út 15. 11. 2022 v 13:56 odesílatel Maurizio Lombardi
<mlombard@xxxxxxxxxx> napsal:
>
>
> +/**
> + * RETURN VALUE:
> + *
> + * 1 = Login successful
> + * -1 = Login failed
> + * 0 = More PDU exchanges required
> + */
> static int iscsi_target_do_login(struct iscsit_conn *conn, struct iscsi_login *login)
> {
> int pdu_count = 0;
> @@ -1363,12 +1370,13 @@ int iscsi_target_start_negotiation(
> ret = -1;
>
> if (ret < 0) {
> - cancel_delayed_work_sync(&conn->login_work);
> iscsi_target_restore_sock_callbacks(conn);
> iscsi_remove_failed_auth_entry(conn);
> }
> - if (ret != 0)
> + if (ret != 0) {
> + cancel_delayed_work_sync(&conn->login_work);
> iscsi_target_nego_release(conn);
> + }
>
> return ret;
> }
>
If ret == 1 the login was successful, the socket callbacks have been
restored by iscsi_target_do_login()
so we just have to call cancel_delayed_work_sync() against login_work
to fix the race condition.
if ret == -1 the login was unsuccessful, I moved the
cancel_delayed_work_sync() call from the (ret < 0) block
to the (ret != 0) to close a potential race condition (we must restore
the socket callbacks first).
if ret == 0 then iscsi_target_do_login() clears
LOGIN_FLAGS_INITIAL_PDU and passes the control to login_work.
With this version of the patch, nothing has been changed in the
login_work context so it should work exactly as before.
iscsi_target_start_negotiation() is executed only in the login_thread context.
Maurizio
