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>
Rework the logic of the version check used in the
database migration, and make sure
that it is full functional to avoid confusion
at release time.
Changelog-Fixed: database: Correctly identity official release versions for database upgrade.
Reported-by: @urza
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>
I discovered this while reworking the CI workflow, and it seems like
the HTLC queries do not order, while the tests assume a specific
order. This matches sqlite3 which without an explicit ORDER clause
will use insertion order, while postgres does not keep things in
insertion order, thus breaking the assumption. Ordering by `id`
re-establishes that implicit assumption
Changelog-Changed: postgres: Ordering of HTLCs in `listhtlcs` are now ordered by time of creation
Changelog-Removed: JSON-RPC: `fundpsbt`/`utxopsbt` `reserve` must be a number, not bool (deprecated v0.11.0)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This was reported, but the channel was closed. So however we ended
up with a duplicate, we're no *worse* off than we were before migration?
Fixes: #5760
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We have a primary key that is spanning the `in_channel_id` and the
`in_htcl_id`. The latter gets set to NULL when the HTLC and channel
gets deleted, so we coalesce with a random large number that is
unlikely to collide for the primary key.
Ubuntu clang 15.0.2-1 complains:
```
wallet/wallet.c:280:6: error: variable 'i' set but not used
[-Werror,-Wunused-but-set-variable]
int i;
^
wallet/wallet.c:339:6: error: variable 'i' set but not used
[-Werror,-Wunused-but-set-variable]
int i;
^
wallet/wallet.c:4768:9: error: variable 'count' set but not used
[-Werror,-Wunused-but-set-variable]
size_t count;
^
```
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`.
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>
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
We still have an "enum forward_style" for the database, where old-style
forwards can still exist.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Removed: Protocol: we no longer forward HTLCs with legacy onions.
We have to allow them (as otherwise `fees_collected_msat` in getinfo breaks),
but it means that actually, in_htlc_id might be missing in listforwards
(also, out_htlc_id might be missing, which we didn't catch before).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Fixes: #5628
Otherwise what the hook sees is actually a lie, and if it sets it
we might override it.
The side effect is that we add an explicit "forward_to" field, and
allow hooks to override it. This lets a *hook* control channel
choice explicitly.
Changelod-Added: Plugins: `htlc_accepted_hook` return can specify what channel to forward htlc to.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is actually what the autoclean plugin wants, especially since
you can't otherwise delete a payment which has failed then succeeded.
But insist on neither or both being specified, at least for now.
Changelog-Added: JSON-RPC: `delpay` takes optional `groupid` and `partid` parameters to specify exactly what payment to delete.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Using `listfowards` for this wrong; expose this directly if people
care (and unlike listforwards, which could be deleted, we have to
remember these while the channel is still open!).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: JSON-RPC: `listhtlcs` new command to list all known HTLCS.
And document that we never know payment_hash.
Changelog-Added: JSON-RPC: `listforwards` now shows `in_htlc_id` and `out_htlc_id`
Changelog-Changed: JSON-RPC: `listforwards` now never shows `payment_hash`; use `listhtlcs`.
Normally, we'd use the delete_columns function to remove the old
`short_channel_id` string field, *but* we can't do that for sqlite, as
there are other tables with references to it. So add a FIXME to do
it once everyone has upgraded to an sqlite3 which has native support
for column deletion.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Although it's deprecated already (because it stores as string), it's
better to make the name explicit. And create a new helper which stores as BIGINT.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This one directly contains the scids of the channels involved, not
references, so can outlive the channels. As a side-effect, however,
it now never lists `payment_hash`. Having it listed (via join) is not
possible as it is a *string* in the channels table, and difficult
anyway because of channel aliases.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
contrib/giantnode.py shows we're spending a lot of time looking up
payments by payment_hash: sendpays does it to see if we've already
paid, and bookkeeper does it in listsendpays:
```
- 94.52% 0.00% lightningd lightningd [.] read_json
- 94.52% read_json
- 94.48% parse_request
- 94.46% plugin_hook_call_rpc_command
- plugin_hook_call_
- rpc_command_hook_final
- 94.46% command_exec
- 49.08% json_sendpay
- 49.01% send_payment
- 48.86% send_payment_core
- 48.84% wallet_payment_list
- 48.80% db_step
+ db_sqlite3_step
- 45.38% json_listsendpays
- 45.36% wallet_payment_list
- 45.30% db_step
+ 45.30% db_sqlite3_step
```
This doesn't actually make much of a difference, so see next patch.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This allows GDB to print values, but also allows us to use them in
'case' statements. This wasn't allowed before because they're not
constant terms.
This also made it clear there's a clash between two error codes,
so move one.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Changed: JSON-RPC: Error code from bcli plugin changed from 400 to 500.
First, merge the _ahf_ and non-ahf interfaces.
Second, remove the always-NULL txs->cmd field.
Then, add optional id_prefix for bitcoind_sendrawx, so if it's
triggered by a command (e.g. "withdraw") it's shown correctly in logs.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is cleaner because, the `remote_addr` and `discovered_ip` are
related but two different things.
Within connectd and lightningd we use the peers `remote_addr` feature
to validate (and guess a port) to be used for IP discovery.
Also when a peer reports us a `remote_addr`, this is given to the plugin API
via the `peer_connected` hook. The network port here is not modified for
godd reason! This can be used i.e. to detect if we are behind a NAT.
But once lightningd figures enough peers report the same `remote_addr`,
it sets the port to the selected network and tells gossipd to use that for
`node_announcement` updates.
Hence, within gossipd, there is no (should not be) `remote_addr`.
Changelog-None
This is a good sanity check that users understand that if they upgrade
to master mid-cycle they can't go back!
Suggested-by: @wtogami
Changelog-Added: Config: `--database-upgrade=true` required if a non-release version wants to (irrevocably!) upgrade the db.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
