The motivation of this change is to move away from tying signing to be
specifically for ed25519. The reason being that the protocol will want move
towards two different kinds of signing – the node signing artifacts, and an
author signing artifacts.
This change captures the former, node signing, by introducing a `Device` type
that is the `NodeId` – a PublicKey underneath the hood – and a signing
mechanism. This allows the replacement of the `Signer` trait being used – which
always assumed a `PublicKey`. Instead, the `Device` is constructed with
`NodeId`.
In `radicle-cob`, a signer is expected to implement
`signature::Signer<ExtendedSignature>`, and everywhere in `radicle`,
`radicle-node`, `radicle-cli`, and `radicle-remote-helper` is expected the
signer is expected to implement `signature::Signer<Signature>`.
A `Device` implements both of these but only requires
`signature::Signer<Signature>` to do so – since an `ExtendedSignature` is
essentially `(PublicKey, Signature)`, and the `NodeId` of the `Device` can be
used.
The aim of this change is to make the `Doc` type more safe to use by approaching
the design via [Parse don't validate][[0]] approach.
The problem with the previous approach was that all field were `pub` and thus a
`Doc<Verified>` could easily be mutated and serialized. Granted, the code that
used the serialization would tend to verify the `Doc` first, however, this
approach *ensures* that only a verified `Doc` can be serialized. It also meant
that trying to add new data that would follow the parse approach would require
more generic parameters on top of the existing `PhantomData` parameter, i.e. we
need to do something like: `Doc<RawField, V> -> Doc<ValidField, V>`.
The new approach splits the type into two separate types: `RawDoc` and `Doc`.
The former is allowed to be mutated at will, and uses types that are less
strict. The latter is the valid type that can only be constructed by validating
a `RawDoc` (or the `initial` constructor). The `Doc` type's fields can then only
be accessed by read-only methods.
Solves the problems above by only allowing mutations to `RawDoc`, as well as,
new fields being added to `RawDoc` which are then validated via
`RawDoc::verified`.
[0]: https://lexi-lambda.github.io/blog/2019/11/05/parse-don-t-validate/
We ensure that a `rad/root` ref is included in the signed refs file
under `rad/sigrefs` for all remotes. This prevents a certain kind of
"grafting" attack where signed refs can be copied between repositories,
by having the peer sign over the identity root together with the data refs.
When verifying signed refs, we ensure that the ref is present and points
to an identity branch root that matches the repository identity containing
the signed refs.
Alternatives: lots of alternatives were considered, but this one doesn't
introduce any changes to the signing. The `rad/id/root` name was
considered but is invalid due to `rad/id`.
Though refs of private repositories are not announced publicly, they can
sometimes be relayed to nodes that are not in the allow list.
We fix this by always checking the visibility of a repository before
sending a refs announcement of it to a peer.
Discovered-by: Adrian Duke <adrian.duke@gmail.com>
In the current design it's possible for one peer to fill the fetch queue
so that fetches from other peers are delayed.
To improve fairness, we move to a queue per peer. We then try to dequeue
from all peers in a random order for good measure.
Previously, our advertized inventory was not always correct: when we
stopped seeding a repository, it was still advertised as part of our
inventory until node restart, because the `Storage` type doesn't have
access to seeding policies.
In this change, we remove the concept of inventory from `Storage` and
make the authoritative place for it the routing table in our database.
To make this work, we have to add our node to the database on profile
creation, to not violate the foreign key constraint on the routing
table. Hence, the tests are changed to include our alias which is now
always available.
We use the *staggered broadcast* technique when relaying gossip messages
to reduce message amplification. In our basic simulation, it brings
amplification down from 3.5 to 1.5.
The idea is simply to accumulate gossip messages and relay them after a fixed
interval of time. This has two effects:
1) Old messages that haven't been sent yet are replaced by newer ones,
so only one message is sent after the delay.
2) Because nodes have their own intervals, messages are not all sent at
the same time, which reduces the chance of messages crossing paths.
For now we only turn this on for inventory messages, since this is the
least likely to be noticed by users, and also the most problematic
currently.
If our inventory hasn't updated, don't issue a new inventory
announcement, simply use the previous message.
This avoids identical inventories propagating over the network and using
up resources.
When deciding whether or not to fetch a repo based on a refs
announcement, we need to know whether our signed refs are stale or not,
compared to the ones in the announcement. This can be quite expensive as
it requires multiple reads from Git's ODB, which often results in
packfile loading.
To remedy this, we introduce a refs cache in our node database that is
consulted when an announcement is received.
We also remove the logic that generates `info` messages, since they are
rarely needed and introduced more complexity.
Additionally, we revert commit 3ad2b4431f,
which introduced logic to minimze wants and haves with additional logic,
since it potentially also increases the load on the ODB.
Instead of dropping the existing connection(s), we drop the same
connection on both sides, by using a simple rule that yields the same
result whether the node sees the connection as inbound or outbound.
The trick is to use something that both nodes agree on, for instance
whos public key is greater than the other, and use that information
to close the same connection on both sides.
Remove the borrow of `rid` to fix the clippy warning.
Signed-off-by: Fintan Halpenny <fintan.halpenny@gmail.com>
X-Clacks-Overhead: GNU Terry Pratchett
Instead of re-loading the full inventory on `rad init`, we simply add
the new repository to the in-memory cache.
We also keep track of whether a fetch was a full clone, to know when
to update our routing table.
Implement caching for the `Issue` and `Patch` COB. This is achieved by
having the `Transaction::initial` and `Transaction::commit` methods
take a cache which can update the newly created/updated object. The
`Store::remove` method also takes the cache for removing the object
from the cache, as well as the repository.
This meant that the `*Mut` types for the respective COBs are required
to carry a cache alongside the repository store. Identities will not
be cached, and so this is always set to `NoCache` -- which performs no
caching, and always succeeds.
To perform cache reading for issues and patches, the `cache::Issues`
and `cache::Patches` traits are introduced. They capture the minimal
amount of methods that are needed for the `rad issue` and `rad patch`
CLI commands.
The implementor for each of these traits is there respective `Cache`
types, which combines their backing repository store and the SQLite
database. They both provide convenience methods:
- `create` (and `draft` for patches)
- `get_mut`
- `remove`
which can be used instead of their repository store counterparts.
All uses of the repository stores (where needed) are replaced with
`Cache` types.
Signed-off-by: Fintan Halpenny <fintan.halpenny@gmail.com>
X-Clacks-Overhead: GNU Terry Pratchett
Dequeued fetches were always fetching all refs, instead of fetching the
specific refs that were initially requested. We fix that by storing the
refs to fetch in the fetch queue.
Previously, we might ban a node if we're having trouble connecting to
it, but this could be due to internet connectivity issues.
We introduce a penalty score for nodes, that increases if there are
connection issues or the node misbehaves, and decreases on successful
connections.
When connected to new nodes, we sort them so that low-penalty nodes are
attempted first.
Instead of having a routing database, and an addresses database,
consolidate that in a single database called `node.db`.
We also simplify database access, by providing a single type for opening
the database, and use traits to access the various functions.
We create a new `seed::Store` that has seed-related methods split out of
the `address::Store`.
Note that tracking is still in its own database currently, as it isn't
ephemeral state. This makes it easy to delete the new `node.db` without
risk of losing any important information.
To avoid renaming all database methods, we create a `Stores` wrapper
that allows us to access routing, addresses, seeds etc. without type
ambiguity.
This is a breaking change and requires `addresses.db` and `routing.db`
to be deleted.
Currently, if refs are updated while the node is stopped, these refs are
not immediately announced when the node starts, due to the node being
unaware of the updates.
To remedy this, on startup, we compare the refs we have in storage with
the refs we announced to the network. Since only our own refs could have
changed while the node was stopped, this change only applies to owned
refs. If we detect that storage has changed while the node was stopped,
we create ref announcements for the updated repos and store them in our
gossip database. These announcements will be included with other
historical gossip announcements when nodes connect to us.
To keep track of what was announced, we store our local node's sync
status in the database. This is also useful when using the `rad sync status`
command, as it tells us what oid our local node is at, and whether it
is aware of our local changes yet.
In addition to this, we make a small fix: previously, inventory and ref
announcements originating from our local node were not stored in the
announcement database. We ensure that they are via the new `announce`
method.
An informational message for the protocol is useful for sending
messages to a node that is not required to be relayed.
This message is introduced by starting with an ACK message for a set
of references being synced. This fixes an issue for announcing to a
node that is already synced, since it can now ACK that it received the
message and replies that it is synced.
Signed-off-by: Fintan Halpenny <fintan.halpenny@gmail.com>
X-Clacks-Overhead: GNU Terry Pratchett
It can be useful to return the `RefsAt` that was announced for this
node, for example, to compare it to node events.
Change the `Handle` method to return `RefsAt` so that when it is
called it returns the local node's `RefsAt`.
Signed-off-by: Fintan Halpenny <fintan.halpenny@gmail.com>
X-Clacks-Overhead: GNU Terry Pratchett
The previous system allowed the same RID to be fetched from multiple
nodes concurrently, which caused issues, especially when cloning.
Worker code assumes that there is only one fetch per repository going on
at any given time. This change ensures that.
We keep a global fetch queue, instead of having a queue per session. We
also enqueue fetches if a session is at capacity. Finally, we allow
multiple fetch requests to receive results for the same fetch.
Previously, we would process any valid inventory or refs announcement.
This could cause problems for keeping sync statuses, as we could get
a refs announcement from a node that isn't in our database.
In practice, there is no good reason why we should receive a refs or
inventory announcement before the announcer's node announcement, unless
nodes are not following the protocol.
Implementing this change meant that many of the tests needed updating,
as it wasn't assumed that a node announcement was needed before
receiving inventories etc. from a peer.
We also had to implement a basic form of node address banning to make
the tests pass after this change.
The RefsAnnouncement previous announced its complete set of
`rad/sigrefs`. This was useful in that verification can occur on the
receiveing side immediately by verifying the refs and proceeding with
the fetch -- note that it did nothing with the ref information it
received.
This can be minimised to only advertising the namespace, i.e. the
NID, and the new SHA of the `rad/sigrefs`. This is represented by a
new message type, `RefsAt`, which contains the NID and the Oid.
For checking if an announcement is stale or fresh, the receiving
side can check the existence of the SHA in its Git repository.
The tests are appropriately modified to use NIDs and Oids.
To prepare for a follow-up change, the announcement data is also
threaded through for the fetch protocol to focus fetching only the
advertised set of Oids.
N.B.: because the refs aren't communicated, the verification happens
when fetching since the data needs to be fetched before
verifying. This is a trade-off with minimising the message payload for
verifying earlier.
Signed-off-by: Fintan Halpenny <fintan.halpenny@gmail.com>
X-Clacks-Overhead: GNU Terry Pratchett
The fetch V2 improvements introduced a `UserInfo` type that is used to
set the Git config's name and email when cloning a repository. This
`UserInfo` can also be used when creating a new `Repository`.
The `Storage` type now takes a `UserInfo` so that it can be passed to
`Repository::create`, for creating a new repository.
For testing, this requires adding a new `fixtures::user` so that
testing `Storage` instances can be easily constructed.
It also means that the `radicle-cli` tests all changed with new SHAs
since the name and email are used as part of the hashing input.
N.b. the order of the timelines while testing the identity cob
changed. It's not clear why that is.
Signed-off-by: Fintan Halpenny <fintan.halpenny@gmail.com>
X-Clacks-Overhead: GNU Terry Pratchett
Wire up the `radicle-fetch` crate to the fetch logic in the
`radicle-node` worker.
Any required traits for storage, identity, tracking, and transport
used by `radicle-fetch`.
The worker adds a new module to run the `git upload-pack` process and
pipes the `stdout`/`stdin` to the respective channels.
The `Worker` type required adding a `FetchConfig` to easily allow the
configuration of fetches and passing the signer and tracking store
location through to the fetch `Handle`.
The channels code is adapted to have a new writer that always flushes
when a call to `write_all` is made. This is necessary since the
Gitoxide code never calls `flush`, and so it will hang since no data
is sent until `flush` is called.
Since `UserInfo` is used to set up a newly cloned repository, this
changed some of the SHA1 hashes output in the CLI test examples. It
was also necessary to update the CLI's clone code to be deterministic
in its output for tests.
Signed-off-by: Fintan Halpenny <fintan.halpenny@gmail.com>
X-Clacks-Overhead: GNU Terry Pratchett
Make the validation logic customisable by extracting it out into a
separate trait that can be defined by other crates.
The Repository validation logic stays the same.
Signed-off-by: Fintan Halpenny <fintan.halpenny@gmail.com>
X-Clacks-Overhead: GNU Terry Pratchett
Introduces a new COB type to store repository identity documents.
The reason for this change is to:
1. Simplify the code, as the identity document logic resembled a COB,
yet it had custom logic. This allows existing COB code to be used for
identities.
2. Make identity document update logic more flexible, since COB actions
can be added in the future.
3. Re-purpose existing tools around COBs to work on identities, eg. `rad
cob`.
4. Unify the concept of an identity change proposal, with regular identity
changes. This means we can remove the `id.proposal` COB in favor of
using the `id` COB itself.
Notes
-----
* Each repository has one Identity COB.
* The `Proposal` COB has been repurposed into the `Identity` COB.
* Identity documents are stored as *embeds* inside the Identity COB
actions.
* The action that contains new document versions is called `revision`,
just like the Patches COB.
* The namespaced `rad/id` ref is a symbolic reference to that Identity
COB.
* The canonical `rad/id` ref is a direct reference to the commit in the
Identity COB that contains the latest *accepted* revision of the
document.
* All commands for managing identities have been folded into `rad id`.
Hence `rad delegate` and `rad edit` are removed.
* The concept of "rebasing" an identity document is gone.
* The `rad id` output has been updated to match the style of other
commands.
* When a revision has enough signatures, it is automatically adopted as
the current identity, there is no longer the need for a "commit"
action.
* When an identity revision is proposed, and the current identity has a
threshold of `1`, that identity is automatically accepted due to the
above point.
* The idea of "verifying a peer's identity branch" no longer applies, as
COBs cannot be verified one history at a time.
* Since the root commit of the Identity COB does not have a "Resource"
to point to (it would have to point to itself, which is impossible),
we've made the resource id optional for COBs.
When a fetch is user-requested, a timeout can be supplied that is passed
down to the worker.
When a fetch is service-requested, a default timeout is used.
Private repos are implemented by extending the identity document with a
`visibility` attribute, that can either be `"public"` (default) or
`"private"`.
In case of `private` visibility, only the delegates are allowed to view
the repo, as well as any DIDs added to the allow list.
To implement repo visibility, we simply block fetches from and
announcements to peers for whom the repo should remain invisible.
Private repos are also not announced in the `inventory` message, since
the full list of peers that *may* have the repo is retrievable from the
repo identity. This could cause errors if eg. a peer who is allowed to
view the repo doesn't actually have it. However this is an ok trade-off
for now to keep the complexity low. For repos to truly be private, it's
important that the RIDs don't leak either.
Finally, we modify `radicle-httpd` for now to only list public repos.
Eventually, we would want to change this depending on whether an allowed
peer is authenticated with the service or not.
---
It's also worth mentioning why this approach was taken, vs. end-to-end
encryption. The reasons are as follows:
1. Nodes that do not have access to a private repo will generally not want to
replicate encrypted data that they cannot examine or use.
2. The chosen solution is trivial, while encrypting git objects isn't.
3. Performance of the chosen solution is much better, there is no
overhead.
4. Privacy of the chosen solution is better: RIDs are never leaked, and
neither is the existence of a private repo, nor who has access to it.
There is one downside: Paying for storage of private repos is no better
in terms of privacy than what GitHub offers. Hosting providers will have
access to your private repos, if this solution is used.
Only return once the connection is established or failed.
It's still not perfect, as several failure scenarios are not handled,
but it's an improvement.
Lorenz Leutgeb
Fintan Halpenny
cloudhead
Alexis Sellier