feat(fcntl): implement F_SETSIG/F_GETSIG#1880
feat(fcntl): implement F_SETSIG/F_GETSIG#1880jiuyue486 wants to merge 2 commits intoDragonOS-Community:masterfrom
Conversation
Implement the F_SETSIG/F_GETSIG fcntl commands so userspace can pick the signal delivered for asynchronous I/O notifications. arg == 0 keeps the SIGIO default; values in (0, SIGRTMAX] override it. - Add a per-File `fasync_signum` (AtomicI32), preserved across `try_clone`. - Validate the raw `usize` arg against `SIGRTMAX` to avoid 64-bit values whose low 32 bits accidentally fall in [0, SIGRTMAX] from being silently accepted. Look up the fd first so EBADF wins over EINVAL, matching Linux `fs/fcntl.c`. - Cover defaults, set/get round-trip, the SIGRTMAX boundary, invalid values, and invalid fd in a new `fcntl_signal` dunitest.
github-actions
Bot
added
enhancement
| ra_state: Mutex<FileReadaheadState>, | ||
| /// 异步 I/O 通知信号编号（F_SETSIG/F_GETSIG）。 | ||
| /// 0 表示默认 SIGIO；范围 [0, SIGRTMAX]。 | ||
| fasync_signum: AtomicI32, |
There was a problem hiding this comment.
这个不应该直接放到 File 结构体下 用原子变量保护吧
应该做一个 FileOwner 结构体，和 pid 一起用 Mutex 保护
|
@codex review |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: e0b2863a4e
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| if arg > Signal::SIGRTMAX as usize { | ||
| return Err(SystemError::EINVAL); | ||
| } | ||
| file.set_fasync_signum(arg as i32); |
There was a problem hiding this comment.
F_SETSIG now returns success and persists arg via file.set_fasync_signum(...), but the notification path still hardcodes SIGIO_OR_POLL in FAsyncItems::send_sigio_to_process (kernel/src/filesystem/vfs/fasync.rs). As a result, programs that set a nonzero signal (especially realtime signals with SA_SIGINFO) will continue to receive SIGIO instead of the configured signal, so the new API behavior is observably incorrect despite passing F_GETSIG round-trips.
Useful? React with 👍 / 👎.
4 participants
Implement the F_SETSIG/F_GETSIG fcntl commands so userspace can pick
the signal delivered for asynchronous I/O notifications. arg == 0
keeps the SIGIO default; values in (0, SIGRTMAX] override it.
fasync_signum(AtomicI32), preserved acrosstry_clone.usizearg againstSIGRTMAXto avoid 64-bitvalues whose low 32 bits accidentally fall in [0, SIGRTMAX] from
being silently accepted. Look up the fd first so EBADF wins over
EINVAL, matching Linux
fs/fcntl.c.values, and invalid fd in a new
fcntl_signaldunitest.