2022-08-03 15:14:44

by Leo Yan

[permalink] [raw]
Subject: [PATCH v1] perf test: Introduce script for data symbol parsing

This commit introduces a shell script for data symbol parsing.

The testing is designed a data structure with 64-byte alignment, and it
has two fields "data1" and "data2", and other fields are reserved.

Using "perf mem" command, it records and reports memory samples for a
workload with 1 second duration. If have no any memory sample for the
data structure "buf1", it reports failure; and any memory accessing for
the data structure is not for "data1" and "data2" filed, it also means
the wrong data symbol parsing and returns failure.

Signed-off-by: Leo Yan <[email protected]>
---
tools/perf/tests/shell/test_data_symbol.sh | 94 ++++++++++++++++++++++
1 file changed, 94 insertions(+)
create mode 100755 tools/perf/tests/shell/test_data_symbol.sh

diff --git a/tools/perf/tests/shell/test_data_symbol.sh b/tools/perf/tests/shell/test_data_symbol.sh
new file mode 100755
index 000000000000..7039dae4e087
--- /dev/null
+++ b/tools/perf/tests/shell/test_data_symbol.sh
@@ -0,0 +1,94 @@
+#!/bin/bash
+# Check perf data symbol parsing
+
+# SPDX-License-Identifier: GPL-2.0
+# Leo Yan <[email protected]>, 2022
+
+skip_if_no_mem_event() {
+ perf mem record -e list 2>&1 | egrep -q 'available' && return 0
+ return 2
+}
+
+skip_if_no_mem_event || exit 2
+
+# skip if there's no compiler
+if ! [ -x "$(command -v cc)" ]; then
+ echo "skip: no compiler, install gcc"
+ exit 2
+fi
+
+TEST_PROGRAM_SOURCE=$(mktemp /tmp/__perf_test.program.XXXXX.c)
+TEST_PROGRAM=$(mktemp /tmp/__perf_test.program.XXXXX)
+PERF_DATA=$(mktemp /tmp/__perf_test.perf.data.XXXXX)
+
+check_result() {
+ # The memory report format is as below:
+ # 99.92% ... [.] buf1+0x38 ...
+ result=$(perf mem report -i ${PERF_DATA} --stdio 2>&1 | egrep "buf1" | \
+ awk '{ for (i = 1; i <= NF; i++) { if ($i ~ /^buf1/) { print $i; break; } } }')
+
+ # Testing is failed if has no any sample for "buf1"
+ [ -z "$result" ] && return 1
+
+ while IFS= read -r line; do
+ # The "data1" and "data2" fields in structure "buf1" have
+ # offset "0x0" and "0x38", returns failure if detect any
+ # other offset value.
+ if [ "$line" != "buf1+0x0" ] && [ "$line" != "buf1+0x38" ]; then
+ return 1
+ fi
+ done <<< "$result"
+
+ return 0
+}
+
+cleanup_files()
+{
+ echo "Cleaning up files..."
+ rm -f ${PERF_DATA}
+ rm -f ${TEST_PROGRAM_SOURCE}
+ rm -f ${TEST_PROGRAM}
+}
+
+trap cleanup_files exit term int
+
+# compile test program
+cat << EOF > $TEST_PROGRAM_SOURCE
+#include <stdio.h>
+#include <stdlib.h>
+#include <unistd.h>
+
+typedef struct _buf {
+ long data1;
+ long reserved[6];
+ long data2;
+} buf __attribute__((aligned (64)));
+
+static buf buf1;
+
+int main() {
+ int i;
+ for (;;) {
+ buf1.data1++;
+ buf1.data2 += buf1.data1;
+ }
+ return 0;
+}
+EOF
+
+echo "Compiling test program..."
+cc $TEST_PROGRAM_SOURCE -o $TEST_PROGRAM || exit 1
+
+echo "Recording workload..."
+perf mem record --all-user -o ${PERF_DATA} -- $TEST_PROGRAM &
+PERFPID=$!
+
+sleep 1
+
+kill $PERFPID
+wait $PERFPID
+# test program may leave an orphan process running the workload
+killall $(basename $TEST_PROGRAM)
+
+check_result "$variable_name"
+exit $?
--
2.25.1



2022-08-03 23:08:12

by Ian Rogers

[permalink] [raw]
Subject: Re: [PATCH v1] perf test: Introduce script for data symbol parsing

On Wed, Aug 3, 2022 at 8:07 AM Leo Yan <[email protected]> wrote:
>
> This commit introduces a shell script for data symbol parsing.
>
> The testing is designed a data structure with 64-byte alignment, and it
> has two fields "data1" and "data2", and other fields are reserved.
>
> Using "perf mem" command, it records and reports memory samples for a
> workload with 1 second duration. If have no any memory sample for the
> data structure "buf1", it reports failure; and any memory accessing for
> the data structure is not for "data1" and "data2" filed, it also means
> the wrong data symbol parsing and returns failure.
>
> Signed-off-by: Leo Yan <[email protected]>
> ---
> tools/perf/tests/shell/test_data_symbol.sh | 94 ++++++++++++++++++++++
> 1 file changed, 94 insertions(+)
> create mode 100755 tools/perf/tests/shell/test_data_symbol.sh
>
> diff --git a/tools/perf/tests/shell/test_data_symbol.sh b/tools/perf/tests/shell/test_data_symbol.sh
> new file mode 100755
> index 000000000000..7039dae4e087
> --- /dev/null
> +++ b/tools/perf/tests/shell/test_data_symbol.sh
> @@ -0,0 +1,94 @@
> +#!/bin/bash
> +# Check perf data symbol parsing
> +
> +# SPDX-License-Identifier: GPL-2.0
> +# Leo Yan <[email protected]>, 2022
> +
> +skip_if_no_mem_event() {
> + perf mem record -e list 2>&1 | egrep -q 'available' && return 0
> + return 2
> +}
> +
> +skip_if_no_mem_event || exit 2
> +
> +# skip if there's no compiler
> +if ! [ -x "$(command -v cc)" ]; then
> + echo "skip: no compiler, install gcc"
> + exit 2
> +fi
> +
> +TEST_PROGRAM_SOURCE=$(mktemp /tmp/__perf_test.program.XXXXX.c)
> +TEST_PROGRAM=$(mktemp /tmp/__perf_test.program.XXXXX)
> +PERF_DATA=$(mktemp /tmp/__perf_test.perf.data.XXXXX)
> +
> +check_result() {
> + # The memory report format is as below:
> + # 99.92% ... [.] buf1+0x38 ...
> + result=$(perf mem report -i ${PERF_DATA} --stdio 2>&1 | egrep "buf1" | \
> + awk '{ for (i = 1; i <= NF; i++) { if ($i ~ /^buf1/) { print $i; break; } } }')
> +
> + # Testing is failed if has no any sample for "buf1"
> + [ -z "$result" ] && return 1
> +
> + while IFS= read -r line; do
> + # The "data1" and "data2" fields in structure "buf1" have
> + # offset "0x0" and "0x38", returns failure if detect any
> + # other offset value.
> + if [ "$line" != "buf1+0x0" ] && [ "$line" != "buf1+0x38" ]; then
> + return 1
> + fi
> + done <<< "$result"
> +
> + return 0
> +}
> +
> +cleanup_files()
> +{
> + echo "Cleaning up files..."
> + rm -f ${PERF_DATA}
> + rm -f ${TEST_PROGRAM_SOURCE}
> + rm -f ${TEST_PROGRAM}
> +}
> +
> +trap cleanup_files exit term int
> +
> +# compile test program
> +cat << EOF > $TEST_PROGRAM_SOURCE
> +#include <stdio.h>
> +#include <stdlib.h>
> +#include <unistd.h>
> +
> +typedef struct _buf {
> + long data1;
> + long reserved[6];
> + long data2;
> +} buf __attribute__((aligned (64)));
> +
> +static buf buf1;
> +
> +int main() {
> + int i;
> + for (;;) {
> + buf1.data1++;
> + buf1.data2 += buf1.data1;
> + }
> + return 0;
> +}
> +EOF
> +
> +echo "Compiling test program..."
> +cc $TEST_PROGRAM_SOURCE -o $TEST_PROGRAM || exit 1
> +
> +echo "Recording workload..."
> +perf mem record --all-user -o ${PERF_DATA} -- $TEST_PROGRAM &
> +PERFPID=$!

I wonder here you could do:
TESTPID=$(ps -o pid -h --ppid "$PERFPID")

> +sleep 1
> +
> +kill $PERFPID
> +wait $PERFPID
> +# test program may leave an orphan process running the workload
> +killall $(basename $TEST_PROGRAM)

And then here kill $TESTPID to avoid the killall which worries me a little.

Thanks,
Ian

> +
> +check_result "$variable_name"
> +exit $?
> --
> 2.25.1
>

2022-08-04 01:40:57

by Leo Yan

[permalink] [raw]
Subject: Re: [PATCH v1] perf test: Introduce script for data symbol parsing

Hi Ian,

On Wed, Aug 03, 2022 at 04:06:04PM -0700, Ian Rogers wrote:

[...]

> > +echo "Compiling test program..."
> > +cc $TEST_PROGRAM_SOURCE -o $TEST_PROGRAM || exit 1
> > +
> > +echo "Recording workload..."
> > +perf mem record --all-user -o ${PERF_DATA} -- $TEST_PROGRAM &
> > +PERFPID=$!
>
> I wonder here you could do:
> TESTPID=$(ps -o pid -h --ppid "$PERFPID")

I tried this command and found "$TESTPID" is empty string. So the
test command doesn't create child process.

> > +sleep 1
> > +
> > +kill $PERFPID
> > +wait $PERFPID
> > +# test program may leave an orphan process running the workload
> > +killall $(basename $TEST_PROGRAM)
>
> And then here kill $TESTPID to avoid the killall which worries me a little.

Yeah, I referred the code from another script test_arm_spe_fork.sh,
this script doesn't fork new child process. So here I get over
complexity, will drop "killall $(basename $TEST_PROGRAM)" in next
version.

Thanks a lot for reviewing!

Leo