This is a fairly direct translation. Even so, it should be faster in
most cases, and and we can do more sophisticated things if we want.
This also handles disabled channels better.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Changed: plugins: `pay` will now try disabled channels as a last resort.
Fixes: #3926
(probably)
Changelog-Fixed: pay: Also limit the number of splits if the payee seems to have a low number of channels that can enter it, given the max-concurrent-htlcs limit.
The routehints paymod shares the storage of the array d->routehints and
p->invoice->routes, but once it operates, it possibly leaves it as a stale
pointer to memory it used to have.
Since other paymods may be interested in the invoice details, including
the routehints in the invoice, we should ensure the p->invoice->routes
remains valid whenever we try mutating that array.
As revealed by the failure of tests in #3936, where we ended up trying
to send a partial payment using legacy style, we are not handling
style properly.
1. BOLT9 has features, so we can *know* that the destination supports
MPP. We may not have seen a node_announcement.
2. We can't assume that nodes inside routehints support TLV.
3. We can't assume direct peers support TLV.
The keysend code tried to fix this up, so I'm not sure that this caused
the issue in #3968, though.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Fixed: `pay` will now make reliable multi-part payments to nodes it doesn't have a node_announcement for.
Using `waitblockheight 0` is a very slightly faster query than `getinfo`.
Also, avoid querying blockheight for child payments (allow `waitblockheight`
paymod to provide the blockheight returned from the `waitblockheight`, and
just resample the starting blockheight from the parent).
Changelog-None: pointless micro-optimization
This was checked with `gcc -S -O2` to see how an optimized build
would compile the function.
The original code completed calls into each child (and the `.s`
file showed that GCC 9.x was not smart enough to do early-out).
This modification explicitly does early-out, and avoids call-return
stack overhead for the common case where a payment is an ancestor
of a long line of single-child payments due to retrying.
Changelog-None: pointless micro-optimization
This is the simplest possible fix: increase the target amount until we get
the desired number of parts, while still bucketizing payments together that
are in approximately the same size.
The current logic puts all payments that are in the range x < amount <= 16*x
in the same bucket, making them harder to distinguish.
Changelog-Fixed: pay: The `presplit` modifier now supports large payments without exhausting the available HTLCs.
Reported-by: ZmnSCPxj
Signed-off-by: Christian Decker <@cdecker>
Changelog-Fixed: pay: Correct a case where we put the sub-payment value instead of the *total* value in the `total_msat` field of a multi-part payment.
Only advance through routehints if no route was found at all, or if the
estimated capacity at the routehint is lower than the amount that we
have to send through the routehint.
Changelog-Fixed: pay: Be less aggressive with forgetting routehints.
It's not all that rare to do these operations, and requiring annotations
for it is a little painful.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Arguably a low-priority bug since no current node ever generates routehints longer
than one hop.
However, it is possible as an edge case, if the destination is directly accessible
*and* supports multiple channels, that we route through the destination, one of the
*other* channels it has not in the routehint, to the entry point, and then through
the routehint.
This change removes the risk of the above edge case.
Changelog-None: arguably a low-priority bug.
We'd previously take the failed attempt and estimate the failing channel's
capacity at 3/4 of the attempted amount, which is rather aggressive. This
reduces this aggressiveness to use the exact amount tried, but excluding on
equality. This still skips attempting the same route with the same amount, but
also permits attempts that are in the range [3/4, 1] of the failed attempt
amount to still be attempted.
While not directly necessary, it still feeds the `listpays` result, and so we
should pass it along if we can, so we don't have to rely solely on the
`amount_sent` field, which includes the fees.
Reported-by: Rusty Russell <@rustyrussell>
The shortcut in the retry_mod that we can skip retrying if getroute fails or
we have no result is only valid if the parameters don't change. As we iterate
through the routehints the parameters change, and so we must signal to the
retry_mod that it can retry even in those cases.
The child payments will sometimes depend on the step of the parent, and making
sure that the parent state is correct before we create the children is
therefore important.
This uses @cdecker's idea of excluding the routehinted channel from the route,
and also consumes the route hints as it goes so that it makes progress.
I don't know if this is correct, but it reliably passes tests/test_pay.py::test_tlv_or_legacy
now.
We store an offset of the current routehint in the modifier data. It gets
incremented on retry, and it gets reset to 0 on split. This is because once we
split we have a different amount and a previously unusable routehint becomes
usable again.
We have two places we need to do that now: in the root payment after we
checked if the destination is reachable, and in any other payment directly in
the initialization-step callback.
It was spread over the step callback, but we only need to initialize the
routehints array there, child-payments can just inherit most of the information.
This does two things: it checks if the destination of the payment is at all
reachable without routehints, and if it is it adds a direct attempt as option
to the routehints in the form of a NULL routehint. It also simplifies the
selection of the routehint since the direct case is no longer special, instead
we just return a NULL routehint as if it were a normal routehint.
The adaptive MPP test was showing an issue with always using a routehint, even
when it wasn't necessary: we would insist on routhing to the entrypoint of the
routehint, even through the actual destination. If a channel on that loop
would result being over capacity we'd slam below 0, and then increase again by
unapplying the route. The solution really is not to insist on routing through
a routehint, so we implement random skipping of routehints, and we rotate them
if we have multiples.
As the hints get new fields added it is easy to forget to amend one of the
places we create them, since we already have an update method let's use that
to handle all additions to the array of known channel hints.
We were removing the current hint from the list and not inheriting the current
routehint, so we'd be forgetting a hint at each retry. Now we keep the array
unchanged in the root, and simply skip the ones that are not usable given the
current information we have about the channels (in the form of channel_hints).
Fixes#3861
This may be related to the issue #3862, however the water was muddied by it
being the wrong error to return, and the node should not expect this courtesy
feature to be present at all...
There is little point in trying to split if the resulting HTLCs exceed the
maximum number of HTLCs we can add to our channels. So abort if a split would
result in more HTLCs than our channels can support.
The presplit modifier could end up exceeding the maximum number of HTLCs we
can add to a channel right out the gate, so we switch to a dynamic presplit if
that is the case. The presplit will now at most use 1/3rd of the available
HTLCs on the channels if the normal split would exceed the number of availabe
HTLCs. And we also abort early if we don't have a sufficient HTLCs available.
It turns out that by aggressively splitting payments we may end up exhausting
the number of HTLCs we can add to a channel quickly. By tracking the number of
HTLCs we can still add, and excluding the channels to which we cannot add any
more we increase the route diversity, and avoid quickly exhausting the HTLC
budget.
In the next commit we'll also implement an early abort if we've exhausted all
channels, so we don't end up splitting indefinitely and we can also optimize
the initial split to not run afoul of that limit.
This PR includes the fix discussed on PR #3855. This fix was tested with the use case described inside the issue and worked.
Fixes: #3855
Changelog-None
The code had incorrect assertions, partially because it didn't clearly
distinguish errors from the final node (which, barring blockheight issues,
mean a complete failre) and intermediate nodes.
In particular, we can't trust the *values*, so we need to distinguish
these by the *sender*.
If a route is of length 2 (A, B):
- erring_index == 0 means us complaining about channel A.
- erring_index == 1 means A.node complaining about channel B.
- erring_index == 2 means the final destination node B.node.
This is particularly of note because Travis does NOT run test_pay_routeboost!
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Rusty Russell
ZmnSCPxj jxPCSnmZ
Christian Decker
Vincenzo Palazzo
Vincent