https://github.com/lightningnetwork/lightning-rfc/pull/877 talks about
removing this restriction (only Electrum actually enforced it on
receive), so start by allowing creation of giant invoices, though
we mark them as requiring mpp.
Changelog-Changed: JSON-RPC: `invoice` now allows creation of giant invoices (>= 2^32 msat)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This mainly helps our CI under valgrind, which starts a fresh instance
and immediately calls the invoice command. This can cause the topology
plugin to try to access the gossmap file before it's created.
We can also move the gossmap reading in topology to init time.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This involves removing some fields from the now-misnamed routing.h
datastructures, and various internal messages.
One non-obvious change is to our "keepalive" logic which refreshes
channels every 13 days: instead of using the 'enabled' flag on the
last channel broadcast to decide whether to refresh it, we use the
local connected status directly.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This turned into a more extensive cleanup than intended. The previous
warnings were overlapping and confusing, especially now MPP is the norm.
*warning_capacity* is now the "even under best circumstances, we don't
have enough incoming capacity", which is really what
warning_mpp_capacity was trying to say (no longer printed).
*warning_offline* and *warning_deadends* are only given if adding such
peers would have helped give capacity (i.e. not if *warning_capacity*
is set). The new *warning_private_unused* tells you that we would
have sufficient capacity, but we refused to expose private channels.
The test cases have been enhanced to cover the new warnings.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: JSON-RPC: `invoice` now gives `warning_private_unused` if unused unannounced channels could have provided sufficient capacity.
Changelog-Changed: JSON-RPC: `invoice` warnings are now better defined, and `warning_mpp_capacity` is no longer included (since `warning_capacity` covers that).
Temporarily rename old getroute to getrouteold (we will remove this).
Changelog-Changed: JSON-RPC: `getroute` is now implemented in a plugin.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Prior to this, sending a v1 address (or, in fact, any random crap!)
would cause the unsupporting node to unilaterally close.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
An "active" channel may still be CHANNELD_AWAITING_LOCKIN, so have ->scid NULL.
You can only trigger this by trying to sendpay to the node using a
manual route, since routing would never use such a channel.
```
lightningd: FATAL SIGNAL 11 (version v0.10.0-319-g81cbc20-modded)
0x55e79d194e17 send_backtrace
common/daemon.c:39
0x55e79d194ec1 crashdump
common/daemon.c:52
0x7fce2d79920f ???
???:0
0x7fce2d8e16f7 ???
???:0
0x55e79d2019eb tal_dup_
ccan/ccan/tal/tal.c:801
0x55e79d14e1d9 immediate_routing_failure
lightningd/pay.c:365
0x55e79d14fe91 send_payment_core
lightningd/pay.c:1022
0x55e79d150995 send_payment
lightningd/pay.c:1180
0x55e79d151975 json_sendpay
lightningd/pay.c:1462
```
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Annotating the htlc in `listpeers` with their current status, and
which plugin is currently holding on to them with their
`htlc_accepted` hook can help us debug where plugins may go wrong.
Changelog-Added: jsonrpc: HTLCs in `listpeers` are now annotated with a status if they are waiting on an `htlc_accepted` hook of a plugin.
We don't actually set desired_type yet, but this handles it.
Changelog-EXPERIMENTAL: Protocol: we can now upgrade old channels to `option_static_remotekey` from https://github.com/lightningnetwork/lightning-rfc/pull/868
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Currently we abuse openingd and dualopend to do this, but connectd already
has the ability to talk to peers, so it's more efficient.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Undocumented (caught by json schema!) if we discard channel because it
wasn't open yet, then close returned the empty object. Make it return
a new type in this case.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: JSONRPC: `close` returns `type` "unopened" if it simply discards channel instead of empty object.
Tor v2 hidden services have been deprecated for a while:
https://blog.torproject.org/v2-deprecation-timeline .
This prevents user from being able to set them in the configuration
and to connect to them while still letting us be able to parse them
for gossip.
Changelog-Deprecated: lightningd: v2 Tor addresses. Use v3. See https://blog.torproject.org/v2-deprecation-timeline.
Signed-off-by: Antoine Poinsot <darosior@protonmail.com>
If the peer is offline when we see the funding txid, we don't actually
update the channel's info. Here, we move it up to where the scid is set,
so that we always update the channel's funding_txid to the correct
(mined) information.
We set the timeout on first HTLC, but didn't clear it if that HTLC failed.
It's saner to have a per-HTLC timeout (since that's what it is!) and
also our timer infra is specially coded to scale approximately infinitely so
trying to optimize this is vastly premature.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Fixed: Protocol: We would sometimes gratuitously disconnect 30 seconds after an HTLC failed.
1. We don't need to check for NULL before tal_count(NULL).
2. Use of json_for_each_arr iterator is probably better.
3. Weird indent fixed.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Trying to put all the disconnect logic into the same path was a dumb
idea. If you asked to reconnect but passed in an 'unsaved' channel, we
would not call the 'reconnect' code.
Instead, we make a differentiation between "unsaved" channels
(ones that we haven't received commitment tx for) and handle the
disconnect for these separate from where we want to do a reconnect.
We were freeing the payload, which is then subsequently freed by the
plugin_hook caller. Whoops.
Now we pass through to the callback function and just clean up neatly.
------------------------------- Valgrind errors --------------------------------
Valgrind error file: valgrind-errors.406602
==406602== Invalid read of size 8
==406602== at 0x12AC93: openchannel2_hook_cb (dual_open_control.c:669)
==406602== by 0x12AF0A: openchannel2_hook_deserialize (dual_open_control.c:721)
==406602== by 0x16EF0E: plugin_hook_callback (plugin_hook.c:186)
==406602== by 0x169746: plugin_response_handle (plugin.c:514)
==406602== by 0x169959: plugin_read_json_one (plugin.c:620)
==406602== by 0x169B23: plugin_read_json (plugin.c:665)
==406602== by 0x1F4076: next_plan (io.c:59)
==406602== by 0x1F4C5B: do_plan (io.c:407)
==406602== by 0x1F4C9D: io_ready (io.c:417)
==406602== by 0x1F6F35: io_loop (poll.c:445)
==406602== by 0x13D48D: io_loop_with_timers (io_loop_with_timers.c:24)
==406602== by 0x143388: main (lightningd.c:1111)
==406602== Address 0x75e7418 is 56 bytes inside a block of size 3,520 free'd
==406602== at 0x483CA3F: free (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
==406602== by 0x204FB0: del_tree (tal.c:421)
==406602== by 0x20527E: tal_free (tal.c:486)
==406602== by 0x122D68: delete_channel (channel.c:124)
==406602== by 0x129291: channel_disconnect (dual_open_control.c:63)
==406602== by 0x129364: channel_close_conn (dual_open_control.c:82)
==406602== by 0x131CF6: peer_please_disconnect (connect_control.c:304)
==406602== by 0x131DEB: connectd_msg (connect_control.c:326)
==406602== by 0x172023: sd_msg_read (subd.c:509)
==406602== by 0x1F4076: next_plan (io.c:59)
==406602== by 0x1F4C5B: do_plan (io.c:407)
==406602== by 0x1F4C9D: io_ready (io.c:417)
==406602== Block was alloc'd at
==406602== at 0x483B7F3: malloc (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
==406602== by 0x204A39: allocate (tal.c:250)
==406602== by 0x204FFA: tal_alloc_ (tal.c:428)
==406602== by 0x123165: new_unsaved_channel (channel.c:209)
==406602== by 0x130D34: peer_start_dualopend (dual_open_control.c:2985)
==406602== by 0x15BD2A: peer_connected_hook_final (peer_control.c:1105)
==406602== by 0x16F2E5: plugin_hook_call_ (plugin_hook.c:275)
==406602== by 0x15BF5C: plugin_hook_call_peer_connected (peer_control.c:1155)
==406602== by 0x15C16C: peer_connected (peer_control.c:1208)
==406602== by 0x131E3B: connectd_msg (connect_control.c:332)
==406602== by 0x172023: sd_msg_read (subd.c:509)
==406602== by 0x171842: read_fds (subd.c:310)
Fixes: #4494
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: config: New option `log-timestamps` allow disabling of timestamp prefix in logs.
Since plugins will start sending them soon, and they are likely to get
it wrong sometimes, be a bit more lenient, warn them in the logs
instead and then make sure it doesn't accidentally work anyway.
A plugin might subscribe to a notification topic that is only
registered by another plugin later, so push the check to that
consistency check phase where we do hook ordering as well.
We will eventually start emitting and dispatching custom notifications
from plugins just like we dispatch internal notifications. In order to
get reasonable error messages we need to make sure that the topics
plugins are asking for were correctly registered. When doing this we
don't really care about whether the plugin that registered the
notification is still alive or not (it might have died, but
subscribers should stay up and running), so we keep a list of all
topics attached to the `struct plugins` which gathers global plugin
information.
This avoids subdaemons complaining about malformed messages from us,
or doing the completely wrong thing, if they are really the wrong
version.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
You still shouldn't do this (you could get some transient failures),
but at least you have a decent chance if you reinstall over a running
daemon, instead of getting confusing internal errors if message
formats have changed.
Changelog-Added: lightningd: we now try to restart if subdaemons are upgraded underneath us.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Fixes: #4346
We generally hang things off our JSON response (this pattern predates
tmpctx!) but sometimes it gets reported as a memleak. I'd prefer not
to mark JSON responses as "notleak", since they can be allocated for
a while), so use tmpctx here.
```
E ValueError:
E Node errors:
E Global errors:
E - Node /tmp/ltests-spnausnb/test_htlc_out_timeout_1/lightning-1/ has memory leaks: [
E {
E "backtrace": [
E "ccan/ccan/tal/tal.c:442 (tal_alloc_)",
E "ccan/ccan/tal/tal.c:471 (tal_alloc_arr_)",
E "wallet/wallet.c:1775 (wallet_state_change_get)",
E "lightningd/peer_control.c:922 (json_add_channel)",
E "lightningd/peer_control.c:1424 (json_add_peer)",
E "lightningd/peer_control.c:1454 (json_listpeers)",
E "lightningd/jsonrpc.c:643 (command_exec)",
E "lightningd/jsonrpc.c:767 (rpc_command_hook_final)",
E "lightningd/plugin_hook.c:275 (plugin_hook_call_)",
E "lightningd/jsonrpc.c:855 (plugin_hook_call_rpc_command)",
E "lightningd/jsonrpc.c:942 (parse_request)",
E "lightningd/jsonrpc.c:1033 (read_json)",
E "ccan/ccan/io/io.c:59 (next_plan)",
E "ccan/ccan/io/io.c:435 (io_do_always)",
E "ccan/ccan/io/poll.c:300 (handle_always)",
E "ccan/ccan/io/poll.c:377 (io_loop)",
E "lightningd/io_loop_with_timers.c:24 (io_loop_with_timers)",
E "lightningd/lightningd.c:1097 (main)"
E ],
E "label": "wallet/wallet.c:1775:struct state_change_entry[]",
E "parents": [
E "common/json_stream.c:29:struct json_stream",
E "ccan/ccan/io/io.c:91:struct io_conn",
E "lightningd/lightningd.c:116:struct lightningd"
E ],
E "value": "0x55c6b02150b8"
E }
E ]
```
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The main change which affects us is that 2016 blocks to forget a channel
is a fixed number in the spec; we make this clear by renaming the
(developer-only) max_funding_unconfirmed to dev_max_funding_unconfirmed
and making it compile DEVELOPER only.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This means remembering the connection direction. We also use the address to try
to reconnect, which we shouldn't bother with if they connect to us.
For peers from the database, we currently always save the addr: we shouldn't really
do this if they connected to us, since it's not useful for reconnecting (we don't
show the addr in JSON reply to listpeers unless we're connected, so it's only an
internal issue). This is left for future work.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This matters: if we connected, the address is probably usable for future connections.
But if they connected, the port is probably not (but the IP address may be).
Changelog-Added: JSON-RPC: `connect` returns "direction" ("in": they iniatated, or "out": we initiated)
Changelog-Added: plugins: `peer_connected` hook and `connect` notifications have "direction" field.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
You can now activate dual-funded channels using the
`--experimental-dual-fund` flag
Changelog-Changed: Config: `--experimental-dual-fund` runtime flag will enable dual-funded protocol on this node
As @fiatjaf points out we were notifying before we were actually set
on accepting, since the hook could also still reject. Switched them
around does and calling the notification only once it's been decided
is the correct thing to do.
Changelog-Fixed: plugin: The `invoice_payment` notification was being sent before the hook was called, which could still abort it.
Suggested-by: Fiatjaf <@fiatjaf>
Signed-off-by: Christian Decker <@cdecker>
There are perfectly valid reasons for us to not have a command on return
(something went boom while sending them our sigs and we've now gotten
their sigs during a reconnect and subsequently broadcast the tx)
This can result in us logging a warning if we've 1) dropped their sigs
response, 2) only us (the opener) added inputs, 3) and we broadcast on
their reconnect (when they retransmit their sigs)
Otherwise, we might find an address other than the one given and
the user might think that address worked.
Fixes: #4185
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: JSON-RPC: `connect` returns `address` it actually connected to
And update all the in-tree callers.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Deprecated: JSON-RPC: `fundchannel_complete` `txid` and `txout` parameters (use `psbt`)
Requiring the user to calculate the txid of the PSBT is a horrible, bad,
no-good idea.
Doesn't deprecate yet, so I can test that this path works while
multifundchannel still uses it.
Fixes: #4416 (at least for future users!)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: JSON-RPC: `fundchannel_complete` takes a psbt parameter.