| Commit message (Expand) | Author | Age |
|---|
| ... | |
| * | | | | | Send END cells on bufferevent tunneled directory conns•••Our old code correctly called bufferevent_flush() on linked
connections to make sure that the other side got an EOF event... but
it didn't call bufferevent_flush() when the connection wasn't
hold_open_until_flushed. Directory connections don't use
hold_open_until_flushed, so the linked exit connection never got an
EOF, so they never sent a RELAY_END cell to the client, and the
client never concluded that data had arrived.
The solution is to make the bufferevent_flush() code apply to _all_
closing linked conns whose partner is not already marked for close.
| Nick Mathewson | 2010-10-13 |
| * | | | | | Revert accidentally-committed changes to routerlist.c•••This reverts part of commit a0c1c2ac012fded493c0d8c49fe57e56373b061f.
| Nick Mathewson | 2010-10-13 |
| * | | | | | Use connection_mark_and_flush consistently. | Nick Mathewson | 2010-10-13 |
| * | | | | | Remember to re-install inbuf/outbuf callbacks on ssl bufferevents•••If we don't, we will (among other bad things) never update
lastread/lastwritten, and so flood the network with keepalives.
| Nick Mathewson | 2010-10-12 |
| * | | | | | Actually call connection_tls_finish_handshake() with bufferevents•••First start of a fix for bug2001, but my test network still isn't
working: the client and the server send each other VERSIONS cells,
but never notice that they got them.
| Nick Mathewson | 2010-10-12 |
| * | | | | | Log OpenSSL errors coming from bufferevent_openssl | Nick Mathewson | 2010-10-11 |
| * | | | | | Fix a crash bug when serving microdescs on a bufferevent. | Nick Mathewson | 2010-10-07 |
| * | | | | | Remove a spurious .z from microdesc fetch URLs so they no logner end with .z.z | Nick Mathewson | 2010-10-07 |
| * | | | | | Add more documentation to microdesc.c, plus another assert. | Nick Mathewson | 2010-10-07 |
| * | | | | | Add a more verbose log message to try to catch bug #2022. | Nick Mathewson | 2010-10-07 |
| * | | | | | Unit tests for microdescriptor cache•••May help with tracking down bug #2022
| Nick Mathewson | 2010-10-07 |
| |/ / / / |
|
| * | | | | Merge branch 'maint-0.2.2' | Roger Dingledine | 2010-10-01 |
| |\ \ \ \
| | |/ /
| |/| | |
|
| | * | | | fix comment | Roger Dingledine | 2010-10-01 |
| | * | | | log when we guess our ip address, not just when we fail | Roger Dingledine | 2010-10-01 |
| * | | | | Merge branch 'tor-fw-squashed2'•••Conflicts:
src/common/util.c
| Nick Mathewson | 2010-09-30 |
| |\ \ \ \ |
|
| | * | | | | Start tor-fw-helper in the background, and log whatever it outputs | Steven Murdoch | 2010-09-30 |
| * | | | | | Merge remote branch 'origin/maint-0.2.2' | Nick Mathewson | 2010-09-30 |
| |\ \ \ \ \
| |/ / / /
|/| / / /
| |/ / / |
|
| | * | | | Note an XXX about potential overflow | Sebastian Hahn | 2010-09-30 |
| | * | | | Use an upper and lower bound for bridge weights•••When picking bridges (or other nodes without a consensus entry (and
thus no bandwidth weights)) we shouldn't just trust the node's
descriptor. So far we believed anything between 0 and 10MB/s, where 0
would mean that a node doesn't get any use from use unless it is our
only one, and 10MB/s would be a quite siginficant weight. To make this
situation better, we now believe weights in the range from 20kB/s to
100kB/s. This should allow new bridges to get use more quickly, and
means that it will be harder for bridges to see almost all our traffic.
| Sebastian Hahn | 2010-09-30 |
| * | | | | Whitespace fixes on recent merges to master. | Nick Mathewson | 2010-09-30 |
| * | | | | Merge branch 'maint-0.2.2' | Roger Dingledine | 2010-09-30 |
| |\| | | |
|
| | * | | | Merge commit 'mikeperry/bug1772' into maint-0.2.2 | Roger Dingledine | 2010-09-30 |
| | |\ \ \ |
|
| | | * | | | Nominaly lower the minimum timeout value to 1500.•••This won't change any behavior, since it will still be rounded back
up to 2seconds, but should reduce the chances of some extra warns.
| Mike Perry | 2010-09-29 |
| | * | | | | Merge branch 'bug1772' into maint-0.2.2 | Roger Dingledine | 2010-09-29 |
| | |\ \ \ \ |
|
| | | * | | | | fix two casts | Roger Dingledine | 2010-09-29 |
| | | |/ / / |
|
| | | * | | | Comment network liveness and change detection behavior. | Mike Perry | 2010-09-29 |
| | | * | | | no measurement circs if not enough build times•••In the first 100 circuits, our timeout_ms and close_ms
are the same. So we shouldn't transition circuits to purpose
CIRCUIT_PURPOSE_C_MEASURE_TIMEOUT, since they will just timeout again
next time we check.
| Roger Dingledine | 2010-09-29 |
| | | * | | | refactor and recomment; no actual changes | Roger Dingledine | 2010-09-29 |
| | | * | | | Cap the circuit build timeout to the max time we've seen.•••Also, cap the measurement timeout to 2X the max we've seen.
| Mike Perry | 2010-09-29 |
| | | * | | | Do away with the complexity of the network liveness detection.•••We really should ignore any timeouts that have *no* network activity for their
entire measured lifetime, now that we have the 95th percentile measurement
changes. Usually this is up to a minute, even on fast connections.
| Mike Perry | 2010-09-29 |
| | | * | | | Fix state checks on liveness handling.•••If we really want all this complexity for these stages here, we need to handle
it better for people with large timeouts. It should probably go away, though.
| Mike Perry | 2010-09-29 |
| | | * | | | Fix non-live condition checks.•••Rechecking the timeout condition was foolish, because it is checked on the
same codepath. It was also wrong, because we didn't round.
Also, the liveness check itself should be <, and not <=, because we only have
1 second resolution.
| Mike Perry | 2010-09-29 |
| * | | | | | Merge branch 'maint-0.2.2' | Roger Dingledine | 2010-09-29 |
| |\| | | |
| |_|/ /
|/| | | |
|
| | * | | | Merge commit 'mikeperry/bug1739' into maint-0.2.2 | Roger Dingledine | 2010-09-29 |
| | |\| | |
|
| | | * | | Send control port events for timeouts.•••We now differentiate between timeouts and cutoffs by the REASON string and
the PURPOSE string.
| Mike Perry | 2010-09-29 |
| | * | | | Merge commit 'mikeperry/bug1740' into maint-0.2.2 | Roger Dingledine | 2010-09-29 |
| | |\| | |
|
| | | * | | Only count timeout data for 3 hop circuits.•••Use 4/3 of this timeout value for 4 hop circuits, and use half of it for
canabalized circuits.
| Mike Perry | 2010-09-29 |
| | | |/ |
|
| * | | | Merge remote branch 'origin/maint-0.2.2' | Nick Mathewson | 2010-09-29 |
| |\| | |
|
| | * | | even more comment | Roger Dingledine | 2010-09-28 |
| | * | | improve code comments, based on comments from nick | Roger Dingledine | 2010-09-28 |
| | * | | handle ugly edge case in retrying entrynodes•••Specifically, a circ attempt that we'd launched while the network was
down could timeout after we've marked our entrynodes up, marking them
back down again. The fix is to annotate as bad the OR conns that were
around before we did the retry, so if a circuit that's attached to them
times out we don't do anything about it.
| Roger Dingledine | 2010-09-28 |
| | * | | Actually notice when our last entrynode goes down•••Otherwise we'd never set have_minimum_dir_info to false, so the
"optimistic retry" would never trigger.
| Roger Dingledine | 2010-09-28 |
| | * | | optimistically retry EntryNodes on socks request•••We used to mark all our known bridges up when they're all down and we
get a new socks request. Now do that when we've set EntryNodes too.
| Roger Dingledine | 2010-09-28 |
| | * | | remove a redundant assert | Roger Dingledine | 2010-09-28 |
| | * | | refactor; no actual changes | Roger Dingledine | 2010-09-28 |
| * | | | Remove the has_old_dnsworkers flag. | Nick Mathewson | 2010-09-28 |
| * | | | Merge branch 'maint-0.2.2' | Roger Dingledine | 2010-09-28 |
| |\| | |
|
| | * | | Merge branch 'maint-0.2.1' into maint-0.2.2 | Roger Dingledine | 2010-09-28 |
| | |\ \ |
|
| | | * | | actually retry bridges when your network goes away | Roger Dingledine | 2010-09-28 |
| * | | | | Autodetect the number of CPUs when possible if NumCPUs==0•••This is needed for IOCP, since telling the IOCP backend about all
your CPUs is a good idea. It'll also come in handy with asn's
multithreaded crypto stuff, and for people who run servers without
reading the manual.
| Nick Mathewson | 2010-09-28 |