Note that check-whitespace and check-bolt already do this, so we
can eliminate redundant lines in common/Makefile and bitcoin/Makefile.
We also include the plugin headers in ALL_C_HEADERS so they get
checked.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
I got a corrupt file, which looked like multiple concurrent attempts
to build it. So instead, build it in one command, but also use
VERBOSE so we print correctly with V=1 (and --quiet).
Also move into plugins/ where it logically belongs.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We promised this in 0.8.1!
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Removed: Plugin: Relative plugin paths are not relative to startup (deprecated v0.7.2.1)
PR #3957 improved performance considerably, however we still look over the
entire message for the message separator. If instead we just look in the
incrementally read data, we remove the quadratic behavior for large messages.
This is safe since we then loop over the messages which would drain any
message separator from the buffer before we attempt the next read.
Changelog-Fixed: bcli: Significant speedups for block synchronization
time lightning-cli -R --network=regtest --lightning-dir /tmp/ltests-k8jhvtty/test_pay_stress_1/lightning-1/ listpays > /dev/null
Before:
real 0m42.741s
user 0m0.149s
sys 0m0.016s
After:
real 0m13.674s
user 0m0.131s
sys 0m0.024s
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Fixed: JSON-RPC: significant speedups for plugins which create large JSON replies (e.g. listpays on large nodes).
The jsmn parser is a beautiful piece of code. In particular, you can parse
part of a string, then continue where you left off.
We don't take advantage of this, however, meaning for large JSON objects
we parse them multiple times before finally having enough to complete.
Expose the parser state and tokens through the API, so the caller can pass
them in repeatedly. For the moment, every caller is allocates each time
(except the unit tests).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This lets us handle it the same way we handle builtin commands, and also
lets us warn if they use deprecated apis and allow-deprecated-apis=false.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: Plugins: can now mark their options and commands deprecated.
This allows plugins to choose how to present things in getmanifest.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: plugins: `getmanifest` may now include "allow-deprecated-apis" boolean flag.
Changelog-Deprecated: plugins: `getmanifest` without any parameters; plugins should accept any parameters for future use.
Changelog-Added: New option `--important-plugin` loads a plugin is so important that if it dies, `lightningd` will exit rather than continue. You can still `--disable-plugin` it, however, which trumps `--important-plugin` and it will not be started at all.
`listconfigs` calls were setting the description twice and was using the
pointer to the boolean value as the boolean value, resulting in always
returning `true`.
There's no reason to assign the plugin vars inside the callback, so do
that outside, and the tal_steal() is redundant (the plugin is already
the conn parent).
And reduce duplication by making plugin_conn_finish call plugin_kill:
just make sure we don't call plugin_conn_finish again if plugin_kill
is called externally.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The previous implementation was a bit lazy: in particular, since we didn't
remember the disabled plugins, we would load them on rescan.
Changelog-Changed: config: the `plugin-disable` option works even if specified before the plugin is found.
1. Make the destructor call check_plugins_resolved(),
unless it was uninitialized (`opt_disable_plugin`).
2. Remove redundant list_del (destructor already does it).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
That's more convenient for most callers, which don't need a fmt.
Fixed-by: Darosior <darosior@protonmail.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is what I expected from plugin_kill, and now all the callers do the
equivalent anywat, it's easy.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Instead of calling plugin_kill() and returning, have them
uniformly return an error string or NULL, and have the top
level (plugin_read_json) do the plugin_kill() call.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This means we now clean up options in startup plugins (that was only
done by dynamic code!), and now they both share the 60 second timeout
instead of 20 seconds for dynamic.
For the dynamic case though, it's 60 seconds to both complete
getmanifest and init, which seems fair.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This will allow the dynamic starting code to use them too.
Also lets us move dev_debug_subprocess under #if DEVELOPER.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This will let us unify the startup and runtime-started infrastructure.
Note that there are two kinds of notifications:
1. Starting a single plugin (i.e. `plugin start`)
2. Starting multiple plugins (i.e. `plugin rescan` or `plugin startdir`).
In the latter case, we want the command to complete only once *all*
the plugins are dead/finished.
We also call plugin_kill() in all cases, and correctly return afterwards
(it matters once we use the same paths for dynamic plugins, which don't
cause a fatal error if they don't startup).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The symptom (under heavy load and valgrind) in test_plugin_command:
lightningd: common/json_stream.c:237: json_stream_output_: Assertion `!js->reader' failed.
This is because we try to call `getmanifest` again on `pay` which has not yet
responded to init.
The minimal fix for this is to keep proper state, so we can tell the
difference between "not yet called getmanifest" and "not yet finished
init".
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We use the new function `plugins_free` to define the correct deallocation
order on shutdown, since under normal operation the allocation tree is
organized to allow plugins to terminate and automatically free all dependent
resources. During shutdown the deallocation order is under-defined since
siblings may get freed in any order, but we implicitly rely on them staying
around.
We now track all pending RPC passthrough calls, and terminate them with an
error if the plugin dies.
Changelog-Fixed: JSON-RPC: Pending RPC method calls are now terminated if the handling plugin exits prematurely.
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>
Shows what features we use in various contexts, including those added
by plugins in getmanifest.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: Plugin: `feature_set` object added to `init`
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 cleans up the boutique handling of features, and importantly, it
means that if a plugin says to offer a feature in init, we will now
*accept* that feature.
Changelog-Fixed: Plugins: setting an 'init' feature bit allows us to accept it from peers.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
also: convert the stored int value from 'int' to 's64'
atoi fails silently, returning a zero. instead we use the more robust
strtoll which will allow us fail with an error.
we also make the parsing for bools stricter, only allowing plausibly
boolean values to parse.
We were waiting for both stdin and stdout to close, however that resulted in
us deferring cleanup indefinitely since we did not poll stdout for being
writable most of the time. On the other hand we are almost always polling
the plugin's stdout, so that notifies us as soon as the plugin stops.
Changelog-Fixed: plugin: Plugins no longer linger indefinitely if their process terminates
As a separated commit because it was pre-existent (changelog + xfail test).
This also fix a logical problem in lightningd/plugin_control: we were
assuming a plugin started with 'plugin start' but which did not comport
a 'dynamic' entry in its manifest to be dynamic, though it should have
been treated as static.
Changelog-fixed: plugins: Dynamic C plugins can now be managed when lightningd is up
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Removed: Relative plugin paths are not relative to startup (deprecated v0.7.2.1)
Changelog-Removed: Dummy fields in listforwards (deprecated v0.7.2.1)
This will change the command `listconfigs` output in several ways:
- Deprecated the duplicated "plugin" JSON output by replacing it with
- a "plugins" array with substructures for each plugin with:
- path, name and their options
Changelog-Changed: JSON-RPC: `listconfigs` now structures plugins and include their options
Changelog-Deprecated: JSON-RPC: `listconfigs` duplicated "plugin" paths
Changelog-changed: .lightningd plugins and files moved into <network>/ subdir
Changelog-changed: WARNING: If you don't have a config file, you now may need to specify the network to lightning-cli
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
A log can have a default node_id, which can be overridden on a per-entry
basis. This changes the format of logging, so some tests need rework.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
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>
The fundchannel plugin needs to know how to build a transaction, so we need to
tell it which chainparams to use. Also adds `chainparams` as a global, since
that seems to be the way to do things in plugins.
This moves field initialization into plugins_new(), and
adds a memleak helper to search the request map:
=================================== ERRORS ====================================
___________________ ERROR at teardown of test_plugin_command ___________________
[gw0] linux -- Python 3.7.1 /opt/python/3.7.1/bin/python3.7
> lambda: ihook(item=item, **kwds),
when=when,
)
../../../.local/lib/python3.7/site-packages/flaky/flaky_pytest_plugin.py:306:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
tests/fixtures.py:112: in node_factory
ok = nf.killall([not n.may_fail for n in nf.nodes])
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
self = <utils.NodeFactory object at 0x7f873b245278>, expected_successes = [True]
def killall(self, expected_successes):
"""Returns true if every node we expected to succeed actually succeeded""
unexpected_fail = False
for i in range(len(self.nodes)):
leaks = None
# leak detection upsets VALGRIND by reading uninitialized mem.
# If it's dead, we'll catch it below.
if not VALGRIND:
try:
# This also puts leaks in log.
leaks = self.nodes[i].rpc.dev_memleak()['leaks']
except Exception:
pass
try:
self.nodes[i].stop()
except Exception:
if expected_successes[i]:
unexpected_fail = True
if leaks is not None and len(leaks) != 0:
raise Exception("Node {} has memory leaks: {}".format(
self.nodes[i].daemon.lightning_dir,
> json.dumps(leaks, sort_keys=True, indent=4)
))
E Exception: Node /tmp/ltests-qm87my20/test_plugin_command_1/lightnng-1/ has memory leaks: [
E {
E "backtrace": [
E "ccan/ccan/tal/tal.c:437 (tal_alloc_)",
E "lightningd/jsonrpc.c:1112 (jsonrpc_request_start_)",
E "lightningd/plugin.c:1041 (plugin_config)",
E "lightningd/plugin.c:1072 (plugins_config)",
E "lightningd/plugin.c:846 (plugin_manifest_cb)",
E "lightningd/plugin.c:252 (plugin_response_handle)",
E "lightningd/plugin.c:342 (plugin_read_json_one)",
E "lightningd/plugin.c:367 (plugin_read_json)",
E "ccan/ccan/io/io.c:59 (next_plan)",
E "ccan/ccan/io/io.c:407 (do_plan)",
E "ccan/ccan/io/io.c:417 (io_ready)",
E "ccan/ccan/io/poll.c:445 (io_loop)",
E "lightningd/io_loop_with_timers.c:24 (io_loop_with_tiers)",
E "lightningd/lightningd.c:840 (main)"
E ],
E "label": "lightningd/jsonrpc.c:1112:struct jsonrpc_reques",
E "parents": [
E "lightningd/plugin.c:66:struct plugin",
E "lightningd/lightningd.c:103:struct lightningd"
E ],
E "value": "0x55d6385e4088"
E },
E {
E "backtrace": [
E "ccan/ccan/tal/tal.c:437 (tal_alloc_)",
E "lightningd/jsonrpc.c:1112 (jsonrpc_request_start_)",
E "lightningd/plugin.c:1041 (plugin_config)",
E "lightningd/plugin.c:1072 (plugins_config)",
E "lightningd/plugin.c:846 (plugin_manifest_cb)",
E "lightningd/plugin.c:252 (plugin_response_handle)",
E "lightningd/plugin.c:342 (plugin_read_json_one)",
E "lightningd/plugin.c:367 (plugin_read_json)",
E "ccan/ccan/io/io.c:59 (next_plan)",
E "ccan/ccan/io/io.c:407 (do_plan)",
E "ccan/ccan/io/io.c:417 (io_ready)",
E "ccan/ccan/io/poll.c:445 (io_loop)",
E "lightningd/io_loop_with_timers.c:24 (io_loop_with_tiers)",
E "lightningd/lightningd.c:840 (main)"
E ],
E "label": "lightningd/jsonrpc.c:1112:struct jsonrpc_reques",
E "parents": [
E "lightningd/plugin.c:66:struct plugin",
E "lightningd/lightningd.c:103:struct lightningd"
E ],
E "value": "0x55d6386529d8"
E }
E ]
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Note that we move adding the plugin to the plugins list to the end, otherwise
the hook from logging can examine the (uninitialized) plugin.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This adds a new pair of files : lightningd/plugin_control, along with a new RPC
command : 'plugin'. This command can be used to manage plugins without restarting lightningd:
lightning-cli plugin start helloworld.py
lightning-cli plugin stop helloworld.py
This adds a 'configured' boolean member to the plugin struct so that we can add plugins to ld->plugins' list and differenciate fresh plugins.
This also adds 'plugins_start' so that new plugins can be started without calling 'plugins_init' and running an io loop
This is an old bug, where a plugin can get called while we're shutting
down (and have freed plugins), but it's triggered more reliably by the
new warning notification hook.
For good measure, we also make freeing a plugin self-delete.
Valgrind error file: valgrind-errors.16763
==16886== Invalid read of size 8
==16886== at 0x422919: plugins_notify (plugin.c:1096)
==16886== by 0x413919: notify_warning (notification.c:61)
==16886== by 0x412BDE: logv (log.c:251)
==16886== by 0x412A98: log_ (log.c:311)
==16886== by 0x4044BE: bcli_finished (bitcoind.c:178)
==16886== by 0x459480: destroy_conn (poll.c:244)
==16886== by 0x459499: destroy_conn_close_fd (poll.c:250)
==16886== by 0x4619E1: notify (tal.c:235)
==16886== by 0x461A7E: del_tree (tal.c:397)
==16886== by 0x461AB5: del_tree (tal.c:407)
==16886== by 0x461AB5: del_tree (tal.c:407)
==16886== by 0x461AB5: del_tree (tal.c:407)
==16886== Address 0x634a578 is 40 bytes inside a block of size 352 free'd
==16886== at 0x4C2EDEB: free (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==16886== by 0x461AFD: del_tree (tal.c:416)
==16886== by 0x461FB7: tal_free (tal.c:481)
==16886== by 0x411E0A: main (lightningd.c:841)
==16886== Block was alloc'd at
==16886== at 0x4C2DB8F: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==16886== by 0x4617CE: allocate (tal.c:245)
==16886== by 0x461E4C: tal_alloc_ (tal.c:423)
==16886== by 0x42255E: plugins_new (plugin.c:106)
==16886== by 0x41133D: new_lightningd (lightningd.c:218)
==16886== by 0x411AD4: main (lightningd.c:649)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Move it closer to ccan/json_out, in preparation for using that as a
replacement.
In particular:
1. Add a 'quote' field in json_add_member.
2. json_add_member now always escapes if 'quote' is true.
3. json_member_direct is exposed to allow avoiding of escaping.
4. json_add_hex can use this, so no longer needs to be in json_stream.c.
5. We don't make JSON manually, but always use helpers.
6. We now flush the stream (wake reader) only when we close it, or mark
command as pending.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
- Related Changes for `warning` notification
Add a `bool` type parameter in `log_()` and `lov()`, this `bool` flag
indicates if we should call `warning` notifier.
1) The process of copying `log_book` of every peer to the `log_book` of
`ld` is usually included in `log_()` and `lov()`, and it may lead to
repeated `warning` notification. So a `bool`, which explicitly indicates
if the `warning` notification is disabled during this call, is necessary
.
2) The `LOG_INFO` and `LOG_DEBUG` level don't need to call
warning, so set that `bool` paramater as `FALSE` for these log level and
only set it as `TRUE` for `LOG_UNUAUSL`/`LOG_BROKEN`. As for `LOG_IO`,
it use `log_io()` to log, so we needn't think about notifier for it.
This was deeply surprising to me; there's a difference between a value not being
specified, and it being specified as "".
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
