the original version of the subtype generator emitted '$'
to designate that a field was a subtype; now it's got a different
format:
funding_type,8,num_inputs,2
funding_type,10,input_info,num_inputs*input_info
this patch updates our generator to understand this new format
This is needed so that some csv's can expose their subtype parsing
functions in their header. This gets used in a later PR where
we start replacing manually created 'subtype' definitions with
generated ones.
This doesn't result in a speedup for our benchmark, since we use the
cli tool which does the formatting.
MCP results from 5 runs, min-max(mean +/- stddev):
store_load_msec:33422-36830(35196.2+/-1.2e+03)
vsz_kb:2637488
store_rewrite_sec:36.030000-37.630000(36.794+/-0.52)
listnodes_sec:0.720000-0.950000(0.86+/-0.077)
listchannels_sec:40.300000-41.080000(40.668+/-0.29)
routing_sec:30.440000-31.030000(30.69+/-0.2)
peer_write_all_sec:50.060000-52.800000(51.416+/-0.91)
MCP notable changes from 2 patches ago (>1 stddev):
-listchannels_sec:48.560000-55.680000(52.642+/-2.7)
+listchannels_sec:40.300000-41.080000(40.668+/-0.29)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Plugins don't do it right anyway, and we're about to remove it from
lightningd. Produces same format as json_pp.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We can save significant space by combining both sides: so much that we
can reduce the WIRE_LEN_LIMIT to something sane again.
MCP results from 5 runs, min-max(mean +/- stddev):
store_load_msec:34467-36764(35517.8+/-7.7e+02)
vsz_kb:2637488
store_rewrite_sec:35.310000-36.580000(35.816+/-0.44)
listnodes_sec:1.140000-2.780000(1.596+/-0.6)
listchannels_sec:55.390000-58.110000(56.998+/-0.99)
routing_sec:30.330000-30.920000(30.642+/-0.19)
peer_write_all_sec:50.640000-53.360000(51.822+/-0.91)
MCP notable changes from previous patch (>1 stddev):
-store_rewrite_sec:34.720000-35.130000(34.94+/-0.14)
+store_rewrite_sec:35.310000-36.580000(35.816+/-0.44)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Don't turn them to/from pubkeys implicitly. This means nodeids in the store
don't get converted, but bitcoin keys still do.
MCP results from 5 runs, min-max(mean +/- stddev):
store_load_msec:33934-35251(34531.4+/-5e+02)
vsz_kb:2637488
store_rewrite_sec:34.720000-35.130000(34.94+/-0.14)
listnodes_sec:1.020000-1.290000(1.146+/-0.086)
listchannels_sec:51.110000-58.240000(54.826+/-2.5)
routing_sec:30.000000-33.320000(30.726+/-1.3)
peer_write_all_sec:50.370000-52.970000(51.646+/-1.1)
MCP notable changes from previous patch (>1 stddev):
-store_load_msec:46184-47474(46673.4+/-4.5e+02)
+store_load_msec:33934-35251(34531.4+/-5e+02)
-vsz_kb:2638880
+vsz_kb:2637488
-store_rewrite_sec:46.750000-48.280000(47.512+/-0.51)
+store_rewrite_sec:34.720000-35.130000(34.94+/-0.14)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Outputs CSV. We add some stats for load times in developer mode, so we can
easily read them out.
peer_read_all_sec doesn't work, since we seem to reject about half the
updates for having bad signatures. It's also very slow...
routing fails, for unknown reasons, so that failure is ignored in routing_sec.
Results from 5 runs, min-max(mean +/- stddev):
store_load_msec,vsz_kb,store_rewrite_sec,listnodes_sec,listchannels_sec,routing_sec,peer_write_all_sec
39275-44779(40466.8+/-2.2e+03),2899248,41.010000-44.970000(41.972+/-1.5),2.280000-2.350000(2.304+/-0.025),49.770000-63.390000(59.178+/-5),33.310000-34.260000(33.62+/-0.35),42.100000-44.080000(43.082+/-0.67)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Header from folded patch 'fixup!_tools-bench-gossipd.sh__rough_benchmark_for_gossipd_and_the_million_channels_project-2.patch':
fixup! tools/bench-gossipd.sh: rough benchmark for gossipd and the million channels project
Suggested-by: @niftynei
Header from folded patch 'fixup!_tools-bench-gossipd.sh__rough_benchmark_for_gossipd_and_the_million_channels_project-1.patch':
fixup! tools/bench-gossipd.sh: rough benchmark for gossipd and the million channels project
MCP filename change.
Header from folded patch 'tools-bench-gossipd.sh__dont_print_csv_by_default.patch':
tools/bench-gossipd.sh: don't print CSV by default.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Header from folded patch 'fixup!_tools-bench-gossipd.sh__rough_benchmark_for_gossipd_and_the_million_channels_project.patch':
fixup! tools/bench-gossipd.sh: rough benchmark for gossipd and the million channels project
Make shellcheck happy.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
fixup printing methods in devtools/decodemsg such that TLV's can
now be printed as well. here's how you'd use it:
$ ./devtools/decodemsg --tlv opening_tlv 0120001E020202020202020202020202020202020202020202020202020202020202
> WIRE_OPTION_UPFRONT_SHUTDOWN_SCRIPT (size 32):
> shutdown_scriptpubkey=[020202020202020202020202020202020202020202020202020202020202]
TLV's use var_int's for messages sizes, both internally and
in the top level (you should really stack a var_int inside a var_int!!)
this updates our automagick generator code to understand 'var_ints'
passing back a null TLV was crashing here, because we tried to
dereference a null pointer. instead, we put it into a temporary
struct that we can check for NULL-ness, before assigning to the
passed in pointer.
let's let the fromwire__tlv methods allocate the tlv-objects and
return them. we also want to initialize all of their underlying
messages to NULL, and fail if we discover a duplicate mesage type.
if parsing fails, instead of returning a struct we return NULL.
Suggested-By: @rustyrussell
Since messages in TLV's are optional, the ideal way to deal with
them is to have a 'master struct' object for every defined tlv, where
the presence or lack of a field can be determined via the presence
(or lack thereof) of a struct for each of the optional message
types.
In order to do this, appropriately, we need a struct for every
TLV message. The next commit will make use of these.
Note that right now TLV message structs aren't namespaced to the
TLV they belong to, so there's the potential for collision. This
should be fixed when/where it occurs (should fail to compile).
Add tlv-messages to the general messages set so that their parsing
messages get printed out.
FIXME: figure out how to account for partial message length processing?
Version 1.1 of the lightning-rfc spec introduces TLVs for optional
data fields. This starts the process of updating our auto-gen'd
wireformat parsers to be able to understand TLV fields.
The general way to declare a new TLV field is to add a '+' to the
end of the fieldname. All field type declarations for that TLV set
should be added to a file in the same directory by the name
`gen_<field_name>_csv`.
Note that the FIXME included in this commit is difficult to fix, as
we currently pass in the csv files via stdin (so there's no easy
way to ascertain the originating directory of file)
Otherwise we can't really return a variable sized message with more than 65k
results. This was causing an integer overflow in `listchannels` (see #2504 for
details).
Signed-off-by: Christian Decker <decker.christian@gmail.com>
We were tarring up the build dir, not the destination dir! We did this
for 0.6.3 and nobody noticed :(
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
@cdecker reports that this gives warnings on exit; and we can't suppress
them by setting ASAN_OPTIONS within the binary itself, unfortunately.
So for 0.7, disable it by default. I'll work through the errors for 0.7.1.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
And fix trivial typo in MAKING-RELEASES.md, and date retreival in
build-release.sh and repro-build.sh (real git tags start with v!)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
I tried building zipfile on a fresh clone inside KVM, and got
1. Different times inside the zipfile, since zip seems to save *local* times.
2. A different zipfile order, since zip seems to use filesystem order.
Fix both of these. I don't know if LANG=C is necessary for git
ls-files, but it can't hurt.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
For the moment it's only Ubuntu 18.04.1.
Complete documentation is in the final commit; you can test this using
the prior commit and comparing with my intermediate files and results
at:
https://ozlabs.org/~rusty/clightning-repro
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Basically we tell it that every field ending in '_msat' is a struct
amount_msat, and 'satoshis' is an amount_sat. The exceptions are
channel_update's fee_base_msat which is a u32, and
final_incorrect_htlc_amount's incoming_htlc_amt which is also a
'struct amount_msat'.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
They're generally used pass-by-copy (unusual for C structs, but
convenient they're basically u64) and all possibly problematic
operations return WARN_UNUSED_RESULT bool to make you handle the
over/underflow cases.
The new #include in json.h means we bolt11.c sees the amount.h definition
of MSAT_PER_BTC, so delete its local version.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Christian and I both unwittingly used it in form:
*tal_arr_expand(&x) = tal(x, ...)
Since '=' isn't a sequence point, the compiler can (and does!) cache
the value of x, handing it to tal *after* tal_arr_expand() moves it
due to tal_resize().
The new version is somewhat less convenient to use, but doesn't have
this problem, since the assignment is always evaluated after the
resize.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This causes a compiler warning if we don't do something with the
result (hopefully return immediately!).
We use was_pending() to ignore the result in the case where we
complete a command in a callback (thus really do want to ignore
the result).
This actually fixes one bug: we didn't return after command_fail
in json_getroute with a bad seed value.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Turns out that I should have tested these with a new dependency
instead of just submitting. `sed` was missing the s command.
Signed-off-by: Christian Decker <decker.christian@gmail.com>
This was introduced in ed268d6c, which broke the mocks
generation. This just filters out the invalid sentinel value.
Signed-off-by: Christian Decker <decker.christian@gmail.com>
There were a few reports that upgrading Ubuntu recently caused issues
because we assert that the sqlite3 library version matches the one we
were built with. 'make' doesn't fix this, because it doesn't know the
external libraries have changed.
Fix this harder, with a helper which updates a file every binary depends
on, which gets relinked every time so we detect link changes.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
If we have an array of varlen structures (which require a ctx arg), we
should make that arg the array itself (which was tal_arr()), not the
root context.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We do this a lot, and had boutique helpers in various places. So add
a more generic one; for convenience it returns a pointer to the new
end element.
I prefer the name tal_arr_expand to tal_arr_append, since it's up to
the caller to populate the new array entry.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
tal_count() is used where there's a type, even if it's char or u8, and
tal_bytelen() is going to replace tal_len() for clarity: it's only needed
where a pointer is void.
We shim tal_bytelen() for now.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This requires a tweak to generate-wire.py too, since it always called the
top-level routine 'print_message'.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
gossip_getnodes_entry was used by gossipd for reporting nodes, and for
reporting peers. But the local_features field is only available for peers,
and most other fields are only available from node_announcement.
Note that the connectd change actually means we get less information
about peers: gossipd used to do the node lookup for peers and include the
node_announcement information if it had it.
Since generate_wire.py can't create arrays-of-arrays, we add a 'struct
peer_features' to encapsulate the two feature arrays for each peer, and
for convenience we add it to lightningd/gossip_msg.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Well, it's generated by shachain, so technically it is a sha256, but
that's an internal detail. It's a secret.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We already work around this by using an array with a 0/1 length convention,
but better to be explicit.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is based on @NicolasDorier's excellent proposal for a Dockerfile, sans the
writing of a config file.
Co-authored-by: Nicolas Dorier <nicolas.dorier@gmail.com>
Co-authored-by: Christian Decker <decker.christian@gmail.com>
Signed-off-by: Christian Decker <decker.christian@gmail.com>
If we change an upstream URL, all submodules break. Users would need
to run 'git submodule sync'. Note that the libbacktrace fix was merged
upstream so this is no longer necessary, but it's good for future changes.
Also, stress-testing reveals that git submodule fails locking
'.git/config' when run in paralell. It also segfaults and other
problems.
This is my final attempt to fix submodules; I've wasted far too many
days on obscure problems it creates: I've already lost one copy of my
repo to apparently unfixable submodule preoblems. The next "fix" will
be to simply import the source code so it works properly.
Reported-by: @jsarenik
Fixes: #1543
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Fixes: #1221
We were using `\x0` to match NUL chars in the input (on the
assumption that NUL chars are "impossible" for decent LFS-compliant
systems).
However `\x0` is a GNUism.
Use the `\n` and the newline character, which is supported by (most)
POSIX sed.
We always hand in "NULL" (which means use tal_len on the msg), except
for two places which do that manually for no good reason.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
A convenient alias for char *, though we don't allow control characters
so our logs can't be fooled with embedded \n.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
These are now logically arrays of pointers. This is much more natural,
and gets rid of the horrible utxo array converters.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The pagination causes it to wait for a keypress even with no output
under emacs (complaining about the terminal); we don't want it anyway.
Example output:
Makefile:228:#lighnting!
Identified a likely misspelling of the word "lightning" (see above). Please fix.
Is this warning incorrect? Please teach tools/check-spelling.sh about the exciting new word.
Makefile:230: recipe for target 'check-spelling' failed
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The close_info is needed to re-derive the secret key that is supposed
to be used to sign the input spending the output.
Signed-off-by: Christian Decker <decker.christian@gmail.com>
It's just a sha256_double, but importantly when we convert it to a
string (in type_to_string, which is used in logging) we use
bitcoin_blkid_to_hex() so it's reversed as people expect.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
It's just a sha256_double, but importantly when we convert it to a
string (in type_to_string, which is used in logging) we use
bitcoin_txid_to_hex() so it's reversed as people expect.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
If a structure foo has a optional fields opt1 and opt2, this creates
towire_foo, towire_foo_opt1 and towire_foo_opt2 (since opt2 implies opt1),
similarly for fromwire_*.
This requires the callers to be updated to call the correct routines (eg.
try fromwire_foo_opt2, then fromwire_foo_opt1, then finally fromwire_foo),
but this is a minimal change to the generation code.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We're going to need this, and the PRINTF_FMT(1,2) in front of it caused
mockup.sh to miss the declaration.
We also eliminate the obviously-unused fallback case (which referred
to daemon/*.h).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
update-mocks was broken, since it assumed the daemon/ directory.
We now use "make" directly to build the test file and harvest errors,
and are more robust if it simply doesn't compile (ie. fails, but no
linker errors).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Now in sync with 8ee57b97738b1e9467a1342ca8373d40f0c4aca5.
Our tool doesn't need to convert them any more, but we actually had a
mis-typed field in the HSM which needed fixing.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The Dockerfile is now stored in contrib and built using the Docker
Hub. This allows us to simply pull in the finished image from the hub
instead of having to build it ourself. Should shave off about 2
minutes from the build time.
I also switched to running the individual build and check steps in
their own containers, but on the same volume, so travis can group the
commands and run them independently.
Some of the struct array helpers need to allocate data when
deserializing their fields. The `getnodes` reply is one such example
that allocates the hostname. Since the change to calling array helpers
the getnodes call was broken because it was attempting to allocate off
of the entry, which did not have a tal header, thus failing.
We use the fourth value (size) to determine the type, unless the fifth
value is suppled. That's silly: allow the fourth value to be a typename,
since that's the only reason we care about the size at all!
Unfortunately there are places in the spec where we use a raw fieldname
without '*1' for a length, so we have to distingish this from the
typename case.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Except for the trivial case of u8 arrays, have the generator create
the loop code for the array iteration.
This removes some trivial helpers, and avoids us having to write more.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The wiregen tool was a bit hard to maintain since it was printing all
over the place, mixing template and processing logic. This commit
tears the two apart, externalizes everything that is not a single code
line, and repackages it into templates. Specifically functions are now
their own template and header/implementation files are a template.
Furthermore this simplifies some of the boilerplate of mapping types
to sizes and back again, by extracting them into dicts.
All changes have been verified to produce identical results on the
current wire definitions, except a bit of whitespace changes.
The spec 4af8e1841151f0c6e8151979d6c89d11839b2f65 uses a 32-byte 'channel-id'
field, not to be confused with the 8-byte short ID used by gossip. Rename
appropriately, and update to the new handshake protocol.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
It's awkward to handle them differently. But this change means we
need to expose them to the generated code.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This lets us marshal and unmarshal undefined complex types; the fromwire
function will allocate it for us, so we don't even need to know the size.
This turns out to be really nice for marshalling 'struct crypto_state'.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
For our internal CSV files, we can specify the type explicitly rather
than trying to guess (eg. bool).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Pretty! Takes an int instead of the enum directly, because in the
main daemon we call it via a function pointer, so want them all the
same type.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
It implies tal_count() gives the length. Great for almost all callers which
don't care if there are extra bytes.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This removes some redundancy in creating messages, but also allows
a lazy form or parsing without explicitly checking the type.
A helper fromwire_peektype() is added to look up the type and handle
the too-short-for-type problem.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We emit them into the generated code at appropriate points, but it
would be better if we simply preserved the order they were given in.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is a bit more awkward for large structures, but avoids
indirection for the simpler ones (I copied the structures for the test
code, however). We also remove explicit padding.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
