Fix double-lock in sync_manager.rs

[BurtonQin]

There is a double-lock bug in fn request_now_from_peer: The first lock: https://github.com/Conflux-Chain/conflux-rust/blob/cc2c436a3bd76f62d196cff0d8845efbf1fa149d/core/src/light_protocol/handler/sync/common/sync_manager.rs#L291 Call insert_waiting: https://github.com/Conflux-Chain/conflux-rust/blob/cc2c436a3bd76f62d196cff0d8845efbf1fa149d/core/src/light_protocol/handler/sync/common/sync_manager.rs#L314 The second lock: https://github.com/Conflux-Chain/conflux-rust/blob/cc2c436a3bd76f62d196cff0d8845efbf1fa149d/core/src/light_protocol/handler/sync/common/sync_manager.rs#L146-L148

The fix is to add drop before calling insert_waiting.

Found the bug with my static detector lockbud.

---

This change is 

[Conflux-Dev]

Can one of the admins verify this patch?

[peilun-conflux]

ok to test

[peilun-conflux]

Thanks for pointing this out!

Releasing the lock here introduces a possible race condition, but as I understand the worse case is to trigger some duplicated requests, which is acceptable.

[BurtonQin]

Thanks for pointing this out!

Releasing the lock here introduces a possible race condition, but as I understand the worse case is to trigger some duplicated requests, which is acceptable.

If this is the case, then one possible solution is to replace the call to insert_waiting with a new fn taking &in_flight as a parameter and to require the lock before calling it.

[peilun-conflux]

Replaced by #2630.