MINOR: fd: remove the unneeded running bit from fd_insert()

There's no point taking the running bit in fd_insert() since by
definition there will never be more than one thread inserting the FD,
and that fd_insert() may only be done after the fd was allocated by
the system, indicating the end of use by any other thread.

This will need to be backported to 2.2 to fix an issue.
diff --git a/include/haproxy/fd.h b/include/haproxy/fd.h
index 69eda48..c0a6957 100644
--- a/include/haproxy/fd.h
+++ b/include/haproxy/fd.h
@@ -432,11 +432,8 @@
 /* Prepares <fd> for being polled */
 static inline void fd_insert(int fd, void *owner, void (*iocb)(int fd), unsigned long thread_mask)
 {
-	int locked = fdtab[fd].running_mask != tid_bit;
 	extern void sock_conn_iocb(int);
 
-	if (locked)
-		fd_set_running_excl(fd);
 	fdtab[fd].owner = owner;
 	fdtab[fd].iocb = iocb;
 	fdtab[fd].ev = 0;
@@ -456,8 +453,7 @@
 	/* note: do not reset polled_mask here as it indicates which poller
 	 * still knows this FD from a possible previous round.
 	 */
-	if (locked)
-		fd_clr_running(fd);
+
 	/* the two directions are ready until proven otherwise */
 	fd_may_both(fd);
 	_HA_ATOMIC_ADD(&ha_used_fds, 1);