Re: [PATCH v2] selftests/amd-pstate: Fix broken paths to run workloads in amd-pstate-ut

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hello Mario,

Thanks for taking a look at this patch.

On 9/23/2023 3:52 AM, Mario Limonciello wrote:
On 9/19/2023 05:33, Swapnil Sapkal wrote:
In selftests/amd-pstate, tbench and gitsource microbenchmarks are used to
compare the performance with different governors. In Current
implementation relative path to run `amd_pstate_tracer.py`
broken. Fixed this by using absolute paths.
Also selftests/amd-pstate uses distro `perf` to capture stats while running
these microbenchmarks. Distro `perf` is not working with upstream
kernel. Fixed this by providing an option to give the perf binary path.


This should be multiple separate patches, at least two.

Sure, I will update this in v3.

Signed-off-by: Swapnil Sapkal <swapnil.sapkal@xxxxxxx>
---
  .../x86/amd_pstate_tracer/amd_pstate_trace.py |  2 +-
  .../testing/selftests/amd-pstate/gitsource.sh | 14 +++++++-----
  tools/testing/selftests/amd-pstate/run.sh     | 22 +++++++++++++------
  tools/testing/selftests/amd-pstate/tbench.sh  |  4 ++--
  4 files changed, 27 insertions(+), 15 deletions(-)

diff --git a/tools/power/x86/amd_pstate_tracer/amd_pstate_trace.py b/tools/power/x86/amd_pstate_tracer/amd_pstate_trace.py
index 904df0ea0a1e..2448bb07973f 100755
--- a/tools/power/x86/amd_pstate_tracer/amd_pstate_trace.py
+++ b/tools/power/x86/amd_pstate_tracer/amd_pstate_trace.py
@@ -30,7 +30,7 @@ import getopt
  import Gnuplot
  from numpy import *
  from decimal import *
-sys.path.append('../intel_pstate_tracer')
+sys.path.append(os.path.join(os.path.dirname(__file__), '../intel_pstate_tracer'))
  #import intel_pstate_tracer
  import intel_pstate_tracer as ipt
diff --git a/tools/testing/selftests/amd-pstate/gitsource.sh b/tools/testing/selftests/amd-pstate/gitsource.sh
index 5f2171f0116d..c327444d3506 100755
--- a/tools/testing/selftests/amd-pstate/gitsource.sh
+++ b/tools/testing/selftests/amd-pstate/gitsource.sh
@@ -66,12 +66,15 @@ post_clear_gitsource()
  install_gitsource()
  {
-    if [ ! -d $git_name ]; then
+    if [ ! -d $SCRIPTDIR/$git_name ]; then
+        BACKUP_DIR=$(pwd)
+        cd $SCRIPTDIR
          printf "Download gitsource, please wait a moment ...\n\n"
          wget -O $git_tar $gitsource_url > /dev/null 2>&1
          printf "Tar gitsource ...\n\n"
          tar -xzf $git_tar
+        cd $BACKUP_DIR
      fi
  }
@@ -79,12 +82,13 @@ install_gitsource()
  run_gitsource()
  {
      echo "Launching amd pstate tracer for $1 #$2 tracer_interval: $TRACER_INTERVAL"
-    ./amd_pstate_trace.py -n tracer-gitsource-$1-$2 -i $TRACER_INTERVAL > /dev/null 2>&1 &
+    $SCRIPTDIR/../../../power/x86/amd_pstate_tracer/amd_pstate_trace.py -n tracer-gitsource-$1-$2 -i $TRACER_INTERVAL > /dev/null 2>&1 &
      printf "Make and test gitsource for $1 #$2 make_cpus: $MAKE_CPUS\n"
-    cd $git_name
-    perf stat -a --per-socket -I 1000 -e power/energy-pkg/ /usr/bin/time -o ../$OUTFILE_GIT.time-gitsource-$1-$2.log make test -j$MAKE_CPUS > ../$OUTFILE_GIT-perf-$1-$2.log 2>&1
-    cd ..
+    BACKUP_DIR=$(pwd)
+    cd $SCRIPTDIR/$git_name
+    $PERF stat -a --per-socket -I 1000 -e power/energy-pkg/ /usr/bin/time -o $BACKUP_DIR/$OUTFILE_GIT.time-gitsource-$1-$2.log make test -j$MAKE_CPUS > $BACKUP_DIR/$OUTFILE_GIT-perf-$1-$2.log 2>&1
+    cd $BACKUP_DIR
      for job in `jobs -p`
      do
diff --git a/tools/testing/selftests/amd-pstate/run.sh b/tools/testing/selftests/amd-pstate/run.sh
index de4d8e9c9565..0803e70b04da 100755
--- a/tools/testing/selftests/amd-pstate/run.sh
+++ b/tools/testing/selftests/amd-pstate/run.sh
@@ -8,9 +8,11 @@ else
      FILE_MAIN=DONE
  fi
-source basic.sh
-source tbench.sh
-source gitsource.sh
+SCRIPTDIR=`dirname "$0"`
+
+source $SCRIPTDIR/basic.sh
+source $SCRIPTDIR/tbench.sh
+source $SCRIPTDIR/gitsource.sh

Rather than hardcoding $SCRIPTDIR/../../../power/x86/amd_pstate_tracer/amd_pstate_trace.py in multiple places, how about you delcare a new variable here?

TRACER=$SCRIPTDIR/../../../power/x86/amd_pstate_tracer/amd_pstate_trace.py

Then all the commands can be called with $TRACER instead.

Furthermore if there needs to be flexibility for multiple calling paths in the future (say in tree vs in filessytem) then you can easily add another case right here that is:

if [ ! -f $TRACER ]; then
     TRACER=foo/bar/baz
fi

This is a good suggestion. I will update this in v3.

  # amd-pstate-ut only run on x86/x86_64 AMD systems.
  ARCH=$(uname -m 2>/dev/null | sed -e 's/i.86/x86/' -e 's/x86_64/x86/')
@@ -22,6 +24,7 @@ OUTFILE=selftest
  OUTFILE_TBENCH="$OUTFILE.tbench"
  OUTFILE_GIT="$OUTFILE.gitsource"
+PERF=/usr/bin/perf
  SYSFS=
  CPUROOT=
  CPUFREQROOT=
@@ -149,8 +152,9 @@ help()
           gitsource: Gitsource testing.>]
      [-t <tbench time limit>]
      [-p <tbench process number>]
-    [-l <loop times for tbench>]
+    [-l <loop times for tbench/gitsource>]
      [-i <amd tracer interval>]
+    [-b <perf binary>]
      [-m <comparative test: acpi-cpufreq>]
      \n"
      exit 2
@@ -158,7 +162,7 @@ help()
  parse_arguments()
  {
-    while getopts ho:c:t:p:l:i:m: arg
+    while getopts ho:c:t:p:l:i:b:m: arg
      do
          case $arg in
              h) # --help
@@ -189,6 +193,10 @@ parse_arguments()
                  TRACER_INTERVAL=$OPTARG
                  ;;
+            b) # --perf-binary
+                PERF=`realpath $OPTARG`
+                ;;
+
              m) # --comparative-test
                  COMPARATIVE_TEST=$OPTARG
                  ;;
@@ -202,8 +210,8 @@ parse_arguments()
  command_perf()
  {
-    if ! command -v perf > /dev/null; then
-        echo $msg please install perf. >&2
+    if ! $PERF -v; then
+        echo $msg please install perf or provide perf binary path as argument >&2
          exit $ksft_skip
      fi
  }
diff --git a/tools/testing/selftests/amd-pstate/tbench.sh b/tools/testing/selftests/amd-pstate/tbench.sh
index 49c9850341f6..70e5863e74ea 100755
--- a/tools/testing/selftests/amd-pstate/tbench.sh
+++ b/tools/testing/selftests/amd-pstate/tbench.sh
@@ -64,11 +64,11 @@ post_clear_tbench()
  run_tbench()
  {
      echo "Launching amd pstate tracer for $1 #$2 tracer_interval: $TRACER_INTERVAL"
-    ./amd_pstate_trace.py -n tracer-tbench-$1-$2 -i $TRACER_INTERVAL > /dev/null 2>&1 &
+    $SCRIPTDIR/../../../power/x86/amd_pstate_tracer/amd_pstate_trace.py -n tracer-tbench-$1-$2 -i $TRACER_INTERVAL > /dev/null 2>&1 &
      printf "Test tbench for $1 #$2 time_limit: $TIME_LIMIT procs_num: $PROCESS_NUM\n"
      tbench_srv > /dev/null 2>&1 &
-    perf stat -a --per-socket -I 1000 -e power/energy-pkg/ tbench -t $TIME_LIMIT $PROCESS_NUM > $OUTFILE_TBENCH-perf-$1-$2.log 2>&1
+    $PERF stat -a --per-socket -I 1000 -e power/energy-pkg/ tbench -t $TIME_LIMIT $PROCESS_NUM > $OUTFILE_TBENCH-perf-$1-$2.log 2>&1
      pid=`pidof tbench_srv`
      kill $pid

--
Thanks and Regards,
Swapnil



[Index of Archives]     [Linux Wireless]     [Linux Kernel]     [ATH6KL]     [Linux Bluetooth]     [Linux Netdev]     [Kernel Newbies]     [Share Photos]     [IDE]     [Security]     [Git]     [Netfilter]     [Bugtraq]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux ATA RAID]     [Samba]     [Device Mapper]

  Powered by Linux