memory leak in pathbias_count_first_hop()
found on moria5 running git master 64676d05714a8248) as a relay:
==28065== 1 bytes in 1 blocks are definitely lost in loss record 2 of 28
==28065== at 0x4C244E8: malloc (vg_replace_malloc.c:236)
==28065== by 0x5F59C61: strdup (strdup.c:43)
==28065== by 0x211C1D: _tor_strdup (util.c:241)
==28065== by 0x2123DB: rate_limit_log (util.c:1721)
==28065== by 0x184E3E: circuit_finish_handshake (circuitbuild.c:2737)
==28065== by 0x138D31: connection_edge_process_relay_cell (relay.c:1216)
==28065== by 0x139A6B: circuit_receive_relay_cell (relay.c:203)
==28065== by 0x19AA7D: command_process_cell (command.c:576)
==28065== by 0x1BF716: connection_or_process_cells_from_inbuf (connection_or.c:1872)
==28065== by 0x1B2623: connection_handle_read (connection.c:2723)
==28065== by 0x11CA50: conn_read_callback (main.c:706)
==28065== by 0x52C9343: event_base_loop (in /usr/lib/libevent-1.4.so.2.1.3)Are there other cases where our 'rate limit' idiom breaks down and we don't free stuff?
To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information