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.
7 - connection.c:553 fails when "make check" is run!
8 (check if this could be memory corruption).
12 - API-DESIGN: network/connection/client/server: need to communicate
13 scheduler's "reason" code for failures to callbacks of the kind
14 used by the notify_transmit_ready-family: shutdown and timeout
15 may require entirely different actions, and currently we
16 cannot tell the difference! [need 4th argument!]
17 - container_bloomfilter: improve efficiency (see FIXME)
18 - server: inefficient memmove
19 - client: should do exponential back-off (starting at 1ms,
20 bounded by 1s) when connection failed (in addition to
21 half-time-to-deadline retry at the end)
22 - Windows: use events instead of pipes to signal select()s [Nils]
23 - only connect() sockets that are ready (select()) [Nils]
24 [On W32, we need to select after calling socket before
27 - transport_api: support forcing disconnects through low quotas!
28 (required for working F2F support!)
29 - API: consider having core provide deadline information for each message
30 (likely important for DV plugin which wants to loop back!)
31 - implement transport API to pretty-print transport address
32 + transport_api extension (API extension!)
33 + service-transport extension (protocol extension)
34 - add calls to statistics in various places
35 - implement gnunet-transport (transport configurator / tester)
36 - UPnP-based IP detection
37 (Note: build library always, build service when libxml2/etc. are available)
38 - instantly filter addresses from *other* peers that
39 are *equal* to our own address + port (i.e., localhost:2086). We
40 no longer filter those for outgoing (helps with loopback testing
41 and keeps the code clean), but we should filter strictly *impossible*
42 incoming addresses! This is for efficiency, not correctness.
43 - We currently are happy to take any address told to us in a WELCOME
44 to our set of addresses; we should have some minimal threshold-based
45 scheme, limiting both the total number of addresses that we accept
46 this way as well as requiring multiple confirmations; also, we
47 should possibly try to confirm that the given address works for
48 us ourselves (loopback-style) before adding it to the list
50 + we may be able to simplify WELCOME messages (no need to add
51 addresses there anymore, but may help to learn them there anyway...).
52 + we probably want some kind of voting/counting for learning IP addresses
53 (maybe including IP addresses in ads proportional to how often others
54 report them? we at least need some protection against >64k HELLOs!),
55 + provide a way to give the user a list of "learned" IP addresses and
56 a way to easily "veto" addresses off the list!
57 => If MiM attacker uses vetoed address, blacklist the specific IP for
58 the presumed neighbour!
59 - not sure current way of doing ACKs works well-enough
60 with unreliable transports where the ACK maybe lost;
61 the "is_new" check would then possibly prevent future
62 ACKs to be delivered, all while we're happily
63 receiving messages from that peer! Worse, the other
64 peer won't generate another ACK since it thinks we're
65 connected just fine...
67 + How necessary is ACKing in the first place? (alternatives?)
68 + Should we transmit ACKs in response to every HELLO? (would that
69 fully address the problem?)
70 - latency measurements implemented in the transport
71 plugins makes it only work for bi-di transports
72 and results in code replication
73 - should latency be included in the ReceiveCallback and
74 NotifyConnect or passed on request?
75 - FIXME's with latency being simply set to 0 in a few places
76 - Memory leak (running valgrind --trace-children=yes on test_transport_api:
77 ==28393== 16 bytes in 1 blocks are indirectly lost in loss record 1 of 5
78 ==28393== at 0x4C2260E: malloc (vg_replace_malloc.c:207)
79 ==28393== by 0x52343E3: GNUNET_xmalloc_unchecked_ (common_allocation.c:62)
80 ==28393== by 0x5234389: GNUNET_xmalloc_ (common_allocation.c:53)
81 ==28393== by 0x524458A: GNUNET_NETWORK_socket_create_from_accept (network.c:289)
82 ==28393== by 0x524B2DA: ??? (server.c:332)
83 ==28393== by 0x524A4C7: ??? (scheduler.c:425)
84 ==28393== by 0x524A73D: GNUNET_SCHEDULER_run (scheduler.c:510)
85 ==28393== by 0x524FF8C: GNUNET_SERVICE_run (service.c:1326)
86 ==28393== by 0x405500: main (gnunet-service-transport.c:2645)
88 ==28393== 65,744 (65,728 direct, 16 indirect) bytes in 1 blocks are definitely lost in loss record 5 of 5
89 ==28393== at 0x4C2260E: malloc (vg_replace_malloc.c:207)
90 ==28393== by 0x52343E3: GNUNET_xmalloc_unchecked_ (common_allocation.c:62)
91 ==28393== by 0x5234389: GNUNET_xmalloc_ (common_allocation.c:53)
92 ==28393== by 0x524473E: GNUNET_NETWORK_socket_create_from_accept (network.c:323)
93 (rest of trace identical)
95 - [./transport/gnunet-service-transport.c:173]: (style) struct or union member 'TransportPlugin::rebuild' is never used
96 - [./transport/plugin_transport_tcp.c:391]: (style) struct or union member 'Plugin::address_update_task' is never used
99 - [./fs/gnunet-service-fs.c:208]: (style) struct or union member 'LocalGetContext::results_bf_size' is never used
100 - [./fs/gnunet-service-fs.c:501]: (style) struct or union member 'PendingRequest::used_pids_size' is never used
101 - [./fs/gnunet-service-fs.c:654]: (style) struct or union member 'ConnectedPeer::last_client_replies' is never used
102 - [./fs/gnunet-service-fs.c:669]: (style) struct or union member 'ConnectedPeer::avg_delay' is never used
103 - [./fs/gnunet-service-fs.c:675]: (style) struct or union member 'ConnectedPeer::avg_priority' is never used
104 - [./fs/gnunet-service-fs.c:688]: (style) struct or union member 'ConnectedPeer::pending_requests' is never used
105 - [./fs/gnunet-service-fs.c:694]: (style) struct or union member 'ConnectedPeer::last_p2p_replies_woff' is never used
106 - [./fs/gnunet-service-fs.c:700]: (style) struct or union member 'ConnectedPeer::last_client_replies_woff' is never used
109 - [./topology/gnunet-daemon-topology.c:94]: (style) struct or union member 'PeerList::last_hello_sent' is never used
110 - while running the topology test with valgrind (--trace-children=yes), I get:
112 Nov 03 22:38:57 topology DEBUG I am peer `AJ5M'==4186== Syscall param socketcall.send(msg) points to uninitialised byte(s)
113 ==4186== at 0x4164BF1: send (socket.S:64)
114 ==4186== by 0x404CC1F: transmit_ready (connection.c:1393)
115 ==4186== by 0x4063C3B: run_ready (scheduler.c:451)
116 ==4186== by 0x40640AE: GNUNET_SCHEDULER_run (scheduler.c:575)
117 ==4186== by 0x406090A: GNUNET_PROGRAM_run (program.c:196)
118 ==4186== by 0x804B1CA: main (gnunet-daemon-topology.c:1250)
119 ==4186== Address 0x46e33b0 is 136 bytes inside a block of size 65,664 alloc'd
120 ==4186== at 0x4024C1C: malloc (vg_replace_malloc.c:195)
121 ==4186== by 0x40476F6: GNUNET_xmalloc_unchecked_ (common_allocation.c:61)
122 ==4186== by 0x404768E: GNUNET_xmalloc_ (common_allocation.c:52)
123 ==4186== by 0x404BB22: GNUNET_CONNECTION_create_from_connect (connection.c:887)
124 ==4186== by 0x40460B8: do_connect (client.c:233)
125 ==4186== by 0x404610C: GNUNET_CLIENT_connect (client.c:259)
126 ==4186== by 0x402C6D5: GNUNET_CORE_connect (core_api.c:857)
127 ==4186== by 0x804B118: run (gnunet-daemon-topology.c:1217)
128 ==4186== by 0x4060498: program_main (program.c:80)
129 ==4186== by 0x4063C3B: run_ready (scheduler.c:451)
130 ==4186== by 0x40640AE: GNUNET_SCHEDULER_run (scheduler.c:575)
131 ==4186== by 0x406090A: GNUNET_PROGRAM_run (program.c:196)
139 - auto-generate "defaults.conf" using gnunet-setup from "config.scm"
140 - integrate all options into "config.scm"
141 - change config-file writing to exclude options set to default values
143 - implement exponential back-off for service restarts
144 - better tracking of which config changes actually need to cause process restarts by ARM.
145 - have way to specify dependencies between services (to manage ARM restarts better)
147 - code currently notifies clients about "encrypted" connections being up well before
148 we get the encrypted PONG; sometimes this may be OK (for topology killing
149 unwanted connnections), but of course not in general. I suspect we want
150 to signal on PONG and have topology hook directly into transport to
151 kill plaintext connections before they have a chance to become encrypted
152 (may require minor hack in transport API)
153 - [./core/gnunet-service-core.c:469]: (style) struct or union member 'Neighbour::message_queue_size' is never used
154 - [./core/test_core_api_start_only.c:50]: (style) struct or union member 'PeerContext::id' is never used
157 - have gnunet-peerinfo print actual host addresses again
158 - add option to gnunet-peerinfo to modify trust value
160 - finish postgres implementation; simplify other SQLs using new stats
162 - Better SSL-support for MHD
165 - active reply route caching design & implementation of service,
168 - implement advertising of hostlist URL
169 - implement learning of hostlist URLs
171 - consider changing API for peer-group termination to
172 call continuation when done