Took me a while (stressing under valgrind) to reproduce this,
then longer to figure out how it happened.
Turns out io_new_conn() can fail if the init function fails.
In our case, this can happen if connect() immediately returns
an error (inside io_connect). But we've already set the finish
function, which (if this was the last address), will free connect,
making the assignment `connect->conn = ...` write to a freed address.
Either way, if it fails, try_connect_one_addr() has taken care to
update connect->conn, or free connect, and the caller should not do it.
Here's the valgrind trace:
```
==384981== Invalid write of size 8
==384981== at 0x11127C: try_connect_one_addr (connectd.c:880)
==384981== by 0x112BA1: destroy_io_conn (connectd.c:708)
==384981== by 0x141459: destroy_conn (poll.c:244)
==384981== by 0x14147F: destroy_conn_close_fd (poll.c:250)
==384981== by 0x149EB9: notify (tal.c:240)
==384981== by 0x149F8B: del_tree (tal.c:402)
==384981== by 0x14A51A: tal_free (tal.c:486)
==384981== by 0x140036: io_close (io.c:450)
==384981== by 0x1400B3: do_plan (io.c:401)
==384981== by 0x140134: io_ready (io.c:423)
==384981== by 0x141A57: io_loop (poll.c:445)
==384981== by 0x112CB0: main (connectd.c:1703)
==384981== Address 0x4d67020 is 64 bytes inside a block of size 160 free'd
==384981== at 0x483CA3F: free (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
==384981== by 0x14A020: del_tree (tal.c:421)
==384981== by 0x14A51A: tal_free (tal.c:486)
==384981== by 0x1110C5: try_connect_one_addr (connectd.c:806)
==384981== by 0x112BA1: destroy_io_conn (connectd.c:708)
==384981== by 0x141459: destroy_conn (poll.c:244)
==384981== by 0x14147F: destroy_conn_close_fd (poll.c:250)
==384981== by 0x149EB9: notify (tal.c:240)
==384981== by 0x149F8B: del_tree (tal.c:402)
==384981== by 0x14A51A: tal_free (tal.c:486)
==384981== by 0x140036: io_close (io.c:450)
==384981== by 0x1405DC: io_connect_ (io.c:345)
==384981== Block was alloc'd at
==384981== at 0x483B7F3: malloc (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
==384981== by 0x149CF1: allocate (tal.c:250)
==384981== by 0x14A3C6: tal_alloc_ (tal.c:428)
==384981== by 0x1114F2: try_connect_peer (connectd.c:1526)
==384981== by 0x111717: connect_to_peer (connectd.c:1558)
==384981== by 0x1124F5: recv_req (connectd.c:1627)
==384981== by 0x1188B2: handle_read (daemon_conn.c:31)
==384981== by 0x13FBCB: next_plan (io.c:59)
==384981== by 0x140076: do_plan (io.c:407)
==384981== by 0x140113: io_ready (io.c:417)
==384981== by 0x141A57: io_loop (poll.c:445)
==384981== by 0x112CB0: main (connectd.c:1703)
```
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Fixed: Occasional crash in connectd due to use-after-free
Fixes: #4343
If we're already attempting to connect to a peer, we would ignore
new connection requests. This is problematic if your node has bad
connection details for the node -- you can't update it while inflight.
This patch appends new connection suggestions to the list of connections
to try.
Fixes#4154
1. Rename memleak_enter_allocations to memleak_find_allocations.
2. Unify scanning for pointers into memleak_remove_region / memleak_remove_pointer.
3. Document the functions.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is simple, and we now can multifundchannel to every node on testnet
(one simply hangs once we connect).
Changelog-Fixed: Protocol: We now hang up if peer doesn't respond to init message after 60 seconds.
This avoids overwriting the ones in git, and generally makes things neater.
We have convenience headers wire/peer_wire.h and wire/onion_wire.h to
avoid most #ifdefs: simply include those.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Note that other directories were explicitly depending on the generated
file, instead of relying on their (already existing) dependency on
$(LIGHTNINGD_HSM_CLIENT_OBJS), so we remove that.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This simplifies our test matrix, as we never have to handle talking
to peers that specify one but not the other.
This is particularly important for option_anchor_outputs which
assumes option_static_remotekey.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
@thestick613 noticed that since tor version below 0.3.2.2-alpha
will not support V3 ed25519 address formats, the error handling
is not that helpful in the error message from cli.
So now we add an hint.
Changelog-None:
Signed-off-by: Saibato <saibato.naga@pm.me>
connectd/connectd.h; Add helper function to update conn error list
Signed-off-by: Saibato <saibato.naga@pm.me>
common/onion is going to need to use this for the case where it finds a blinding
seed inside the TLV. But how it does ecdh is daemon-specific.
We already had this problem for devtools/gossipwith, which supplied a
special hsm_do_ecdh(). This just makes it more general.
So we create a generic ecdh() interface, with a specific implementation
which subdaemons and lightningd can use.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
It's almost always "their_features" and "our_features" respectively, so
make those names clear.
Suggested-by: @cdecker
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Turns out that unnecessary: all callers can access the feature_set,
so make it much more like a normal primitive.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This will help with the next patch, where we wean off using a global
for features: connectd.c has access to the feature bits.
Since connectd might now want to send a message, it needs the crypto_state
non-const, which makes this less trivial than it would otherwise be.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This restriction was removed from the spec as of
86c2ebcc5973a4133d3ce4d80ae1c203061a1646.
We also fix up some strange formatting in that part of the documentation.
Changelog-changed: We now announce multiple addresses of the same type, if given.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is a common thing to do, so create a macro.
Unfortunately, it still needs the type arg, because the paramter may
be const, and the return cannot be, and C doesn't have a general
"(-const)" cast.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The `init_featurebits` are computed at startup, and then cached
indefinitely. They are then used whenever a new `init` handshake is performed.
We could add a new message to push updates to `connectd` whenever a plugin is
added or removed, but that's up for discussion.
Before this patch we used `int` for error codes. The problem with
`int` is that we try to pass it to/from wire and the size of `int` is
not defined by the standard. So a sender with 4-byte `int` would write
4 bytes to the wire and a receiver with 2-byte `int` (for example) would
read just 2 bytes from the wire.
To resolve this:
* Introduce an error code type with a known size:
`typedef s32 errcode_t`.
* Change all error code macros to constants of type `errcode_t`.
Constants also play better with gdb - it would visualize the name of
the constant instead of the numeric value.
* Change all functions that take error codes to take the new type
`errcode_t` instead of `int`.
* Introduce towire / fromwire functions to send / receive the newly added
type `errcode_t` and use it instead of `towire_int()`.
In addition:
* Remove the now unneeded `towire_int()`.
* Replace a hardcoded error code `-2` with a new constant
`INVOICE_EXPIRED_DURING_WAIT` (903).
Changelog-Changed: The waitinvoice command would now return error code 903 to designate that the invoice expired during wait, instead of the previous -2
Add towire_int() and fromwire_int() functions to "(de)serialize"
"int". This will only work as long as both the caller of towire_int()
and the caller of fromwire_int() use the same in-memory representation
of signed integers and have the same sizeof(int).
Changelog-None
We want to have a static Tor service created from a blob bound to
our node on cmdline
Changelog-added: persistent Tor address support
Changelog-added: allow the Tor inbound service port differ from 9735
Signed-off-by: Saibato <saibato.naga@pm.me>
Add base64 encode/decode to common
We need this to encode the blob for the tor service
Signed-off-by: Saibato <saibato.naga@pm.me>
Even if it is on startup only once ...
Like @bitcoin-software indicated the expected UX should be in
line with what a user expects the software will do
so we should not dns if we say so with a flag that suggest that.
Changelog-Fixed: We disable all dns even on startup the scan for bogus dns servers, if --always-use-proxy is set true
Signed-off-by: Saibato <saibato.naga@pm.me>
I was wondering why TAGS was missing some functions, and finally
tracked it down: PRINTF_FMT() confuses etags if it's at the start
of a function, and it ignores the rest of the file.
So we put PRINTF_FMT at the end, but that doesn't work for
*definitions*, only *declarations*. So we remove it from definitions
and add gratuitous declarations in the few static places.1
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We were using `i` as index variable in two nested loops. This works as long as
the DNS seed resolves to a single address, but will crash if the node has both
an A as well as an AAAA entry, at which point we'll try to index the hostname
without a matching entry.
Signed-off-by: Christian Decker <@cdecker>
This is mainly an internal-only change, especially since we don't
offer any globalfeatures.
However, LND (as of next release) will offer global features, and also
expect option_static_remotekey to be a *global* feature. So we send
our (merged) feature bitset as both global and local in init, and fold
those bitsets together when we get an init msg.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
With enable-autotor-v2 defined in cmdline the default behavior to create
v3 onions with the tor service call, is set to v2 onions.
Signed-off-by: Saibato <saibato.naga@pm.me>
This encoding scheme is no longer just used for short_channel_ids, so make
the names more generic.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
It's generally clearer to have simple hardcoded numbers with an
#if DEVELOPER around it, than apparent variables which aren't, really.
Interestingly, our pruning test was always kinda broken: we have to pass
two cycles, since l2 will refresh the channel once to avoid pruning.
Do the more obvious thing, and cut the network in half and check that
l1 and l3 time out.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Rather than reaching into data structures, let them register their own
callbacks. This avoids us having to expose "memleak_remove_xxx"
functions, and call them manually.
Under the hood, this is done by having a specially-named tal child of
the thing we want to assist, containing the callback.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Keeping the uintmap ordering all the broadcastable messages is expensive:
130MB for the million-channels project. But now we delete obsolete entries
from the store, we can have the per-peer daemons simply read that sequentially
and stream the gossip itself.
This is the most primitive version, where all gossip is streamed;
successive patches will bring back proper handling of timestamp filtering
and initial_routing_sync.
We add a gossip_state field to track what's happening with our gossip
streaming: it's initialized in gossipd, and currently always set, but
once we handle timestamps the per-peer daemon may do it when the first
filter is sent.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We used to generate this in the caller, then save it in case we needed
to retry. We're about to change the message we send to lightningd, so
we'll need to regenerate it every time; just hand all the extra args
into peer_connected() and we can generate the `connect_peer_connected`
msg there.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
1diff --git a/connectd/connectd.c b/connectd/connectd.c
index 94fe50b56..459c9ac63 100644
I tried to just do gossipd, but it was uncontainable, so this ended up being
a complete sweep.
We didn't get much space saving in gossipd, even though we should save
24 bytes per node.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Rusty Russell
niftynei
Saibato
Sebastian Falbesoner
darosior
Christian Decker
Vasil Dimov