diff options
Diffstat (limited to 'tools/perf/tests/shell/lib')
| -rw-r--r-- | tools/perf/tests/shell/lib/coresight.sh | 132 | ||||
| -rw-r--r-- | tools/perf/tests/shell/lib/perf_json_output_lint.py | 99 | ||||
| -rw-r--r-- | tools/perf/tests/shell/lib/perf_metric_validation.py | 574 | ||||
| -rw-r--r-- | tools/perf/tests/shell/lib/perf_metric_validation_rules.json | 398 | ||||
| -rw-r--r-- | tools/perf/tests/shell/lib/probe.sh | 13 | ||||
| -rw-r--r-- | tools/perf/tests/shell/lib/probe_vfs_getname.sh | 33 | ||||
| -rw-r--r-- | tools/perf/tests/shell/lib/stat_output.sh | 170 | ||||
| -rw-r--r-- | tools/perf/tests/shell/lib/waiting.sh | 78 |
8 files changed, 1497 insertions, 0 deletions
diff --git a/tools/perf/tests/shell/lib/coresight.sh b/tools/perf/tests/shell/lib/coresight.sh new file mode 100644 index 0000000000..6c3d34ec64 --- /dev/null +++ b/tools/perf/tests/shell/lib/coresight.sh @@ -0,0 +1,132 @@ +# SPDX-License-Identifier: GPL-2.0 +# Carsten Haitzler <carsten.haitzler@arm.com>, 2021 + +# This is sourced from a driver script so no need for #!/bin... etc. at the +# top - the assumption below is that it runs as part of sourcing after the +# test sets up some basic env vars to say what it is. + +# This currently works with ETMv4 / ETF not any other packet types at thi +# point. This will need changes if that changes. + +# perf record options for the perf tests to use +PERFRECMEM="-m ,16M" +PERFRECOPT="$PERFRECMEM -e cs_etm//u" + +TOOLS=$(dirname $0) +DIR="$TOOLS/$TEST" +BIN="$DIR/$TEST" +# If the test tool/binary does not exist and is executable then skip the test +if ! test -x "$BIN"; then exit 2; fi +DATD="." +# If the data dir env is set then make the data dir use that instead of ./ +if test -n "$PERF_TEST_CORESIGHT_DATADIR"; then + DATD="$PERF_TEST_CORESIGHT_DATADIR"; +fi +# If the stat dir env is set then make the data dir use that instead of ./ +STATD="." +if test -n "$PERF_TEST_CORESIGHT_STATDIR"; then + STATD="$PERF_TEST_CORESIGHT_STATDIR"; +fi + +# Called if the test fails - error code 1 +err() { + echo "$1" + exit 1 +} + +# Check that some statistics from our perf +check_val_min() { + STATF="$4" + if test "$2" -lt "$3"; then + echo ", FAILED" >> "$STATF" + err "Sanity check number of $1 is too low ($2 < $3)" + fi +} + +perf_dump_aux_verify() { + # Some basic checking that the AUX chunk contains some sensible data + # to see that we are recording something and at least a minimum + # amount of it. We should almost always see Fn packets in just about + # anything but certainly we will see some trace info and async + # packets + DUMP="$DATD/perf-tmp-aux-dump.txt" + perf report --stdio --dump -i "$1" | \ + grep -o -e I_ATOM_F -e I_ASYNC -e I_TRACE_INFO > "$DUMP" + # Simply count how many of these packets we find to see that we are + # producing a reasonable amount of data - exact checks are not sane + # as this is a lossy process where we may lose some blocks and the + # compiler may produce different code depending on the compiler and + # optimization options, so this is rough just to see if we're + # either missing almost all the data or all of it + ATOM_FX_NUM=$(grep -c I_ATOM_F "$DUMP") + ASYNC_NUM=$(grep -c I_ASYNC "$DUMP") + TRACE_INFO_NUM=$(grep -c I_TRACE_INFO "$DUMP") + rm -f "$DUMP" + + # Arguments provide minimums for a pass + CHECK_FX_MIN="$2" + CHECK_ASYNC_MIN="$3" + CHECK_TRACE_INFO_MIN="$4" + + # Write out statistics, so over time you can track results to see if + # there is a pattern - for example we have less "noisy" results that + # produce more consistent amounts of data each run, to see if over + # time any techinques to minimize data loss are having an effect or + # not + STATF="$STATD/stats-$TEST-$DATV.csv" + if ! test -f "$STATF"; then + echo "ATOM Fx Count, Minimum, ASYNC Count, Minimum, TRACE INFO Count, Minimum" > "$STATF" + fi + echo -n "$ATOM_FX_NUM, $CHECK_FX_MIN, $ASYNC_NUM, $CHECK_ASYNC_MIN, $TRACE_INFO_NUM, $CHECK_TRACE_INFO_MIN" >> "$STATF" + + # Actually check to see if we passed or failed. + check_val_min "ATOM_FX" "$ATOM_FX_NUM" "$CHECK_FX_MIN" "$STATF" + check_val_min "ASYNC" "$ASYNC_NUM" "$CHECK_ASYNC_MIN" "$STATF" + check_val_min "TRACE_INFO" "$TRACE_INFO_NUM" "$CHECK_TRACE_INFO_MIN" "$STATF" + echo ", Ok" >> "$STATF" +} + +perf_dump_aux_tid_verify() { + # Specifically crafted test will produce a list of Tread ID's to + # stdout that need to be checked to see that they have had trace + # info collected in AUX blocks in the perf data. This will go + # through all the TID's that are listed as CID=0xabcdef and see + # that all the Thread IDs the test tool reports are in the perf + # data AUX chunks + + # The TID test tools will print a TID per stdout line that are being + # tested + TIDS=$(cat "$2") + # Scan the perf report to find the TIDs that are actually CID in hex + # and build a list of the ones found + FOUND_TIDS=$(perf report --stdio --dump -i "$1" | \ + grep -o "CID=0x[0-9a-z]\+" | sed 's/CID=//g' | \ + uniq | sort | uniq) + # No CID=xxx found - maybe your kernel is reporting these as + # VMID=xxx so look there + if test -z "$FOUND_TIDS"; then + FOUND_TIDS=$(perf report --stdio --dump -i "$1" | \ + grep -o "VMID=0x[0-9a-z]\+" | sed 's/VMID=//g' | \ + uniq | sort | uniq) + fi + + # Iterate over the list of TIDs that the test says it has and find + # them in the TIDs found in the perf report + MISSING="" + for TID2 in $TIDS; do + FOUND="" + for TIDHEX in $FOUND_TIDS; do + TID=$(printf "%i" $TIDHEX) + if test "$TID" -eq "$TID2"; then + FOUND="y" + break + fi + done + if test -z "$FOUND"; then + MISSING="$MISSING $TID" + fi + done + if test -n "$MISSING"; then + err "Thread IDs $MISSING not found in perf AUX data" + fi +} diff --git a/tools/perf/tests/shell/lib/perf_json_output_lint.py b/tools/perf/tests/shell/lib/perf_json_output_lint.py new file mode 100644 index 0000000000..ea55d5ea1c --- /dev/null +++ b/tools/perf/tests/shell/lib/perf_json_output_lint.py @@ -0,0 +1,99 @@ +#!/usr/bin/python +# SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause) +# Basic sanity check of perf JSON output as specified in the man page. + +import argparse +import sys +import json + +ap = argparse.ArgumentParser() +ap.add_argument('--no-args', action='store_true') +ap.add_argument('--interval', action='store_true') +ap.add_argument('--system-wide-no-aggr', action='store_true') +ap.add_argument('--system-wide', action='store_true') +ap.add_argument('--event', action='store_true') +ap.add_argument('--per-core', action='store_true') +ap.add_argument('--per-thread', action='store_true') +ap.add_argument('--per-cache', action='store_true') +ap.add_argument('--per-die', action='store_true') +ap.add_argument('--per-node', action='store_true') +ap.add_argument('--per-socket', action='store_true') +ap.add_argument('--file', type=argparse.FileType('r'), default=sys.stdin) +args = ap.parse_args() + +Lines = args.file.readlines() + +def isfloat(num): + try: + float(num) + return True + except ValueError: + return False + + +def isint(num): + try: + int(num) + return True + except ValueError: + return False + +def is_counter_value(num): + return isfloat(num) or num == '<not counted>' or num == '<not supported>' + +def check_json_output(expected_items): + checks = { + 'aggregate-number': lambda x: isfloat(x), + 'core': lambda x: True, + 'counter-value': lambda x: is_counter_value(x), + 'cgroup': lambda x: True, + 'cpu': lambda x: isint(x), + 'cache': lambda x: True, + 'die': lambda x: True, + 'event': lambda x: True, + 'event-runtime': lambda x: isfloat(x), + 'interval': lambda x: isfloat(x), + 'metric-unit': lambda x: True, + 'metric-value': lambda x: isfloat(x), + 'metricgroup': lambda x: True, + 'node': lambda x: True, + 'pcnt-running': lambda x: isfloat(x), + 'socket': lambda x: True, + 'thread': lambda x: True, + 'unit': lambda x: True, + } + input = '[\n' + ','.join(Lines) + '\n]' + for item in json.loads(input): + if expected_items != -1: + count = len(item) + if count != expected_items and count >= 1 and count <= 6 and 'metric-value' in item: + # Events that generate >1 metric may have isolated metric + # values and possibly other prefixes like interval, core, + # aggregate-number, or event-runtime/pcnt-running from multiplexing. + pass + elif count != expected_items and count >= 1 and count <= 5 and 'metricgroup' in item: + pass + elif count != expected_items: + raise RuntimeError(f'wrong number of fields. counted {count} expected {expected_items}' + f' in \'{item}\'') + for key, value in item.items(): + if key not in checks: + raise RuntimeError(f'Unexpected key: key={key} value={value}') + if not checks[key](value): + raise RuntimeError(f'Check failed for: key={key} value={value}') + + +try: + if args.no_args or args.system_wide or args.event: + expected_items = 7 + elif args.interval or args.per_thread or args.system_wide_no_aggr: + expected_items = 8 + elif args.per_core or args.per_socket or args.per_node or args.per_die or args.per_cache: + expected_items = 9 + else: + # If no option is specified, don't check the number of items. + expected_items = -1 + check_json_output(expected_items) +except: + print('Test failed for input:\n' + '\n'.join(Lines)) + raise diff --git a/tools/perf/tests/shell/lib/perf_metric_validation.py b/tools/perf/tests/shell/lib/perf_metric_validation.py new file mode 100644 index 0000000000..50a34a9cc0 --- /dev/null +++ b/tools/perf/tests/shell/lib/perf_metric_validation.py @@ -0,0 +1,574 @@ +#SPDX-License-Identifier: GPL-2.0 +import re +import csv +import json +import argparse +from pathlib import Path +import subprocess + +class Validator: + def __init__(self, rulefname, reportfname='', t=5, debug=False, datafname='', fullrulefname='', workload='true', metrics=''): + self.rulefname = rulefname + self.reportfname = reportfname + self.rules = None + self.collectlist:str = metrics + self.metrics = self.__set_metrics(metrics) + self.skiplist = set() + self.tolerance = t + + self.workloads = [x for x in workload.split(",") if x] + self.wlidx = 0 # idx of current workloads + self.allresults = dict() # metric results of all workload + self.allignoremetrics = dict() # metrics with no results or negative results + self.allfailtests = dict() + self.alltotalcnt = dict() + self.allpassedcnt = dict() + self.allerrlist = dict() + + self.results = dict() # metric results of current workload + # vars for test pass/failure statistics + self.ignoremetrics= set() # metrics with no results or negative results, neg result counts as a failed test + self.failtests = dict() + self.totalcnt = 0 + self.passedcnt = 0 + # vars for errors + self.errlist = list() + + # vars for Rule Generator + self.pctgmetrics = set() # Percentage rule + + # vars for debug + self.datafname = datafname + self.debug = debug + self.fullrulefname = fullrulefname + + def __set_metrics(self, metrics=''): + if metrics != '': + return set(metrics.split(",")) + else: + return set() + + def read_json(self, filename: str) -> dict: + try: + with open(Path(filename).resolve(), "r") as f: + data = json.loads(f.read()) + except OSError as e: + print(f"Error when reading file {e}") + sys.exit() + + return data + + def json_dump(self, data, output_file): + parent = Path(output_file).parent + if not parent.exists(): + parent.mkdir(parents=True) + + with open(output_file, "w+") as output_file: + json.dump(data, + output_file, + ensure_ascii=True, + indent=4) + + def get_results(self, idx:int = 0): + return self.results[idx] + + def get_bounds(self, lb, ub, error, alias={}, ridx:int = 0) -> list: + """ + Get bounds and tolerance from lb, ub, and error. + If missing lb, use 0.0; missing ub, use float('inf); missing error, use self.tolerance. + + @param lb: str/float, lower bound + @param ub: str/float, upper bound + @param error: float/str, error tolerance + @returns: lower bound, return inf if the lower bound is a metric value and is not collected + upper bound, return -1 if the upper bound is a metric value and is not collected + tolerance, denormalized base on upper bound value + """ + # init ubv and lbv to invalid values + def get_bound_value (bound, initval, ridx): + val = initval + if isinstance(bound, int) or isinstance(bound, float): + val = bound + elif isinstance(bound, str): + if bound == '': + val = float("inf") + elif bound in alias: + vall = self.get_value(alias[ub], ridx) + if vall: + val = vall[0] + elif bound.replace('.', '1').isdigit(): + val = float(bound) + else: + print("Wrong bound: {0}".format(bound)) + else: + print("Wrong bound: {0}".format(bound)) + return val + + ubv = get_bound_value(ub, -1, ridx) + lbv = get_bound_value(lb, float('inf'), ridx) + t = get_bound_value(error, self.tolerance, ridx) + + # denormalize error threshold + denormerr = t * ubv / 100 if ubv != 100 and ubv > 0 else t + + return lbv, ubv, denormerr + + def get_value(self, name:str, ridx:int = 0) -> list: + """ + Get value of the metric from self.results. + If result of this metric is not provided, the metric name will be added into self.ignoremetics and self.errlist. + All future test(s) on this metric will fail. + + @param name: name of the metric + @returns: list with value found in self.results; list is empty when value is not found. + """ + results = [] + data = self.results[ridx] if ridx in self.results else self.results[0] + if name not in self.ignoremetrics: + if name in data: + results.append(data[name]) + elif name.replace('.', '1').isdigit(): + results.append(float(name)) + else: + self.ignoremetrics.add(name) + return results + + def check_bound(self, val, lb, ub, err): + return True if val <= ub + err and val >= lb - err else False + + # Positive Value Sanity check + def pos_val_test(self): + """ + Check if metrics value are non-negative. + One metric is counted as one test. + Failure: when metric value is negative or not provided. + Metrics with negative value will be added into the self.failtests['PositiveValueTest'] and self.ignoremetrics. + """ + negmetric = dict() + pcnt = 0 + tcnt = 0 + rerun = list() + for name, val in self.get_results().items(): + if val < 0: + negmetric[name] = val + rerun.append(name) + else: + pcnt += 1 + tcnt += 1 + if len(rerun) > 0 and len(rerun) < 20: + second_results = dict() + self.second_test(rerun, second_results) + for name, val in second_results.items(): + if name not in negmetric: continue + if val >= 0: + del negmetric[name] + pcnt += 1 + + self.failtests['PositiveValueTest']['Total Tests'] = tcnt + self.failtests['PositiveValueTest']['Passed Tests'] = pcnt + if len(negmetric.keys()): + self.ignoremetrics.update(negmetric.keys()) + negmessage = ["{0}(={1:.4f})".format(name, val) for name, val in negmetric.items()] + self.failtests['PositiveValueTest']['Failed Tests'].append({'NegativeValue': negmessage}) + + return + + def evaluate_formula(self, formula:str, alias:dict, ridx:int = 0): + """ + Evaluate the value of formula. + + @param formula: the formula to be evaluated + @param alias: the dict has alias to metric name mapping + @returns: value of the formula is success; -1 if the one or more metric value not provided + """ + stack = [] + b = 0 + errs = [] + sign = "+" + f = str() + + #TODO: support parenthesis? + for i in range(len(formula)): + if i+1 == len(formula) or formula[i] in ('+', '-', '*', '/'): + s = alias[formula[b:i]] if i+1 < len(formula) else alias[formula[b:]] + v = self.get_value(s, ridx) + if not v: + errs.append(s) + else: + f = f + "{0}(={1:.4f})".format(s, v[0]) + if sign == "*": + stack[-1] = stack[-1] * v + elif sign == "/": + stack[-1] = stack[-1] / v + elif sign == '-': + stack.append(-v[0]) + else: + stack.append(v[0]) + if i + 1 < len(formula): + sign = formula[i] + f += sign + b = i + 1 + + if len(errs) > 0: + return -1, "Metric value missing: "+','.join(errs) + + val = sum(stack) + return val, f + + # Relationships Tests + def relationship_test(self, rule: dict): + """ + Validate if the metrics follow the required relationship in the rule. + eg. lower_bound <= eval(formula)<= upper_bound + One rule is counted as ont test. + Failure: when one or more metric result(s) not provided, or when formula evaluated outside of upper/lower bounds. + + @param rule: dict with metric name(+alias), formula, and required upper and lower bounds. + """ + alias = dict() + for m in rule['Metrics']: + alias[m['Alias']] = m['Name'] + lbv, ubv, t = self.get_bounds(rule['RangeLower'], rule['RangeUpper'], rule['ErrorThreshold'], alias, ridx=rule['RuleIndex']) + val, f = self.evaluate_formula(rule['Formula'], alias, ridx=rule['RuleIndex']) + if val == -1: + self.failtests['RelationshipTest']['Failed Tests'].append({'RuleIndex': rule['RuleIndex'], 'Description':f}) + elif not self.check_bound(val, lbv, ubv, t): + lb = rule['RangeLower'] + ub = rule['RangeUpper'] + if isinstance(lb, str): + if lb in alias: + lb = alias[lb] + if isinstance(ub, str): + if ub in alias: + ub = alias[ub] + self.failtests['RelationshipTest']['Failed Tests'].append({'RuleIndex': rule['RuleIndex'], 'Formula':f, + 'RangeLower': lb, 'LowerBoundValue': self.get_value(lb), + 'RangeUpper': ub, 'UpperBoundValue':self.get_value(ub), + 'ErrorThreshold': t, 'CollectedValue': val}) + else: + self.passedcnt += 1 + self.failtests['RelationshipTest']['Passed Tests'] += 1 + self.totalcnt += 1 + self.failtests['RelationshipTest']['Total Tests'] += 1 + + return + + + # Single Metric Test + def single_test(self, rule:dict): + """ + Validate if the metrics are in the required value range. + eg. lower_bound <= metrics_value <= upper_bound + One metric is counted as one test in this type of test. + One rule may include one or more metrics. + Failure: when the metric value not provided or the value is outside the bounds. + This test updates self.total_cnt and records failed tests in self.failtest['SingleMetricTest']. + + @param rule: dict with metrics to validate and the value range requirement + """ + lbv, ubv, t = self.get_bounds(rule['RangeLower'], rule['RangeUpper'], rule['ErrorThreshold']) + metrics = rule['Metrics'] + passcnt = 0 + totalcnt = 0 + faillist = list() + failures = dict() + rerun = list() + for m in metrics: + totalcnt += 1 + result = self.get_value(m['Name']) + if len(result) > 0 and self.check_bound(result[0], lbv, ubv, t) or m['Name'] in self.skiplist: + passcnt += 1 + else: + failures[m['Name']] = result + rerun.append(m['Name']) + + if len(rerun) > 0 and len(rerun) < 20: + second_results = dict() + self.second_test(rerun, second_results) + for name, val in second_results.items(): + if name not in failures: continue + if self.check_bound(val, lbv, ubv, t): + passcnt += 1 + del failures[name] + else: + failures[name] = val + self.results[0][name] = val + + self.totalcnt += totalcnt + self.passedcnt += passcnt + self.failtests['SingleMetricTest']['Total Tests'] += totalcnt + self.failtests['SingleMetricTest']['Passed Tests'] += passcnt + if len(failures.keys()) != 0: + faillist = [{'MetricName':name, 'CollectedValue':val} for name, val in failures.items()] + self.failtests['SingleMetricTest']['Failed Tests'].append({'RuleIndex':rule['RuleIndex'], + 'RangeLower': rule['RangeLower'], + 'RangeUpper': rule['RangeUpper'], + 'ErrorThreshold':rule['ErrorThreshold'], + 'Failure':faillist}) + + return + + def create_report(self): + """ + Create final report and write into a JSON file. + """ + alldata = list() + for i in range(0, len(self.workloads)): + reportstas = {"Total Rule Count": self.alltotalcnt[i], "Passed Rule Count": self.allpassedcnt[i]} + data = {"Metric Validation Statistics": reportstas, "Tests in Category": self.allfailtests[i], + "Errors":self.allerrlist[i]} + alldata.append({"Workload": self.workloads[i], "Report": data}) + + json_str = json.dumps(alldata, indent=4) + print("Test validation finished. Final report: ") + print(json_str) + + if self.debug: + allres = [{"Workload": self.workloads[i], "Results": self.allresults[i]} for i in range(0, len(self.workloads))] + self.json_dump(allres, self.datafname) + + def check_rule(self, testtype, metric_list): + """ + Check if the rule uses metric(s) that not exist in current platform. + + @param metric_list: list of metrics from the rule. + @return: False when find one metric out in Metric file. (This rule should not skipped.) + True when all metrics used in the rule are found in Metric file. + """ + if testtype == "RelationshipTest": + for m in metric_list: + if m['Name'] not in self.metrics: + return False + return True + + # Start of Collector and Converter + def convert(self, data: list, metricvalues:dict): + """ + Convert collected metric data from the -j output to dict of {metric_name:value}. + """ + for json_string in data: + try: + result =json.loads(json_string) + if "metric-unit" in result and result["metric-unit"] != "(null)" and result["metric-unit"] != "": + name = result["metric-unit"].split(" ")[1] if len(result["metric-unit"].split(" ")) > 1 \ + else result["metric-unit"] + metricvalues[name.lower()] = float(result["metric-value"]) + except ValueError as error: + continue + return + + def _run_perf(self, metric, workload: str): + tool = 'perf' + command = [tool, 'stat', '-j', '-M', f"{metric}", "-a"] + wl = workload.split() + command.extend(wl) + print(" ".join(command)) + cmd = subprocess.run(command, stderr=subprocess.PIPE, encoding='utf-8') + data = [x+'}' for x in cmd.stderr.split('}\n') if x] + return data + + + def collect_perf(self, workload: str): + """ + Collect metric data with "perf stat -M" on given workload with -a and -j. + """ + self.results = dict() + print(f"Starting perf collection") + print(f"Long workload: {workload}") + collectlist = dict() + if self.collectlist != "": + collectlist[0] = {x for x in self.collectlist.split(",")} + else: + collectlist[0] = set(list(self.metrics)) + # Create metric set for relationship rules + for rule in self.rules: + if rule["TestType"] == "RelationshipTest": + metrics = [m["Name"] for m in rule["Metrics"]] + if not any(m not in collectlist[0] for m in metrics): + collectlist[rule["RuleIndex"]] = [",".join(list(set(metrics)))] + + for idx, metrics in collectlist.items(): + if idx == 0: wl = "true" + else: wl = workload + for metric in metrics: + data = self._run_perf(metric, wl) + if idx not in self.results: self.results[idx] = dict() + self.convert(data, self.results[idx]) + return + + def second_test(self, collectlist, second_results): + workload = self.workloads[self.wlidx] + for metric in collectlist: + data = self._run_perf(metric, workload) + self.convert(data, second_results) + + # End of Collector and Converter + + # Start of Rule Generator + def parse_perf_metrics(self): + """ + Read and parse perf metric file: + 1) find metrics with '1%' or '100%' as ScaleUnit for Percent check + 2) create metric name list + """ + command = ['perf', 'list', '-j', '--details', 'metrics'] + cmd = subprocess.run(command, stdout=subprocess.PIPE, stderr=subprocess.PIPE, encoding='utf-8') + try: + data = json.loads(cmd.stdout) + for m in data: + if 'MetricName' not in m: + print("Warning: no metric name") + continue + name = m['MetricName'].lower() + self.metrics.add(name) + if 'ScaleUnit' in m and (m['ScaleUnit'] == '1%' or m['ScaleUnit'] == '100%'): + self.pctgmetrics.add(name.lower()) + except ValueError as error: + print(f"Error when parsing metric data") + sys.exit() + + return + + def remove_unsupported_rules(self, rules): + new_rules = [] + for rule in rules: + add_rule = True + for m in rule["Metrics"]: + if m["Name"] in self.skiplist or m["Name"] not in self.metrics: + add_rule = False + break + if add_rule: + new_rules.append(rule) + return new_rules + + def create_rules(self): + """ + Create full rules which includes: + 1) All the rules from the "relationshi_rules" file + 2) SingleMetric rule for all the 'percent' metrics + + Reindex all the rules to avoid repeated RuleIndex + """ + data = self.read_json(self.rulefname) + rules = data['RelationshipRules'] + self.skiplist = set([name.lower() for name in data['SkipList']]) + self.rules = self.remove_unsupported_rules(rules) + pctgrule = {'RuleIndex':0, + 'TestType':'SingleMetricTest', + 'RangeLower':'0', + 'RangeUpper': '100', + 'ErrorThreshold': self.tolerance, + 'Description':'Metrics in percent unit have value with in [0, 100]', + 'Metrics': [{'Name': m.lower()} for m in self.pctgmetrics]} + self.rules.append(pctgrule) + + # Re-index all rules to avoid repeated RuleIndex + idx = 1 + for r in self.rules: + r['RuleIndex'] = idx + idx += 1 + + if self.debug: + #TODO: need to test and generate file name correctly + data = {'RelationshipRules':self.rules, 'SupportedMetrics': [{"MetricName": name} for name in self.metrics]} + self.json_dump(data, self.fullrulefname) + + return + # End of Rule Generator + + def _storewldata(self, key): + ''' + Store all the data of one workload into the corresponding data structure for all workloads. + @param key: key to the dictionaries (index of self.workloads). + ''' + self.allresults[key] = self.results + self.allignoremetrics[key] = self.ignoremetrics + self.allfailtests[key] = self.failtests + self.alltotalcnt[key] = self.totalcnt + self.allpassedcnt[key] = self.passedcnt + self.allerrlist[key] = self.errlist + + #Initialize data structures before data validation of each workload + def _init_data(self): + + testtypes = ['PositiveValueTest', 'RelationshipTest', 'SingleMetricTest'] + self.results = dict() + self.ignoremetrics= set() + self.errlist = list() + self.failtests = {k:{'Total Tests':0, 'Passed Tests':0, 'Failed Tests':[]} for k in testtypes} + self.totalcnt = 0 + self.passedcnt = 0 + + def test(self): + ''' + The real entry point of the test framework. + This function loads the validation rule JSON file and Standard Metric file to create rules for + testing and namemap dictionaries. + It also reads in result JSON file for testing. + + In the test process, it passes through each rule and launch correct test function bases on the + 'TestType' field of the rule. + + The final report is written into a JSON file. + ''' + if not self.collectlist: + self.parse_perf_metrics() + self.create_rules() + for i in range(0, len(self.workloads)): + self.wlidx = i + self._init_data() + self.collect_perf(self.workloads[i]) + # Run positive value test + self.pos_val_test() + for r in self.rules: + # skip rules that uses metrics not exist in this platform + testtype = r['TestType'] + if not self.check_rule(testtype, r['Metrics']): + continue + if testtype == 'RelationshipTest': + self.relationship_test(r) + elif testtype == 'SingleMetricTest': + self.single_test(r) + else: + print("Unsupported Test Type: ", testtype) + self.errlist.append("Unsupported Test Type from rule: " + r['RuleIndex']) + self._storewldata(i) + print("Workload: ", self.workloads[i]) + print("Total metrics collected: ", self.failtests['PositiveValueTest']['Total Tests']) + print("Non-negative metric count: ", self.failtests['PositiveValueTest']['Passed Tests']) + print("Total Test Count: ", self.totalcnt) + print("Passed Test Count: ", self.passedcnt) + + self.create_report() + return sum(self.alltotalcnt.values()) != sum(self.allpassedcnt.values()) +# End of Class Validator + + +def main() -> None: + parser = argparse.ArgumentParser(description="Launch metric value validation") + + parser.add_argument("-rule", help="Base validation rule file", required=True) + parser.add_argument("-output_dir", help="Path for validator output file, report file", required=True) + parser.add_argument("-debug", help="Debug run, save intermediate data to files", action="store_true", default=False) + parser.add_argument("-wl", help="Workload to run while data collection", default="true") + parser.add_argument("-m", help="Metric list to validate", default="") + args = parser.parse_args() + outpath = Path(args.output_dir) + reportf = Path.joinpath(outpath, 'perf_report.json') + fullrule = Path.joinpath(outpath, 'full_rule.json') + datafile = Path.joinpath(outpath, 'perf_data.json') + + validator = Validator(args.rule, reportf, debug=args.debug, + datafname=datafile, fullrulefname=fullrule, workload=args.wl, + metrics=args.m) + ret = validator.test() + + return ret + + +if __name__ == "__main__": + import sys + sys.exit(main()) + + + diff --git a/tools/perf/tests/shell/lib/perf_metric_validation_rules.json b/tools/perf/tests/shell/lib/perf_metric_validation_rules.json new file mode 100644 index 0000000000..eb6f59e018 --- /dev/null +++ b/tools/perf/tests/shell/lib/perf_metric_validation_rules.json @@ -0,0 +1,398 @@ +{ + "SkipList": [ + "tsx_aborted_cycles", + "tsx_transactional_cycles", + "C2_Pkg_Residency", + "C6_Pkg_Residency", + "C1_Core_Residency", + "C6_Core_Residency", + "tma_false_sharing", + "tma_remote_cache", + "tma_contested_accesses" + ], + "RelationshipRules": [ + { + "RuleIndex": 1, + "Formula": "a+b", + "TestType": "RelationshipTest", + "RangeLower": "c", + "RangeUpper": "c", + "ErrorThreshold": 5.0, + "Description": "Intel(R) Optane(TM) Persistent Memory(PMEM) bandwidth total includes Intel(R) Optane(TM) Persistent Memory(PMEM) read bandwidth and Intel(R) Optane(TM) Persistent Memory(PMEM) write bandwidth", + "Metrics": [ + { + "Name": "pmem_memory_bandwidth_read", + "Alias": "a" + }, + { + "Name": "pmem_memory_bandwidth_write", + "Alias": "b" + }, + { + "Name": "pmem_memory_bandwidth_total", + "Alias": "c" + } + ] + }, + { + "RuleIndex": 2, + "Formula": "a+b", + "TestType": "RelationshipTest", + "RangeLower": "c", + "RangeUpper": "c", + "ErrorThreshold": 5.0, + "Description": "DDR memory bandwidth total includes DDR memory read bandwidth and DDR memory write bandwidth", + "Metrics": [ + { + "Name": "memory_bandwidth_read", + "Alias": "a" + }, + { + "Name": "memory_bandwidth_write", + "Alias": "b" + }, + { + "Name": "memory_bandwidth_total", + "Alias": "c" + } + ] + }, + { + "RuleIndex": 3, + "Formula": "a+b", + "TestType": "RelationshipTest", + "RangeLower": "100", + "RangeUpper": "100", + "ErrorThreshold": 5.0, + "Description": "Total memory read accesses includes memory reads from last level cache (LLC) addressed to local DRAM and memory reads from the last level cache (LLC) addressed to remote DRAM.", + "Metrics": [ + { + "Name": "numa_reads_addressed_to_local_dram", + "Alias": "a" + }, + { + "Name": "numa_reads_addressed_to_remote_dram", + "Alias": "b" + } + ] + }, + { + "RuleIndex": 4, + "Formula": "a", + "TestType": "SingleMetricTest", + "RangeLower": "0.125", + "RangeUpper": "", + "ErrorThreshold": "", + "Description": "", + "Metrics": [ + { + "Name": "cpi", + "Alias": "a" + } + ] + }, + { + "RuleIndex": 5, + "Formula": "", + "TestType": "SingleMetricTest", + "RangeLower": "0", + "RangeUpper": "1", + "ErrorThreshold": 5.0, + "Description": "Ratio values should be within value range [0,1)", + "Metrics": [ + { + "Name": "loads_per_instr", + "Alias": "" + }, + { + "Name": "stores_per_instr", + "Alias": "" + }, + { + "Name": "l1d_mpi", + "Alias": "" + }, + { + "Name": "l1d_demand_data_read_hits_per_instr", + "Alias": "" + }, + { + "Name": "l1_i_code_read_misses_with_prefetches_per_instr", + "Alias": "" + }, + { + "Name": "l2_demand_data_read_hits_per_instr", + "Alias": "" + }, + { + "Name": "l2_mpi", + "Alias": "" + }, + { + "Name": "l2_demand_data_read_mpi", + "Alias": "" + }, + { + "Name": "l2_demand_code_mpi", + "Alias": "" + } + ] + }, + { + "RuleIndex": 6, + "Formula": "a+b+c+d", + "TestType": "RelationshipTest", + "RangeLower": "100", + "RangeUpper": "100", + "ErrorThreshold": 5.0, + "Description": "Sum of TMA level 1 metrics should be 100%", + "Metrics": [ + { + "Name": "tma_frontend_bound", + "Alias": "a" + }, + { + "Name": "tma_bad_speculation", + "Alias": "b" + }, + { + "Name": "tma_backend_bound", + "Alias": "c" + }, + { + "Name": "tma_retiring", + "Alias": "d" + } + ] + }, + { + "RuleIndex": 7, + "Formula": "a+b", + "TestType": "RelationshipTest", + "RangeLower": "c", + "RangeUpper": "c", + "ErrorThreshold": 5.0, + "Description": "Sum of the level 2 children should equal level 1 parent", + "Metrics": [ + { + "Name": "tma_fetch_latency", + "Alias": "a" + }, + { + "Name": "tma_fetch_bandwidth", + "Alias": "b" + }, + { + "Name": "tma_frontend_bound", + "Alias": "c" + } + ] + }, + { + "RuleIndex": 8, + "Formula": "a+b", + "TestType": "RelationshipTest", + "RangeLower": "c", + "RangeUpper": "c", + "ErrorThreshold": 5.0, + "Description": "Sum of the level 2 children should equal level 1 parent", + "Metrics": [ + { + "Name": "tma_branch_mispredicts", + "Alias": "a" + }, + { + "Name": "tma_machine_clears", + "Alias": "b" + }, + { + "Name": "tma_bad_speculation", + "Alias": "c" + } + ] + }, + { + "RuleIndex": 9, + "Formula": "a+b", + "TestType": "RelationshipTest", + "RangeLower": "c", + "RangeUpper": "c", + "ErrorThreshold": 5.0, + "Description": "Sum of the level 2 children should equal level 1 parent", + "Metrics": [ + { + "Name": "tma_memory_bound", + "Alias": "a" + }, + { + "Name": "tma_core_bound", + "Alias": "b" + }, + { + "Name": "tma_backend_bound", + "Alias": "c" + } + ] + }, + { + "RuleIndex": 10, + "Formula": "a+b", + "TestType": "RelationshipTest", + "RangeLower": "c", + "RangeUpper": "c", + "ErrorThreshold": 5.0, + "Description": "Sum of the level 2 children should equal level 1 parent", + "Metrics": [ + { + "Name": "tma_light_operations", + "Alias": "a" + }, + { + "Name": "tma_heavy_operations", + "Alias": "b" + }, + { + "Name": "tma_retiring", + "Alias": "c" + } + ] + }, + { + "RuleIndex": 11, + "Formula": "a+b+c", + "TestType": "RelationshipTest", + "RangeLower": "100", + "RangeUpper": "100", + "ErrorThreshold": 5.0, + "Description": "The all_requests includes the memory_page_empty, memory_page_misses, and memory_page_hits equals.", + "Metrics": [ + { + "Name": "memory_page_empty_vs_all_requests", + "Alias": "a" + }, + { + "Name": "memory_page_misses_vs_all_requests", + "Alias": "b" + }, + { + "Name": "memory_page_hits_vs_all_requests", + "Alias": "c" + } + ] + }, + { + "RuleIndex": 12, + "Formula": "a-b", + "TestType": "RelationshipTest", + "RangeLower": "0", + "RangeUpper": "", + "ErrorThreshold": 5.0, + "Description": "CPU utilization in kernel mode should always be <= cpu utilization", + "Metrics": [ + { + "Name": "cpu_utilization", + "Alias": "a" + }, + { + "Name": "cpu_utilization_in_kernel_mode", + "Alias": "b" + } + ] + }, + { + "RuleIndex": 13, + "Formula": "a-b", + "TestType": "RelationshipTest", + "RangeLower": "0", + "RangeUpper": "", + "ErrorThreshold": 5.0, + "Description": "Total L2 misses per instruction should be >= L2 demand data read misses per instruction", + "Metrics": [ + { + "Name": "l2_mpi", + "Alias": "a" + }, + { + "Name": "l2_demand_data_read_mpi", + "Alias": "b" + } + ] + }, + { + "RuleIndex": 14, + "Formula": "a-b", + "TestType": "RelationshipTest", + "RangeLower": "0", + "RangeUpper": "", + "ErrorThreshold": 5.0, + "Description": "Total L2 misses per instruction should be >= L2 demand code misses per instruction", + "Metrics": [ + { + "Name": "l2_mpi", + "Alias": "a" + }, + { + "Name": "l2_demand_code_mpi", + "Alias": "b" + } + ] + }, + { + "RuleIndex": 15, + "Formula": "b+c+d", + "TestType": "RelationshipTest", + "RangeLower": "a", + "RangeUpper": "a", + "ErrorThreshold": 5.0, + "Description": "L3 data read, rfo, code misses per instruction equals total L3 misses per instruction.", + "Metrics": [ + { + "Name": "llc_mpi", + "Alias": "a" + }, + { + "Name": "llc_data_read_mpi_demand_plus_prefetch", + "Alias": "b" + }, + { + "Name": "llc_rfo_read_mpi_demand_plus_prefetch", + "Alias": "c" + }, + { + "Name": "llc_code_read_mpi_demand_plus_prefetch", + "Alias": "d" + } + ] + }, + { + "RuleIndex": 16, + "Formula": "a", + "TestType": "SingleMetricTest", + "RangeLower": "0", + "RangeUpper": "8", + "ErrorThreshold": 0.0, + "Description": "Setting generous range for allowable frequencies", + "Metrics": [ + { + "Name": "uncore_freq", + "Alias": "a" + } + ] + }, + { + "RuleIndex": 17, + "Formula": "a", + "TestType": "SingleMetricTest", + "RangeLower": "0", + "RangeUpper": "8", + "ErrorThreshold": 0.0, + "Description": "Setting generous range for allowable frequencies", + "Metrics": [ + { + "Name": "cpu_operating_frequency", + "Alias": "a" + } + ] + } + ] +}
\ No newline at end of file diff --git a/tools/perf/tests/shell/lib/probe.sh b/tools/perf/tests/shell/lib/probe.sh new file mode 100644 index 0000000000..5aa6e2ec57 --- /dev/null +++ b/tools/perf/tests/shell/lib/probe.sh @@ -0,0 +1,13 @@ +#!/bin/bash +# SPDX-License-Identifier: GPL-2.0 +# Arnaldo Carvalho de Melo <acme@kernel.org>, 2017 + +skip_if_no_perf_probe() { + perf probe 2>&1 | grep -q 'is not a perf-command' && return 2 + return 0 +} + +skip_if_no_perf_trace() { + perf trace -h 2>&1 | grep -q -e 'is not a perf-command' -e 'trace command not available' && return 2 + return 0 +} diff --git a/tools/perf/tests/shell/lib/probe_vfs_getname.sh b/tools/perf/tests/shell/lib/probe_vfs_getname.sh new file mode 100644 index 0000000000..bf4c1fb71c --- /dev/null +++ b/tools/perf/tests/shell/lib/probe_vfs_getname.sh @@ -0,0 +1,33 @@ +#!/bin/sh +# Arnaldo Carvalho de Melo <acme@kernel.org>, 2017 + +perf probe -l 2>&1 | grep -q probe:vfs_getname +had_vfs_getname=$? + +cleanup_probe_vfs_getname() { + if [ $had_vfs_getname -eq 1 ] ; then + perf probe -q -d probe:vfs_getname* + fi +} + +add_probe_vfs_getname() { + add_probe_verbose=$1 + if [ $had_vfs_getname -eq 1 ] ; then + line=$(perf probe -L getname_flags 2>&1 | grep -E 'result.*=.*filename;' | sed -r 's/[[:space:]]+([[:digit:]]+)[[:space:]]+result->uptr.*/\1/') + perf probe -q "vfs_getname=getname_flags:${line} pathname=result->name:string" || \ + perf probe $add_probe_verbose "vfs_getname=getname_flags:${line} pathname=filename:ustring" + fi +} + +skip_if_no_debuginfo() { + add_probe_vfs_getname -v 2>&1 | grep -E -q "^(Failed to find the path for the kernel|Debuginfo-analysis is not supported)|(file has no debug information)" && return 2 + return 1 +} + +# check if perf is compiled with libtraceevent support +skip_no_probe_record_support() { + if [ $had_vfs_getname -eq 1 ] ; then + perf record --dry-run -e $1 2>&1 | grep "libtraceevent is necessary for tracepoint support" && return 2 + return 1 + fi +} diff --git a/tools/perf/tests/shell/lib/stat_output.sh b/tools/perf/tests/shell/lib/stat_output.sh new file mode 100644 index 0000000000..3cc158a643 --- /dev/null +++ b/tools/perf/tests/shell/lib/stat_output.sh @@ -0,0 +1,170 @@ +#!/bin/bash +# SPDX-License-Identifier: GPL-2.0 + +# Return true if perf_event_paranoid is > $1 and not running as root. +function ParanoidAndNotRoot() +{ + [ "$(id -u)" != 0 ] && [ "$(cat /proc/sys/kernel/perf_event_paranoid)" -gt $1 ] +} + +# $1 name $2 extra_opt +check_no_args() +{ + echo -n "Checking $1 output: no args " + perf stat $2 true + commachecker --no-args + echo "[Success]" +} + +check_system_wide() +{ + echo -n "Checking $1 output: system wide " + if ParanoidAndNotRoot 0 + then + echo "[Skip] paranoid and not root" + return + fi + perf stat -a $2 true + commachecker --system-wide + echo "[Success]" +} + +check_system_wide_no_aggr() +{ + echo -n "Checking $1 output: system wide no aggregation " + if ParanoidAndNotRoot 0 + then + echo "[Skip] paranoid and not root" + return + fi + perf stat -A -a --no-merge $2 true + commachecker --system-wide-no-aggr + echo "[Success]" +} + +check_interval() +{ + echo -n "Checking $1 output: interval " + perf stat -I 1000 $2 true + commachecker --interval + echo "[Success]" +} + +check_event() +{ + echo -n "Checking $1 output: event " + perf stat -e cpu-clock $2 true + commachecker --event + echo "[Success]" +} + +check_per_core() +{ + echo -n "Checking $1 output: per core " + if ParanoidAndNotRoot 0 + then + echo "[Skip] paranoid and not root" + return + fi + perf stat --per-core -a $2 true + commachecker --per-core + echo "[Success]" +} + +check_per_thread() +{ + echo -n "Checking $1 output: per thread " + if ParanoidAndNotRoot 0 + then + echo "[Skip] paranoid and not root" + return + fi + perf stat --per-thread -a $2 true + commachecker --per-thread + echo "[Success]" +} + +check_per_cache_instance() +{ + echo -n "Checking $1 output: per cache instance " + if ParanoidAndNotRoot 0 + then + echo "[Skip] paranoid and not root" + return + fi + perf stat --per-cache -a $2 true + commachecker --per-cache + echo "[Success]" +} + +check_per_die() +{ + echo -n "Checking $1 output: per die " + if ParanoidAndNotRoot 0 + then + echo "[Skip] paranoid and not root" + return + fi + perf stat --per-die -a $2 true + commachecker --per-die + echo "[Success]" +} + +check_per_node() +{ + echo -n "Checking $1 output: per node " + if ParanoidAndNotRoot 0 + then + echo "[Skip] paranoid and not root" + return + fi + perf stat --per-node -a $2 true + commachecker --per-node + echo "[Success]" +} + +check_per_socket() +{ + echo -n "Checking $1 output: per socket " + if ParanoidAndNotRoot 0 + then + echo "[Skip] paranoid and not root" + return + fi + perf stat --per-socket -a $2 true + commachecker --per-socket + echo "[Success]" +} + +# The perf stat options for per-socket, per-core, per-die +# and -A ( no_aggr mode ) uses the info fetched from this +# directory: "/sys/devices/system/cpu/cpu*/topology". For +# example, socket value is fetched from "physical_package_id" +# file in topology directory. +# Reference: cpu__get_topology_int in util/cpumap.c +# If the platform doesn't expose topology information, values +# will be set to -1. For example, incase of pSeries platform +# of powerpc, value for "physical_package_id" is restricted +# and set to -1. Check here validates the socket-id read from +# topology file before proceeding further + +FILE_LOC="/sys/devices/system/cpu/cpu*/topology/" +FILE_NAME="physical_package_id" + +function check_for_topology() +{ + if ! ParanoidAndNotRoot 0 + then + socket_file=`ls $FILE_LOC/$FILE_NAME | head -n 1` + [ -z $socket_file ] && { + echo 0 + return + } + socket_id=`cat $socket_file` + [ $socket_id == -1 ] && { + echo 1 + return + } + fi + echo 0 +} diff --git a/tools/perf/tests/shell/lib/waiting.sh b/tools/perf/tests/shell/lib/waiting.sh new file mode 100644 index 0000000000..bdd5a7c715 --- /dev/null +++ b/tools/perf/tests/shell/lib/waiting.sh @@ -0,0 +1,78 @@ +#!/bin/sh +# SPDX-License-Identifier: GPL-2.0 + +tenths=date\ +%s%1N + +# Wait for PID $1 to have $2 number of threads started +# Time out after $3 tenths of a second or 5 seconds if $3 is "" +wait_for_threads() +{ + tm_out=$3 ; [ -n "${tm_out}" ] || tm_out=50 + start_time=$($tenths) + while [ -e "/proc/$1/task" ] ; do + th_cnt=$(find "/proc/$1/task" -mindepth 1 -maxdepth 1 -printf x | wc -c) + if [ "${th_cnt}" -ge "$2" ] ; then + return 0 + fi + # Wait at most tm_out tenths of a second + if [ $(($($tenths) - start_time)) -ge $tm_out ] ; then + echo "PID $1 does not have $2 threads" + return 1 + fi + done + return 1 +} + +# Wait for perf record -vvv 2>$2 with PID $1 to start by looking at file $2 +# It depends on capturing perf record debug message "perf record has started" +# Time out after $3 tenths of a second or 5 seconds if $3 is "" +wait_for_perf_to_start() +{ + tm_out=$3 ; [ -n "${tm_out}" ] || tm_out=50 + echo "Waiting for \"perf record has started\" message" + start_time=$($tenths) + while [ -e "/proc/$1" ] ; do + if grep -q "perf record has started" "$2" ; then + echo OK + break + fi + # Wait at most tm_out tenths of a second + if [ $(($($tenths) - start_time)) -ge $tm_out ] ; then + echo "perf recording did not start" + return 1 + fi + done + return 0 +} + +# Wait for process PID %1 to exit +# Time out after $2 tenths of a second or 5 seconds if $2 is "" +wait_for_process_to_exit() +{ + tm_out=$2 ; [ -n "${tm_out}" ] || tm_out=50 + start_time=$($tenths) + while [ -e "/proc/$1" ] ; do + # Wait at most tm_out tenths of a second + if [ $(($($tenths) - start_time)) -ge $tm_out ] ; then + echo "PID $1 did not exit as expected" + return 1 + fi + done + return 0 +} + +# Check if PID $1 is still running after $2 tenths of a second +# or 0.3 seconds if $2 is "" +is_running() +{ + tm_out=$2 ; [ -n "${tm_out}" ] || tm_out=3 + start_time=$($tenths) + while [ -e "/proc/$1" ] ; do + # Check for at least tm_out tenths of a second + if [ $(($($tenths) - start_time)) -gt $tm_out ] ; then + return 0 + fi + done + echo "PID $1 exited prematurely" + return 1 +} |