summaryrefslogtreecommitdiff
path: root/crates
Commit message (Collapse)AuthorAgeFilesLines
...
* | | tor-hsservice: key expiry: Code motion and reindentingIan Jackson2024-01-241-26/+22
| | | | | | | | | | | | | | | Now the remove_if_expire closure, which does the actual work, is lexically outside the macro, and the macro is trivial.
* | | tor-hsservice: key expiry: Introduce HsTimePeriodKeySpecifier traitIan Jackson2024-01-241-10/+49
| | | | | | | | | | | | | | | | | | This allows us to have a closure containing runtime-polymorphic code, reducing monomorphisation and moving code out of a macro into a closure.
* | | tor-hsservice: key expiry: Rely on limiting publisher key scanIan Jackson2024-01-241-2/+6
| | | | | | | | | | | | | | | | | | | | | | | | Our pattern is now supposed to ensure that we don't see any irrelevant keys. So if we do, that's a bug. (The code layout is getting increasingly odd. We'll sort that out along with some code motion later.)
* | | tor-hsservice: key expiry: Limit publisher key scanIan Jackson2024-01-241-5/+5
| | | | | | | | | | | | | | | | | | | | | Only examine keys for this HS. This avoids thinking about every key for every HS for every netdir change. That's quadratic in the number of HS's.
* | | tor-hsservice: key expiry: clippy followupIan Jackson2024-01-241-3/+2
| | | | | | | | | | | | | | | These are now the last uses of these values, so they need the "no clone" form.
* | | tor-hsservice: key expiry: Do publisher expiry from reactorIan Jackson2024-01-243-131/+10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | In principle there is a small race with doing this in a separate task: the reactor and the sweeper might process the new netdir at different times. For example, if two netdir updates come in quick succession, and the sweeper is slower, the sweeper might still be running with the previous TPs as the publisher is creating keys in the new TPs. Theoretically, the sweeper might delete keys the reactor has just generated. This is also considerably simpler.
* | | tor-hsservice: key expiry: Break out KeystoreSweeper::expire_keysIan Jackson2024-01-242-43/+55
| | | | | | | | | | | | We're going to change where this is called.
* | | tor-netdir: Make hs_dirs_{upload, download} take separate args instead of tuple.Gabriela Moldovan2024-01-245-5/+10
| | | | | | | | | | | | Prompted by https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/1904#note_2987777
* | | tor-netdir, tor-hscrypto: Add function for computing SRV period offset.Gabriela Moldovan2024-01-246-13/+45
| | | | | | | | | | | | Part of #1166
* | | tor-hscrypto: Remove TimePeriodOffset, offset_within_period().Gabriela Moldovan2024-01-243-27/+3
| | | | | | | | | | | | | | | | | | These are now unused. Part of #1166
* | | tor-hsservice: Generate revision counter using the start of SRV period (fmt).Gabriela Moldovan2024-01-241-2/+1
| | |
* | | tor-hsservice: Generate revision counter using the start of SRV period.Gabriela Moldovan2024-01-243-22/+20
| | | | | | | | | | | | Closes #1166
* | | tor-netdir: Simplify `hs_dirs_upload` by only returning the `Relay`.Gabriela Moldovan2024-01-244-24/+19
| | | | | | | | | | | | | | | | | | | | | | | | The publisher is the only user of `hs_dirs_upload`. It turns out it never actually uses the first element of the yielded `Item`s, so we can simplify `hs_dir_upload` to only return the HsDir `Relay`s. Part of #1166
* | | tor-hsservice: Store HsDirParams in the publisher's TimePeriodContext.Gabriela Moldovan2024-01-241-10/+19
| | | | | | | | | | | | Part of #1166
* | | tor-netdir: Make hs_all_time_periods return HsDirParams.Gabriela Moldovan2024-01-244-11/+14
| | | | | | | | | | | | | | | | | | | | | | | | | | | The descriptor publisher uses this function to obtain the list of relevant time periods. It will soon also need to know the `srv_lifespan` associated with each time period, so we change this function to return `HsDirParams`. Part of #1166
* | | Merge branch 'less_experimental' into 'main'gabi-2502024-01-245-9/+12
|\ \ \ | | | | | | | | | | | | | | | | Make lower-level "hs-service" features non-experimental See merge request tpo/core/arti!1908
| * | | Remove dependency from hsrproxy to tor-proto::experimental-apiNick Mathewson2024-01-231-1/+1
| | | | | | | | | | | | | | | | It would appear nothing was using this.
| * | | Mark lower-level hs-service features non-experimentalNick Mathewson2024-01-234-8/+11
| | | |
* | | | Merge branch 'fslock-guard' into 'main'Nick Mathewson2024-01-233-0/+322
|\ \ \ \ | |/ / / |/| / / | |/ / | | | fslock-guard: sketch implementation See merge request tpo/core/arti!1900
| * | fslockguard: try to document windows assumptionsNick Mathewson2024-01-231-0/+41
| | | | | | | | | | | | | | | These are necessarily a bit hand-wavey, but I think they summarize what we are assuming.
| * | Remove an XXX: fslock does indeed use O_CLOEXEC.Nick Mathewson2024-01-231-1/+0
| | |
| * | fslockguard: implement a delete_lock_file function.Nick Mathewson2024-01-232-3/+30
| | |
| * | Use fslock-arti-forkNick Mathewson2024-01-232-20/+7
| | |
| * | fslock-guard: Write down the locking protocol on UnixIan Jackson2024-01-231-0/+65
| | | | | | | | | | | | | | | | | | Text from here https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/1900#note_2986683 with a few minor fixes.
| * | fslock-guard: Rename os modulesIan Jackson2024-01-231-2/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | These two modules are key to the implementation of the locking protocols. We must define the locking protocol in terms of underlying OS semantics (since Unix and Windows have different fs concepts and different concurrency semantics) and therefore, although we are sharing some code between the implementations, these modules are what defines the two protocols.
| * | Give a formal semantics for the fslock operations.Ian Jackson2024-01-221-0/+15
| | |
| * | fslock-guard: sketch implementationNick Mathewson2024-01-213-0/+188
| | | | | | | | | | | | | | | | | | | | | This is not yet "correct", since it will rely on https://github.com/brunoczim/fslock/pull/15 (Conceivably, it might be better to make the `fslock` crate rm-safe.)
* | | Merge branch 'hsdirparams' into 'main'gabi-2502024-01-234-15/+44
|\ \ \ | |_|/ |/| | | | | | | | | | | | | | Expose SRV lifespan info from netdir Closes #1254 See merge request tpo/core/arti!1903
| * | netdir: Make hs_dirs_upload() yield &HsDirParams.Nick Mathewson2024-01-232-2/+4
| | | | | | | | | | | | Closes #1254.
| * | Expose HsDirParams, and give it accessors.Nick Mathewson2024-01-232-2/+24
| | |
| * | hsservice: Make HsDirParams include the SRV lifespan.Nick Mathewson2024-01-232-11/+16
| | | | | | | | | | | | This is part of #1254.
* | | tor-hsservice: Move compute_subcredentials to RendRequestContext.Gabriela Moldovan2024-01-223-103/+100
| | | | | | | | | | | | Part of #1242
* | | tor-hsservice: Do not store the subcredentials in RendRequestContext.Gabriela Moldovan2024-01-223-21/+22
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Previously, the subcredentials were computed in `IptEstablisher::launch` and stored in `RendRequestContext`. This caused long-running services to report errors like: ``` WARN tor_hsservice::helpers: Problem while accepting rendezvous request: error: Could not process INTRODUCE request: Introduction handshake was invalid: Circuit-extension handshake authentication failed ``` for clients using newer subcredentials than the ones the service had at the time the IPT was established. Fixes #1242
* | | tor-hsservice: Add an error type for subcredential lookup failures.Gabriela Moldovan2024-01-222-0/+6
|/ / | | | | | | | | | | | | | | The subcredential lookup will be moved to `IntroRequest::decrypt_from_introduce2`. The error returned on failure is going to be `IntroRequestError::Subcredentials`. Part of #1242
* | Merge branch 'shutdown_on_drop' into 'main'gabi-2502024-01-195-14/+27
|\ \ | | | | | | | | | | | | | | | | | | Clarify shutdown behavior when RemoteOnionService is dropped. Closes #1238 and #1236 See merge request tpo/core/arti!1899
| * | Document when (most) spawned tasks will be canceled.Nick Mathewson2024-01-182-0/+6
| | |
| * | hsservice: Temporarily remove .pause() (formerly stop()).Nick Mathewson2024-01-182-0/+3
| | | | | | | | | | | | | | | We don't need to implement this for our first release of onion services, but we shouldn't ship a function that calls todo!().
| * | Clarify usage and behavior of shutdown_tx.Nick Mathewson2024-01-181-6/+7
| | |
| * | Clarify IptEstablisher shutdown rulesNick Mathewson2024-01-181-8/+11
| | | | | | | | | | | | | | | | | | There was no actual bug here; just some missing comments. Closes #1236.
* | | Merge branch 'tolerant_intro_established' into 'main'gabi-2502024-01-191-17/+12
|\| | | |/ |/| | | | | | | | | Do not reject INTRO_ESTABLISHED messages with extensions Closes #1238 See merge request tpo/core/arti!1898
| * Do not reject INTRO_ESTABLISHED messages with extensionsNick Mathewson2024-01-181-17/+12
| | | | | | | | | | | | | | | | | | | | | | | | | | The spec says that we should ignore unrecognized extensions. Per discussion at torspec#249, this is still correct. Additionally, this commit moves responsibility for checking INTRO_ESTABLISHED messages into IptMsgHandler::handle_msg, to make sure that the circuit is torn down as soon as possible on a bad reply. (There's nothing to check yet, but there will be once we start sending extensions that expect a reply.) Closes #1238.
* | Mark ShutdownStatus as must_useNick Mathewson2024-01-181-0/+1
|/ | | | Followup from !1895.
* Merge branch 'publisher-shut-down' into 'main'Nick Mathewson2024-01-182-7/+13
|\ | | | | | | | | tor-hsservice: The publisher should exit when the IPT manager says so. See merge request tpo/core/arti!1895
| * tor-hsservice: The publisher should exit when the IPT manager says so.Gabriela Moldovan2024-01-182-7/+13
| | | | | | | | | | | | The publisher needs to shut down when `IptPublisherView::sawait_update()` returns `None`, not pause the uploads.
* | Merge branch 'downgrade-todo' into 'main'Nick Mathewson2024-01-181-1/+1
|\ \ | | | | | | | | | | | | | | | | | | tor-hsservice: Reference #1226 instead of #1219 in a TODO. Closes #1219 See merge request tpo/core/arti!1896
| * | tor-hsservice: Reference #1226 instead of #1219 in a TODO.Gabriela Moldovan2024-01-181-1/+1
| |/ | | | | | | | | | | | | | | `note_publication_attempt` can only fail: * due to an internal error, in which case there is no point in retrying * if `PublishIptSet::save` fails, i.e. if we fail to write to persistent storage (this is #1226)
* | tor-hsservice: Store the StateMgr inside OnionServiceState.Gabriela Moldovan2024-01-183-25/+61
| | | | | | | | | | | | This introduces an internal `OnionServiceStateMgr` trait, which enables us to store the `StateMgr` inside the `OnionServiceState` (without having to parameterize `OnionServiceState` on `S: StateMgr`).
* | tor-hsservice: Add TODO about testing arti hss.Gabriela Moldovan2024-01-181-0/+1
| |
* | tor-hsservice: Add a test for OnionService::onion_name.Gabriela Moldovan2024-01-181-0/+31
| |
* | arti-client: Move state_dir and mistrust handling to separate function.Gabriela Moldovan2024-01-181-11/+19
| | | | | | | | This reduces code duplication.
generated by cgit v1.2.3 (git 2.46.0) at 2026-08-27 23:16:00 +0000