Fix double freed in remove_self_from_waiters()

[vicLin8712]

Fix double freed in remove_self_from_waiters()

free() is already invoked in list_remove(). Freeing the same pointer twice can crash the program with an invalid free error or corrupt the heap.

Ensure a single memory-release path by removing the outer free() call.

close #57

---

Summary by cubic

Removed redundant free in remove_self_from_waiters() to prevent a double-free when removing a waiter. list_remove() already frees the node, avoiding crashes and memory corruption in the mutex waiters list.

^{Written for commit 4948444. Summary will update automatically on new commits.}

[jserv]

Can you revisit codebase to eliminate the similar misuse?

[vicLin8712]

Can you revisit codebase to eliminate the similar misuse?

Hi, Jserv,

There is no remaining similar misuse after this PR merged. This PR is the last misuse case.

[jserv]

I defer to @HeatCrab for confirmation.

[HeatCrab]

Hi, I've reviewed the entire codebase for similar misuse patterns and confirmed that this is the only remaining instance that needs to be fixed.
LGTM.

[jserv]

Append Close #57 at the end of git commit message.

[visitorckw]

The code change itself LGTM.

I'm not sure if this description was generated by an LLM, but "prevent doubly function call' sounds a bit strange to me. Calling a function twice isn't an issue, whereas calling free() twice on the same pointer definitely is.

Also, did you omit the user impact from the commit message because it's already described in the issue? I think it's short enough to be included here. The commit message should be self-contained without requiring the reader to check external links.

One last thing: please avoid starting with "This commit". We know it's a commit. Please use the imperative mood instead.

[vicLin8712]

The code change itself LGTM.

I'm not sure if this description was generated by an LLM, but "prevent doubly function call' sounds a bit strange to me. Calling a function twice isn't an issue, whereas calling free() twice on the same pointer definitely is.

Also, did you omit the user impact from the commit message because it's already described in the issue? I think it's short enough to be included here. The commit message should be self-contained without requiring the reader to check external links.

One last thing: please avoid starting with "This commit". We know it's a commit. Please use the imperative mood instead.

Hi @visitorckw, thanks for your feedback, and I've updated the new commit message.

If you still have further suggestions, please let me know.

[jserv]

Thank @vicLin8712 for contributing!