK_TEST_ISAL=0}
export SPDK_TEST_ISAL
: ${SPDK_TEST_ISCSI=0}
export SPDK_TEST_ISCSI
: ${SPDK_TEST_ISCSI_INITIATOR=0}
export SPDK_TEST_ISCSI_INITIATOR
: ${SPDK_TEST_NVME=0}
export SPDK_TEST_NVME
: ${SPDK_TEST_NVME_CLI=0}
export SPDK_TEST_NVME_CLI
: ${SPDK_TEST_NVME_CUSE=0}
export SPDK_TEST_NVME_CUSE
: ${SPDK_TEST_NVMF=0}
export SPDK_TEST_NVMF
: ${SPDK_TEST_NVMF_TRANSPORT="rdma"}
export SPDK_TEST_NVMF_TRANSPORT
: ${SPDK_TEST_RBD=0}
export SPDK_TEST_RBD
: ${SPDK_TEST_VHOST=0}
export SPDK_TEST_VHOST
: ${SPDK_TEST_BLOCKDEV=0}
export SPDK_TEST_BLOCKDEV
: ${SPDK_TEST_IOAT=0}
export SPDK_TEST_IOAT
: ${SPDK_TEST_BLOBFS=0}
export SPDK_TEST_BLOBFS
: ${SPDK_TEST_VHOST_INIT=0}
export SPDK_TEST_VHOST_INIT
: ${SPDK_TEST_PMDK=0}
export SPDK_TEST_PMDK
: ${SPDK_TEST_LVOL=0}
export SPDK_TEST_LVOL
: ${SPDK_TEST_JSON=0}
export SPDK_TEST_JSON
: ${SPDK_TEST_REDUCE=0}
export SPDK_TEST_REDUCE
: ${SPDK_TEST_VPP=0}
export SPDK_TEST_VPP
: ${SPDK_RUN_ASAN=0}
export SPDK_RUN_ASAN
: ${SPDK_RUN_UBSAN=0}
export SPDK_RUN_UBSAN
: ${SPDK_RUN_INSTALLED_DPDK=0}
export SPDK_RUN_INSTALLED_DPDK
: ${SPDK_RUN_NON_ROOT=0}
export SPDK_RUN_NON_ROOT
: ${SPDK_TEST_CRYPTO=0}
export SPDK_TEST_CRYPTO
: ${SPDK_TEST_FTL=0}
export SPDK_TEST_FTL
: ${SPDK_TEST_OCF=0}
export SPDK_TEST_OCF
: ${SPDK_TEST_FTL_EXTENDED=0}
export SPDK_TEST_FTL_EXTENDED
: ${SPDK_TEST_VMD=0}
export SPDK_TEST_VMD
: ${SPDK_TEST_OPAL=0}
export SPDK_TEST_OPAL
: ${SPDK_AUTOTEST_X=true}
export SPDK_AUTOTEST_X
: ${SPDK_TEST_RAID5=0}
export SPDK_TEST_RAID5
: ${SPDK_TEST_URING=0}
export SPDK_TEST_URING
# Export PYTHONPATH with addition of RPC framework. New scripts can be created
# specific use cases for tests.
export PYTHONPATH=$PYTHONPATH:$rootdir/scripts
# Don't create Python .pyc files. When running with sudo these will be
# created with root ownership and can cause problems when cleaning the repository.
export PYTHONDONTWRITEBYTECODE=1
# Export flag to skip the known bug that exists in librados
# Bug is reported on ceph bug tracker with number 24078
export ASAN_OPTIONS=new_delete_type_mismatch=0
export UBSAN_OPTIONS='halt_on_error=1:print_stacktrace=1:abort_on_error=1'
# Export LeakSanitizer option to use suppression file in order to prevent false positives
# and known leaks in external executables or libraries from showing up.
asan_suppression_file="/var/tmp/asan_suppression_file"
sudo rm -rf "$asan_suppression_file"
cat << EOL >> "$asan_suppression_file"
# ASAN has some bugs around thread_local variables. We have a destructor in place
# to free the thread contexts, but ASAN complains about the leak before those
# destructors have a chance to run. So suppress this one specific leak using
# LSAN_OPTIONS.
leak:spdk_fs_alloc_thread_ctx
test/asan: preload ASAN for fio ASAN needs to be LD_PRELOADed before SPDK fio_plugin in order to analyze its code. Just adding that will report any issues in fio binary as well as the fio_plugin. To prevent known fio leaks from affecting the results, a suppression list for LeakSanitizer (used in conjunction with ASAN). At this time the suppression list contains known leaks for fio 3.3. The list might need adjustments as fio version is updated. Side note. Even though it is possible to specify directory to ignore ("leak:/usr/src/fio/"). Which in theory should suppress any leaks in fio. It has side effect of hiding SPDK leaks as well, since the fio_plugins leaks are seen as coming from /usr/src/fio/ioengines.c. See below for examples of each suppressed error: Direct leak of 42 byte(s) in 4 object(s) allocated from: #0 0x7f9d52f3e320 in strdup (/lib64/libasan.so.5+0x3b320) #1 0x41f267 in get_new_job /usr/src/fio/init.c:490 Direct leak of 914936 byte(s) in 10397 object(s) allocated from: #0 0x7f74422e8ea6 in __interceptor_calloc (/lib64/libasan.so.5+0x10dea6) #1 0x46402e in log_io_piece /usr/src/fio/iolog.c:214 Direct leak of 608 byte(s) in 19 object(s) allocated from: #0 0x7f74422e8ca8 in __interceptor_malloc (/lib64/libasan.so.5+0x10dca8) #1 0x44c4e1 in add_to_dump_list /usr/src/fio/parse.c:1039 #2 0x44c4e1 in parse_option /usr/src/fio/parse.c:1098 Direct leak of 173 byte(s) in 20 object(s) allocated from: #0 0x7f744227153d in strdup (/lib64/libasan.so.5+0x9653d) #1 0x44b50d in __handle_option /usr/src/fio/parse.c:718 Indirect leak of 111925528 byte(s) in 1271881 object(s) allocated from: #0 0x7f74422e8ea6 in __interceptor_calloc (/lib64/libasan.so.5+0x10dea6) #1 0x46402e in log_io_piece /usr/src/fio/iolog.c:214 Indirect leak of 171 byte(s) in 19 object(s) allocated from: #0 0x7f744227153d in strdup (/lib64/libasan.so.5+0x9653d) #1 0x44c4ed in add_to_dump_list /usr/src/fio/parse.c:1040 #2 0x44c4ed in parse_option /usr/src/fio/parse.c:1098 Indirect leak of 167 byte(s) in 19 object(s) allocated from: #0 0x7f744227153d in strdup (/lib64/libasan.so.5+0x9653d) #1 0x44c502 in add_to_dump_list /usr/src/fio/parse.c:1042 #2 0x44c502 in parse_option /usr/src/fio/parse.c:1098 Change-Id: I9b5811993508421be50b12af160645c77ea93d7e Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/456315 Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-05-30 07:29:06 +00:00
# Suppress known leaks in fio project
leak:$CONFIG_FIO_SOURCE_DIR/parse.c
leak:$CONFIG_FIO_SOURCE_DIR/iolog.c
leak:$CONFIG_FIO_SOURCE_DIR/init.c
leak:$CONFIG_FIO_SOURCE_DIR/filesetup.c
leak:fio_memalign
leak:spdk_fio_io_u_init
test/asan: preload ASAN for fio ASAN needs to be LD_PRELOADed before SPDK fio_plugin in order to analyze its code. Just adding that will report any issues in fio binary as well as the fio_plugin. To prevent known fio leaks from affecting the results, a suppression list for LeakSanitizer (used in conjunction with ASAN). At this time the suppression list contains known leaks for fio 3.3. The list might need adjustments as fio version is updated. Side note. Even though it is possible to specify directory to ignore ("leak:/usr/src/fio/"). Which in theory should suppress any leaks in fio. It has side effect of hiding SPDK leaks as well, since the fio_plugins leaks are seen as coming from /usr/src/fio/ioengines.c. See below for examples of each suppressed error: Direct leak of 42 byte(s) in 4 object(s) allocated from: #0 0x7f9d52f3e320 in strdup (/lib64/libasan.so.5+0x3b320) #1 0x41f267 in get_new_job /usr/src/fio/init.c:490 Direct leak of 914936 byte(s) in 10397 object(s) allocated from: #0 0x7f74422e8ea6 in __interceptor_calloc (/lib64/libasan.so.5+0x10dea6) #1 0x46402e in log_io_piece /usr/src/fio/iolog.c:214 Direct leak of 608 byte(s) in 19 object(s) allocated from: #0 0x7f74422e8ca8 in __interceptor_malloc (/lib64/libasan.so.5+0x10dca8) #1 0x44c4e1 in add_to_dump_list /usr/src/fio/parse.c:1039 #2 0x44c4e1 in parse_option /usr/src/fio/parse.c:1098 Direct leak of 173 byte(s) in 20 object(s) allocated from: #0 0x7f744227153d in strdup (/lib64/libasan.so.5+0x9653d) #1 0x44b50d in __handle_option /usr/src/fio/parse.c:718 Indirect leak of 111925528 byte(s) in 1271881 object(s) allocated from: #0 0x7f74422e8ea6 in __interceptor_calloc (/lib64/libasan.so.5+0x10dea6) #1 0x46402e in log_io_piece /usr/src/fio/iolog.c:214 Indirect leak of 171 byte(s) in 19 object(s) allocated from: #0 0x7f744227153d in strdup (/lib64/libasan.so.5+0x9653d) #1 0x44c4ed in add_to_dump_list /usr/src/fio/parse.c:1040 #2 0x44c4ed in parse_option /usr/src/fio/parse.c:1098 Indirect leak of 167 byte(s) in 19 object(s) allocated from: #0 0x7f744227153d in strdup (/lib64/libasan.so.5+0x9653d) #1 0x44c502 in add_to_dump_list /usr/src/fio/parse.c:1042 #2 0x44c502 in parse_option /usr/src/fio/parse.c:1098 Change-Id: I9b5811993508421be50b12af160645c77ea93d7e Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/456315 Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-05-30 07:29:06 +00:00
# Suppress leaks in libiscsi
leak:libiscsi.so
EOL
# Suppress leaks in libfuse3
echo "leak:libfuse3.so" >> "$asan_suppression_file"
export LSAN_OPTIONS=suppressions="$asan_suppression_file"
export DEFAULT_RPC_ADDR="/var/tmp/spdk.sock"
if [ -z "$DEPENDENCY_DIR" ]; then
export DEPENDENCY_DIR=/home/sys_sgsw
else
export DEPENDENCY_DIR
fi
# Export location of where all the SPDK binaries are
export SPDK_BIN_DIR="$rootdir/build/bin"
export SPDK_EXAMPLE_DIR="$rootdir/build/examples"
# pass our valgrind desire on to unittest.sh
if [ $SPDK_RUN_VALGRIND -eq 0 ]; then
export valgrind=''
fi
if [ "$(uname -s)" = "Linux" ]; then
MAKE="make"
MAKEFLAGS=${MAKEFLAGS:--j$(nproc)}
DPDK_LINUX_DIR=/usr/share/dpdk/x86_64-default-linuxapp-gcc
if [ -d $DPDK_LINUX_DIR ] && [ $SPDK_RUN_INSTALLED_DPDK -eq 1 ]; then
WITH_DPDK_DIR=$DPDK_LINUX_DIR
fi
# Override the default HUGEMEM in scripts/setup.sh to allocate 8GB in hugepages.
export HUGEMEM=8192
elif [ "$(uname -s)" = "FreeBSD" ]; then
MAKE="gmake"
MAKEFLAGS=${MAKEFLAGS:--j$(sysctl -a | grep -E -i 'hw.ncpu' | awk '{print $2}')}
DPDK_FREEBSD_DIR=/usr/local/share/dpdk/x86_64-native-bsdapp-clang
if [ -d $DPDK_FREEBSD_DIR ] && [ $SPDK_RUN_INSTALLED_DPDK -eq 1 ]; then
WITH_DPDK_DIR=$DPDK_FREEBSD_DIR
fi
# FreeBSD runs a much more limited set of tests, so keep the default 2GB.
export HUGEMEM=2048
else
echo "Unknown OS \"$(uname -s)\""
exit 1
fi
if [ -z "$output_dir" ]; then
if [ -z "$rootdir" ] || [ ! -d "$rootdir/../output" ]; then
output_dir=.
else
output_dir=$rootdir/../output
fi
export output_dir
fi
TEST_MODE=
for i in "$@"; do
case "$i" in
--iso)
TEST_MODE=iso
;;
--transport=*)
TEST_TRANSPORT="${i#*=}"
;;
--sock=*)
TEST_SOCK="${i#*=}"
;;
esac
done
scripts/rpc.py: add daemon mode Add rpc_cmd() bash command that sends rpc command to an rpc.py instance permanently running in background. This makes sending RPC commands even 17 times faster. We make use of bash coprocesses - a builtin bash feature that allow starting background processes with stdin and stdout connected to pipes. rpc.py will block trying to read stdin, effectively being always "ready" to read an RPC command. The background rpc.py is started with a new --server flag that's described as: > Start listening on stdin, parse each line as a regular > rpc.py execution and create a separate connection for each command. > Each command's output ends with either **STATUS=0 if the > command succeeded or **STATUS=1 if it failed. > --server is meant to be used in conjunction with bash > coproc, where stdin and stdout are named pipes and can be > used as a faster way to send RPC commands. As a part of this patch I'm attaching a sample test that runs the following rpc commands first with the regular rpc.py, then the new rpc_cmd() function. ``` time { bdevs=$($rpc bdev_get_bdevs) [ "$(jq length <<< "$bdevs")" == "0" ] malloc=$($rpc bdev_malloc_create 8 512) bdevs=$($rpc bdev_get_bdevs) [ "$(jq length <<< "$bdevs")" == "1" ] $rpc bdev_passthru_create -b "$malloc" -p Passthru0 bdevs=$($rpc bdev_get_bdevs) [ "$(jq length <<< "$bdevs")" == "2" ] $rpc bdev_passthru_delete Passthru0 $rpc bdev_malloc_delete $malloc bdevs=$($rpc bdev_get_bdevs) [ "$(jq length <<< "$bdevs")" == "0" ] } ``` Regular rpc.py: ``` real 0m1.477s user 0m1.289s sys 0m0.139s ``` rpc_cmd(): ``` real 0m0.085s user 0m0.025s sys 0m0.006s ``` autotest_common.sh will now spawn an rpc.py daemon if it's not running yet, and it will offer rpc_cmd() function to quickly send RPC commands. If the command is invalid or SPDK returns with error, the bash function will return a non-zero code and may trigger ERR trap just like a regular rpc.py instance. Pipes have major advantage over e.g. unix domain sockets - the pipes will be automatically closed once the owner process exits. This means we can create a named pipe in autotest_common.sh, open it, then start rpc.py in background and never worry about it again - it will be closed automatically once the test exits. It doesn't even matter if the test is executed manually in isolation, or as a part of the entire autotest. (check_so_deps.sh needs to be modified not to wait for *all* background processes to finish, but just the ones it started) Change-Id: If0ded961b7fef3af3837b44532300dee8b5b4663 Signed-off-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com> Signed-off-by: Michal Berger <michalx.berger@intel.com> Signed-off-by: Pawel Kaminski <pawelx.kaminski@intel.com> Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/621 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2019-06-21 07:38:47 +00:00
# start rpc.py coprocess if it's not started yet
if [[ -z $RPC_PIPE_PID ]] || ! kill -0 "$RPC_PIPE_PID" &> /dev/null; then
scripts/rpc.py: add daemon mode Add rpc_cmd() bash command that sends rpc command to an rpc.py instance permanently running in background. This makes sending RPC commands even 17 times faster. We make use of bash coprocesses - a builtin bash feature that allow starting background processes with stdin and stdout connected to pipes. rpc.py will block trying to read stdin, effectively being always "ready" to read an RPC command. The background rpc.py is started with a new --server flag that's described as: > Start listening on stdin, parse each line as a regular > rpc.py execution and create a separate connection for each command. > Each command's output ends with either **STATUS=0 if the > command succeeded or **STATUS=1 if it failed. > --server is meant to be used in conjunction with bash > coproc, where stdin and stdout are named pipes and can be > used as a faster way to send RPC commands. As a part of this patch I'm attaching a sample test that runs the following rpc commands first with the regular rpc.py, then the new rpc_cmd() function. ``` time { bdevs=$($rpc bdev_get_bdevs) [ "$(jq length <<< "$bdevs")" == "0" ] malloc=$($rpc bdev_malloc_create 8 512) bdevs=$($rpc bdev_get_bdevs) [ "$(jq length <<< "$bdevs")" == "1" ] $rpc bdev_passthru_create -b "$malloc" -p Passthru0 bdevs=$($rpc bdev_get_bdevs) [ "$(jq length <<< "$bdevs")" == "2" ] $rpc bdev_passthru_delete Passthru0 $rpc bdev_malloc_delete $malloc bdevs=$($rpc bdev_get_bdevs) [ "$(jq length <<< "$bdevs")" == "0" ] } ``` Regular rpc.py: ``` real 0m1.477s user 0m1.289s sys 0m0.139s ``` rpc_cmd(): ``` real 0m0.085s user 0m0.025s sys 0m0.006s ``` autotest_common.sh will now spawn an rpc.py daemon if it's not running yet, and it will offer rpc_cmd() function to quickly send RPC commands. If the command is invalid or SPDK returns with error, the bash function will return a non-zero code and may trigger ERR trap just like a regular rpc.py instance. Pipes have major advantage over e.g. unix domain sockets - the pipes will be automatically closed once the owner process exits. This means we can create a named pipe in autotest_common.sh, open it, then start rpc.py in background and never worry about it again - it will be closed automatically once the test exits. It doesn't even matter if the test is executed manually in isolation, or as a part of the entire autotest. (check_so_deps.sh needs to be modified not to wait for *all* background processes to finish, but just the ones it started) Change-Id: If0ded961b7fef3af3837b44532300dee8b5b4663 Signed-off-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com> Signed-off-by: Michal Berger <michalx.berger@intel.com> Signed-off-by: Pawel Kaminski <pawelx.kaminski@intel.com> Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/621 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2019-06-21 07:38:47 +00:00
coproc RPC_PIPE { "$rootdir/scripts/rpc.py" --server; }
exec {RPC_PIPE_OUTPUT}<&${RPC_PIPE[0]} {RPC_PIPE_INPUT}>&${RPC_PIPE[1]}
# all descriptors will automatically close together with this bash
# process, this will make rpc.py stop reading and exit gracefully
fi
if [ $SPDK_TEST_VPP -eq 1 ]; then
VPP_PATH="/usr/local/src/vpp-19.04/build-root/install-vpp_debug-native/vpp/"
export LD_LIBRARY_PATH=${LD_LIBRARY_PATH}:${VPP_PATH}/lib/
export PATH=${PATH}:${VPP_PATH}/bin/
fi
function set_test_storage() {
[[ -v testdir ]] || return 0
local requested_size=$1 # bytes
local mount target_dir
local -A mounts fss sizes avails uses
local source fs size avail mount use
local storage_fallback storage_candidates
storage_fallback=/tmp/spdk
storage_candidates=(
"$testdir"
"$storage_fallback/tests/${testdir##*/}"
"$storage_fallback"
)
if [[ -n $ADD_TEST_STORAGE ]]; then
# List of dirs|mounts separated by whitespaces
storage_candidates+=($ADD_TEST_STORAGE)
fi
if [[ -n $DEDICATED_TEST_STORAGE ]]; then
# Single, dedicated dir|mount
storage_candidates=("$DEDICATED_TEST_STORAGE")
fi
mkdir -p "${storage_candidates[@]}"
# add some headroom - 64M
requested_size=$((requested_size + (64 << 20)))
while read -r source fs size use avail _ mount; do
mounts["$mount"]=$source fss["$mount"]=$fs
avails["$mount"]=$((avail * 1024)) sizes["$mount"]=$((size * 1024))
uses["$mount"]=$((use * 1024))
done < <(df -T | grep -v Filesystem)
printf '* Looking for test storage...\n' >&2
local target_space new_size
for target_dir in "${storage_candidates[@]}"; do
# FreeBSD's df is lacking the --output arg
# mount=$(df --output=target "$target_dir" | grep -v "Mounted on")
mount=$(df "$target_dir" | awk '$1 !~ /Filesystem/{print $6}')
target_space=${avails["$mount"]}
if ((target_space == 0 || target_space < requested_size)); then
continue
fi
if ((target_space >= requested_size)); then
# For in-memory fs, and / make sure our requested size won't fill most of the space.
if [[ ${fss["$mount"]} == tmpfs ]] || [[ ${fss["$mount"]} == ramfs ]] || [[ $mount == / ]]; then
new_size=$((uses["$mount"] + requested_size))
if ((new_size * 100 / sizes["$mount"] > 95)); then
continue
fi
fi
fi
export SPDK_TEST_STORAGE=$target_dir
printf '* Found test storage at %s\n' "$SPDK_TEST_STORAGE" >&2
return 0
done
printf '* Test storage is not available\n'
return 1
}
function get_config_params() {
xtrace_disable
config_params='--enable-debug --enable-werror'
# for options with dependencies but no test flag, set them here
if [ -f /usr/include/infiniband/verbs.h ]; then
config_params+=' --with-rdma'
fi
if [ $(uname -s) == "FreeBSD" ]; then
intel="hw.model: Intel"
cpu_vendor=$(sysctl -a | grep hw.model | cut -c 1-15)
else
intel="GenuineIntel"
cpu_vendor=$(grep -i 'vendor' /proc/cpuinfo --max-count=1)
fi
lib/idxd: add low level idxd library Module, etc., will follow. Notes: * IDXD is an Intel silicon feature available in future Intel CPUs. Initial development is being done on a simulator. Once HW is available and the code fully tested the experimental label will be lifted. Spec can be found here: https://software.intel.com/en-us/download/intel-data-streaming-accelerator-preliminary-architecture-specification * The current implementation will only work with VFIO. * DSA has a number of engines that can be grouped based on application need such as type of memory being served or QoS. Engines are processing units and are assigned to groups. Work queues are on device structures that act as front-end groups for queueing descriptors. Full details on what is configurable & how will come in later doc patches. * There is a finite number of work queue slots that are divided amongst the number of desired work queues in some fashion (ie evenly). * SW (outside of the idxd lib) is required to manage flow control, to not over-run the work queues.This is provided in the accel plug-in module. The upper layers use public API to manage this. * Work queue submissions are done with a 64 byte atomic instruction * The design here creates a set of descriptor rings per channel that match the size of the work queues. Then, an spdk_bit_array is used to make sure we don't overrun a queue. If there are not slots available, the operation is put on a linked list to be retried later from the poller. * As we need to support any number of channels (we can't limit ourselves to the number of work queues) we need to dynamically size/resize our per channel descriptor rings based on the number of current channels. This is done from upper layers via public API into the lib. * As channels are created, the total number of work queue slots is divided across the channels evenly. Same thing when they are destroyed, remaining channels with see the ring sizes increase. This is done from upper layers via public API into the lib. * The sim has 64 total work queue entries (WQE) that get dolled out to the work queues (WQ) evenly. Signed-off-by: paul luse <paul.e.luse@intel.com> Change-Id: I899bbeda3cef3db05bea4197b8757e89dddb579d Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1809 Community-CI: Mellanox Build Bot Reviewed-by: Jim Harris <james.r.harris@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com> Reviewed-by: Vitaliy Mysak <vitaliy.mysak@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2020-04-10 15:29:01 +00:00
if [[ "$cpu_vendor" != *"$intel"* ]]; then
config_params+=" --without-idxd"
else
config_params+=" --with-idxd"
fi
if [[ -d $CONFIG_FIO_SOURCE_DIR ]]; then
config_params+=" --with-fio=$CONFIG_FIO_SOURCE_DIR"
fi
if [ -d ${DEPENDENCY_DIR}/vtune_codes ]; then
config_params+=' --with-vtune='${DEPENDENCY_DIR}'/vtune_codes'
fi
if [ -d /usr/include/iscsi ]; then
libiscsi_version=$(grep LIBISCSI_API_VERSION /usr/include/iscsi/iscsi.h | head -1 | awk '{print $3}' | awk -F '(' '{print $2}' | awk -F ')' '{print $1}')
if [ $libiscsi_version -ge 20150621 ]; then
config_params+=' --with-iscsi-initiator'
fi
fi
if [ $SPDK_TEST_UNITTEST -eq 0 ]; then
config_params+=' --disable-unit-tests'
fi
if [ $SPDK_TEST_NVME_CUSE -eq 1 ]; then
config_params+=' --with-nvme-cuse'
fi
# for options with both dependencies and a test flag, set them here
if [ -f /usr/include/libpmemblk.h ] && [ $SPDK_TEST_PMDK -eq 1 ]; then
config_params+=' --with-pmdk'
fi
if [ -f /usr/include/libpmem.h ] && [ $SPDK_TEST_REDUCE -eq 1 ]; then
if [ $SPDK_TEST_ISAL -eq 1 ]; then
config_params+=' --with-reduce'
fi
fi
if [ -d /usr/include/rbd ] && [ -d /usr/include/rados ] && [ $SPDK_TEST_RBD -eq 1 ]; then
config_params+=' --with-rbd'
fi
if [ $SPDK_TEST_VPP -eq 1 ]; then
config_params+=" --with-vpp=${VPP_PATH}"
fi
# for options with no required dependencies, just test flags, set them here
if [ $SPDK_TEST_CRYPTO -eq 1 ]; then
config_params+=' --with-crypto'
fi
if [ $SPDK_TEST_OCF -eq 1 ]; then
config_params+=" --with-ocf"
fi
if [ $SPDK_RUN_UBSAN -eq 1 ]; then
config_params+=' --enable-ubsan'
fi
if [ $SPDK_RUN_ASAN -eq 1 ]; then
config_params+=' --enable-asan'
fi
if [ "$(uname -s)" = "Linux" ]; then
config_params+=' --enable-coverage'
fi
if [ $SPDK_TEST_ISAL -eq 0 ]; then
config_params+=' --without-isal'
fi
if [ $SPDK_TEST_BLOBFS -eq 1 ]; then
if [[ -d /usr/include/fuse3 ]] || [[ -d /usr/local/include/fuse3 ]]; then
config_params+=' --with-fuse'
fi
fi
if [ $SPDK_TEST_RAID5 -eq 1 ]; then
config_params+=' --with-raid5'
fi
# Check whether liburing library header exists
if [ -f /usr/include/liburing/io_uring.h ] && [ $SPDK_TEST_URING -eq 1 ]; then
config_params+=' --with-uring'
fi
# By default, --with-dpdk is not set meaning the SPDK build will use the DPDK submodule.
# If a DPDK installation is found in a well-known location though, WITH_DPDK_DIR will be
# set which will override the default and use that DPDK installation instead.
if [ -n "$WITH_DPDK_DIR" ]; then
config_params+=" --with-dpdk=$WITH_DPDK_DIR"
fi
echo "$config_params"
xtrace_restore
}
scripts/rpc.py: add daemon mode Add rpc_cmd() bash command that sends rpc command to an rpc.py instance permanently running in background. This makes sending RPC commands even 17 times faster. We make use of bash coprocesses - a builtin bash feature that allow starting background processes with stdin and stdout connected to pipes. rpc.py will block trying to read stdin, effectively being always "ready" to read an RPC command. The background rpc.py is started with a new --server flag that's described as: > Start listening on stdin, parse each line as a regular > rpc.py execution and create a separate connection for each command. > Each command's output ends with either **STATUS=0 if the > command succeeded or **STATUS=1 if it failed. > --server is meant to be used in conjunction with bash > coproc, where stdin and stdout are named pipes and can be > used as a faster way to send RPC commands. As a part of this patch I'm attaching a sample test that runs the following rpc commands first with the regular rpc.py, then the new rpc_cmd() function. ``` time { bdevs=$($rpc bdev_get_bdevs) [ "$(jq length <<< "$bdevs")" == "0" ] malloc=$($rpc bdev_malloc_create 8 512) bdevs=$($rpc bdev_get_bdevs) [ "$(jq length <<< "$bdevs")" == "1" ] $rpc bdev_passthru_create -b "$malloc" -p Passthru0 bdevs=$($rpc bdev_get_bdevs) [ "$(jq length <<< "$bdevs")" == "2" ] $rpc bdev_passthru_delete Passthru0 $rpc bdev_malloc_delete $malloc bdevs=$($rpc bdev_get_bdevs) [ "$(jq length <<< "$bdevs")" == "0" ] } ``` Regular rpc.py: ``` real 0m1.477s user 0m1.289s sys 0m0.139s ``` rpc_cmd(): ``` real 0m0.085s user 0m0.025s sys 0m0.006s ``` autotest_common.sh will now spawn an rpc.py daemon if it's not running yet, and it will offer rpc_cmd() function to quickly send RPC commands. If the command is invalid or SPDK returns with error, the bash function will return a non-zero code and may trigger ERR trap just like a regular rpc.py instance. Pipes have major advantage over e.g. unix domain sockets - the pipes will be automatically closed once the owner process exits. This means we can create a named pipe in autotest_common.sh, open it, then start rpc.py in background and never worry about it again - it will be closed automatically once the test exits. It doesn't even matter if the test is executed manually in isolation, or as a part of the entire autotest. (check_so_deps.sh needs to be modified not to wait for *all* background processes to finish, but just the ones it started) Change-Id: If0ded961b7fef3af3837b44532300dee8b5b4663 Signed-off-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com> Signed-off-by: Michal Berger <michalx.berger@intel.com> Signed-off-by: Pawel Kaminski <pawelx.kaminski@intel.com> Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/621 Tested-by: SPDK CI Jenkins <sys_sgci@intel.com> Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Jim Harris <james.r.harris@intel.com>
2019-06-21 07:38:47 +00:00
function rpc_cmd() {
xtrace_disable
local rsp rc
echo "$@" >&$RPC_PIPE_INPUT
while read -t 5 -ru $RPC_PIPE_OUTPUT rsp; do
if [[ $rsp == "**STATUS="* ]]; then
break
fi
echo "$rsp"
done
rc=${rsp#*=}
xtrace_restore
[[ $rc == 0 ]]
}
function rpc_cmd_simple_data_json() {
local elems="$1[@]" elem
local -gA jq_out=()
local jq val
local lvs=(
"uuid"
"name"
"base_bdev"
"total_data_clusters"
"free_clusters"
"block_size"
"cluster_size"
)
local bdev=(
"name"
"aliases[0]"
"block_size"
"num_blocks"
"uuid"
"product_name"
)
[[ -v $elems ]] || return 1
for elem in "${!elems}"; do
jq="${jq:+$jq,\"\\n\",}\"$elem\",\" \",.[0].$elem"
done
jq+=',"\n"'
shift
while read -r elem val; do
jq_out["$elem"]=$val
done < <(rpc_cmd "$@" | jq -jr "$jq")
((${#jq_out[@]} > 0)) || return 1
}
# invert error code of any command and also trigger ERR on 0 (unlike bash ! prefix)
function NOT() {
if "$@"; then
return 1
fi
}
function timing() {
direction="$1"
testname="$2"
now=$(date +%s)
if [ "$direction" = "enter" ]; then
export timing_stack="${timing_stack};${now}"
export test_stack="${test_stack};${testname}"
else
touch "$output_dir/timing.txt"
child_time=$(grep "^${test_stack:1};" $output_dir/timing.txt | awk '{s+=$2} END {print s}')
start_time=$(echo "$timing_stack" | sed -e 's@^.*;@@')
timing_stack=$(echo "$timing_stack" | sed -e 's@;[^;]*$@@')
elapsed=$((now - start_time - child_time))
echo "${test_stack:1} $elapsed" >> $output_dir/timing.txt
test_stack=$(echo "$test_stack" | sed -e 's@;[^;]*$@@')
fi
}
function timing_enter() {
xtrace_disable
timing "enter" "$1"
xtrace_restore
}
function timing_exit() {
xtrace_disable
timing "exit" "$1"
xtrace_restore
}
function timing_finish() {
flamegraph='/usr/local/FlameGraph/flamegraph.pl'
if [ -x "$flamegraph" ]; then
"$flamegraph" \
--title 'Build Timing' \
--nametype 'Step:' \
--countname seconds \
$output_dir/timing.txt \
> $output_dir/timing.svg
fi
}
function create_test_list() {
xtrace_disable
# First search all scripts in main SPDK directory.
completion=$(grep -shI -d skip --include="*.sh" -e "run_test " $rootdir/*)
# Follow up with search in test directory recursively.
completion+=$(grep -rshI --include="*.sh" --exclude="autotest_common.sh" -e "run_test " $rootdir/test)
printf "%s" "$completion" | grep -v "#" \
| sed 's/^.*run_test/run_test/' | awk '{print $2}' \
| sed 's/\"//g' | sort > $output_dir/all_tests.txt || true
xtrace_restore
}
function gdb_attach() {
gdb -q --batch \
-ex 'handle SIGHUP nostop pass' \
-ex 'handle SIGQUIT nostop pass' \
-ex 'handle SIGPIPE nostop pass' \
-ex 'handle SIGALRM nostop pass' \
-ex 'handle SIGTERM nostop pass' \
-ex 'handle SIGUSR1 nostop pass' \
-ex 'handle SIGUSR2 nostop pass' \
-ex 'handle SIGCHLD nostop pass' \
-ex 'set print thread-events off' \
-ex 'cont' \
-ex 'thread apply all bt' \
-ex 'quit' \
--tty=/dev/stdout \
-p $1
}
function process_core() {
ret=0
while IFS= read -r -d '' core; do
exe=$(eu-readelf -n "$core" | grep psargs | sed "s/.*psargs: \([^ \'\" ]*\).*/\1/")
if [[ ! -f "$exe" ]]; then
exe=$(eu-readelf -n "$core" | grep -oP -m1 "$exe.+")
fi
echo "exe for $core is $exe"
if [[ -n "$exe" ]]; then
if hash gdb &> /dev/null; then
gdb -batch -ex "thread apply all bt full" $exe $core
fi
cp $exe $output_dir
fi
mv $core $output_dir
chmod a+r $output_dir/$core
ret=1
done < <(find . -type f \( -name 'core\.?[0-9]*' -o -name '*.core' \) -print0)
return $ret
}
function process_shm() {
type=$1
id=$2
if [ "$type" = "--pid" ]; then
id="pid${id}"
elif [ "$type" = "--id" ]; then
id="${id}"
else
echo "Please specify to search for pid or shared memory id."
return 1
fi
shm_files=$(find /dev/shm -name "*.${id}" -printf "%f\n")
if [[ -z $shm_files ]]; then
echo "SHM File for specified PID or shared memory id: ${id} not found!"
return 1
fi
for n in $shm_files; do
tar -C /dev/shm/ -cvzf $output_dir/${n}_shm.tar.gz ${n}
done
return 0
}
function waitforlisten() {
# $1 = process pid
if [ -z "$1" ]; then
exit 1
fi
local rpc_addr="${2:-$DEFAULT_RPC_ADDR}"
echo "Waiting for process to start up and listen on UNIX domain socket $rpc_addr..."
# turn off trace for this loop
xtrace_disable
local ret=0
local i
for ((i = 40; i != 0; i--)); do
# if the process is no longer running, then exit the script
# since it means the application crashed
if ! kill -s 0 $1; then
echo "ERROR: process (pid: $1) is no longer running"
ret=1
break
fi
if $rootdir/scripts/rpc.py -t 1 -s "$rpc_addr" rpc_get_methods &> /dev/null; then
break
fi
sleep 0.5
done
xtrace_restore
if ((i == 0)); then
echo "ERROR: timeout while waiting for process (pid: $1) to start listening on '$rpc_addr'"
ret=1
fi
return $ret
}
function waitfornbd() {
local nbd_name=$1
local i
for ((i = 1; i <= 20; i++)); do
if grep -q -w $nbd_name /proc/partitions; then
break
else
sleep 0.1
fi
done
# The nbd device is now recognized as a block device, but there can be
# a small delay before we can start I/O to that block device. So loop
# here trying to read the first block of the nbd block device to a temp
# file. Note that dd returns success when reading an empty file, so we
# need to check the size of the output file instead.
for ((i = 1; i <= 20; i++)); do
dd if=/dev/$nbd_name of="$SPDK_TEST_STORAGE/nbdtest" bs=4096 count=1 iflag=direct
size=$(stat -c %s "$SPDK_TEST_STORAGE/nbdtest")
rm -f "$SPDK_TEST_STORAGE/nbdtest"
if [ "$size" != "0" ]; then
return 0
else
sleep 0.1
fi
done
return 1
}
function waitforbdev() {
local bdev_name=$1
local i
for ((i = 1; i <= 20; i++)); do
if $rpc_py bdev_get_bdevs | jq -r '.[] .name' | grep -qw $bdev_name; then
return 0
fi
if $rpc_py bdev_get_bdevs | jq -r '.[] .aliases' | grep -qw $bdev_name; then
return 0
fi
sleep 0.1
done
return 1
}
function make_filesystem() {
local fstype=$1
local dev_name=$2
local i=0
local force
if [ $fstype = ext4 ]; then
force=-F
else
force=-f
fi
while ! mkfs.${fstype} $force ${dev_name}; do
if [ $i -ge 15 ]; then
return 1
fi
i=$((i + 1))
sleep 1
done
return 0
}
function killprocess() {
# $1 = process pid
if [ -z "$1" ]; then
exit 1
fi
if kill -0 $1; then
if [ $(uname) = Linux ]; then
process_name=$(ps --no-headers -o comm= $1)
else
process_name=$(ps -c -o command $1 | tail -1)
fi
if [ "$process_name" = "sudo" ]; then
# kill the child process, which is the actual app
# (assume $1 has just one child)
local child
child="$(pgrep -P $1)"
echo "killing process with pid $child"
kill $child
else
echo "killing process with pid $1"
kill $1
fi
# wait for the process regardless if its the dummy sudo one
# or the actual app - it should terminate anyway
wait $1
else
# the process is not there anymore
echo "Process with pid $1 is not found"
exit 1
fi
}
function iscsicleanup() {
echo "Cleaning up iSCSI connection"
iscsiadm -m node --logout || true
iscsiadm -m node -o delete || true
rm -rf /var/lib/iscsi/nodes/*
}
function stop_iscsi_service() {
if cat /etc/*-release | grep Ubuntu; then
service open-iscsi stop
else
service iscsid stop
fi
}
function start_iscsi_service() {
if cat /etc/*-release | grep Ubuntu; then
service open-iscsi start
else
service iscsid start
fi
}
function rbd_setup() {
# $1 = monitor ip address
# $2 = name of the namespace
if [ -z "$1" ]; then
echo "No monitor IP address provided for ceph"
exit 1
fi
if [ -n "$2" ]; then
if ip netns list | grep "$2"; then
NS_CMD="ip netns exec $2"
else
echo "No namespace $2 exists"
exit 1
fi
fi
if hash ceph; then
export PG_NUM=128
export RBD_POOL=rbd
export RBD_NAME=foo
$NS_CMD $rootdir/scripts/ceph/stop.sh || true
$NS_CMD $rootdir/scripts/ceph/start.sh $1
$NS_CMD ceph osd pool create $RBD_POOL $PG_NUM || true
$NS_CMD rbd create $RBD_NAME --size 1000
fi
}
function rbd_cleanup() {
if hash ceph; then
$rootdir/scripts/ceph/stop.sh || true
rm -f /var/tmp/ceph_raw.img
fi
}
function nvme_cli_build() {
if [[ -z "${DEPENDENCY_DIR}" ]]; then
echo DEPENDENCY_DIR not defined!
exit 1
fi
spdk_nvme_cli="${DEPENDENCY_DIR}/nvme-cli"
if [[ ! -d $spdk_nvme_cli ]]; then
echo "nvme-cli repository not found at $spdk_nvme_cli; skipping tests."
exit 1
fi
if ! grep -q "DEF_VER=v1.6" $spdk_nvme_cli/NVME-VERSION-GEN; then
echo "SPDK supports only \"spdk/nvme-cli\" project on \"spdk-1.6\" branch."
exit 1
fi
# Build against the version of SPDK under test
pushd $spdk_nvme_cli
git clean -dfx
rm -f "$spdk_nvme_cli/spdk"
ln -sf "$rootdir" "$spdk_nvme_cli/spdk"
make -j$(nproc) LDFLAGS="$(make -s -C $spdk_nvme_cli/spdk ldflags)"
popd
}
function _start_stub() {
# Disable ASLR for multi-process testing. SPDK does support using DPDK multi-process,
# but ASLR can still be unreliable in some cases.
# We will reenable it again after multi-process testing is complete in kill_stub().
# Save current setting so it can be restored upon calling kill_stub().
_randomize_va_space=$(< /proc/sys/kernel/randomize_va_space)
echo 0 > /proc/sys/kernel/randomize_va_space
$rootdir/test/app/stub/stub $1 &
stubpid=$!
echo Waiting for stub to ready for secondary processes...
while ! [ -e /var/run/spdk_stub0 ]; do
# If stub dies while we wait, bail
[[ -e /proc/$stubpid ]] || return 1
sleep 1s
done
echo done.
}
function start_stub() {
if ! _start_stub "$@"; then
echo "stub failed" >&2
return 1
fi
}
function kill_stub() {
if [[ -e /proc/$stubpid ]]; then
kill $1 $stubpid
wait $stubpid
fi 2> /dev/null || :
rm -f /var/run/spdk_stub0
# Re-enable ASLR now that we are done with multi-process testing
# Note: "1" enables ASLR w/o randomizing data segments, "2" adds data segment
# randomizing and is the default on all recent Linux kernels
echo "${_randomize_va_space:-2}" > /proc/sys/kernel/randomize_va_space
}
function run_test() {
if [ $# -le 1 ]; then
echo "Not enough parameters"
echo "usage: run_test test_name test_script [script_params]"
exit 1
fi
xtrace_disable
local test_name="$1"
shift
if [ -n "$test_domain" ]; then
export test_domain="${test_domain}.${test_name}"
else
export test_domain="$test_name"
fi
timing_enter $test_name
echo "************************************"
echo "START TEST $test_name"
echo "************************************"
xtrace_restore
time "$@"
xtrace_disable
echo "************************************"
echo "END TEST $test_name"
echo "************************************"
timing_exit $test_name
export test_domain=${test_domain%"$test_name"}
if [ -n "$test_domain" ]; then
export test_domain=${test_domain%?}
fi
if [ -z "$test_domain" ]; then
echo "top_level $test_name" >> $output_dir/test_completions.txt
else
echo "$test_domain $test_name" >> $output_dir/test_completions.txt
fi
xtrace_restore
}
function skip_run_test_with_warning() {
echo "WARNING: $1"
echo "Test run may fail if run with autorun.sh"
echo "Please check your $rootdir/test/common/skipped_tests.txt"
}
function print_backtrace() {
# if errexit is not enabled, don't print a backtrace
[[ "$-" =~ e ]] || return 0
local args=("${BASH_ARGV[@]}")
xtrace_disable
echo "========== Backtrace start: =========="
echo ""
for i in $(seq 1 $((${#FUNCNAME[@]} - 1))); do
local func="${FUNCNAME[$i]}"
local line_nr="${BASH_LINENO[$((i - 1))]}"
local src="${BASH_SOURCE[$i]}"
local bt="" cmdline=()
if [[ -f $src ]]; then
bt=$(nl -w 4 -ba -nln $src | grep -B 5 -A 5 "^${line_nr}[^0-9]" \
| sed "s/^/ /g" | sed "s/^ $line_nr /=> $line_nr /g")
fi
# If extdebug set the BASH_ARGC[i], try to fetch all the args
if ((BASH_ARGC[i] > 0)); then
# Use argc as index to reverse the stack
local argc=${BASH_ARGC[i]} arg
for arg in "${args[@]::BASH_ARGC[i]}"; do
cmdline[argc--]="[\"$arg\"]"
done
args=("${args[@]:BASH_ARGC[i]}")
fi
echo "in $src:$line_nr -> $func($(
IFS=","
printf '%s\n' "${cmdline[*]:-[]}"
))"
echo " ..."
echo "${bt:-backtrace unavailable}"
echo " ..."
done
echo ""
echo "========== Backtrace end =========="
xtrace_restore
return 0
}
function waitforserial() {
local i=0
local nvme_device_counter=1
if [[ -n "$2" ]]; then
nvme_device_counter=$2
fi
while [ $(lsblk -l -o NAME,SERIAL | grep -c $1) -lt $nvme_device_counter ]; do
[ $i -lt 15 ] || break
i=$((i + 1))
echo "Waiting for devices"
sleep 1
done
if [[ $(lsblk -l -o NAME,SERIAL | grep -c $1) -lt $nvme_device_counter ]]; then
return 1
fi
return 0
}
function waitforserial_disconnect() {
local i=0
while lsblk -o NAME,SERIAL | grep -q -w $1; do
[ $i -lt 15 ] || break
i=$((i + 1))
echo "Waiting for disconnect devices"
sleep 1
done
if lsblk -l -o NAME | grep -q -w $1; then
return 1
fi
return 0
}
function waitforblk() {
local i=0
while ! lsblk -l -o NAME | grep -q -w $1; do
[ $i -lt 15 ] || break
i=$((i + 1))
sleep 1
done
if ! lsblk -l -o NAME | grep -q -w $1; then
return 1
fi
return 0
}
function waitforblk_disconnect() {
local i=0
while lsblk -l -o NAME | grep -q -w $1; do
[ $i -lt 15 ] || break
i=$((i + 1))
sleep 1
done
if lsblk -l -o NAME | grep -q -w $1; then
return 1
fi
return 0
}
function waitforfile() {
local i=0
while [ ! -e $1 ]; do
[ $i -lt 200 ] || break
i=$((i + 1))
sleep 0.1
done
if [ ! -e $1 ]; then
return 1
fi
return 0
}
function fio_config_gen() {
local config_file=$1
local workload=$2
local bdev_type=$3
local fio_dir=$CONFIG_FIO_SOURCE_DIR
if [ -e "$config_file" ]; then
echo "Configuration File Already Exists!: $config_file"
return 1
fi
if [ -z "$workload" ]; then
workload=randrw
fi
touch $1
cat > $1 << EOL
[global]
thread=1
group_reporting=1
direct=1
norandommap=1
percentile_list=50:99:99.9:99.99:99.999
time_based=1
ramp_time=0
EOL
if [ "$workload" == "verify" ]; then
cat <<- EOL >> $config_file
verify=sha1
verify_backlog=1024
rw=randwrite
EOL
# To avoid potential data race issue due to the AIO device
# flush mechanism, add the flag to serialize the writes.
# This is to fix the intermittent IO failure issue of #935
if [ "$bdev_type" == "AIO" ]; then
if [[ $($fio_dir/fio --version) == *"fio-3"* ]]; then
echo "serialize_overlap=1" >> $config_file
fi
fi
elif [ "$workload" == "trim" ]; then
echo "rw=trimwrite" >> $config_file
else
echo "rw=$workload" >> $config_file
fi
}
function fio_bdev() {
# Setup fio binary cmd line
local fio_dir=$CONFIG_FIO_SOURCE_DIR
local bdev_plugin="$rootdir/build/fio/spdk_bdev"
test/asan: preload ASAN for fio ASAN needs to be LD_PRELOADed before SPDK fio_plugin in order to analyze its code. Just adding that will report any issues in fio binary as well as the fio_plugin. To prevent known fio leaks from affecting the results, a suppression list for LeakSanitizer (used in conjunction with ASAN). At this time the suppression list contains known leaks for fio 3.3. The list might need adjustments as fio version is updated. Side note. Even though it is possible to specify directory to ignore ("leak:/usr/src/fio/"). Which in theory should suppress any leaks in fio. It has side effect of hiding SPDK leaks as well, since the fio_plugins leaks are seen as coming from /usr/src/fio/ioengines.c. See below for examples of each suppressed error: Direct leak of 42 byte(s) in 4 object(s) allocated from: #0 0x7f9d52f3e320 in strdup (/lib64/libasan.so.5+0x3b320) #1 0x41f267 in get_new_job /usr/src/fio/init.c:490 Direct leak of 914936 byte(s) in 10397 object(s) allocated from: #0 0x7f74422e8ea6 in __interceptor_calloc (/lib64/libasan.so.5+0x10dea6) #1 0x46402e in log_io_piece /usr/src/fio/iolog.c:214 Direct leak of 608 byte(s) in 19 object(s) allocated from: #0 0x7f74422e8ca8 in __interceptor_malloc (/lib64/libasan.so.5+0x10dca8) #1 0x44c4e1 in add_to_dump_list /usr/src/fio/parse.c:1039 #2 0x44c4e1 in parse_option /usr/src/fio/parse.c:1098 Direct leak of 173 byte(s) in 20 object(s) allocated from: #0 0x7f744227153d in strdup (/lib64/libasan.so.5+0x9653d) #1 0x44b50d in __handle_option /usr/src/fio/parse.c:718 Indirect leak of 111925528 byte(s) in 1271881 object(s) allocated from: #0 0x7f74422e8ea6 in __interceptor_calloc (/lib64/libasan.so.5+0x10dea6) #1 0x46402e in log_io_piece /usr/src/fio/iolog.c:214 Indirect leak of 171 byte(s) in 19 object(s) allocated from: #0 0x7f744227153d in strdup (/lib64/libasan.so.5+0x9653d) #1 0x44c4ed in add_to_dump_list /usr/src/fio/parse.c:1040 #2 0x44c4ed in parse_option /usr/src/fio/parse.c:1098 Indirect leak of 167 byte(s) in 19 object(s) allocated from: #0 0x7f744227153d in strdup (/lib64/libasan.so.5+0x9653d) #1 0x44c502 in add_to_dump_list /usr/src/fio/parse.c:1042 #2 0x44c502 in parse_option /usr/src/fio/parse.c:1098 Change-Id: I9b5811993508421be50b12af160645c77ea93d7e Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/456315 Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-05-30 07:29:06 +00:00
# Preload AddressSanitizer library to fio if fio_plugin was compiled with it
local asan_lib
asan_lib=$(ldd $bdev_plugin | grep libasan | awk '{print $3}')
test/asan: preload ASAN for fio ASAN needs to be LD_PRELOADed before SPDK fio_plugin in order to analyze its code. Just adding that will report any issues in fio binary as well as the fio_plugin. To prevent known fio leaks from affecting the results, a suppression list for LeakSanitizer (used in conjunction with ASAN). At this time the suppression list contains known leaks for fio 3.3. The list might need adjustments as fio version is updated. Side note. Even though it is possible to specify directory to ignore ("leak:/usr/src/fio/"). Which in theory should suppress any leaks in fio. It has side effect of hiding SPDK leaks as well, since the fio_plugins leaks are seen as coming from /usr/src/fio/ioengines.c. See below for examples of each suppressed error: Direct leak of 42 byte(s) in 4 object(s) allocated from: #0 0x7f9d52f3e320 in strdup (/lib64/libasan.so.5+0x3b320) #1 0x41f267 in get_new_job /usr/src/fio/init.c:490 Direct leak of 914936 byte(s) in 10397 object(s) allocated from: #0 0x7f74422e8ea6 in __interceptor_calloc (/lib64/libasan.so.5+0x10dea6) #1 0x46402e in log_io_piece /usr/src/fio/iolog.c:214 Direct leak of 608 byte(s) in 19 object(s) allocated from: #0 0x7f74422e8ca8 in __interceptor_malloc (/lib64/libasan.so.5+0x10dca8) #1 0x44c4e1 in add_to_dump_list /usr/src/fio/parse.c:1039 #2 0x44c4e1 in parse_option /usr/src/fio/parse.c:1098 Direct leak of 173 byte(s) in 20 object(s) allocated from: #0 0x7f744227153d in strdup (/lib64/libasan.so.5+0x9653d) #1 0x44b50d in __handle_option /usr/src/fio/parse.c:718 Indirect leak of 111925528 byte(s) in 1271881 object(s) allocated from: #0 0x7f74422e8ea6 in __interceptor_calloc (/lib64/libasan.so.5+0x10dea6) #1 0x46402e in log_io_piece /usr/src/fio/iolog.c:214 Indirect leak of 171 byte(s) in 19 object(s) allocated from: #0 0x7f744227153d in strdup (/lib64/libasan.so.5+0x9653d) #1 0x44c4ed in add_to_dump_list /usr/src/fio/parse.c:1040 #2 0x44c4ed in parse_option /usr/src/fio/parse.c:1098 Indirect leak of 167 byte(s) in 19 object(s) allocated from: #0 0x7f744227153d in strdup (/lib64/libasan.so.5+0x9653d) #1 0x44c502 in add_to_dump_list /usr/src/fio/parse.c:1042 #2 0x44c502 in parse_option /usr/src/fio/parse.c:1098 Change-Id: I9b5811993508421be50b12af160645c77ea93d7e Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/456315 Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-05-30 07:29:06 +00:00
LD_PRELOAD="$asan_lib $bdev_plugin" "$fio_dir"/fio "$@"
}
function fio_nvme() {
# Setup fio binary cmd line
local fio_dir=$CONFIG_FIO_SOURCE_DIR
local nvme_plugin="$rootdir/build/fio/spdk_nvme"
test/asan: preload ASAN for fio ASAN needs to be LD_PRELOADed before SPDK fio_plugin in order to analyze its code. Just adding that will report any issues in fio binary as well as the fio_plugin. To prevent known fio leaks from affecting the results, a suppression list for LeakSanitizer (used in conjunction with ASAN). At this time the suppression list contains known leaks for fio 3.3. The list might need adjustments as fio version is updated. Side note. Even though it is possible to specify directory to ignore ("leak:/usr/src/fio/"). Which in theory should suppress any leaks in fio. It has side effect of hiding SPDK leaks as well, since the fio_plugins leaks are seen as coming from /usr/src/fio/ioengines.c. See below for examples of each suppressed error: Direct leak of 42 byte(s) in 4 object(s) allocated from: #0 0x7f9d52f3e320 in strdup (/lib64/libasan.so.5+0x3b320) #1 0x41f267 in get_new_job /usr/src/fio/init.c:490 Direct leak of 914936 byte(s) in 10397 object(s) allocated from: #0 0x7f74422e8ea6 in __interceptor_calloc (/lib64/libasan.so.5+0x10dea6) #1 0x46402e in log_io_piece /usr/src/fio/iolog.c:214 Direct leak of 608 byte(s) in 19 object(s) allocated from: #0 0x7f74422e8ca8 in __interceptor_malloc (/lib64/libasan.so.5+0x10dca8) #1 0x44c4e1 in add_to_dump_list /usr/src/fio/parse.c:1039 #2 0x44c4e1 in parse_option /usr/src/fio/parse.c:1098 Direct leak of 173 byte(s) in 20 object(s) allocated from: #0 0x7f744227153d in strdup (/lib64/libasan.so.5+0x9653d) #1 0x44b50d in __handle_option /usr/src/fio/parse.c:718 Indirect leak of 111925528 byte(s) in 1271881 object(s) allocated from: #0 0x7f74422e8ea6 in __interceptor_calloc (/lib64/libasan.so.5+0x10dea6) #1 0x46402e in log_io_piece /usr/src/fio/iolog.c:214 Indirect leak of 171 byte(s) in 19 object(s) allocated from: #0 0x7f744227153d in strdup (/lib64/libasan.so.5+0x9653d) #1 0x44c4ed in add_to_dump_list /usr/src/fio/parse.c:1040 #2 0x44c4ed in parse_option /usr/src/fio/parse.c:1098 Indirect leak of 167 byte(s) in 19 object(s) allocated from: #0 0x7f744227153d in strdup (/lib64/libasan.so.5+0x9653d) #1 0x44c502 in add_to_dump_list /usr/src/fio/parse.c:1042 #2 0x44c502 in parse_option /usr/src/fio/parse.c:1098 Change-Id: I9b5811993508421be50b12af160645c77ea93d7e Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com> Reviewed-on: https://review.gerrithub.io/c/spdk/spdk/+/456315 Reviewed-by: Ben Walker <benjamin.walker@intel.com> Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com> Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
2019-05-30 07:29:06 +00:00
# Preload AddressSanitizer library to fio if fio_plugin was compiled with it
asan_lib=$(ldd $nvme_plugin | grep libasan | awk '{print $3}')
LD_PRELOAD="$asan_lib $nvme_plugin" "$fio_dir"/fio "$@"
}
function get_lvs_free_mb() {
local lvs_uuid=$1
local lvs_info
local fc
local cs
lvs_info=$($rpc_py bdev_lvol_get_lvstores)
fc=$(jq ".[] | select(.uuid==\"$lvs_uuid\") .free_clusters" <<< "$lvs_info")
cs=$(jq ".[] | select(.uuid==\"$lvs_uuid\") .cluster_size" <<< "$lvs_info")
# Change to MB's
free_mb=$((fc * cs / 1024 / 1024))
echo "$free_mb"
}
function get_bdev_size() {
local bdev_name=$1
local bdev_info
local bs
local nb
bdev_info=$($rpc_py bdev_get_bdevs -b $bdev_name)
bs=$(jq ".[] .block_size" <<< "$bdev_info")
nb=$(jq ".[] .num_blocks" <<< "$bdev_info")
# Change to MB's
bdev_size=$((bs * nb / 1024 / 1024))
echo "$bdev_size"
}
function autotest_cleanup() {
$rootdir/scripts/setup.sh reset
$rootdir/scripts/setup.sh cleanup
if [ $(uname -s) = "Linux" ]; then
if grep -q '#define SPDK_CONFIG_IGB_UIO_DRIVER 1' $rootdir/include/spdk/config.h; then
[[ -e /sys/module/igb_uio ]] && rmmod igb_uio
else
modprobe -r uio_pci_generic
fi
fi
rm -rf "$asan_suppression_file"
}
function freebsd_update_contigmem_mod() {
if [ $(uname) = FreeBSD ]; then
kldunload contigmem.ko || true
if [ -n "$WITH_DPDK_DIR" ]; then
echo "Warning: SPDK only works on FreeBSD with patches that only exist in SPDK's dpdk submodule"
cp -f "$WITH_DPDK_DIR/kmod/contigmem.ko" /boot/modules/
cp -f "$WITH_DPDK_DIR/kmod/contigmem.ko" /boot/kernel/
cp -f "$WITH_DPDK_DIR/kmod/nic_uio.ko" /boot/modules/
cp -f "$WITH_DPDK_DIR/kmod/nic_uio.ko" /boot/kernel/
else
cp -f "$rootdir/dpdk/build/kmod/contigmem.ko" /boot/modules/
cp -f "$rootdir/dpdk/build/kmod/contigmem.ko" /boot/kernel/
cp -f "$rootdir/dpdk/build/kmod/nic_uio.ko" /boot/modules/
cp -f "$rootdir/dpdk/build/kmod/nic_uio.ko" /boot/kernel/
fi
fi
}
function get_nvme_name_from_bdf() {
blkname=()
nvme_devs=$(lsblk -d --output NAME | grep "^nvme") || true
if [ -z "$nvme_devs" ]; then
return
fi
for dev in $nvme_devs; do
link_name=$(readlink /sys/block/$dev/device/device) || true
if [ -z "$link_name" ]; then
link_name=$(readlink /sys/block/$dev/device)
fi
bdf=$(basename "$link_name")
if [ "$bdf" = "$1" ]; then
blkname+=($dev)
fi
done
printf '%s\n' "${blkname[@]}"
}
function get_nvme_ctrlr_from_bdf() {
bdf_sysfs_path=$(readlink -f /sys/class/nvme/nvme* | grep "$1/nvme/nvme")
if [[ -z "$bdf_sysfs_path" ]]; then
return
fi
printf '%s\n' "$(basename $bdf_sysfs_path)"
}
function opal_revert_cleanup() {
$SPDK_BIN_DIR/spdk_tgt &
spdk_tgt_pid=$!
waitforlisten $spdk_tgt_pid
# OPAL test only runs on the first NVMe device
# So we just revert the first one here
bdf=$($rootdir/scripts/gen_nvme.sh --json | jq -r '.config[].params | select(.name=="Nvme0").traddr')
$rootdir/scripts/rpc.py bdev_nvme_attach_controller -b "nvme0" -t "pcie" -a $bdf
# Ignore if this fails.
$rootdir/scripts/rpc.py bdev_nvme_opal_revert -b nvme0 -p test || true
killprocess $spdk_tgt_pid
}
# Get BDF addresses of all NVMe drives currently attached to
# uio-pci-generic or vfio-pci
function get_nvme_bdfs() {
xtrace_disable
bdfs=$(jq -r .config[].params.traddr <<< $($rootdir/scripts/gen_nvme.sh --json))
if [[ -z $bdfs ]]; then
echo "No devices to test on!"
exit 1
fi
echo "$bdfs"
xtrace_restore
}
# Same as function above, but just get the first disks BDF address
function get_first_nvme_bdf() {
head -1 <<< "$(get_nvme_bdfs)"
}
function nvme_namespace_revert() {
$rootdir/scripts/setup.sh
sleep 1
bdfs=$(get_nvme_bdfs)
$rootdir/scripts/setup.sh reset
sleep 1
for bdf in $bdfs; do
nvme_ctrlr=/dev/$(get_nvme_ctrlr_from_bdf ${bdf})
if [[ -z "$nvme_ctrlr" ]]; then
continue
fi
# Check Optional Admin Command Support for Namespace Management
oacs=$(nvme id-ctrl ${nvme_ctrlr} | grep oacs | cut -d: -f2)
oacs_ns_manage=$((oacs & 0x8))
if [[ "$oacs_ns_manage" -ne 0 ]]; then
# This assumes every NVMe controller contains single namespace,
# encompassing Total NVM Capacity and formatted as 512 block size.
# 512 block size is needed for test/vhost/vhost_boot.sh to
# succesfully run.
unvmcap=$(nvme id-ctrl ${nvme_ctrlr} | grep unvmcap | cut -d: -f2)
if [[ "$unvmcap" -eq 0 ]]; then
# All available space already used
continue
fi
tnvmcap=$(nvme id-ctrl ${nvme_ctrlr} | grep tnvmcap | cut -d: -f2)
blksize=512
size=$((tnvmcap / blksize))
nvme detach-ns ${nvme_ctrlr} -n 0xffffffff -c 0 || true
nvme delete-ns ${nvme_ctrlr} -n 0xffffffff || true
nvme create-ns ${nvme_ctrlr} -s ${size} -c ${size} -b ${blksize}
nvme attach-ns ${nvme_ctrlr} -n 1 -c 0
nvme reset ${nvme_ctrlr}
waitforblk "${nvme_ctrlr}n1"
fi
done
}
# Define temp storage for all the tests. Look for 2GB at minimum
set_test_storage "${TEST_MIN_STORAGE_SIZE:-$((1 << 31))}"
set -o errtrace
shopt -s extdebug
trap "trap - ERR; print_backtrace >&2" ERR
PS4=' \t \$ '
if $SPDK_AUTOTEST_X; then
# explicitly enable xtraces, overriding any tracking information.
unset XTRACE_DISABLED
unset XTRACE_NESTING_LEVEL
set -x
xtrace_enable
else
xtrace_restore
fi