xpay is relying on the destructor to send another request. This means
that it doesn't actually submit the request until *next time* we wake.
This has been in xpay from the start, but it is not noticeable until
xpay stops subscribing to every command on the rpc_command hook.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Basically, `devtools/reduce-includes.sh */*.c`.
Build time from make clean (RUST=0) (includes building external libs):
Before:
real 0m38.944000-40.416000(40.1131+/-0.4)s
user 3m6.790000-17.159000(15.0571+/-2.8)s
sys 0m35.304000-37.336000(36.8942+/-0.57)s
After:
real 0m37.872000-39.974000(39.5466+/-0.59)s
user 3m1.211000-14.968000(12.4556+/-3.9)s
sys 0m35.008000-36.830000(36.4143+/-0.5)s
Build time after touch config.vars (RUST=0):
Before:
real 0m19.831000-21.862000(21.5528+/-0.58)s
user 2m15.361000-30.731000(28.4798+/-4.4)s
sys 0m21.056000-22.339000(22.0346+/-0.35)s
After:
real 0m18.384000-21.307000(20.8605+/-0.92)s
user 2m5.585000-26.843000(23.6017+/-6.7)s
sys 0m19.650000-22.003000(21.4943+/-0.69)s
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Each header should only include the other headers it needs to compile;
`devtools/reduce-includes.sh */*.h` does this. The C files then need
additional includes if they don't compile.
And remove the entirely useless wire/onion_wire.h, which only serves to include wire/onion_wiregen.h.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Note that we create a destructor on the command to reset request->cmd
pointer if the cmd is freed (so we know not to call the callback).
But attaching hundreds of thousands of them is slow: it's a
single-linked list, which is iterated in several places.
But that's redundant: the request is now allocated off the cmd, so freeing the command
will free the request anyway.
Hacking in something to print progress to a file, here's the number of
requests processed every 10 seconds before and after:
Before:
$ while sleep 10; do wc -l /tmp/bkpr-progress; done
181529 /tmp/bkpr-progress
195994 /tmp/bkpr-progress
207083 /tmp/bkpr-progress
226336 /tmp/bkpr-progress
234319 /tmp/bkpr-progress
241514 /tmp/bkpr-progress
247421 /tmp/bkpr-progress
255292 /tmp/bkpr-progress
261367 /tmp/bkpr-progress
269085 /tmp/bkpr-progress
276953 /tmp/bkpr-progress
282233 /tmp/bkpr-progress
286193 /tmp/bkpr-progress
290930 /tmp/bkpr-progress
295276 /tmp/bkpr-progress
301086 /tmp/bkpr-progress
After:
169505 /tmp/bkpr-progress
196010 /tmp/bkpr-progress
219370 /tmp/bkpr-progress
235671 /tmp/bkpr-progress
244242 /tmp/bkpr-progress
255362 /tmp/bkpr-progress
265636 /tmp/bkpr-progress
276966 /tmp/bkpr-progress
284451 /tmp/bkpr-progress
288836 /tmp/bkpr-progress
296578 /tmp/bkpr-progress
304571 /tmp/bkpr-progress
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We're going to be using this instead of our internal db.
I also made json_out_obj() take the str arg, as it didn't and I
expected it to.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
After much thought and mis-steps, I chose a simple solution: open another fd
for sync comms. It's almost impossible to know what state the async one is in.
jsonrpc_request_sync() is enhanced to return a valid tal object, as the current
behaviour of returning a pointer to inside an array was surprising.
Changelog-Changed: libplugin: you can now call the synchronous API functions at any time (not just in the init callback).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
All the core notifications changed over to wrapping the notification
fields in an object with the name of the notification, but notifications
from plugins were missed.
Changelog-Added: Plugins: `channel_hint_update`, `pay_failure` and `pay_success` notifications now have objects of the same name containing the expected fields.
Changelog-Deprecated: Plugins: `channel_hint_update`, `pay_failure` and `pay_success` notification fields outside the same-named object.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
A log event LOG_TRACE submitted by a plugin was being logged as
**BROKEN** by lightningd before this commit.
Changelog-Fixed: plugins can now log events under the LOG_TRACE flag.
Signed-off-by: Lagrang3 <lagrang3@protonmail.com>
This happens with autoclean, which does a datastore request then frees
the parent command without waiting for a response (see clean_finished).
This leaks a trace, and causes a crash if the pointer is later reused.
My solution is to create a trace variant which declares the trace key
to be a tal ptr and then we can clean up in the destructor if this happens.
This fixes the issue for me.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Fixed: autoclean: fixed occasional crash when tracepoints compiled in.
Turns out we weren't wiring them through! And libplugin wasn't reading them anyway.
Changelog-Fixed: lightningd: tell plugins our bolt12 features (so our bolt12 invoices explicitly allow MPP).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Default goes to stderr for LOG_UNUSUAL and higher.
We have to whitelist more cases in map_catchup so we don't spam the logs
with perfectly-expected (but ignored) messages though.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Because we initalized plugin->io_rpc_conn *after* calling plugin->init,
send_outreq would do a (harmless, in our case) wakeup on an uninitialized address:
```
==1164079== Conditional jump or move depends on uninitialised value(s)
==1164079== at 0x1628FC: backend_wake (poll.c:227)
==1164079== by 0x160B98: io_wake (io.c:384)
==1164079== by 0x1160A8: ld_rpc_send (libplugin.c:255)
==1164079== by 0x1187E0: send_outreq (libplugin.c:1099)
==1164079== by 0x115041: init (xpay.c:1620)
```
Solution is simple: set plugin->io_rpc_conn to NULL, and don't wake it in this case.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The `send_outreq` function is a good place to suspend and resume
traces, since these are usually the places where we hand off control
back to the `io_loop`. This assumes that we do not continue doing
heavy liftin after we have queued an `outreq` call, but that is most
likely the case anyway. This frees us from having to track suspensions
whenever we call the RPC from a plugin.
Without knowing what method was called, we can't have useful general logging
methods, so go through the pain of adding "const char *method" everywhere,
and add:
1. ignore_and_complete - we're done when jsonrpc returned
2. log_broken_and_complete - we're done, but emit BROKEN log.
3. plugin_broken_cb - if this happens, fail the plugin.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
When we used to allow cmd to be NULL, we had to hand the plugin
everywhere. We no longer do.
1. Various jsonrpc_ functions no longer need the plugin arg.
2. send_outreq no longer needs a plugin arg.
3. The init function takes a command, not a plugin.
4. Remove command_deprecated_in_nocmd_ok.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is cleaner: everything can now be associated with a command
context.
You're supposed to eventually dispose of it using timer_complete().
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Sometimes we want to clean up *after* a command has completed, but
we're moving to a model where all libplugin operations require a
`struct command`. This adds `aux_command` to create an
independent-lifetime command with the same id.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This avoids globals (and means memleak traverses the variables!): we
only change over the test plugin though, to avoid unnecessary churn.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Currently make a plugin that do reportings of logs on
a services like graphana is not possible. So this commit
include the possibility to write a plugin that do the report
of this analisys.
Changelog-Added: core: notify plugins when a log line is emitted.
Signed-off-by: Vincenzo Palazzo <vincenzopalazzodev@gmail.com>
This means we can see the values in listconfigs, even if we haven't set
them yet.
In particular, we now see the following:
* autoclean-cycle.value_int=3600
* bitcoin-rpcclienttimeout.value_int=60
* bitcoin-retry-timeout.value_int=60
* funder-max-their-funding.value_str=4294967295sat
* funder-per-channel-min.value_str=10000sat
* funder-reserve-tank.value_str=0sat
* funder-fund-probability.value_int=100
Changelog-Changed: plugins: libplugin now shows plugin option default values (where they're non-trivial)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We need to pass through setconfig in check mode, then we need to have libplugin
add (and use!) a `check_only` parameter to all option setting.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: JSON-RPC: `check` `setconfig` on plugin options can now check the config value would be accepted.
We don't thoroughly handle `check setconfig`: it would be good to
allow this to do further checking!
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We want to extend it to plugins, and we want it to be allowed to be async for more power,
so rather than not completing the cmd if we're checking, do it in command_check_done()
and call it.
This is cleaner than the special case we had before, and allows check to us all the
normal jsonrpc mechanisms, especially async requests (which we'll need if we want to
hand check requests to plugins!).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Uncommited channels are missing several fields which would normally be
populated by an active channel. This simply skips them for the purposes
of finding a route. The particular culprit was:
{
"peer_id": "038cd9f3679d5b39bb2105978467918d549572de472f07dd729e37c7a6377d41d5",
"peer_connected": true,
"state": "OPENINGD",
"owner": "lightning_openingd",
"opener": "local",
"to_us_msat": 8317559000,
"total_msat": 8317559000,
"features": [
"option_static_remotekey",
"option_anchors_zero_fee_htlc_tx"
]
}
Fixes#7197 - SEGV in direct_pay_listpeerchannels when field private missing
Changelog-Fixed: Fixed crash in pay plugin caused by parsing uncommitted dual open channels
And we don't need to handle 0.9 lightningd which didn't include
allow-deprecated-apis in getmanifest call.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
