1 This file lists minor work items (also possibly called "known bugs").
2 We are not tracking them in Mantis yet since there are too many and no
3 sane end-user should care about this codebase yet anyway.
6 - connection.c:553 fails when "make check" is run!
7 (check if this could be memory corruption).
11 - container_bloomfilter: improve efficiency (see FIXME)
12 - Windows: use events instead of pipes to signal select()s [Nils]
13 - only connect() sockets that are ready (select()) [Nils]
14 [On W32, we need to select after calling socket before
16 - server: inefficient memmove
17 - client: should do exponential back-off (starting at 1ms,
18 bounded by 1s) when connection failed (in addition to
19 half-time-to-deadline retry at the end)
21 - transport_api: support forcing disconnects through low quotas!
22 (required for working F2F support!)
23 - API: consider having core provide deadline information for each message
24 (likely important for DV plugin which wants to loop back!)
25 - implement transport API to pretty-print transport address
26 + transport_api extension (API extension!)
27 + service-transport extension (protocol extension)
28 - add calls to statistics in various places
29 - implement gnunet-transport (transport configurator / tester)
30 - UPnP-based IP detection
31 (Note: build library always, build service when libxml2/etc. are available)
32 - instantly filter addresses from *other* peers that
33 are *equal* to our own address + port (i.e., localhost:2086). We
34 no longer filter those for outgoing (helps with loopback testing
35 and keeps the code clean), but we should filter strictly *impossible*
36 incoming addresses! This is for efficiency, not correctness.
37 - We currently are happy to take any address told to us in a WELCOME
38 to our set of addresses; we should have some minimal threshold-based
39 scheme, limiting both the total number of addresses that we accept
40 this way as well as requiring multiple confirmations; also, we
41 should possibly try to confirm that the given address works for
42 us ourselves (loopback-style) before adding it to the list
44 + we may be able to simplify WELCOME messages (no need to add
45 addresses there anymore, but may help to learn them there anyway...).
46 + we probably want some kind of voting/counting for learning IP addresses
47 (maybe including IP addresses in ads proportional to how often others
48 report them? we at least need some protection against >64k HELLOs!),
49 + provide a way to give the user a list of "learned" IP addresses and
50 a way to easily "veto" addresses off the list!
51 => If MiM attacker uses vetoed address, blacklist the specific IP for
52 the presumed neighbour!
53 - not sure current way of doing ACKs works well-enough
54 with unreliable transports where the ACK maybe lost;
55 the "is_new" check would then possibly prevent future
56 ACKs to be delivered, all while we're happily
57 receiving messages from that peer! Worse, the other
58 peer won't generate another ACK since it thinks we're
59 connected just fine...
61 + How necessary is ACKing in the first place? (alternatives?)
62 + Should we transmit ACKs in response to every HELLO? (would that
63 fully address the problem?)
64 - latency measurements implemented in the transport
65 plugins makes it only work for bi-di transports
66 and results in code replication
67 - should latency be included in the ReceiveCallback and
68 NotifyConnect or passed on request?
69 - FIXME's with latency being simply set to 0 in a few places
70 - Memory leak (running valgrind --trace-children=yes on test_transport_api:
71 ==28393== 16 bytes in 1 blocks are indirectly lost in loss record 1 of 5
72 ==28393== at 0x4C2260E: malloc (vg_replace_malloc.c:207)
73 ==28393== by 0x52343E3: GNUNET_xmalloc_unchecked_ (common_allocation.c:62)
74 ==28393== by 0x5234389: GNUNET_xmalloc_ (common_allocation.c:53)
75 ==28393== by 0x524458A: GNUNET_NETWORK_socket_create_from_accept (network.c:289)
76 ==28393== by 0x524B2DA: ??? (server.c:332)
77 ==28393== by 0x524A4C7: ??? (scheduler.c:425)
78 ==28393== by 0x524A73D: GNUNET_SCHEDULER_run (scheduler.c:510)
79 ==28393== by 0x524FF8C: GNUNET_SERVICE_run (service.c:1326)
80 ==28393== by 0x405500: main (gnunet-service-transport.c:2645)
82 ==28393== 65,744 (65,728 direct, 16 indirect) bytes in 1 blocks are definitely lost in loss record 5 of 5
83 ==28393== at 0x4C2260E: malloc (vg_replace_malloc.c:207)
84 ==28393== by 0x52343E3: GNUNET_xmalloc_unchecked_ (common_allocation.c:62)
85 ==28393== by 0x5234389: GNUNET_xmalloc_ (common_allocation.c:53)
86 ==28393== by 0x524473E: GNUNET_NETWORK_socket_create_from_accept (network.c:323)
87 (rest of trace identical)
89 - [./transport/gnunet-service-transport.c:173]: (style) struct or union member 'TransportPlugin::rebuild' is never used
90 - [./transport/plugin_transport_tcp.c:391]: (style) struct or union member 'Plugin::address_update_task' is never used
93 - [./fs/gnunet-service-fs.c:208]: (style) struct or union member 'LocalGetContext::results_bf_size' is never used
94 - [./fs/gnunet-service-fs.c:501]: (style) struct or union member 'PendingRequest::used_pids_size' is never used
95 - [./fs/gnunet-service-fs.c:654]: (style) struct or union member 'ConnectedPeer::last_client_replies' is never used
96 - [./fs/gnunet-service-fs.c:669]: (style) struct or union member 'ConnectedPeer::avg_delay' is never used
97 - [./fs/gnunet-service-fs.c:675]: (style) struct or union member 'ConnectedPeer::avg_priority' is never used
98 - [./fs/gnunet-service-fs.c:688]: (style) struct or union member 'ConnectedPeer::pending_requests' is never used
99 - [./fs/gnunet-service-fs.c:694]: (style) struct or union member 'ConnectedPeer::last_p2p_replies_woff' is never used
100 - [./fs/gnunet-service-fs.c:700]: (style) struct or union member 'ConnectedPeer::last_client_replies_woff' is never used
103 - [./topology/gnunet-daemon-topology.c:94]: (style) struct or union member 'PeerList::last_hello_sent' is never used
104 - while running the topology test with valgrind (--trace-children=yes), I get:
106 Nov 03 22:38:57 topology DEBUG I am peer `AJ5M'==4186== Syscall param socketcall.send(msg) points to uninitialised byte(s)
107 ==4186== at 0x4164BF1: send (socket.S:64)
108 ==4186== by 0x404CC1F: transmit_ready (connection.c:1393)
109 ==4186== by 0x4063C3B: run_ready (scheduler.c:451)
110 ==4186== by 0x40640AE: GNUNET_SCHEDULER_run (scheduler.c:575)
111 ==4186== by 0x406090A: GNUNET_PROGRAM_run (program.c:196)
112 ==4186== by 0x804B1CA: main (gnunet-daemon-topology.c:1250)
113 ==4186== Address 0x46e33b0 is 136 bytes inside a block of size 65,664 alloc'd
114 ==4186== at 0x4024C1C: malloc (vg_replace_malloc.c:195)
115 ==4186== by 0x40476F6: GNUNET_xmalloc_unchecked_ (common_allocation.c:61)
116 ==4186== by 0x404768E: GNUNET_xmalloc_ (common_allocation.c:52)
117 ==4186== by 0x404BB22: GNUNET_CONNECTION_create_from_connect (connection.c:887)
118 ==4186== by 0x40460B8: do_connect (client.c:233)
119 ==4186== by 0x404610C: GNUNET_CLIENT_connect (client.c:259)
120 ==4186== by 0x402C6D5: GNUNET_CORE_connect (core_api.c:857)
121 ==4186== by 0x804B118: run (gnunet-daemon-topology.c:1217)
122 ==4186== by 0x4060498: program_main (program.c:80)
123 ==4186== by 0x4063C3B: run_ready (scheduler.c:451)
124 ==4186== by 0x40640AE: GNUNET_SCHEDULER_run (scheduler.c:575)
125 ==4186== by 0x406090A: GNUNET_PROGRAM_run (program.c:196)
133 - auto-generate "defaults.conf" using gnunet-setup from "config.scm"
134 - integrate all options into "config.scm"
135 - change config-file writing to exclude options set to default values
137 - implement exponential back-off for service restarts
138 - better tracking of which config changes actually need to cause process restarts by ARM.
139 - have way to specify dependencies between services (to manage ARM restarts better)
141 - code currently notifies clients about "encrypted" connections being up well before
142 we get the encrypted PONG; sometimes this may be OK (for topology killing
143 unwanted connnections), but of course not in general. I suspect we want
144 to signal on PONG and have topology hook directly into transport to
145 kill plaintext connections before they have a chance to become encrypted
146 (may require minor hack in transport API)
147 - [./core/gnunet-service-core.c:469]: (style) struct or union member 'Neighbour::message_queue_size' is never used
148 - [./core/test_core_api_start_only.c:50]: (style) struct or union member 'PeerContext::id' is never used
151 - have gnunet-peerinfo print actual host addresses again
152 - add option to gnunet-peerinfo to modify trust value
154 - finish postgres implementation; simplify other SQLs using new stats
156 - Better SSL-support for MHD
159 - active reply route caching design & implementation of service,
162 - implement advertising of hostlist URL
163 - implement learning of hostlist URLs
165 - consider changing API for peer-group termination to
166 call continuation when done