Cross-interpreter memory leak in _xidata_release when originating interpreter is destroyed
还没有人认领这个 Issue。
- 主要语言
- Python
- 星标
- 77.2k
- 派生
- 35.9k
- PR 合并指标
- PR 指标待抓取
描述
Bug report
Bug description:
Bug report
Bug summary
There is a known memory leak in Python/crossinterp.c related to the teardown of sub-interpreters when they share data.
When a sub-interpreter packages data into _PyXIData_t and passes it to a receiving interpreter, the receiver is expected to call _PyXIData_Release to free it. However, if the originating interpreter is destroyed before the receiver calls release, the original interpreter's memory context is lost.
In _xidata_release (Python/crossinterp.c around line 1049), the code correctly identifies this state:
PyInterpreterState *interp = _PyInterpreterState_LookUpID(
_PyXIData_INTERPID(xidata));
if (interp == NULL) {
// The interpreter was already destroyed.
// This function shouldn't have been called.
// XXX Someone leaked some memory...
Because the originating interpreter is gone, Python intentionally leaks xidata->obj to avoid a use-after-free segfault.
Proposed Solution Architecture
To resolve this without crashing, we would need to track cross-interpreter data at the interpreter level:
- Modify
PyInterpreterStateto include a thread-safe registry tracking all active_PyXIData_tinstances created by that interpreter. - Register instances during
_PyCode_GetXIDataand deregister them during_xidata_release. - Modify
Py_EndInterpreterteardown to check this registry. If outstanding shared data exists, the shutting-down interpreter must either block until it is released, or forcefully invalidate/free the shared data (using anis_invalidatedflag) before tearing down its ownobmallocstate.
I'm opening this issue to get feedback on the proposed architecture before starting a Pull Request.
CPython versions tested on:
main branch
CPython versions tested on:
CPython main branch
Operating systems tested on:
Windows
Linked PRs
- gh-153745
贡献指南
从这里开始
- 先读完整个 Issue,再读项目的贡献指南。
- 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
- Fork 仓库,在一个分支上完成修改。
- 提交 Pull Request,并在描述里引用这个 Issue 编号。
调研方向
从 Python/crossinterp.c 中的 _xidata_release 附近开始,然后跟踪 _PyCode_GetXIData 和 Py_EndInterpreter 的 teardown 路径。在继续之前查看链接的 gh-153745 工作,因为此 issue 请求架构方面的反馈,并未定义确定的实现方案或完成测试。
由索引模型根据 Issue 内容生成。
评估
- 技术栈
- c, python
- 领域
- backend
- Issue 类型
- 缺陷
- 难度
- 5/5
- 预计耗时
- 一周以上
- 活跃度
- 停滞
- 描述清晰度
- 需要澄清
- 新手友好度
- 25/100