aboutsummaryrefslogtreecommitdiff
path: root/src/cadet
Commit message (Collapse)AuthorAge
...
* handle KX AUTH failure by retransmitting KXChristian Grothoff2017-01-30
|
* pass connection tunnel identifier to channel when receiving messages, so ↵Christian Grothoff2017-01-30
| | | | that performance can be matched
* have tunnel tell channel which connection it used for transmission, so we ↵Christian Grothoff2017-01-30
| | | | can track connection performance by observing ACKs on the channel-level
* Save next pointer in case mqm gets freedDavid Barksdale2017-01-29
|
* Don't pass NULL to destroy_routeDavid Barksdale2017-01-29
|
* added path desirability calculationsChristian Grothoff2017-01-29
|
* add path desirability calculationsChristian Grothoff2017-01-29
|
* remove route also from routesChristian Grothoff2017-01-29
|
* fix shutdown sequenceChristian Grothoff2017-01-29
|
* fix task cleanup on CTRL-DChristian Grothoff2017-01-29
|
* of course we need 'i' here, not 'off'Christian Grothoff2017-01-29
|
* sanitize paths to eliminate loops before using them; fix off-by-one causing ↵Christian Grothoff2017-01-29
| | | | a peer to be listed twice
* note on desChristian Grothoff2017-01-29
|
* collect misc. TODO items into one big listChristian Grothoff2017-01-29
|
* introduce buffering options on the route levelChristian Grothoff2017-01-29
|
* optimize mqm_head scans by avoiding constantly scanning over definitively ↵Christian Grothoff2017-01-29
| | | | non-ready entries
* fix ready_cb notifications to ensure they exactly happen only when neededChristian Grothoff2017-01-29
|
* implement route timeouts at 3x keepalive frequencyChristian Grothoff2017-01-29
|
* implement basics of connection management for tunnels (in particular ↵Christian Grothoff2017-01-29
| | | | destruction of bad connections, eventual selection of new ones
* separate connection DLL into a list for ready connections and a list of busy ↵Christian Grothoff2017-01-29
| | | | connections, and count them separately; make iterator tolerate modifications to current element of DLL during iteration
* provide stricter assurances about when we call ready_cb, avoiding supurious ↵Christian Grothoff2017-01-29
| | | | calls that merely reinforce what tunnel already knows about the connection state
* when applicable, do connection selection outside of send_kx*Christian Grothoff2017-01-29
|
* complete new KX implementation, including KX_AUTH defense against ↵Christian Grothoff2017-01-28
| | | | KX-injection based reset attacks
* KEY_REKEY state was dead in old code, remove in new code; refactor KX logic ↵Christian Grothoff2017-01-28
| | | | in preparation for KX_AUTH
* slightly better documentation for KX logicChristian Grothoff2017-01-28
|
* also count messages dropped due to slow clients as allowable losses in ↵Christian Grothoff2017-01-27
| | | | unreliable mode
* if in-order unreliable, advance to next message even if we skip a midChristian Grothoff2017-01-27
|
* fix comments, fix testChristian Grothoff2017-01-27
|
* indentation fixesChristian Grothoff2017-01-27
|
* unset XDG environment variable during testcases, as they can screw things up ↵Christian Grothoff2017-01-27
| | | | badly
* add #include to avoid warningChristian Grothoff2017-01-27
|
* allow for message drops by CADET in unreliable modeChristian Grothoff2017-01-27
|
* add statistics for packets dropped by cadet due to full bufferChristian Grothoff2017-01-27
|
* introduce concept of unverified_kx, as a step towards having KX_AUTH ↵Christian Grothoff2017-01-27
| | | | implemented properly
* pass only CadetTunnelAxolotl if it suffices, preparation for having ↵Christian Grothoff2017-01-27
| | | | ambiguous KX context
* trying to make KX logic slightly more readableChristian Grothoff2017-01-27
|
* implement tie-breaking in case both peers establish a connection to each ↵Christian Grothoff2017-01-27
| | | | other via the same path at the same time, so that only one connection survives
* use 'futures' bitfield in ACKs properly, revisit unbuffered/out-of-order ↵Christian Grothoff2017-01-26
| | | | transmission
* implement random packet drop option, fix retransmission logicChristian Grothoff2017-01-26
|
* fix log callsChristian Grothoff2017-01-25
|
* make new cadet implementation the default: all tests pass (on my system); ↵Christian Grothoff2017-01-25
| | | | however, implementation is NOT complete
* implement keepalivesChristian Grothoff2017-01-25
|
* preparations for keepalivesChristian Grothoff2017-01-25
|
* cleanupsChristian Grothoff2017-01-25
|
* handle ACKs past client downChristian Grothoff2017-01-25
|
* typo in commentChristian Grothoff2017-01-25
|
* towards proper DATA_ACK handlingChristian Grothoff2017-01-25
|
* towards proper DATA_ACK handlingChristian Grothoff2017-01-25
|
* handle ancient/future duplicate payload properlyChristian Grothoff2017-01-25
|
* fix comment, this was implementedChristian Grothoff2017-01-25
|