summaryrefslogtreecommitdiff
path: root/crates/tor-circmgr/src/usage.rs
Commit message (Collapse)AuthorAgeFilesLines
* tor-circmgr: Remove AbstractSpec and FakeSpec.Wesley Aptekar-Cassels2024-09-161-17/+58
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | AbstractSpec and FakeSpec actually make testing more difficult, since they prevent using FakeBuilder in code that relies on the concrete TargetCircUsage and SupportedCircUsage types. Removing them means FakeBuilder can be used in more places, and also means that the test code is closer to the real code, since TargetCircUsage and SupportedCircUsage are now exercised directly in more tests. This did require making one change to a test, which I think was previously testing behaviour that was true for FakeSpec but not for the real code: The mgr::test::isolated test previously asserted that, in the case where three circuits were requested, two with isolation and one without, the non-isolated circuit would be shared with one of the isolated circuits. This was allowed by the FakeSpec::supports function. However, in the actual code, the path is as follows: * AbstractCircMgr::get_or_launch * AbstractCircMgr::prepare_action * CircList::find_open * AbstractSpec::find_supported * abstract_spec_find_supported * OpenEntry::supports * SupportedCircUsage::supports * StreamIsolation::compatible_same_type StreamIsolation::compatible_same_type checks owner_type, which is always zero for non-isolated streams and always non-zero for isolated streams, meaning that a isolated stream will never be compatible with a non-isolated stream. The seems like desirable behaviour, so I simply modified the test to make four connections, two isolated and two not, and checked that the isolated streams never share any circuits, and that the two non-isolated streams use the same circuit. As far as I can tell, this is the intended behaviour in the existing code.
* tor-netdir: Allow access to the `ConsensusBuilder` when building test netdirs.Gabriela Moldovan2024-09-091-2/+2
| | | | | This allows us to set SRVs for example (needed because by default, the test `NetDir` is built from a consensus that doesn't contain any SRVs).
* tor-circmgr: make `GuardMgr` mandatorySteven Engler2024-08-121-34/+30
| | | | | | | Functions that took `Option<&GuardMgr>` now take only `&GuardMgr`. Three unit tests were removed that covered behaviour when no guard manager was set.
* tor-circmg: prepare tests for runtime requirementSteven Engler2024-08-121-130/+135
| | | | | | This wraps some unit tests with `tor_rtcompat::test_with_all_runtimes!`. This is its own commit to get the indentation changes out of the way and declutter the following commit.
* tor-guardmgr: Store whether we're running an onion svc.Gabriela Moldovan2024-04-111-0/+2
| | | | | | The `VanguardMgr` needs to know whether arti is running as an onion service or not, in order to figure out which of the `vanguards_enabled` and `vanguards_hs_service` modes to use.
* tor-guardmgr: Give VanguardMgr a handle to the runtime (fmt).Gabriela Moldovan2024-04-051-2/+12
|
* tor-guardmgr: Give VanguardMgr a handle to the runtime.Gabriela Moldovan2024-04-051-5/+6
|
* Remove temporary functions in CheckedRelay.Nick Mathewson2024-03-281-6/+6
|
* tor-circmgr: Rename pick_path to pick_path_with_vanguards.Gabriela Moldovan2024-03-201-1/+1
| | | | | | When the `vanguards` feature is enabled, we define `pick_path_with_vanguards` instead of `pick_path`, rather than conditionally defining 2 different versions of the same function.
* tor-circmgr: Give pick_path a VanguardMgr (fmt).Gabriela Moldovan2024-03-201-4/+36
|
* tor-circmgr: Give pick_path a VanguardMgr.Gabriela Moldovan2024-03-201-6/+26
|
* tor-circmgr: Tell HsPathBuilder what kind of circuit stub is needed (fmt).Gabriela Moldovan2024-03-201-5/+2
|
* tor-circmgr: Tell HsPathBuilder what kind of circuit stub is needed.Gabriela Moldovan2024-03-201-2/+8
|
* tor-circmgr: Move pick_path out of AnonymousCircuitBuilder.Gabriela Moldovan2024-03-141-3/+1
| | | | This enables us to make `AnonymousCircuitBuilder` private.
* tor-circmgr: Remove unused require_stability.Gabriela Moldovan2024-03-141-3/+0
|
* tor-circmgr: Use HsPathBuilder instead of ExitPathBuilder.Gabriela Moldovan2024-03-141-6/+7
|
* tor-circmgr: Implement AnonymousPathBuilder for ExitPathBuilder.Gabriela Moldovan2024-03-141-1/+3
|
* Annotate remaining "TODO #504" comments explaining why they are there.Nick Mathewson2024-03-121-4/+14
|
* circmgr: Rename "any exit" code to reflect its real purposeNick Mathewson2024-03-121-1/+1
| | | | | (These functionalities are for onion service circuits, and don't actually need to go to an exit at all.)
* New crate to define a high-level API for relay selectionNick Mathewson2024-03-121-45/+2
|
* Mark code with "TODO #504" and "TODO #789".Nick Mathewson2024-02-221-0/+5
| | | | | | | | I've done this by looking for every non-test instance of pick_relays or pick_n_relays, and for every non-test usage of any non-ID-related method on Relay or UncheckedRelay. Part of #504.
* Teach *CircUsage structs about stability.Nick Mathewson2024-02-201-12/+57
| | | | | | | | | | | This lets us check long_lived_ports only when we are constructing a TargetCircUsage, and lets us remember whether or not the circuit we've built is suitable for non-long-lived ports. It also lets us decide whether existing circuits are compatible with new requests. Closes #1100.
* Add a couple of TODOs about Stable-based path selection.Nick Mathewson2024-02-201-0/+2
|
* When building paths, require the Stable flag as appropriateNick Mathewson2024-02-201-2/+12
| | | | | | | | | | | | The Stable flag means that circuits build through a given relay are not likely to get closed precipitously. (Currently, the authorities guess this by looking at the relay's mean time between failures.) Generally speaking, paths must be Stable if: * They are going to get used for a rendezvous point. * They are going to get used for an introduction point. * They are going to get used for exiting to a member of long_lived_ports.
* Downgrade some messages to traceIan Jackson2023-11-291-2/+2
| | | | | These messages are very verbose and I doubt anyone will want them, usually, even when debugging.
* arti-client: add exit selection with GeoIP country codeseta2023-09-181-6/+89
| | | | | | | | | | | | | | | | | | | | This threads the country codes work through the rest of the codebase: - `tor-dirmgr` will now enable GeoIP with the embedded database when the `geoip` future is enabled - This can be extended later using the `DirMgrConfig` to allow specifying a custom database; this is not done here, though - `tor-circmgr`'s `SupportedCircUsage` and `TargetCircUsage` fields gain new `country_code` members to allow filtering circuits by country - These are `()` in builds where the `geoip` feature is not enabled -- doing it this way means we don't have to copy and paste huge swathes of code, since we can't use `#[cfg]` in patterns - `ExitPathBuilder` gains (hacked-in) support for choosing a relay with the correct country code - Due to the lack of conjuction, we just copy and paste a small bit, pending further refactoring - `StreamPrefs` now lets you specify a country code, letting embedders make use of the feature
* circmgr: Mark a number of internal APIs as hs-only.Nick Mathewson2023-04-061-0/+2
| | | | | | This resolves a few dead-code warnings. Closes #801.
* Add some missing importsIan Jackson2023-03-281-1/+4
| | | | | | | | Now nailing-cargo +stable clippy -p tor-hsclient --all-features --all-targets actually works. squash! Add some missing imports
* circmgr: Make a separate SupportedCircUsage for HS stuffNick Mathewson2023-03-221-3/+6
| | | | | This uncovered a bug: NoUsage wasn't correct for Hs circuits because of its behavior with channel_usage().
* circmgr: Add a usage for building the beginning of an hs circNick Mathewson2023-03-221-0/+25
| | | | | This only builds the first 3 hops. It can be extended to a fourth hop later -- or not, depending on the circuit kind.
* circmgr: make some imports conditionalNick Mathewson2022-11-181-2/+1
| | | | | This resolves an "unused import" warning when bridges are not enabled.
* Remove a spurious log that I committed by mistake. Whoops.Nick Mathewson2022-11-151-1/+0
|
* Implement more cases for `DirSpecificTarget`.Nick Mathewson2022-11-151-1/+22
| | | | | | Without this fix, any attempt to actually use a circuit for a `DirSpecificTarget` will fail, since supports() will say false, and `restrict_mut()` will say `NotSupported`.
* tor-circmgr: Make get_or_launch_dir_specific feature-specificIan Jackson2022-10-211-2/+9
|
* circmgr: Plumbing for direct to target directory circuitsIan Jackson2022-10-201-1/+13
| | | | | The target is identified by Into<OwnedChanTarget>. We introduce corresponding TargetCircUsage and SupportedCircUsage.
* channel usage: Rename CU::UserTraffic from CU::ExitIan Jackson2022-08-171-1/+1
| | | | | As suggested in https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/657#note_2827246
* Move ChannelUsage from tor_proto to tor_chanmgrIan Jackson2022-08-171-1/+1
| | | | | | | | | | | Replace Channel::note_usage with Channel::engage_padding_activities, which unconditionally causes the channel to (start to) do netflow padding things. The condition now lives in chanmgr. Addresses https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/657#note_2826094
* Provide ChannelUsage and plumb it all the way downIan Jackson2022-08-161-0/+11
| | | | | | | | | | | | | Channel padding depends on what the channel is being used for. We therefore need to let the channel code know this information. The implementation of the per-channel padding control logic will be in the new note_usage function, which for now is simply a stub. A future commit will introduce a `PaddingControlState` which lives in the channel frontend; consult the doc comment for that type to see why the plumbing through the channel manager terminates in the channel frontend.
* tor-netdir: Collapse by_id and by_relay_id into a single fn.Nick Mathewson2022-08-101-4/+5
| | | | | | There are some downstream changes required for this to work, but they are all just unit tests that could no longer infer the type of an Ed25519 key.
* Avoid a simple usage of ed_identity() in circmgr testsNick Mathewson2022-08-021-2/+1
|
* tor-linkspec: Refactor out traits to represent a relay's ID set.Nick Mathewson2022-08-021-1/+1
| | | | | | | | | | | | | | We want the set of identities supported by a relay to be extensible in the future with minimal fuss; we'd also like to make working with these ID sets more convenient. To handle that, this commit adds a new trait for "Something that has the same IDs as a relay" and a new object for "an owned representation of a relay's IDs." This commit introduces a similar trait for "Something with a list of SocketAddr, like a relay has." There's no owned equivelent for that, since Vec<SocketAddr> is already a thing. Closes #428.
* tor-netdir: testnet: Make construct_netdir infallible (rustfmt)Ian Jackson2022-06-131-3/+1
| | | | | Run rustfmt. Separate commit to make review of the substantive commit easier.
* tor-netdir: testnet: Make construct_netdir infallibleIan Jackson2022-06-131-1/+0
| | | | | This is a *lot* of unwraps. The function takes no parameters and is used only for testing. It ought to be infallible.
* Use testing_rng() in tests throughout our crates.Nick Mathewson2022-06-021-2/+3
| | | | | | This only affects uses of thread_rng(), and affects them all more or less indiscriminately. One test does not work with ARTI_TEST_PRNG=deterministic; the next commit will fix it.
* Resolve the new `derive_partial_eq_without_eq` lint.Nick Mathewson2022-05-231-1/+1
| | | | It's a little overzealous sometimes, but it's mostly to the good.
* squash! Bump every crate's edition to 2021.Nick Mathewson2022-04-251-1/+0
| | | | | Remove all `use` statements for `TryFrom` and `TryInto`. These are now redundant in Rust 2021.
* circmgr: Avoid a race condition in circuit usage restrictionNick Mathewson2022-04-041-5/+7
| | | | | | | | | | | | | | | | | | | | | | We were treating restrict_mut() failures as internal errors, and using internal errors to represent them. But in fact, these failures are entirely possible based on timing. Here's how it happens: * Two different circuit requests arrive at the same time, and both notice a pending circuit that they could use. * The pending circuit completes; both pending requests are notified. * The first request calls restrict_mut(), and restricts the request in such a way that the second couldn't use it. * The second request calls restrict_mut(), and gets a failure. Because of this issue, we treat these errors as transient failures and just wait for another circuit. Closes #427. (This is not a breaking API change, since `AbstractSpec` is a crate-private trait.)
* Remove allow(clippy::disallowed_methods) lint.Nick Mathewson2022-03-301-2/+0
|
* Merge branch 'no-system-time' into 'main'eta2022-03-301-7/+13
|\ | | | | | | | | | | | | Don't use SystemTime::now() Closes #306 See merge request tpo/core/arti!365
| * use wallclock where possible in teststrinity-1686a2022-02-261-0/+2
| |
generated by cgit v1.2.3 (git 2.46.0) at 2026-08-27 05:33:18 +0000