Commit bf30943c authored by David Goulet's avatar David Goulet 🐼
Browse files

Merge branch 'maint-0.4.7'

parents 4d3f42a6 eee35adf
Loading
Loading
Loading
Loading

changes/ticket40649

0 → 100644
+4 −0
Original line number Diff line number Diff line
  o Minor bugfixes (relay):
    - Do not propagate either forward or backward a DESTROY remote reason when
      closing a circuit so to avoid a possible side channel. Fixes bug 40649;
      bugfix on 0.1.2.4-alpha.
+7 −4
Original line number Diff line number Diff line
@@ -656,9 +656,11 @@ command_process_destroy_cell(cell_t *cell, channel_t *chan)
  if (!CIRCUIT_IS_ORIGIN(circ) &&
      chan == TO_OR_CIRCUIT(circ)->p_chan &&
      cell->circ_id == TO_OR_CIRCUIT(circ)->p_circ_id) {
    /* the destroy came from behind */
    /* The destroy came from behind so nullify its p_chan. Close the circuit
     * with a DESTROYED reason so we don't propagate along the path forward the
     * reason which could be used as a side channel. */
    circuit_set_p_circid_chan(TO_OR_CIRCUIT(circ), 0, NULL);
    circuit_mark_for_close(circ, reason|END_CIRC_REASON_FLAG_REMOTE);
    circuit_mark_for_close(circ, END_CIRC_REASON_DESTROYED);
  } else { /* the destroy came from ahead */
    circuit_set_n_circid_chan(circ, 0, NULL);
    if (CIRCUIT_IS_ORIGIN(circ)) {
@@ -666,9 +668,10 @@ command_process_destroy_cell(cell_t *cell, channel_t *chan)
    } else {
      /* Close the circuit so we stop queuing cells for it and propagate the
       * DESTROY cell down the circuit so relays can stop queuing in-flight
       * cells for this circuit which helps with memory pressure. */
       * cells for this circuit which helps with memory pressure. We do NOT
       * propagate the remote reason so not to create a side channel. */
      log_debug(LD_OR, "Received DESTROY cell from n_chan, closing circuit.");
      circuit_mark_for_close(circ, reason | END_CIRC_REASON_FLAG_REMOTE);
      circuit_mark_for_close(circ, END_CIRC_REASON_DESTROYED);
    }
  }
}