| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| ... | |
| |
|
|
| |
Signed-off-by: David Goulet <[email protected]>
|
| |
|
|
| |
Signed-off-by: David Goulet <[email protected]>
|
| |
|
|
|
|
| |
No need to have these comments in the implementation of the Trait.
Signed-off-by: David Goulet <[email protected]>
|
| |
|
|
|
|
|
| |
Put code for test inside the test module instead of gated by a
cfg(test).
Signed-off-by: David Goulet <[email protected]>
|
| |
|
|
|
|
|
|
| |
In unit tests, we set a 1000 by default but else, we don't set a default
so our implementation doesn't forget to set the right value from the
consensus.
Signed-off-by: David Goulet <[email protected]>
|
| |
|
|
| |
Signed-off-by: David Goulet <[email protected]>
|
| |
|
|
|
|
|
| |
Don't put in the comments, let the top-level module comments do that
job.
Signed-off-by: David Goulet <[email protected]>
|
| |
|
|
|
|
|
|
| |
Instead of having the congestion.rs test module public, make a
test_utils.rs file that contains code for other unit tests within the
congestion module.
Signed-off-by: David Goulet <[email protected]>
|
| |
|
|
|
|
|
| |
Turns out that RoundTripEstimatorParams is the same exact thing so one
less data structure.
Signed-off-by: David Goulet <[email protected]>
|
| |
|
|
| |
Signed-off-by: David Goulet <[email protected]>
|
| |
|
|
| |
Signed-off-by: David Goulet <[email protected]>
|
| |
|
|
|
|
|
|
|
| |
We also pass along congestion signals that may or may not be used by our
congestion control algorithm.
Part of #534
Signed-off-by: David Goulet <[email protected]>
|
| |
|
|
|
|
|
|
|
|
| |
It is official, congestion control is now used at this commit by the
circuit reactor making circuit/sendme.rs unused. Will be removed with
another commit.
Related #534
Signed-off-by: David Goulet <[email protected]>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
The congestion control parameters are created from the consensus
parameters (netparams) and then put into the CircParameters object that
is then passed down the tor-proto crate.
Because different parameters are selected depending on the circuit type
(onion vs exit vs sbws), a CircuitType enum is introduced for the sole
purpose of being used to select the right parameters.
Related #534
Signed-off-by: David Goulet <[email protected]>
|
| |
|
|
|
|
|
|
|
|
|
| |
The big one! This is the Vegas algorithm implementation that is hooked
to the CongestionControl object by implementing the algorithm trait.
Still, at this commit, nothing is being used by the circuit reactor yet.
Related #534
Signed-off-by: David Goulet <[email protected]>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add the top level CongestionControl object that will allow the circuit
reactor to use it in order to decide if a cell can be sent or not. In
order to be used, it is configured with an algorithm that implements the
CongestionControlAlgorithm trait.
The Fixed Window algorithm is also added which essentially implements
the SENDME logic as arti knows it today with a fixed window size.
The SENDME code has been refactor in two different logical steps in
order to accomodate the future Vegas algorithm for which the congestion
window logic and SENDME validity is seperated.
There is now a SENDME validator that takes care of tracking the tags
(authenticated SENDMEs) and validating them upon reception. Then, if
valid, the window management is passed down the congestion control
algorithm, at this commit, FixedWindow object.
Related #534
Signed-off-by: David Goulet <[email protected]>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This commit adds the congestion window object, a round trip estimator
(RTT) and a state enum. These 3 entities are used by congestion control
in a generic way that is they are passed and used by any algorithm.
At this commit, they are not used hence the allow deadcode attribute for
now in order to minimize the build warnings.
We also introduce the params.rs file containing the parameters, taken
from consensus, used to configure these objects. They will be exposed to
the tor-cirmgr crate to build the CircParameters. More will come.
This also introduces the congestion/ directory that will contain more
code in future commits.
Related #534
Signed-off-by: David Goulet <[email protected]>
|
| |
|
|
|
|
|
|
| |
This moves the `NetParameters -> KistParams` conversion to
`tor-chanmgr`.
Prompted by
https://gitlab.torproject.org/tpo/core/arti/-/merge_requests/2706#note_3147557
|
| | |
|
| | |
|
| |
|
|
|
| |
This renames UnsupportedStreamOpsHandle to NoOpStreamOpsHandle for
clarity (the old name kind of sounded like the name of an error type).
|
| | |
|
| |
|
|
|
| |
This will enable us to update the channel's KIST configuration whenever
there is a change in the consensus or config.
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
| |
This is unfortunately necessary, because after the channel handshake, we
need to give the channel reactor a `StreamOps` handle to the underlying
stream.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Needed for cases where we wrap an object that implements `StreamOps` in
an external type, thereby losing access to the `StreamOps`
functionality. For example, during the channel handshake, we `.split()`
the stream that implements `StreamOps`, which leaves us with a
`SplitSink` and a `SplitStream`, neither of which implement `StreamOps`.
Getting a handle to the underlying object that implements `StreamOps`
(for example, a file handle) *before* the stream is `.split()` enables
us to use `StreamOps` to manipulate the underlying split stream.
This commit also introduces a special `UnsupportedStreamOpsHandle`,
which is a type that implements `StreamOps`, but always returns an
error. This type is meant to simplify error handling and usage, and is
meant to be used in cases where `StreamOps` is not supported. TODO: the
name of this type is pretty confusing (it's very similar to
`UnsupportedStreamOp`, which is an error type), and should probably be
renamed to something else (`NoOpStreamOpsHandle`,
`BrokenStreamOpsHandle`, `DummyStreamOpsHandle` come to mind...).
Note: this changes the `StreamOps` trait to be slightly different from
what I originally envisioned in !2660 and #1769
|
| |
|
|
|
|
|
|
| |
Note: this commit makes `tor-proto` depend on `tor-netdir`
(because it adds a `KistParams` type that is buildable from
`NetParameters`, which is defined in `tor-netdir`).
Closes #1729
|
| |
|
|
| |
(We don't add it to the handful of unit tests that don't use an executor.)
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
| |
This had a TODO about returning a "status indication instead of just ()"
so this commit adds some status indication that we can expand later if
needed.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
The affected crates follow our regular versioning.
They all get bumped to 0.26.0.
Done with
```
for crate in $(./maint/list_crates |grep '^arti-\|tor-' ); do
cargo set-version --bump minor -p $crate;
done
```
|
| |\
| |
| |
| |
| | |
clippy: deny `mod_module_files`
See merge request tpo/core/arti!2689
|
| | |
| |
| |
| |
| |
| | |
Denies 'mod.rs' files for consistency.
https://rust-lang.github.io/rust-clippy/master/index.html#mod_module_files
|
| |/ |
|
| |\
| |
| |
| |
| | |
Simple dependency upgrades for upcoming release
See merge request tpo/core/arti!2685
|
| | | |
|
| | | |
|
| |/ |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
