We can get weird errors when we try to load a database of a different
from. Just slap a git version in there for now.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Pierre points out that we don't handle this, and it can happen due
to race; the spec says we are not supposed to send PKT_CLOSE with
uncommitted changes.
Closes: #29
Reported-by: Pierre-Marie Padiou
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This means running 3 bitcoinds, which is slow enough to start on my laptop
that I need to increase the startup wait for 30 to 60 seconds, and similarly
the test.sh check loop.
Before: real 13m42.868s
After: real 8m19.563s (make -j3)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Waiting until lightningd is up is too long: do a --version test in setup,
and then check that all reported versions match later on.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Otherwise if they reconnect, we hit the assert in recv_body:
assert(!peer->inpkt);
Found by testing on my build box *without* valgrind (so it was fast
enough to do this).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Running on my build machine, without valgrind, it managed to exchange
closing sigs before restart, and spotted this bug.
Fixes: #76
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
broadcast_remainder() does two things: get the error message for the
previous transaction, and send the next one (shrinking the array).
But it has two bugs:
1) It logs results on the tx at the end of the array, which is the one
it is *about* to send, and
2) The initial caller (rebroadcast_txs) hands it the complete array,
so the first tx gets broadcast twice.
The correct thing to do is to strip the array, then send the tail for
the next callback. And use nicely-named vars to help document what
we're doing.
Reported-by: Christian Decker
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Now broadcast_tx() doesn't take ownership of the tx, make sure callers
free; a bit of refactoring to make it clear when we're making a new tx
vs. accessing an existing one, to make this clearer.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is in preparation for the next step.
Note that we now don't add it to the linked list of txs we've send
until after it's sent by the immediate callback; this means it won't
get broadcast by the timer until after it's been done by broadcast_tx.
Also, this means we no longer steal the tx in broadcast_tx(); but we'll fix
up the leaks 4 patches later.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We don't simply parent them on the object, we use a dummy object which tells
us not to call the callback if freed.
This would be better fixed by rewriting ccan/io to handle tal_free() of
a conn; then we could simply parent the entire thing and forget about it.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Testnet fees are all over the place: don't close a connection due to
that.
Closes: #59
Reported-by: Thomas Daede <daede003@umn.edu>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This fixes half of #59 (the part which crashes). It doesn't fix the
fact that we should never be doing this for testnet.
Reported-by: Thomas Daede <daede003@umn.edu>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Allocate the route off the current command, not dstate. And in the
case where the route is somehow not via a peer, don't leak memory.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Generally, the pattern is: everything returned is allocated off the return
value, which is the only thing allocated off the context. And it's always
freed.
Also, tal_free() returns NULL, so it's useful for one-line error
cleanups.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This shouldn't happen, but if our RNG is busted and get a duplicate,
r would be NULL.
(scan-build found this)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
`listinvoice` was not handling numeric labels correctly if they are not
passed in as JSON strings (which was the behavior of
`lightning-cli`). So now we accept both string labels as well as numeric
labels.
The JSON-RPC was non-standard compliant in that it omitted the required
`jsonrpc` entry and it was returning both `error` as well as
`result`. This fixes both of these issues.
Fixes#50. Was causing a segfault because it was creating an empty route
and was trying to extract the first hop as next hop. Routes through self
can still be created manually, but `sendpay` would still refuse to act
on them due to the missing self-link.
We're going to wean off IRC, but as a quick fix, only announce 0-60 seconds
after we see a join, or every 6 hours.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We've seen intermittant failures on testnet, so disable sending feechanges
for now: we're completely changing it for 0.6 anyway, due to Milan Spec.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Make this always the first packet after auth. That means there's no
reliance on whether a node remembers an aborted connection.
It also gives us a place to put version bits.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This gives much better errors, and allows us to return the peer id.
Closes: #37
Reported-by: Glenn Willen
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
My lightning node was *way* behind, and I couldn't figure out why it
wasn't seeing channel establishment.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
In particular, remove fee ranges on testnet (too unreliable) and accept
a single confirm.
(Note that an earlier version of this had a bug when there was no
config file, this version includes the fix).
Closes: #40
Reported-by: Glenn Willen
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Firstly, we need to update the staging fee amount when we queue a change.
Secondly we need to remove completed fee updates, otherwise we hit a
database constraint that peer & state are unique.
Reported-by: Christian Decker
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We were out by 1000, and also derived it from the previous, not current
state.
Reported-by: Christian Decker
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We store peers in the database for STATE_INIT, but they don't reconnect
properly. We should not forget STATE_INIT dropped peers, but use some
timeout mechanism if we can't reconnect to clean up.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Three days of on and off debugging, before I realized my server was talking
to a non-testnet bitcoind. There was a bitcoind on that machine running
on testnet, but it uses the same dir and config, so the --bitcoin-datadir
option couldn't help.
This is more certain: specify whether we're testnet on every single query.
Now we can skip the attempt to parse bitcoin.conf, too.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Fix dff50c3a5f2224592d135ad9cc6c11be89e26def: if multiple blocks come between
polling, we'll get called once for the new tip.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
getnodes returns an object containing a single array of 'nodes'. Each
element contains the node's ID, its hostname and its port. If
unknown (because we haven't seen a node announcement yet) then the port
is 0 and the hostname is null.
lightningd now uses a WHOIS query on itself to learn its external IP
address and announces that on the channel with the NODE message. It also
tracks other nodes in the routing table.
Refactored the signature verification to reuse it for both CHAN and NODE
messages.
'getchannels' returns a 'channels' array containing an object for each
known channel. Each channel object represents one direction of a
bidirectional channel, with a from and a to node ID along with the fees
for that direction. This matched the internal storage of channels and
allows unbalanced fees for each direction.
We *should* be in a state which accepts it (could happen with reorg),
and there's no reason to test for greater than depth since we must process
blocks in order.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Currently we get the odd message "Own anchor has insufficient funds".
Reported-by: Christian Decker
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
NO_VALGRIND= daemon/test/test.sh --normal --restart
lightning-cli: Connecting to 'lightning-rpc': Connection refused
lightning-cli: Connecting to 'lightning-rpc': Connection refused
lightning-cli: Connecting to 'lightning-rpc': Connection refused
lightning-cli: Connecting to 'lightning-rpc': Connection refused
lightning-cli: Connecting to 'lightning-rpc': Connection refused
This is expected: it happens when node3 is restarting. Redirect
errors to /dev/null.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Added channel announcement serialization and parsing, as well as the
entrypoints for the IRC peer discovery. Announcements are signed by the
sending endpoint and signatures are verified before adding the channels
to the local view of the topology. We do not yet verify the existence of
the anchor transaction.
It's not in a transaction in one caller, so wrap that.
This removes some more error handling code.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Rename the structs to match (and remove dev-echo).
This makes it clear that they're not the normal API.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We need some ordering to deliver them to the JSON "waitinvoice" command;
we use a counter where 0 means "unpaid".
We keep two lists now, one for unpaid and one for paid invoices.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We need some way to reflect the tradeoff between the possible delay if
a payment gets stuck, and the fees charged by nodes. This adds a risk
factor which reflects the probability that a node goes down, and the
cost associated with losing access to our funds for a given time.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
I had some nonsensical columns, eg "bool ours", but sqlite3 pretty much
ignores them. Use macros so mistakes are harder to make.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is important when we put payments in the database: they need to be
updated atomically as the HTLC is.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is important when we put payments in the database: they need to be
updated atomically as the HTLC is.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We had enum channel_side (OURS, THEIRS) for which end of a channel we
had, and htlc_side (LOCAL, REMOTE) for who proposed the HTLC.
Combine these both into simply "enum side".
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
In particular, make sure B can just afford it, then have the A add a
HTLC which means B can no longer afford the fees, and A should cover
it.
We do this by modifying the previous overlapping-fail test, but we
need to have B offer it the htlc before A does: racy in the normal
autocommit case. So we do a manual commit here, always.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
When they propose an HTLC to us, they need to be able to cover both it,
and the associated fees. When it gets acked and applied to them, however,
they may no longer be able to afford the fees; this is OK and expected.
So add a flag to say whether they can dig into fees or not: without
this patch the code calls fatal() on the next patch which tests it.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We create a logging object when we connect, then carry it through. If
it comes from the database, we just use the peerid as the log prefix.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
If we haven't received their closing signature yet, we might try to
send the closing packet anyway (and segfault). Make sure we have
their signature before trying that.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is less convenient to use, but makes far more sense for a real
user (like a wallet). It can ask about the route, then decide whether
to use it or not.
This will make even more sense once we add a parameter to control how
long we let the HTLC be delayed for, so a client can query for high,
medium and low tolerances and compare results.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We stopped automatically retransmitting locally-generated add/removes
after a reconnect, but this breaks the "pay" interface as it stands.
The correct solution to this is to make the pay interface idempotent:
you can trigger it as many times as you want and it will only succeed
once.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
If we've not relayed a failure yet (ie. we relayed it instantly, but it
wasn't confirmed), we need to redo it.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
It's not currently encrypted, but at least you get some idea now why
an HTLC failed. We (ab)use HTTP error codes for the moment.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
These low level commands we restarted on reconnect for ease of
testing. Don't do that, and check that we're connected when those
commands occur.
This introduces subtle issues with --manual-commit --reconnect: restarting
node1 also forgets uncommitted things from node2, requiring reordering for
some tests.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We capture the output in case we need to resubmit the command after restarting,
but we weren't printing it out on failure (set -e means we'd stop immediately).
As a side-effect of this change, we don't restart after failed
commands, which caused another bug: we were writing the 2->3 route to
the config file, but not restarting again, so we lost the route.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
If we send an HTLC #1, then get disconnected before a confirm, we will
forget it. But we've incremented peer->htlc_id_counter, so when we offer
it again we'll make it HTLC #2, which is non-consecutive.
To make this clear, we always start htlc ids at 0 now. That revealed
the bugs handled in the previous patch.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We don't have an ordering of HTLCs between peers: we don't know
whether your HTLC 0 or my HTLC 0 occurred first. This matters,
as we play them all back to reconstruct state (probably overkill anyway).
So we add force_* operators which don't do bounds checks, and do
bounds checks at the end. We also note that we don't need to apply
fee changes: that should be in the database already.
Also relax db constraints: IDs are not unique, they are unique per
side (we can both have HTLC #0).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The first hop is being stripped from computed routes, however the
first channel of the route is being used to get our peer address. This
results in segfaults if the route is just one hop, i.e., has no first
channel to get the peer's address from. Fixed by simply using an
existing pointer to our peer.
I originally overloaded struct htlc for this, as they go through the
same states, but separating them turned out to be clearer.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This allows hardcoded routes in the config file, which is required until
we get route advertisements.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This makes more sense eventually: we may know the network addresses of
many peers, not just those we're connecting to. So keep a mapping, and
update it when we successfully connect outwards.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Testing this revealed that we can't just reconnect when we have something to
send, as we might be NATed; we should try to reconnect anyway.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We add a "dev-restart" command which causes the daemon to close fds
and exec itself. Then we do it after every command, with the caveat
that we always send a commit before newhtlc, because if not committed,
that is forgotten. Fulfillhtlc and failhtlc get resent, since they're
idempotent.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
1. Fix #ifdef DEBUG code in signature.c so it compiles.
2. Don't set peer->closing.our_script in queue_pkt_close_shutdown: it's
assigned in caller already.
3. Wrap setting of htlc's rval in set_htlc_rval() function.
4. Log where we were when unexpected packet comes in.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
More of a pure allocator, for when we load peers from db. Also moves
shachain_init out of secrets and into new_peer where it logically
belongs.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
So if there are no HTLCs, and the receiver can't spend anyway, don't
sign. This has the added benefit that no two signed commitment
transactions will ever be identical (the revocation preimage changes).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This gives us a clear way to indicate "invalid", and also sqlite3 stores
signed 64-bit numbers, so it's clearer this way.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
It fits in a u32, but we mix it with other values which could cause
overflow, so let's just use u64 everywhere.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is dumb, since one side will never succeed. But in future when
there is a method for nodes to broadcast their public address (or send
their address inline to connected nodes), either side should try to
connect.
Importantly though, there are places which will queue packets at
various times (eg. HTLC timeout), so we need to clear the queue just
before re-transmitting, not when disconnecting.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
To do this we keep an order counter so we know how to retransmit. We
could simply keep old packets, but this is a little clearer for now.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Caught because we generated an HTLCs which had already expired, since
we didn't know the latest block. Other errors are certainly possible,
so it's safest to load the entire thing before going live.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This avoids us having to query it when we create anchor transaction, and
lets us always use dynamic fee information.
The config options for max and min are now percentages, rather than absolute.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We no longer need it anywhere. This simplifies things to the point where
we might as well just not include dust outputs as we go, rather than
explicitly removing them, which gets rid of remove_dust.c as well.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Similar to the way we derive which outputs are which for old transactions
we steal, we derive them even for their current transaction.
We keep track of this information in peer->closing.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
At the moment, for our or their unilateral close, we create a resolved[]
entry for our output, their output, and each HTLC, in cstate order. Some
of these outputs might not exist (too small), so it's actually better
to simply keep a resolved[] entry for each of the tx's actual outputs.
(We already changed the steal resolved[] array to work like this, but
these are trickier, since we rely on that order if we need to fulfill an
on-chain HTLC).
It also helps as we are weaning off knowing the cstate and permutation
mapping for each commitment transaction.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We want to stop keeping old commitment information (except the minimal
txid to commitment-number mapping). One place we currently use it is
after sending a commitment signature, and before we've received the
revocation for the old commitment. For this duration, there are two
valid commitment transactions.
So we store "their_prev_revocation_hash" explicitly for this duration.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Rusty Russell
Glenn Willen
Christian Decker