Open
Milestone
Tor: 0.3.2.x-final
All issues for this milestone are closed. You may close this milestone now.
Unstarted Issues (open and unassigned)
0
Ongoing Issues (open and assigned)
0
Completed Issues (closed)
360
- older gccs don't accept -Wunused-const-variable
-
defect: Spec for net/listeners/* doesn't covers HTTPTunnelPort directive or ExtORPort
#26647 032-backport 033-backport 034-backport 035-triaged-in-20180711 component::core tor/tor controller easy fast-fix intro milestone::Tor: 0.3.2.x-final owner::nickm priority::medium reporter::pyhedgehog resolution::implemented reviewer::teor severity::minor status::closed tor-httptunnel tor-spec type::defect
-
HSv3 throws many "Tried connecting to router at [IP:port], but RSA identity key was not as expected"
#26627 032-backport 033-backport 034-backport 035-must 035-roadmap-proposed 035-triaged-in-20180711 certs component::core tor/tor ed25519 fast-fix handshake milestone::Tor: 0.3.2.x-final owner::teor priority::medium resolution::fixed reviewer::asn severity::normal status::closed tor-hs tor-relay type::defect version::tor 0.3.2.4-alpha
- ValueError: "blue" isn't a valid color - color_interface is not working
- Tor not generating ADDRMAP Events for Externally Resolved HTTP(S) Requests
- Tor startup crash with Sandbox 1 in torrc.d - sandbox_intern_string(): Bug: No interned sandbox parameter found
- Can not use nyx with Tor 0.3.2.10 to display like ARM
- configure test for libevent fails with c99
- Intermittent test failures for hs_service/service_intro_point
- tor_assertion_failed_(): Bug: ../src/or/hs_service.c:1985: rotate_all_descriptors: Assertion service->desc_current failed; aborting.
- explicitly mention email address to contact for rejected relays
- Look at PRT_HSREND, not PRT_HSDIR, for supports_v3_rendezvous_point
- [armhf, tor] strange "your system clock just jumped" warnings
- memory leak in test-scheduler unit tests
- Bug: src/or/hs_cache.c:628: cache_store_as_client: Non-fatal assertion !(cache_entry->desc->plaintext_data.revision_counter > client_desc->desc->plaintext_data.revision_counter) failed
- I keep getting this error. "OpenSSL version from headers does not match the version we're running with"
- Travis CI environment change breaks clang builds
- Added -1 signatures to consensus
- Generate a new fallback list and backport it
- rust: protover.rs needs retain_hash_collection to build with older rustc versions
- Document IPv6Exit in the sample torrcs
- sched: KISTLite should set an upper limit to write on the outbuf
- sched: Store the circuit ID instead of the full DESTROY cell in the destroy queue
- sched: In KIST, the extra_space kernel value needs to be allowed to be negative
- Rust builds fail on macOS with linker error
- test_hs_common.c:974:33: warning: declaration of 'time' shadows a global declaration
- Stop initialising rust git submodules, travis does this automatically
- hs_service_intro_circ_has_closed: Assertion desc failed
- sched: Fix integer overflow for KIST
- Make signal handlers optional, for starting Tor in-process
- Make it easy to debug restart-in-process
- Hypothesis: Some exit nodes are tampering/blocking DNS request
- Non-fatal assertion !((*diff)->entry == NULL) failed in cdm_diff_ht_purge at ../src/or/consdiffmgr.c:329. Stack trace
- TorBrowser 7.5 a8 takes multiple minutes to connect
- scheduler_release_channel: Non-fatal assertion
- Confusing log message "Can't get entropy from getrandom()"
- Mark all bridges as up on application activity
- Set the hsdir_spread_store parameter to 4 (or maybe even 5)
- fails to build with libseccomp-dev installed on arm64
- add ipv6 dirauth address
- Memory leak in config/check_bridge_distribution_setting_not_a_bridge
- Crash: died: Caught signal 11 [crash from rend_consider_services_intro_points]
- configure libevent leaks
- hs-v3: Change "hsdir-interval" to "hsdir_interval" to match the spec
- There needs to be documentation on what kernel versions the KIST Scheduler will run on
- undefined reference to evdns_shutdown
- Fuzzing: check_signature mock can be fooled by a really short key.
- control: HS_DESC event failed upload sends back the wrong Action
- spec: ADD_ONION syntax is not reflecting the code
- vote: Voting schedule is not updated before voting
- Log the actual bandwidth total when logging "Generated weighted bandwidths"
- Generated a networkstatus consensus we couldn't parse
- errors I get after I installed 0.3.2.3 on my relay
- Recalculating voting schedule should be called first when setting a new consensus
- I get this error "Looks like our kernel doesn't have the support for KIST anymore." on my relay
- Memory leak in v3 hsdesc parsing with empty encrypted data.
- consdiffmgr.c:329: cdm_diff_ht_purge: Non-fatal assertion
- Make supported version timeline easier to find
- spec: Update dir-spec.txt with HS v3 consensus param
- man: Mention that HS v3 num intro points is 0 to 20, not 0 to 10.
- Test IPv6-only clients using microdescriptors in make-test-network-all
- evdns_callback(): Bug: eventdns returned no addresses or error
- We still do client-side caching. We just don't use the cache.
- Log warning: Inconsistent ed25519 identities in the nodelist
- Say "At most once" for exit policy summaries in microdescriptors
- Check for ed25519 key is inverted in pick_intro_point()
- Add a new HSDir protocol version that definitely doesn't have the 0.3.0.7 bug
- control-spec.txt should document which "signals" don't actually exist as Unix signals
- LOG_PROTOCOL_WARN can call get_options() during an options transition.
- Add bastet to the list of default authorities
- Excessive I learned some more directory information. [...] log message at startup
- Make sure v3 single onion services and v3 onion service clients only send IPv4 addresses
- Use exponential backoff with jitter and/or tune its parameters
- When I host a V3 hidden service using macOs server it requires me to have "https" in front of the onion so that it lets me connect
- Can't build Tor 0.3.2.2-alpha with mingw32 on Windows 7
- sched: KIST doesn't free the socket info hashtable on free_all()
- hs-v3: If a client can't launch a desc fetch, it shouldn't try to open IP/RP circuits
- hs-v3: Client request with missing dirinfo will always timeout
- test config/include_no_permission fails when run as root
- tor's .gitlab-ci.yml doesn't have the same behaviour as our .travis.yml
- don't unconditionally mirror in .gitlab-ci.yml
- sched: Implement a SCHED_BUG() that prints more information
- [warn] tor_bug_occurred_: Bug: src/common/buffers.c, etc.
- inconsistent/redundant handling of hs_ed25519_public_key file
- scripts/test/coverage tries to mv .gcov files to /
- improve documentation on how we use gcov
- Somebody should profile a Tor 0.3.1.7 relay
- Tor tells me "Scheduler type KIST has been enabled" multiple times
- Bug: scheduler_kist.c:520: kist_scheduler_schedule: Non-fatal assertion !((diff < 0)) failed.
- Add resource on writing Rust FFI to doc/HACKING/GettingStartedRust.md
- leak in test_util_decompress_dos_impl()
- kist: Channel in waiting_to_write goes missing until a destroy cell
- Tor kinda _is_ an http proxy now...
- kist on 0.3.2.1-alpha-dev beats its head against a wall trying to flush a conn that's closed
- hs-v3: With multiple SOCKS connection to the same .onion, don't trigger a descriptor fetch for each connection
- Say how many other nodes we're missing descriptors for
- Say how many primary entry guards we're missing descriptors for
- hs: Continue to improve logging in both HS and circuit subsystems
- Type-check struct members that are passed to confparse
- prop224: Fake client auth lines do not actually provide obfuscation
- Make exit flag depend on ports 80 and 443, not 6667
- hs: Cache current time period number and SRV start time
- misspelled reference "UseEntryGuardsAsDirectoryGuards" to obsolete config parameter in manual page
- Avoid assertion failure for state file clock skew
- incorrect clock skew sign convention in or_state_load()
- Improve circuit logging
- hs: Cleanup race between circuit close and free with the HS circuitmap
- Detect homebrew OpenSSL on OSX (was:Fix compilation on macOS)
- Update longclaw's IPv4 address (after it moves)
- Stop assuming that every extend_info contains an IPv4 address in get_lspecs_from_extend_info()
- ext/timeouts/timeout-bitops.c:234: bad shift
- Die more helpfully if Schedulers option isn't compatible with platform
- doc: HiddenServiceVersion man page entry only lists version 2 as supported
- Avoid printf warning in scheduler_ev_add
- Clarify what the client timestamp actually does
- prop224 should become an official specification
- options/validate__transproxy fails on FreeBSD (thanks to the new scheduler)
- run make check-changes in CI
- document changes files release-readiness requirement in doc/HACKING
- keep changes files on master release-ready
- Fix typo(s) in comment(s) in the scheduling system.
- sched: KIST schedule() bad cast on monotime_diff_msec()
- check-spaces.pl should check if there are double semi-colons at the end of a line
- Add FreeBSD ports maintainer to ReleasingTor.md
- We don't need to log our scheduler type so often
- We've defined "don't use kist" as a negative interval, so don't check for -1
- Allow KISTSchedRunInterval to be negative
- Allow the new sched to respond to a new conensus, not the old one.
- NETINFO clock skew detection doesn't work on clients
- prop224: Add changes file for state of prop224 in 0.3.2
- Explain the restrictions on divisor in round*_to_next_multiple_of
- Add a missing call to download_status_get_next_attempt_at()
- remove inappropriate paragraph in GettingStarted.md
- clock_skew_warning should be a bootstrap event
- dir server responses should include a Date: header even when not responding 200
- Add a chutney network with a HS via a bridge to tor's make test-network-all
- Fix some typos in prop 280
- Make scan-build.sh run automatically, and keep an eye on the outputs.
- key-expiration check always returns exitcode 0
- Incorrect output on bad directory ownership
- hs-v3: Client fails to reach service before 90% bootstrap
- hs: Rate limit some logs in hs-v3 subsystem
- Bug: ../src/or/config.c:785: get_options_mutable: Non-fatal assertion ! in_option_validation failed
- v3 single onion services appear to be broken
- Remove legacy(?) intro point warning
- buf_t is in kernel_types.h on macOS (and maybe BSD)
- hs: v3 client descriptor cache entry should be indexed by blinded key
- hs: Downgrade client log warn to info
- circuit: Add a log_info() when marking a circuit for close.
- test: Unit tests memleaks in nodelist/ed_id and hs_common/client_service_hsdir_set_sync
- prop224: Service descriptor uploads race condition
- hs: hs_circuitmap_get_rend_circ_client_side() doesn't consider REND_JOINED purpose
- dir-spec: "protocols" line has been remove (past not future)
- make test: fgets_eagain FAILED
- hs: assert on rend_data when pruning the v2 service list
- Add new Ubuntu packager to ReleasingTor.md
- Remove AllowDotExit
- prop224: Pad RENDEZVOUS1 cell to match legacy cell length
- prop224: HSdir index desynch between client and service
- Circuit dirtiness is inconsistant with MaxCircuitDirtiness
- test: stack-use-after-scope in hs_service/build_update_descriptors
- Add design and coding guidelines for using floating point
- test: test_options_validate__outbound_addresses is broken
- prop224: client can pick super old rendezvous points
- prop224: Service has dead code which removed a feature
- Multiple definitions of `response_handler_args_t` in src/or/directory.h are breaking clang builds
- prop224: Implement a client purge state for NEWNYM
- Using bridges or switching to bridges sometimes does not work with tor 0.3.2
- prop224: HSdir set changed detection fails sometimes
- prop224: Bug: Non-fatal assertion !((strlen(req_key_str) + REND_DESC_ID_V2_LEN_BASE32) < strlen(key)) failed
- There is no HSDir3 flag, but prop224 says there should be
- prop224: Clang compile error + bad log_warn on master
- prop224: Don't build HS desc if we don't have a live consensus
- sandbox: Double free when initializing the HSv3 config
- prop224: Uninitialized hsdir indices if requesting onion before bootstrap completes
- prop224: client-side memleaks on the last hidserv request subsystem
- hs: Memory leak in test hs_descriptor/decode_bad_signature
- prop224: Spammy intro point logs in v2
- test: prop224 client unit tests
- hs: We need to get rid of a descriptor when entering non-overlap mode
- prop224: Implement note_connection_attempt_succeeded() in the client code
- hs: Maybe don't use REND_DESC_ID_V2_LEN_BASE32 as the length for a base32 relay digest id
- prop224: Dump a malformed descriptor in a file and log_warn about it
- hs: Explain why we reset the directory connection timestamp client side
- RELAY_CONNECTED cells responding to RELAY_BEGIN_DIR cells don't have a payload
- BridgeAuth goes offline when it has an expired ed25519_signing_cert
- Add 'GETINFO ip-to-country/available'
- hs: Compiler warnings with clang and mingwcross
- test: Fix prop224 hs unit tests
- Raise MIN_DL_PER_REQUEST to improve download performance
- Bug: Non-fatal assertion ei failed in launch_intro_point_circuits at src/or/hs_service.c:1784
- prop224: Fix coverity reports generated by prop224 service merge (#20657)
- Refactor buffer.c: split and rename functions.
- prop224: Synchronize spec with implementation (August 2017)
- Cannibalized HS circuit don't have their timestamp_dirty updated
- prop224: Rate limit HS descriptor reuploads
- prop224: Do we need to re-encode HS desc everytime we upload it?
- Audit code for swapped ntoh*/hton* calls
- FTBFS again on Hurd
- The circuit timeout prediction is not working properly
- Stop ignoring CircuitIdleTimeout when it's lower than IDLE_TIMEOUT_WHILE_LEARNING
- Broken condition in check_expired_networkstatus_callback()
- Sandbox failure on Debian 8.9 under OpenVZ with kernel version 2.6.32
- connection_ext_or_handle_cmd_useraddr and proposal 196 disagree on the format of ExtORPort USERADDR
- prop224: Intro point aren't transfered between services on HUP
- Y2106 bug in certificate expiration parsing.
- BUG() macros shouldn't be warned as dead-code under coverity.
- Tor sandbox should to reset all of tor browser bundle settings at launch
- Rename hybrid_encrypt functions to prevent accidental use
- test: some prop224 unit tests fail only sometimes
- prop224: Validate received onion addresses on the client side
- Make it clear that RELAY_SENDME cells don't have a payload
- Determine if Tor browser bundle sandbox has been compromised
- The router protocol versions section in dir-spec is out of date
- prop224 should say we use SHA3-256 for rend circuit digests
- prop224 has two section 0.5s
- TAP Hybrid Encryption case 1 is used when the payload is equal to the maximum length
- prop224: Add an introduction point onion key in the descriptor
- disallow tor exec'ing
- Clarify comment about all tor data being encrypted
- Should tor-spec say that nodes MUST NOT use TLS compression?
- NETINFO cells are mandatory, but tor-spec says "may"
- prop224: HS revision counter should persist after service reboot
- Clarify how resolved values are encoded in cells
- What cells can be sent before a VERSIONS cell, and what is their CIRCID_LEN?
- signed integer overflow in unit tests crashes hardened build on 32-bit trusty
- Add link protocol 5 throughout torspec
- clang 4.0 double promotion warnings in clamp_double_to_int64()
- Revise privcount patch series to use trace modules
- Add GitLab CI configs
- Recent tor relays do use canonical connections
- When uploading the first descriptor of a session, call it dirty because "Tor just started"
- The v4 link protocol requires the initiator to set the most significant bit
- NETINFO cells are used for clock skew checks
- tor-spec doesn't say how clients authenticate authorities or fallback directories
- Update tor-spec for ed25519 link authentication keys
- Build tor with Rust code enabled on Windows
- Establish guidelines for using assertions
- Improve documentation for building Tor with Rust
- find_dl_min_and_max_delay's DL_SCHED_DETERMINISTIC case is never used
- Remove or_circuit_t.is_first_hop, because it's not accurate any more
- Refactor circuit_send_next_onionskin() to be less horribly large
- Memory leak in link-handshake/certs_ok_ed25519
- Avoid use of "0" with tor_parse_foo()
- torrc has an optional final newline
- Prop 224 encrypted public key
- Determine Rust support levels for our targeted platforms
- Revise coding standards expectation for tests to be run before review
- Incorrect syscall in some sandbox error messages
- CID 1413651: No retval check in ed25519_donna_blind_public_key()
- Update specs to describe ed25519 point validation and rationale
- prop224: Update spec wrt overlap behavior of HSes
- Relative DataDirectory + RunAsDaemon = Tor can't read or write most of its datadirectory files
- Restricting exit nodes in torrc fails silently
- prop224: Add functions to check for HS v3 support
- prop224: Ignore unparseable v3 introduction points
- tor-spec still says "For a public-key cipher, we use RSA with 1024-bit keys"
- prop224: Disable netflow padding if v3 single onion services are enabled
- SR: Authorities can add a reveal to their own vote, but expect a commit in all votes
- Expose we_fetch_{micro,router_}descriptors on control port
- Document that Sandbox 1 requires linux and seccomp2.
- Assert crash with HSPOST and POSTDESCRIPTOR control port commands
-
Add Travis configs so GitHub forks get CI coverage
#22636 actualpoints::1 best-practice ci component::core tor/tor continuous-integration milestone::Tor: 0.3.2.x-final new-developers owner::patrickod points::.5 priority::high resolution::implemented review-group-21 reviewer::nickm severity::normal status::closed testing travis type::defect unit-testing
- Refactor connection_or_set_state(OPEN) to connection_or_set_state_open()
- Refactor tor's signal handler to avoid undefined behaviour
- ServerDNSAllowBrokenConfig defaults to 1, but TestingTorNetwork overrides it to...1.
- use offsetof() instead of rolling our own
- Clean-up tt_assert() usage in our tests
- Build Tor with clang's -fsanitize=pointer-overflow
- Tor emits inaccurate safesocks warning event whenever you visit a naked IP address
- Add HSv3 chutney networks to tor's make test-network-all
- Mysterious "XXXX RD This is a bug, right?" comment from Nick
- ensure that uint8_t is unsigned char
- Support HTTP CONNECT tunnels as an alternative to SOCKS
- Rip out AUTHDIR_NEWDESCS event?
- authdir_mode_any_nonhidserv() is an obsolete concept
- make test-stem doesn't work
- Prevent pattern of bugs caused by calling get_options() within options_validate() etc
- Is anything undocumented from proposals 158, 162?
- Merge prop220/prop244 to tor-spec/dir-spec
- [warn] channelpadding_compute_time_until_pad_for_netflow(): Bug: Channel padding timeout scheduled 164729ms in the past. Did the monotonic clock just jump?
- Support looking up node by ed25519 identity (backend)
- Make entry relays expire idle client and bridge connections
- Defining TOR_BUILD_TAG and tor_git_revision violates the version spec
- Test that own version passes directory authority checks
- Initial Rust support
- Can't HUP with a relative path in Log line
- Can't have relative DataDirectory with CookieAuthentication enabled
- prop224: Validate ed25519 pubkeys to remove torsion component
- prop224: Load and configure service
- Still had 1 address policies cached at shutdown
- prop224: Groundwork for service implementation
- prop224: Rendezvous circuit e2e crypto
- prop224: Client rendezvous point establishment
- prop224: Client introduction point establishment
- prop224: Client introduction point failure cache
- prop224: Client descriptor cache
- Update release checklist to add step about highlighting certain changes for operators
- test suite triggers Bug: Result does not fit in tor_timegm but does not fail
- "Client asked me to extend back to the previous hop" in small networks
- ClientDNSRejectInternalAddresses should not be deprecated in test networks.
- Fix scan-build warning: true and false branches are identical
- Run clang-scan on tor master
- Clarify "address" in man page: IPv4, IPv6, hostname?
- prop224: Implement HS descriptor fetching
- Remove broken fallback directory mirrors
- Lower the directory overhead for low-bandwidth clients
- Based on measurement of #21205, write/analyze additional proposals and tickets for lowering bw usage for directory stuff
- Write and analyze proposals for fetching consensuses/microdescriptors less frequently?
- Write and analyze proposals for transmitting microdescriptors with less bandwidth
- Write, revise, analyze proposals for ways to use less directory bandwidth
- Restore documentation for the approved-routers file
- Please don't remove ClientDNSRejectInternalAddresses
- [notice] The Tor Directory Consensus has changed how many circuits we must track to detect network failures from 0 to 20.
- prop224: Implement service support.
- Deprecate HTTPProxy and httpproxyauthenticator
- Make sure directory_initiate_request handles pluggable transports correctly
- Revise initial descriptor upload behavior for onion services
- Nickname registration message is confusing
- Relative paths don't work for PidFile and control_auth_cookie
- Clarify our #if{n}def by commenting what they are at the #elif/#else/#endif
- Update DirAuthority man entry for client begindir, no IPv6 DirPort
- Tor should exit if it fails to write its PidFile, under principle of least confusion. Also, maybe Tor should create the directory that the PidFile points to
- Crypto operation counters are unused and obsolete
- Update longclaw's hard-coded IPv6 address
- circuit_package_relay_cell(): Bug: outgoing relay cell sent from src/or/relay.c:701 has n_chan==NULL. Dropping. (on Tor 0.2.8.5-rc )
- Deprecate ReachableDirAddresses and ClientPreferIPv6DirPort
- Error in descriptor download log message
- Cryptic description for download GETINFO options
- Use conditional intervals in heartbeat_callback
- Add "make" targets to generate module callgraph
- Potential heap corruption via `write_escaped_data` in control.c
- Provide timestamps in the CIRC_BW and STREAM_BW events
- Tor should log 1-based hop numbers
- Add IP versions to man page
- Make it clear that Address only works for IPv4
- [Manual] Add some information about sub-domain rules
- Update RSOS Proposal (260)
- Update dir-spec for "reasonably live" consensuses
- Make bootstrapping clients wait before trying an authority
- prop224: Control the number of HSDirs using a consensus parameter
- provide an option to display the expiry date of a given ed25519 signing key
- prop224: Implement client support
- Document PublishServerDescriptor better
- compile warning in test_util.c: In function 'test_util_sscanf'
- Put some actual hsdescs in the unit tests for parsing
- Minor control-spec.txt corrections.
- Clarify in control-spec which relays have ns and md entries and which don't
- tor-spec says additional fields in exitpolicy response are "optional" when they're not
- Design and implement optimizations for socket write limits
- Integrate KIST socket/circuit scheduling
- Implement improved hidden service protocol (prop224)
- connection_handle_event_cb() should handle orconns correctly even when not in OR_CONN_STATE_CONNECTING
- meaningless error message displayed by tor at start up
- Can we speed up bootstrapping by starting circuit building earlier?
- Better error message on GETINFO desc/* when you only have MDs.
- Collect Rob's patch for throttling flows at guards
- Tor warns about public SocksPort addresses twice on startup
- surprising dns responses received from hosts that aren't our resolver
- Give a more appropriate "I'm not an HTTP proxy" message when we get an HTTP request on the control port
Loading
Loading