Do not enable stack tracer for ZFS performance test

Linux ZFS test suite runs with /proc/sys/kernel/stack_tracer_enabled=1,
via zfs.sh script, which has negative performance impact, up to 40%.

Since large stack is a rare issue now, preferred behavior would be:
- making stack tracer an opt-in feature for zfs.sh
- zfs-test.sh enables stack tracer only when requested

Reviewed-by: George Melikov <mail@gmelikov.ru>
Reviewed-by: Richard Elling <Richard.Elling@RichardElling.com>
Reviewed-by: John Kennedy <john.kennedy@delphix.com>
Signed-off-by: Tony Nguyen <tony.nguyen@delphix.com>
#8173
This commit is contained in:
Tony Nguyen
2018-12-07 11:51:42 -07:00
committed by Brian Behlendorf
parent d6496040d9
commit ef57371a92
2 changed files with 21 additions and 6 deletions
+8 -3
View File
@@ -14,6 +14,7 @@ fi
PROG=zfs.sh
VERBOSE="no"
UNLOAD="no"
STACK_TRACER="no"
ZED_PIDFILE=${ZED_PIDFILE:-/var/run/zed.pid}
LDMOD=${LDMOD:-/sbin/modprobe}
@@ -33,7 +34,7 @@ KMOD_ZFS=${KMOD_ZFS:-zfs}
usage() {
cat << EOF
USAGE:
$0 [hvud] [module-options]
$0 [hvudS] [module-options]
DESCRIPTION:
Load/unload the ZFS module stack.
@@ -42,10 +43,11 @@ OPTIONS:
-h Show this message
-v Verbose
-u Unload modules
-S Enable kernel stack tracer
EOF
}
while getopts 'hvu' OPTION; do
while getopts 'hvuS' OPTION; do
case $OPTION in
h)
usage
@@ -57,6 +59,9 @@ while getopts 'hvu' OPTION; do
u)
UNLOAD="yes"
;;
S)
STACK_TRACER="yes"
;;
?)
usage
exit
@@ -192,7 +197,7 @@ stack_clear() {
STACK_MAX_SIZE=/sys/kernel/debug/tracing/stack_max_size
STACK_TRACER_ENABLED=/proc/sys/kernel/stack_tracer_enabled
if [ -e "$STACK_MAX_SIZE" ]; then
if [ "$STACK_TRACER" = "yes" ] && [ -e "$STACK_MAX_SIZE" ]; then
echo 1 >"$STACK_TRACER_ENABLED"
echo 0 >"$STACK_MAX_SIZE"
fi