Ascend/msagent / skills/operator/mssanitizer-basic-usage
mssanitizer-basic-usage
Verified昇腾AscendC算子异常检测工具(msSanitizer)。检测内存越界、内存泄漏、数据竞争、未初始化读取、同步异常等缺陷。触发关键词:mssanitizer、sanitizer、检测工具、内存检测、竞争检测、初始化检测、同步检测
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 24.7 KB |
| evals/evals.json | config | 6.8 KB |
| references/alarms.md | markdown | 29.6 KB |
| references/official_workflow.md | markdown | 25.1 KB |
| references/sample_initcheck/CMakeLists.txt | markdown | 1.2 KB |
| references/sample_initcheck/README.md | markdown | 4.8 KB |
| references/sample_initcheck/initcheck.asc | other | 2.3 KB |
| references/sample_memcheck/CMakeLists.txt | markdown | 1.2 KB |
| references/sample_memcheck/README.md | markdown | 4.9 KB |
| references/sample_memcheck/memcheck.asc | other | 2.4 KB |
| references/sample_racecheck/CMakeLists.txt | markdown | 1.2 KB |
| references/sample_racecheck/README.md | markdown | 6.6 KB |
| references/sample_racecheck/racecheck.asc | other | 2.5 KB |
| references/sample_synccheck/CMakeLists.txt | markdown | 1.2 KB |
| references/sample_synccheck/README.md | markdown | 4.1 KB |
| references/sample_synccheck/synccheck.asc | other | 2.8 KB |
| scripts/parse_mssanitizer_report.py | script | 19.0 KB |
| scripts/run_mssanitizer.py | script | 12.3 KB |
scripts/run_mssanitizer.py
View on GitHub#!/usr/bin/env python3
# -*- coding: utf-8 -*-
# -----------------------------------------------------------------------------------------------------------
# Copyright (c) 2026 Huawei Technologies Co., Ltd.
# This program is free software, you can redistribute it and/or modify it under the terms and conditions of
# CANN Open Software License Agreement Version 2.0 (the "License").
# Please refer to the License for details. You may not use this file except in compliance with the License.
# THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, EITHER EXPRESS OR IMPLIED,
# INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE.
# See LICENSE in the root of the software repository for the full text of the License.
# -----------------------------------------------------------------------------------------------------------
"""run_mssanitizer.py — 批量运行 msSanitizer 四种检测工具
用法:
python run_mssanitizer.py [--output-dir <目录>] [--extra-args "<mssanitizer参数>"] -- <用户自定义命令>
示例:
python run_mssanitizer.py -- ./execute_add_op
python run_mssanitizer.py --output-dir ./add_example -- bash build.sh --run_example AddCustom
python run_mssanitizer.py --extra-args "--leak-check=yes" -- ./execute_add_op
python run_mssanitizer.py -- bash run.sh
python run_mssanitizer.py -- python test_add.py
脚本会依次执行 memcheck → racecheck → initcheck → synccheck,
每个检测的输出实时打印到终端并保存到带时间戳的 .txt 文件。
报告头部会记录真实命令、核心参数与环境信息,供解析脚本生成最终报告。
注意:
- 请在算子仓库根目录下执行本脚本:mssanitizer 会将运行明细日志写入
当前工作目录的 mindstudio_sanitizer_log/,result.log 等产物同样落在
当前工作目录。
- 通过 --output-dir 指定原始报告输出目录(建议为算子目录),默认为当前工作目录。
"""
import logging
import os
import shlex
import subprocess
import sys
from dataclasses import dataclass
from datetime import datetime, timezone
from typing import List, Optional, Tuple
TOOLS = [
{"name": "memcheck", "desc": "内存检测"},
{"name": "racecheck", "desc": "竞争检测"},
{"name": "initcheck", "desc": "未初始化检测"},
{"name": "synccheck", "desc": "同步检测"},
]
@dataclass
class RunContext:
"""检测运行共享上下文"""
output_dir: str
user_cmd: List[str]
user_cmd_str: str
extra_args_strip: str
timestamp: str
env_line: str
log_dir: str
def now() -> datetime:
"""返回带本地时区信息的当前时间"""
return datetime.now(timezone.utc).astimezone()
def now_str(fmt: str = "%Y-%m-%d %H:%M:%S") -> str:
"""返回格式化后的当前时间字符串"""
return now().strftime(fmt)
def _run_capture(argv: List[str], timeout: int = 30) -> str:
"""执行命令并返回标准输出,失败返回空字符串"""
try:
result = subprocess.run(
argv, shell=False, capture_output=True, text=True,
errors="replace", timeout=timeout,
)
return result.stdout.strip()
except (OSError, subprocess.SubprocessError):
return ""
def _read_first_line(path: str) -> str:
"""读取文件首个非空行,失败返回空字符串"""
try:
with open(path, "r", encoding="utf-8", errors="replace") as f:
lines = f.readlines()
except OSError:
return ""
for line in lines:
if line.strip():
return line.strip()
return ""
def collect_env_info() -> dict:
"""尽力采集环境信息,采集失败的项目记为 unknown"""
env = {}
ms_ver = _run_capture(["mssanitizer", "-v"])
if "command not found" in ms_ver or not ms_ver:
ms_ver = "unknown"
env["mssanitizer"] = ms_ver.replace("\n", " / ")
# CANN 版本:依次尝试常见 version.info 位置
cann_ver = ""
candidates = [
os.path.join(os.environ.get("ASCEND_HOME_PATH", ""), "version.info"),
"/usr/local/Ascend/cann/version.info",
"/usr/local/Ascend/version.info",
]
for p in candidates:
if p and os.path.isfile(p):
cann_ver = _read_first_line(p)
if cann_ver:
break
env["cann"] = cann_ver or "unknown"
# 芯片型号:优先 acl 接口,失败记为 unknown
soc = _run_capture(
["python3", "-c", "import acl; print(acl.get_soc_name())"]
)
env["npu"] = soc.splitlines()[0] if soc else "unknown"
env["host"] = _run_capture(["hostname"]) or "unknown"
env["cwd"] = os.getcwd()
env["date"] = now_str()
return env
def parse_args(argv: List[str]) -> Optional[Tuple[str, str, List[str]]]:
"""解析 --output-dir / --extra-args 与 -- 之后的用户命令,不合法返回 None"""
output_dir = None
extra_args = ""
try:
sep_idx = argv.index("--")
except ValueError:
logging.error("错误: 未找到 '--' 分隔符")
logging.error("用法: python run_mssanitizer.py [--output-dir <目录>] "
"[--extra-args \"<mssanitizer参数>\"] -- <用户自定义命令>")
logging.error("示例: python run_mssanitizer.py --output-dir ./add_example -- ./execute_add_op")
return None
pre_args = argv[:sep_idx]
user_cmd = argv[sep_idx + 1:]
# 解析 --output-dir / --extra-args
i = 0
while i < len(pre_args):
if pre_args[i] == "--output-dir" and i + 1 < len(pre_args):
output_dir = pre_args[i + 1]
i += 2
elif pre_args[i] == "--extra-args" and i + 1 < len(pre_args):
extra_args = pre_args[i + 1]
i += 2
else:
logging.warning("警告: 未知参数 — %s", pre_args[i])
i += 1
if not user_cmd:
logging.error("错误: '--' 后未提供用户命令")
logging.error("示例: python run_mssanitizer.py -- ./execute_add_op")
return None
return output_dir, extra_args, user_cmd
def build_real_cmd_str(tool_name: str, extra_args_strip: str, user_cmd_str: str) -> str:
"""构造报告头部记录的真实检测命令字符串"""
cmd = f"mssanitizer --tool={tool_name}"
if extra_args_strip:
cmd += f" {extra_args_strip}"
cmd += f" -- {user_cmd_str}"
return cmd
def build_run_cmd(tool_name: str, extra_args_strip: str, user_cmd: List[str]) -> List[str]:
"""构造最终 mssanitizer 运行命令(shell=False 列表形式)"""
cmd = ["mssanitizer", f"--tool={tool_name}"]
if extra_args_strip:
cmd.extend(shlex.split(extra_args_strip))
cmd.append("--")
cmd.extend(user_cmd)
return cmd
def write_report_header(report_file: str, tool_name: str, real_cmd_str: str,
ctx: RunContext) -> None:
"""将运行命令、核心参数与环境信息写入报告文件头部"""
params = f"--tool={tool_name}"
if ctx.extra_args_strip:
params += f" {ctx.extra_args_strip}"
with open(report_file, "w", encoding="utf-8") as f:
f.write(f"[MSSANITIZER_CMD] {ctx.user_cmd_str}\n")
f.write(f"[MSSANITIZER_REAL_CMD] {real_cmd_str}\n")
f.write(f"[MSSANITIZER_TOOL] {tool_name}\n")
f.write(f"[MSSANITIZER_PARAMS] {params}\n")
f.write(f"[MSSANITIZER_ENV] {ctx.env_line}\n")
f.write(f"[MSSANITIZER_LOGDIR] {ctx.log_dir}\n")
def _stream_output(cmd: List[str], report_file: str) -> int:
"""运行命令,实时将 stdout/stderr 输出到终端并追加到报告文件,返回退出码"""
proc = subprocess.Popen(
cmd, shell=False, stdout=subprocess.PIPE,
stderr=subprocess.STDOUT, text=True, errors="replace",
)
with open(report_file, "a", encoding="utf-8") as f:
for line in proc.stdout:
logging.info("%s", line.rstrip("\n"))
f.write(line)
proc.wait()
return proc.returncode
def count_alarms(report_file: str) -> Tuple[int, int]:
"""统计报告文件中的 ERROR / WARNING 数量"""
error_count = 0
warn_count = 0
if not os.path.exists(report_file):
return error_count, warn_count
with open(report_file, "r", encoding="utf-8", errors="replace") as f:
for line in f:
if line.startswith("====== ERROR:") or line.startswith("[mssanitizer] Error:"):
error_count += 1
elif line.startswith("====== WARNING:") or line.startswith("[mssanitizer] Warning:"):
warn_count += 1
return error_count, warn_count
def run_single_tool(index: int, tool: dict, ctx: RunContext) -> Tuple[str, int, int]:
"""运行单个检测工具,返回 (报告文件路径, ERROR 数, WARNING 数)"""
tool_name = tool["name"]
report_file = os.path.join(ctx.output_dir, f"mssanitizer_origin_{tool_name}_{ctx.timestamp}.txt")
real_cmd_str = build_real_cmd_str(tool_name, ctx.extra_args_strip, ctx.user_cmd_str)
logging.info("[%d/4] %s (%s)", index, tool["desc"], tool_name)
logging.info(" 报告文件: %s", report_file)
logging.info(" 真实命令: %s", real_cmd_str)
logging.info("-" * 40)
write_report_header(report_file, tool_name, real_cmd_str, ctx)
run_cmd = build_run_cmd(tool_name, ctx.extra_args_strip, ctx.user_cmd)
try:
exit_code = _stream_output(run_cmd, report_file)
except FileNotFoundError:
logging.error("错误: 未找到 mssanitizer 命令,请确认工具已安装")
logging.error(" 跳过 %s...", tool["desc"])
return report_file, 0, 0
error_count, warn_count = count_alarms(report_file)
logging.info(" 退出码: %s", exit_code)
logging.info(" ERROR: %d WARNING: %d", error_count, warn_count)
return report_file, error_count, warn_count
def print_summary(total_errors: int, total_warnings: int,
report_files: List[str], log_dir: str) -> None:
"""打印检测完成汇总信息"""
logging.info("=" * 60)
logging.info(" 检测完成")
logging.info(" 结束时间: %s", now_str())
logging.info(" 总计 ERROR: %d WARNING: %d", total_errors, total_warnings)
logging.info("")
logging.info(" 生成的报告文件:")
for report_file in report_files:
if os.path.exists(report_file):
logging.info(" %s (%d bytes)", report_file, os.path.getsize(report_file))
else:
logging.info(" %s (未生成)", report_file)
logging.info("")
logging.info(" 运行明细日志目录: %s", log_dir)
logging.info(" (result.log 仅在手动运行并重定向时生成,同样位于执行目录)")
logging.info("=" * 60)
def main() -> None:
"""主流程"""
logging.basicConfig(level=logging.INFO, format="%(message)s", stream=sys.stdout)
parsed = parse_args(sys.argv[1:])
if parsed is None:
sys.exit(1)
output_dir, extra_args, user_cmd = parsed
if output_dir:
os.makedirs(output_dir, exist_ok=True)
else:
output_dir = os.getcwd()
user_cmd_str = " ".join(user_cmd)
timestamp = now_str("%Y%m%d_%H%M%S")
env_info = collect_env_info()
env_line = "; ".join(f"{k}={v}" for k, v in env_info.items())
log_dir = os.path.join(os.getcwd(), "mindstudio_sanitizer_log")
extra_args_strip = extra_args.strip()
logging.info("=" * 60)
logging.info(" msSanitizer 全量检测")
logging.info(" 用户命令: %s", user_cmd_str)
if extra_args_strip:
logging.info(" 附加参数: %s", extra_args_strip)
logging.info(" 执行目录: %s", env_info["cwd"])
logging.info(" 开始时间: %s", env_info["date"])
logging.info(" 明细日志目录: %s", log_dir)
logging.info("=" * 60)
logging.info("")
ctx = RunContext(
output_dir=output_dir,
user_cmd=user_cmd,
user_cmd_str=user_cmd_str,
extra_args_strip=extra_args_strip,
timestamp=timestamp,
env_line=env_line,
log_dir=log_dir,
)
total_errors = 0
total_warnings = 0
report_files = []
for i, tool in enumerate(TOOLS, 1):
report_file, error_count, warn_count = run_single_tool(i, tool, ctx)
report_files.append(report_file)
total_errors += error_count
total_warnings += warn_count
print_summary(total_errors, total_warnings, report_files, log_dir)
if __name__ == "__main__":
main()
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 193k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 193k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 193k | Updated | Claude CodeCodex | Verified | Install |