Matt Corallo [Fri, 2 Oct 2020 01:32:39 +0000 (21:32 -0400)]
[bindings] Drop one static-lifetime restriction and check success
In general we should stop enforcing that all lifetimes are static
- we may take references from C and its up to reviewing the diff on
the bindings changes and the user(s) to ensure lifetimes are valid.
Also asserts a success criteria that was missed before.
Matt Corallo [Fri, 2 Oct 2020 01:28:57 +0000 (21:28 -0400)]
[bindings] Handle type X = Y aliasing in type resolution
For non-generic type aliases which are meant as convinient aliases
for more complex types, we need to store the aliased type (with all
paths made absolute) and use that in type resolution.
The most code by far is just making all the paths in a type absolute
but its not too bad either.
Matt Corallo [Thu, 1 Oct 2020 22:42:33 +0000 (18:42 -0400)]
[bindings] Avoid guessing whether resolved type is a ref in blocks
In some cases, things which are a Rust Reference (ie slices), we
may still want to map them as a non-reference and need to put a
"mut " in front of the variable name in a function decl. This
worked fine by just checking for the slice case, except that we
are about to add support for type aliases, which no longer match
the naive case.
Instead, we can just have the types module print out the C type and
check if it begins with a '&' to figure out if it is a reference.
Matt Corallo [Fri, 25 Sep 2020 17:25:35 +0000 (13:25 -0400)]
[bindings] Support mapping slices which contain tuples (with refs)
New work upstream puts tuples in slices, which is a very reasonable
thing to expect, however we don't know how to generate conversions
for such objects. Making it more complicated, upstream changes also
include references to things inside such slices, which requires
special handling to avoid creating dangling references.
This adds support for converting such objects, noting that slices
need to be converted first into Vecs which own their underlying
objects and then need to map any reference types into references.
Matt Corallo [Thu, 24 Sep 2020 23:03:05 +0000 (19:03 -0400)]
[bindings] Give Transaction objects a buffer-is-owned flag.
A lot of our container mapping depends on the `is_owned` flag
which we have for in-crate mapped objects to map references and
non-references into the same container type. Transaction was
mapped to two completely different types (a slice and a Vec type),
which led to a number of edge cases in the bindings generation.
Specifically, I spent a few days trying to map
`[(A, &Transaction)]` properly and came up empty - we map slices
into the same types as Vecs (and rely on the `is_owned` flag to
avoid double-free) and the lack of one for `Transaction` would have
required a special-case in numerous functions.
Instead, we just add a flag in `Transaction` to mirror what we do
for in-crate types and check it before free-ing any underlying
memory.
Note that, sadly, because the c_types objects aren't mapped as a
part of our C++ bindings generation, you have to manually call
`Transaction_free()` even in C++.
Matt Corallo [Fri, 25 Sep 2020 16:12:03 +0000 (12:12 -0400)]
[bindings] Add support for `Option<T>` where T is a mapped trait
When mapping an `Option<T>` where T is a mapped trait, we need to
move out of the `*mut T`, however the current generation results in
a `*const T` and a conversion that just takes a reference to the
pointed-to object. This is because the only place this code was
previously used was for slices, which *do* need a reference.
Additionally, we need to know how to convert `Option` containers
which do not contain an opaque type.
Sadly, the easiest way to get the desired result is to add another
special case in container mapping, keeping the current behavior for
slices, but moving out of the pointed-to object for other types.
Matt Corallo [Fri, 25 Sep 2020 18:13:38 +0000 (14:13 -0400)]
Ignore cbindgen version in latest-bindings-in-git check.
Currently the check_binidngs GitHub CI job always fails when there
is a new cbindgen release because the cbindgen version is in the
generated header file. When the new version doesn't change the
generated header except for the version number, we should ignore
the difference.
Matt Corallo [Wed, 23 Sep 2020 01:46:26 +0000 (21:46 -0400)]
[bindings] Use == null() instead of is_null() to avoid ambiguity
When we have a `Trait` wrapped as an `Option<Trait>`, we called
`<*const Trait>.is_null()` which resulted in rustc trying to take
the most braindead option of dereferencing the whole thing and
hitting a recursive dereference since we
`impl Deref<Target=Trait> for Trait` for all our traits.
Instead, we can be explicit and just compare the pointer directly
with `std::ptr::null()` which avoids this.
Matt Corallo [Wed, 16 Sep 2020 21:18:18 +0000 (17:18 -0400)]
[bindings] Include a GenericTypes context in more places
A few places got a None in the previous commit to avoid increasing
the diff size. However, it makes sense to have GenericTypes contexts
there, so we pipe them through the neccessary places.
Matt Corallo [Tue, 15 Sep 2020 22:49:59 +0000 (18:49 -0400)]
[bindings] Push generic resolution into maybe_resolve_path
This pushes down the logic to check if a given Path is, in fact,
a reference to a generic into the common maybe_resolve_path instead
of doing it redundantly in a few different places. Net loss of LoC.
Matt Corallo [Mon, 5 Oct 2020 16:47:08 +0000 (12:47 -0400)]
Fix passing -O1 to build from `cargo test`
In 9e03087d6acbc876a5ad1c9b9d8746bf18d5ca86 we started setting
`opt-level` only on profile.test and not profile.dev. When that
commit was authored I tested only that rustc was being called with
opt-level set in its flags, not that the resulted run ran at the
speed I expected. It seems profile.test isn't applied properly to
dependencies or so, resulting in tests running much slower than
they do at profile.dev.opt-level=1.
Matt Corallo [Sun, 27 Sep 2020 21:52:09 +0000 (17:52 -0400)]
Drop the redundant/broken `ChannelMonitor::get_monitored_outpoints`
In review of the final doc changes in #649, I noticed there
appeared to be redundant monitored-outpoints function in
`ChannelMonitor` - `get_monitored_outpoints()` and
`get_outputs_to_watch()`.
In 6f08779b0439e7e4367a75f4ee88de093dfb68cb,
get_monitored_outpoints() was added, with its behavior largely the
same as today's - only returning the set of remote commitment txn
outputs that we've learned about on-chain. This is clearly not
sufficient, and in 73dce207dd0ea6c3ac57af3ebb8b87ee03e82c9e,
`get_outputs_to_watch` was added which was overly cautious to
ensure nothing was missed. Still, the author of 73dce207dd0ea6c3ac5
(me) seemed entirely unaware of the work in 6f08779b0439e7e4367a75f
(also me), despite the function being the literal next function in
the same file. This is presumably because it was assumed that
`get_monitored_outpoints` referred to oupoints for which we should
monitor for spends of (which is true), while `get_outputs_to_watch`
referred to outpouts which we should monitor for the transaction
containing said output (which is not true), or something of that
nature. Specifically, it is the expected behavior that the only
time we care about `Filter::register_tx` is for the funding
transaction (which we aren't aware of the inputs of), but for all
other transactions we register interest on the basis of an outpoint
in the previous transaction (ie via `Filter::register_output`).
Here we drop the broken-on-day-one `get_monitored_outpoints()`
version, but assert in testing that the values which it would return
are all present in `get_outputs_to_watch()`.
Matt Corallo [Fri, 18 Sep 2020 22:26:12 +0000 (18:26 -0400)]
Fix feerate calculation on closing transactions
This resolves a number of bugs around how we calculate feerates on
closing transactions:
* We previously calculated the weight wrong both by always
counting two outputs instead of counting the number of outputs
that actually exist in the closing transaction and by not
counting the witness redeemscript.
* We use assertions to check the calculated weight matches what we
actually build (with debug_assertions for variable-length sigs).
* As an additional sanity check, we really should check that the
transaction had at least min-relay-fee when we were the channel
initator.
Matt Corallo [Fri, 2 Oct 2020 16:51:25 +0000 (12:51 -0400)]
Actually hold the total_consistency_lock instead of take-and-drop
It was noticed (via clippy) by @casey that we were taking and then
immediately dropping the total_consistency_lock because `let _ =`
doesn't actually bind the response to anything. This appears to be
a consequence of wanting `if let Some(_) =` to not hold a ref to
the contained value at all, but is relatively surprising to me.
Matt Corallo [Wed, 30 Sep 2020 19:18:29 +0000 (15:18 -0400)]
Drop last bits of rescan as its too complicated to be worth having
Previously, we had a concept of "rescaning" blocks when we detected
a need to monitor for a new set of outputs in future blocks while
connecting a block. In such cases, we'd need to possibly learn about
these new spends later in the *same block*, requiring clients who
filter blocks to get a newly-filtered copy of the same block. While
redoing the chain access API, it became increasingly clear this was
an overly complicated API feature, and it seems likely most clients
will not use it anyway.
Further, any client who *does* filter blocks can simply update their
filtering algorithm to include any descendants of matched
transactions in the filter results, avoiding the need for rescan
support entirely.
Thus, it was decided that we'd move forward without rescan support
in #649, however to avoid significant further changes in the
already-large 649, we decided to fully remove support in a
follow-up.
Here, we remove the API features that existed for rescan and fix
the few tests to not rely on it.
After this commit, we now only ever have one possible version of
block connection transactions, making it possible to be
significantly more confident in our test coverage actually
capturing all realistic scenarios.
Jeffrey Czyz [Fri, 7 Aug 2020 17:58:15 +0000 (10:58 -0700)]
Move channelmonitor.rs from ln to chain module
Given the chain::Watch interface is defined in terms of ChannelMonitor
and ChannelMonitorUpdateErr, move channelmonitor.rs from the ln module
to the chain module.
Transaction data from a block may be filtered before it is passed to
block_connected functions, which may need the index of each transaction
within the block. Rather than define each function in terms of a slice
of tuples, define a type alias for the slice where it can be documented.
Jeffrey Czyz [Fri, 7 Aug 2020 22:13:57 +0000 (15:13 -0700)]
Replace WatchEvent usage with get_outputs_to_watch
Outputs to watch are tracked by ChannelMonitor as of 73dce207dd0ea6c3ac57af3ebb8b87ee03e82c9e. Instead of determining new
outputs to watch independently using ChainWatchedUtil, do so by
comparing against outputs already tracked. Thus, ChainWatchedUtil and
WatchEvent are no longer needed.
Jeffrey Czyz [Sat, 29 Aug 2020 23:04:46 +0000 (16:04 -0700)]
Include funding TXO in outputs to watch
The funding TXO was never added to ChannelMonitor's outputs_to_watch in 73dce207dd0ea6c3ac57af3ebb8b87ee03e82c9e. Include it when constructing a
ChannelMonitor.
WatchEventProvider served as a means for replacing ChainWatchInterface.
However, it requires users to explicitly fetch WatchEvents, even if not
interested in them. Replace WatchEventProvider by chain::Filter, which
is an optional member of ChainMonitor. If set, interesting transactions
and output spends are registered such that blocks containing them can be
retrieved from a chain source in an efficient manner.
This is useful when the chain source is not a full node. For Electrum,
it allows for pre-filtered blocks. For BIP157/158, it serves as a means
to match against compact filters.
BlockNotifier was removed in the previous commit, thus ChainListener is
no longer needed. Instead, anything needing chain events should be
notified directly.
BlockNotifier is a convenience for handing blocks to listeners. However,
it requires that each listener conforms to the ChainListener interface.
Additionally, there are only two listeners, ChannelManager and
ChainMonitor, the latter of which may not be used when monitoring
channels remotely. Remove BlockNotifier since it doesn't provide much
value and constrains each listener to a specific interface.
Jeffrey Czyz [Thu, 6 Aug 2020 23:54:13 +0000 (16:54 -0700)]
Remove Key parameter from ChainMonitor
ChainMonitor's template Key parameter was meant to allow supporting
both local monitoring, where Key=OutPoint, and watchtowers, where Key=
(PublicKey, u32). Use OutPoint directly since the watchtower case will
not be supported this way.
ManyChannelMonitor was renamed chain::Watch in the previous commit. Use
a more concise name for an implementation that monitors the chain for
channel activity. Future work will parameterize the struct to allow for
different varieties of persistence. Thus, users usually will be able to
use ChainMonitor directly rather than implementing a chain::Watch that
wraps it.
Rename ManyChannelMonitor to chain::Watch and move to chain/mod.rs,
where chain-related interfaces live. Update the documentation for
clarity and to conform to rustdoc formatting.
ChainWatchInterface was intended as an interface for watching rather
than accessing the chain. Remove get_chain_utxo and add chain::Access
trait for this behavior. Wrap it with an Option in NetGraphMsgHandler in
order to simplify the error interface.
Use of ChainWatchInterface was replaced with WatchEvent in the previous
commit. Remove it from the parameterization of SimpleManyChannelMonitor
since it is no longer needed.
Replace use of ChainWatchInterface with WatchEvent
SimpleManyChannelMonitor is parameterized by ChainWatchInterface to
signal what transactions and outputs to watch for on chain. The
interface has grown to cover chain access (via get_chain_utxo) and block
block filtering (via filter_block and reentered), which has added
complexity for implementations and user (see ChainWatchInterfaceUtil).
Pull the watch functionality out as a first step to eliminating
ChainWatchInterface entirely.
Replace direct uses of BlockNotifier in functional tests with utility
functions. This is in preparation for signaling watch events back via a
refactoring of ManyChannelMonitor and ChainWatchInterface. Those events
will be processed by connect_block.
Change confirm_transaction and connect_blocks to take a Node instead of
a BlockNotifier. This is in preparation for signaling watch events back
via a refactoring of ManyChannelMonitor and ChainWatchInterface.
Jeffrey Czyz [Fri, 26 Jun 2020 17:43:24 +0000 (10:43 -0700)]
Align ChannelMonitor interface with ChainListener
ChannelMonitor has block_connected and block_disconnected methods called
by <SimpleManyChannelMonitor as ChainListener>. Use similar parameters
in ChannelMonitor such that transformations are not needed and the
interface is more closely aligned with ChainListener.
Jeffrey Czyz [Tue, 16 Jun 2020 22:10:17 +0000 (15:10 -0700)]
Remove ChainWatchInterface from BlockNotifier
ChainListeners should be independent of each other, but in practice this
is not the case because ChainWatchInterface introduces a dependency
between them. Push ChainWatchInterface down into the ChainListener
implementations where needed. Update ChainListener's block_connected
method to take a slice of the form &[(usize, &Transaction)] where each
transaction is paired with its position within the block.
Matt Corallo [Fri, 18 Sep 2020 18:46:58 +0000 (14:46 -0400)]
Do not test any block-contents-rescan cases
In anticipation for removing support for users calling
block_connected multiple times for the same block to include all
relevant transactions in the next PR, this commit stops testing
such cases. Specifically, users who filter blocks for relevant
transactions before calling block_connected will need to filter by
including any transactions which spend a previously-matched
transaction in the same block (and we now do so in our own
filtering logic, which is also used in our testing).
Matt Corallo [Fri, 18 Sep 2020 22:01:10 +0000 (18:01 -0400)]
derive(Debug) on Events
We'd previously largely not turned on derive(Debug) on any of our
structs, but not for good reason. Especially for Events objects,
Debug can be a very useful for users to quickly print what they
received from us without having to write out a large match.
Matt Corallo [Mon, 14 Sep 2020 21:39:42 +0000 (17:39 -0400)]
Fix two bugs which access the wrong peer's htlc_minimum_msat value
* Channel::get_counterparty_htlc_minimum_msat() returned
holder_htlc_minimum_msat, which was obviously incorrect.
* ChannelManager::get_channel_update set htlc_minimum_msat to
Channel::get_holder_htlc_minimum_msat(), but the spec explicitly
states we "MUST set htlc_minimum_msat to the minimum HTLC value
(in millisatoshi) that the channel peer will accept." This makes
sense because the reason we're rejecting the HTLC is because our
counterparty's HTLC minimum value is too small for us to send to
them, our own HTLC minimum value plays no role. Further, our
router already expects this - looking at the same directional
channel info as it does fees.
Finally, we add a test in the existing onion router test cases
which fails if either of the above is incorrect (the second issue
discovered in the process of writing the test).
Matt Corallo [Mon, 14 Sep 2020 20:49:05 +0000 (16:49 -0400)]
Move onion failure tests from functional_tests to their own file
They all have a specific structure, so having them in the mess that
is functional_tests isn't really conducive to readability. More
importantly, functional_tests is so big it slows down compilation,
so even dropping a few hundred lines is a win.
Sources of the failure may be multiple in case of distributed watchtower
deployment. In either case, the channel manager must return a final
update asking to its channel monitor(s) to broadcast the lastest state
available. Revocation secret must not be released for the faultive
channel.
In the future, we may return wider type of failures to take more
fine-grained processing decision (e.g if local disk failure and
redudant remote channel copy available channel may still be processed
forward).
Antoine Riard [Thu, 27 Aug 2020 23:48:35 +0000 (19:48 -0400)]
Add test_concurrent_monitor_claim
Watchower Alice receives block 134, broadcasts state X, rejects state Y.
Watchtower Bob accepts state Y, receives blocks 135, broadcasts state Y.
State Y confirms onchain. Alice must be able to claim outputs.
Antoine Riard [Fri, 28 Aug 2020 00:47:02 +0000 (20:47 -0400)]
Implement concurrent broadcast tolerance for distributed watchtowers
With a distrbuted watchtowers deployment, where each monitor is plugged
to its own chain view, there is no guarantee that block are going to be
seen in same order. Watchtower may diverge in their acceptance of a
submitted `commitment_signed` update due to a block timing-out a HTLC
and provoking a subset but yet not seen by the other watchtower subset.
Any update reject by one of the watchtower must block offchain coordinator
to move channel state forward and release revocation secret for previous
state.
In this case, we want any watchtower from the rejection subset to still
be able to claim outputs if the concurrent state, has accepted by the
other subset, is confirming. This improve overall watchtower system
fault-tolerance.
This change stores local commitment transaction unconditionally and fail
the update if there is knowledge of an already signed commitment
transaction (ChannelMonitor.local_tx_signed=true).