We keep the node_id, not a pointer to the peer.
This also means that it might have reconnected while we were in the hook, so make
sure we ignore the result if it's in state PEER_CONNECTED.
And remove the `tal_steal(peer, hook_payload)` which doesn't do anything: the
plugin_hook call steals hook_payload anyway!
Fixes: #5944
It's not likely but possible that the node's settings will shift btw a
start and an RBF; we persist the setting to the database so we don't
lose it.
Right now holding onto it forever is kind of extra but maybe we'll
reuse the setting for splices? idk.
Should this be a channel type??
`openchannel_init` takes a psbt, which we pipe over to dualopend
process.
If the peer requests that they'll only accept confirmed inputs, we need
to go validate those before we continue.
This wires up the harness for this (validation check yet tc)
technically we don't need this info after the channel opens, but for any
subsequent RBF (and maybe splice?) we need to remember what the
open/accept peer signaled
not amazing, since we'll probably call openchannel_update multiple
times per open, but this is the simplest way to confirm that we're
not sending unconfirmed outputs to peer.
This will save a lot of RPC ping/pong when plugins still need to iterate
both, `listpeers` and `listpeerchannels`. When `num_channels` is 0 they
can skip additional calls.
Changelog-Added: RPC `listpeers` output now has `num_channels`.
Though there's already a `createinvoice` command, there are usecases where a
user may want to sign an invoice that they don't yet have the preimage to. For
example, they may have an htlc_accepted plugin that pays to obtain the preimage
from someone else and returns a `{ "result": "resolve", ... }`.
This RPC command addresses this usecase without overly complicating the
semantics of the existing `createinvoice` command.
Changelog-Added: JSON-RPC: `signinvoice` new command to sign BOLT11
invoices.
This allows to accept safely long paths as options
and does not truncate them
as https://github.com/ElementsProject/lightning/issues/5576
described
Changelog-Fixed: cli: accepts long paths as options
Suggested-by: @rustyrussell <rusty@rustcorp.com.au>
Signed-off-by: Vincenzo Palazzo <vincenzopalazzodev@gmail.com>
In v0.11 (71f736678f) we changed lightningd to wait for gossipd to
acknowledge blocks before updating blockheight: this resolved a problem
which lnprototest had where it wanted to know when we'd fully digested
a block.
However, it broke the syncing case: until then we don't even tell
gossipd, so this stayed at zero. We should use the current blockheight
for that corner case!
Fixes: #5894
Changelog-Fixed: JSON-RPC: `getinfo` `blockheight` no longer sits on 0 while we sync with bitcoind the first time.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
When a channel open fails, we use tx-abort instead of warning/error.
This means that the peer won't disconnect! And instead when a new
message arrives, we'll need to rebuild the dualopend subd (if missing).
Makes opens a bit easer to retry (no reconnect needed), as well as keeps
the connection alive for other channels we may have with that peer.
Changelog-Changed: Experimental-Dual-Fund: open failures don't disconnect, but instead fail the opening process
```
----------------------------- Captured stderr call -----------------------------
Sending onchaind an invalid message 03ed00000000000000004e52a9129a66619d6809b1024eb9a0159f173a988f3a5d0bdd2447b4fcc24cef
lightningd: FATAL SIGNAL 6 (version 3c57147-modded)
```
The channel state can also be `FUNDING_SPEND_SEEN` if onchaind is still
starting up.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
In various circumstances we can start a reconnection while one is
already going on. These can stockpile if the node really is unreachable.
Reported-by: @whitslack
Fixes: #5654
Changelog-Fixed: lightningd: we no longer stack multiple reconnection attempts if connections fail.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The leak-detector can't find unconnected_htlcs_in on the stack and
incorrectly flags this as a leak. However, it is appropriately tal
allocated and freed.
Changelog-None
There were cases where address it's empty, and this cases are not right if the
field is considered optional.
This makes it required and add the field also when `--offline` is set.
Changelog-Changed: JSON-RPC: `getinfo` `address` array is always present (though may be empty)
We need to check if the key parameter is an empty array in
`listdatastore` as we do assume an array of at least length 1 in
`wallet.c:5306`.
Signed-off-by: Peter Neuroth <pet.v.ne@gmail.com>
We didn't actually populate them properly, and the real annotations
are on inputs and outputs.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-EXPERIMENTAL: JSON-RPC: `listtransactions` `channel` and `type` field removed at top level.
```
make check-source-bolt CHECK_BOLT_PREFIX="--prefix=BOLT-onion-message" BOLTVERSION=guilt/offers
```
Mainly textual, though I neatened the extra fields check for TLVs with
blinding, and implemented the "no other fields" requirement for
non-final onion message hops.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
```
make check-source-bolt CHECK_BOLT_PREFIX="--prefix=BOLT-route-blinding" BOLTVERSION=guilt/offers
```
Other than textual changes, this does:
1. Ensures we put total_amount_msat in onion final hop (reported by @t-bast).
2. Require that they put total_amount_msat in onion final hop.
3. Return `invalid_onion_blinding` exactly as defined by the spec (i.e. less
aggressive when we're the final hop) (also reported by @t-bast, but I already
knew).
See: #5823
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-EXPERIMENTAL: `offers` breaking blinded payments change (total_amount_sat required, Eclair compat)
You can use rs->nextcase, but we don't always keep that around, so
keep a flag in onion_payload.
We'll use this in the "do we need to return a blinded error code"
patch.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Also, we don't need to pass the total length to the field parsers,
just the length for this field (confusingly, this was called
"data_length").
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This will give the user an option to set a custom port when using
discovered IPs for node_announcents. Without this, only the selected
networks default port can used.
Changelog-Added: Adds --announce-addr-discovered-port config option to set custom port for IP discovery.
This switch was not doing anything useful anymore.
We deprecate it anyways to notify the user about the new switch.
Changelog-Deprecated: The old --disable-ip-discovery config switch
This adds the option to explicitly enable ip-discovery, which maybe
helpful for example when a user wants TOR announced along with
discovered IPs to improve connectivity and have TOR just as a fallback.
Changelog-Added: Adds config switch 'announce-addr-discovered': on/off/auto
After connecting 100,000 peers with one channel each (not all at
once!), we see various places where we exhibit O(N^2) behaviour.
Fix these by keeping a map of id->peer instead of a simple
linked-list.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
```
E Global errors:
E - Node /tmp/ltests-adkwu44c/test_logging_1/lightning-2/ has memory leaks: [
E {
E "backtrace": [
E "ccan/ccan/tal/tal.c:442 (tal_alloc_)",
E "lightningd/peer_control.c:2203 (load_channels_from_wallet)",
E "lightningd/lightningd.c:1105 (main)"
E ],
E "label": "lightningd/peer_control.c:2203:struct htlc_in_map",
E "parents": [
E "lightningd/lightningd.c:107:struct lightningd"
E ],
E "value": "0x55d920a345e8"
E }
E ]
```
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: JSON-RPC: new command `listpeerchannels` now contains information on direct channels with our peers.
Signed-off-by: Vincenzo Palazzo <vincenzopalazzodev@gmail.com>
We're soon going to call json_add_unsaved_channel and
json_add_uncommitted_channel from a new place, where we want the peer
state directly included.
Based on patch by @vincenzopalazzo.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This avoids any confusion between primitive and string ids, and in
particular stops an issue with commando once it starts chaining ids,
that weird ids can be double-escaped and commando will not recognize
the response, leaving the client hanging. It's the client's fault for
using a weird id, but it's still rude (and triggered by our tests!).
It also makes substituting the id in passthrough simpler, FTW.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
When a channel is not yet confirmed onchain and only has an alias,
`getroute` will return the alias under the "channel" key. However, if
you supply that to `sendpay`, it will not be able to find the channel
and will fail since it calls `find_channel_for_htlc_add` and that only
looks for channels by their scid and doesn't consider their alias.
This patch makes `find_channel_for_htlc_add` also consider channel
aliases when looking for channels.
Changelog-Fixed: JSON-RPC: `sendpay` now can send to a short-channel-id alias for the first hop.
We were stressing the servers if node cannot be found. Only do lookup
on manual connect commands.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Protocol: lightningd: Only use DNS server address lookup on manual `connect` commands, not normal reconnection attempts.
This broke BTCPayServer, so revert. I originally (accidentally!)
implemented this such that it broadcast both DNS and IP entries, but
Michael reported earlier that they still don't propagage well, so
simply suppress them.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Fixes: #5795
Changelog-Changeed: Config: `announce-addr-dns` needs to be set to *true* to put DNS names into node announcements, otherwise they are suppressed.
Changelog-Deprecated: Config: `announce-addr-dns` (currently defaults to `false`). This will default to `true` once enough of the network has upgraded to understand DNS entries.
This reverts commit 43b037ab0b.
Nicholas Dorier says BTC Payserver still wants this for another year
or so.
Changelog-Added: JSON-RPC: reverts requirement for "jsonrpc" "2.0" inside requests (still deprecated though, just for a while longer!)
The read_json() call expects len_read to be the amount of *new*
data read. If we call this back without resetting, it will parse
this much random junk in the buffer.
Fixes: #5766
Changelog-Fixed: Plugin: `autoclean` could misperform or get killed due to lightningd's invalid handling of JSON batching.
If we both support large channels, we can actually send giant HTLCs.
This, in turn, fixes pay (which relies on this field!).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Fixed: Plugins: `pay` now knows it can use locally-connected wumbo channels for large payments.
Fixes: #5250Fixes: #5417
This is a mistake that I introduced while I implemented the lightningd custom error for hsmd.
In particular, when the command line parser check
if the file is encrypted make an additional check regarding the existence of the file.
This can be a not useful check, but due to the delicate nature of the hsm file, it is better to check if exist and if it doesn't exist an informative line inside the log is emitted that notifies the user that the file does not exist.
This log may return useful while debugging disaster
(that can happen) to understand that there is something strange (eg. the user moves the hsm file somewhere else).
Fixes https://github.com/ElementsProject/lightning/issues/5719
Changelog-Fixed: lightningd: do not abort while parsing hsm pwd
Signed-off-by: Vincenzo Palazzo <vincenzopalazzodev@gmail.com>
This was found by tests/test_plugin.py::test_important_plugin and
was NOT a flake!
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-None: only just committed
This is a minimal fix: we wait until all plugins reply from init before
continuing. Really large or busy nodes can have other things monopolize
lightningd, then the timer goes off and we blame the plugin (which has
responded, we just haven't read it yet!).
The real answer is to have some timeouts only advance when we're idle,
or have them low-priority so we only activate them when we're idle (this
doesn't apply to all timers: some are probably important!). But
this is a minimal fix for -rc3.
Fixes: https://github.com/ElementsProject/lightning/issues/5736
Changelog-Fixed: plugins: on large/slow nodes we could blame plugins for failing to answer init in time, when we were just slow.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
And document support for it.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: Plugins: `getmanfest` response can contain `nonnumericids` to indicate support for modern string-based JSON request ids.
Changelog-Deprecated: Plugins: numeric JSON request ids: modern ones will be strings (see doc/lightningd-rpc.7.md!)
We had a quadratic check when processing a block, and this just
reduces it to be linear. Combined with the previous fix of adding txs
multiple times, this should speed up block processing considerably.
We were using a quadratic lookup to locate the transactions we
broadcast, so let's use an `htable` instead, allowing for constant
lookups during block processing.
The list of outgoing transactions was growing with every
rebroadcast. Now we just check whether it is already in the list and
don't add it anymore
Changelog-Fixed: ld: Reduce identification of own transactions to not slow down over time, reducing block processing time
Suggested-by: Matt Whitlock <@whitslack>
On testnet I noticed if we can't reach bitcoind for some reason, we'll
keep RBFing our penalty tx ("it didn't go in, RBF harder!!"). Makes
no sense to grossly exceed the amount needed for next block, so simply
cap penalty at 2x "estimatesmartfee 2 CONSERVATIVE".
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is how we put new invoice_requests into the db; this will be used
by a new "invoicerequest" command which replaces "offerout".
The API is now the same as the offers api.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We no longer use offers for "I want to send you money", but we'll use
invoice_requests directly. Create a new table for them, and
associated functions.
The "localofferid" for "pay" and "sendpay" is now "localinvreqid".
This is an experimental-only option, so document the change under
experimental only.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-EXPERIMENTAL: JSON-RPC: `pay` and `sendpay` `localofferid` is now `localinvreqid`.
I know this is an unforgivably large diff, but the spec has changed so
much that most of this amounts to a rewrite.
Some points:
* We no longer have "offer_id" fields, we generate that locally, as all
offer fields are mirrored into invoice_request and then invoice.
* Because of that mirroring, field names all have explicit offer/invreq/invoice
prefixes.
* The `refund_for` fields have been removed from spec: will re-add locally later.
* quantity_min was removed, max == 0 now mean "must specify a quantity".
* I have put recurrence fields back in locally.
This brings us to 655df03d8729c0918bdacac99eb13fdb0ee93345 ("BOLT 12:
add explicit invoice_node_id.")
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
And check it in invoice.c, insead of a hack where we compare against invhash.
Restore checking, too.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The "path" is just a message to ourselves. This meets the minimal
requirement for bolt12 invoices: that there be a blinded path (at
least so we can use the path_id inside in place of "payment_secret").
We expose the method to make this path_id to a common routine: offers
will need this for generating more sophisticated paths.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Instead of doing command_fail() in the else, do it immediately then
unindent the normal path.
No code changes.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We had a scheme where lightningd itself would put a per-node secret in
the blinded path, then we'd tell the caller when it was used. Then it
simply checks the alias to determine if the correct path was used.
But this doesn't work when we start to offer multiple blinded paths.
So go for a far simpler scheme, where the secret is generated (and
stored) by the caller, and hand it back to them.
We keep the split "with secret" or "without secret" API, since I'm
sure callers who don't care about the secret won't check that it
doesn't exist! And without that, someone can use a blinded path for a
different message and get a response which may reveal the node.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We actually want lightningd to create these, since it wants to put the
path_id secret in the last element. So best API is actually a generic
one, rather than separate APIs to create first and last ones.
And really, the more explicit initialization makes the users clearer.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This current spec is not strict enough: we might complain that the
next peer is not connected, for example, which leaks information.
So return WIRE_INVALID_ONION_BLINDING even if we're the first hop
on the path, to be safe.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This makes us match eed2ab0c30ad7f93e3b2641ca9d7ade32f3d121d
("Use `invalid_onion_blinding` everywhere").
1. Numerous typographical changes.
2. Make sure we *always* return WIRE_INVALID_ONION_BLINDING if
we're in a blinded path.
3. Handle p->total_msat correctly (MPP payments).
4. Reorganize blinding handling just like spec order.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Changed: JSON-RPC: `listfunds` now lists coinbase outputs as 'immature' until they're spendable
Changelog-Changed: JSON-RPC: UTXOs aren't spendable while immature
Since the "struct command" is different from plugins and lightningd, we
need an accessor for this to work (the plugin one is a dummy for now!).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Rusty Russell
niftynei
Michael Schmoock
Carl Dong
Vincenzo Palazzo
Alex Myers
Peter Neuroth
Ken Sedgwick
Carl Dong
Christian Decker