summaryrefslogtreecommitdiff
path: root/crates/arti-rpcserver/src
Commit message (Collapse)AuthorAgeFilesLines
...
* | cancel: Document deadlock/panic issue and how to avoid it.Nick Mathewson2025-01-221-2/+19
| |
* | rpc: Make cancel requests uncancellable.Nick Mathewson2025-01-222-27/+56
| | | | | | | | | | | | | | | | | | The current cancel code is prone to deadlock, so the easiest way to solve it appears to be making cancel requests themselves uncancellable. I've included a test to verify the behavior; previously, this test caused a deadlock.
* | cancel: Add a TODO about moving the Future inside the lock.Nick Mathewson2025-01-221-0/+5
| |
* | Cancel: Drop lock before calling waker.Nick Mathewson2025-01-221-0/+1
| |
* | cancel: Remove FusedFuture implementation.Nick Mathewson2025-01-221-12/+1
| | | | | | | | | | | | | | Nothing used it, and it has some semantic complexity. (see https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/2722#note_3149591 )
* | cancel: Use Waker::clone_from to avoid needless clones.Nick Mathewson2025-01-221-1/+9
| | | | | | | | | | | | The Waker::clone_from implementation uses Waker::will_wake to avoid unnecessarily cloning a Waker that it already has a copy of.
* | rpc: Implement the rpc:cancel command.Nick Mathewson2025-01-162-1/+79
| |
* | rpc: Better RpcErrorKind for for RequestCancelled.Nick Mathewson2025-01-161-4/+7
| |
* | rpc: Strengthen guarantees from Cancel.Nick Mathewson2025-01-161-20/+124
|/ | | | | | | | | | | Cancellation is now fallable, which allows us to detect attempts to cancel which cannot work. We now guarantee that when you try to cancel a `Cancel<F>` future, either the cancel operation will succeed, or the future will return (or will have already returned) Ok(), but not both, and not neither. Closes #818.
* Merge branch 'rpc-clean-up-after-auth' into 'main'Nick Mathewson2025-01-151-0/+5
|\ | | | | | | | | rpclib: Clean up after performing cookie auth See merge request tpo/core/arti!2716
| * rpc: More documentation on cookie_continue.Nick Mathewson2025-01-151-0/+5
| |
* | Merge branch 'rpc-renaming-2' into 'main'Nick Mathewson2025-01-152-62/+56
|\ \ | | | | | | | | | | | | | | | | | | rpc: Rename new_stream_handle to new_oneshot_client. Closes #1664 See merge request tpo/core/arti!2715
| * | rpc: Clean up documenation surrounding OneshotClientNick Mathewson2025-01-151-11/+13
| | |
| * | rpc: Rename new_stream_handle to new_oneshot_client.Nick Mathewson2025-01-152-52/+44
| |/ | | | | | | | | | | | | | | This method doesn't actually create a new stream; it creates a single-use client object that can be used with SOCKS to launch a new stream, and capture an RPC object for that stream. Closes #1664.
* / arti-rpcserver: Rename MethodNotFound to NoSuchMethod.Nick Mathewson2025-01-153-4/+4
|/ | | | Closes #1500.
* rpc: Clarify auth-repetition rules.Nick Mathewson2025-01-151-2/+5
|
* rpc: Document cookie messages a little more.Nick Mathewson2025-01-151-1/+14
|
* rpc: Refactor Cookie and UnloadedCookie into a single type.Nick Mathewson2025-01-152-2/+11
|
* rpc: consolodate naming of "inherent" auth.Nick Mathewson2025-01-152-9/+13
| | | | | | | | | We don't want to call this "unix path" anywhere, since it corresponds to _any_ case where the ability to negotiate a successful connection means that the client is authorized. We also don't want to call it "none": The authentication is inherent to the connection, not nonexistent.
* rpc: Tweak cookie protocol to bind both nonces.Nick Mathewson2025-01-151-4/+9
| | | | | | | | | | Previously participants in the cookie protocol only bound the peer nonce in their MACs. With this change, they bind both nonces. This change is _probably_ not necessary for security, but it can't hurt. It follows a general principle that Adam Langley told me a long time ago: you won't regret binding more, but you might regret binding less.
* rpc: Keep Cookie in an Arc.Nick Mathewson2025-01-151-2/+2
| | | | | Since this is a secret value, it's probably best not to copy it all over the place.
* arti-rpcserver: Server side of cookie auth.Nick Mathewson2025-01-153-8/+189
|
* arti-rpcserver: Tell connections what kind of auth to expect.Nick Mathewson2025-01-154-14/+32
|
* arti-rpcserver: move inherent authentication to its own module.Nick Mathewson2025-01-152-69/+83
|
* arti-rpcserver: remove some dead code.Nick Mathewson2025-01-151-57/+0
| | | | | Now that we have a solid idea of how connections happen, it's clear we won't need to enable this negotiation mechanism.
* Implement banner for RPC protocol.Nick Mathewson2025-01-091-2/+12
| | | | Closes #1753
* clippy: deny `mod_module_files`Steven Engler2025-01-061-0/+1
| | | | | | Denies 'mod.rs' files for consistency. https://rust-lang.github.io/rust-clippy/master/index.html#mod_module_files
* add_warnings, *: Allow clippy::needless_lifetimesNick Mathewson2024-12-031-0/+1
| | | | | | | | In 1.83, this warning triggers on many of our crates. We're thinking of fixing them all, but for now, we're going to disable the warning. This is part of #1765.
* Remove echo structsNeel Chauhan2024-10-251-14/+0
|
* arti-rpcserver: Remove echo_on_sessionNeel Chauhan2024-10-251-12/+0
|
* rpc: Tweak release_owned to give an error on second removal of "connection".Nick Mathewson2024-10-231-16/+20
|
* rpc: Use a const for the "connection" object ID.Nick Mathewson2024-10-231-2/+9
|
* rpc: Simplify method-dispatch-bypass logic a bitNick Mathewson2024-10-232-9/+2
| | | | | | | | | | | | | | | This commit removes the separate function for asking whether to bypass the dispatch code. Instead, it gives the "invoke with bypass" function an error to return when no dispatch is warranted, and moves the whole responsibility for method dispatch or non-dispatch back into tor-rpcbase. I had to add an ObjectId argument to `invoke_rpc_method` to make this work, but that's probably a good thing. Additionally, this commit tweaks the derive-deftly macro to prevent you from asking for dispatch bypass on special methods, where it isn't implemented (and doesn't really make sense).
* rpc: Add support for dropping the special "connection" object ID.Nick Mathewson2024-10-231-9/+19
|
* rpc: Move "rpc:release" to be a method on the rpc object.Nick Mathewson2024-10-231-24/+24
| | | | | | | Now, instead of telling the session or the connection to drop the object ID, we tell the object ID to go away. Closes #1663.
* rpcserver: detect and use methods that bypass method dispatch.Nick Mathewson2024-10-231-3/+8
|
* rpc: Remove unimplemented rpc:downgrade method for now.Nick Mathewson2024-10-231-9/+0
| | | | | If and when we implement this, it will likely be different; arti#868 has some thoughts on the implications.
* rpc: Move some object-id-manipulation methods to a new module.Nick Mathewson2024-10-233-50/+63
| | | | | These will have different implementations soon; this is a more logical place for them.
* Merge branch 'arti-rpcserver-comment' into 'main'wesleyac2024-10-211-1/+1
|\ | | | | | | | | Fix typo in arti-rpcserver auth.rs comment See merge request tpo/core/arti!2558
| * Fix typo in arti-rpcserver auth.rs commentNeel Chauhan2024-10-181-1/+1
| |
* | rpc: List missing required features in error dataNick Mathewson2024-10-171-5/+7
| | | | | | | | | | | | | | | | | | | | Our spec says that when the RPC client has said "I require you to have feature X" and we don't have it, we need to include the feature(s) we don't have in an `rpc:unsupported_features` field of our error. Also, add an integration test for this behavior. Closes #1662
* | Remove Rpc* variants from tor_error::ErrorKindNick Mathewson2024-10-173-21/+27
| | | | | | | | | | | | | | | | | | | | | | | | | | These are not regular ErrorKinds, since they can never occur in an error that's meant to be returned from a Rust API like `arti-client`. Instead, they only exist for errors returned from RpcError. (I can't find the place where we discussed this previously, but the rationale is that if an ErrorKind never makes sense in response to something that the user does from Rust, we should never have that be an ErrorKind. The fact that the removed kinds do not actually appear outside the RPC system suggests that this is reasonable.)
* | rpcserver: Wrap InvokeError as internal errorNick Mathewson2024-10-172-6/+8
|/ | | | | | | | | I'm about to remove HasKind from InvokeError, which would otherwise break this code. These errors are all in fact internal errors, since in this context they can only stem from incorrectly formed calls to `invoke_special_method`.
* Merge branch 'mq-prevent' into 'main'Ian Jackson2024-10-151-2/+5
|\ | | | | | | | | | | | | Use clippy to prevent non-mq use of mpsc::channel Closes #1659 See merge request tpo/core/arti!2536
| * Use mpsc_channel_no_memquota in the RPC system (adding a TODO)Ian Jackson2024-10-151-2/+5
| | | | | | | | | | We need to decide whether RPC will participate in memquota. Perhaps it should. But that's for the future.
* | Replace _ => panic!() elsewhereIan Jackson2024-10-151-1/+1
|/
* arti-rpcserver: minimal handling for request.meta.requireNick Mathewson2024-10-072-2/+30
| | | | | | | | | | We have defined a forward-compatibility mechanism for the RPC system, where a request may list one or more required "features", and if any feature is absent, the request will fail. Since no "features" are currently implemented, this code does the bare minimum to implement this mechanism, by rejecting every request with a nonempty "require" field.
* arti-rpcserver: Correctly handle rpc.meta without `updates`Nick Mathewson2024-10-071-0/+1
| | | | | The `update` field in request.meta is optional, so we should allow it to default to `false` when it is absent.
* rpcserver: Remove TODO about optionality of paramsNick Mathewson2024-10-071-3/+0
| | | | | | | The spec no longer says that `params` is optional, so we can remove the TODO about making it optional. Closes #1665.
* rpc: clarify documentation and fix typosNick Mathewson2024-10-022-3/+6
| | | | Based on review from @opara.
generated by cgit v1.2.3 (git 2.46.0) at 2026-08-27 13:13:23 +0000